About Me

My Photo

I previously worked on Virtual Reality and other hardware at Valve.  I currently work at Google[x].

Prior to starting at Valve, I built computer peripherals such as keyboards, mice, and joysticks that were designed to be used inside MRI machines.  My company, Mag Design and Engineering, sold these devices directly to researchers at academic institutions who used them to publish scientific papers in peer-reviewed journals.

After work, I spend time on many different types of projects that usually involve circuit design, machining, material selection, and general fabrication/hacking.  My favorite place to be is my home workshop.

ben dot krasnow at gmail




Sunday, August 24, 2008

DIY liquid nitrogen generator

Maker Faire 2010

I have always been intrigued by science demonstrations using liquid nitrogen, and often made trips to a local welding supply store with my stainless steel vacuum flask to purchase liquid nitrogen and satisfy my cryogenic craving at home. After a few fill-ups, I wondered about the possibility of making liquid nitrogen on demand. Some companies have already produced self-contained liquid nitrogen generators that are designed for small laboratories (http://www.elan2.com/). The Elan2 would be ideal for home experimenters, but the cost is over $10,000, so I decided to build a similar device with less total output, lower purity, and at much lower cost. The device that I built cost less than $500 and produces 1 liter of liquid nitrogen per day.
Nearly all large-scale liquid nitrogen is made by compressing, cooling, and expanding air. This process removes heat from the air and can be repeated until the air liquefies. The condensing gasses are then separated using fractional distillation. This process cannot be easily scaled down because it relies on maintaining a complex, large distillation column to separate nitrogen from the other gasses in air. To avoid using a distillation column, one could use a nitrogen separation device to strip out the nitrogen from air at room temperature. Then, the room temperature nitrogen can be liquefied via the standard compression and expansion method. This is likely the process used in the Elan2 generator. However, it still requires the use of a very high pressure compressor and heat exchanger, extensive insulation and many other custom parts.
Another approach to producing small-scale liquid nitrogen is to use a self-contained cryocooler, which is a specialized refrigeration device that is designed to pump heat across a high temperature differential. In many cases, the devices are specifically designed for small-scale use and designed for spot-cooling in electronics. The benefit of using a cryocooler is that the device requires almost no maintenance and can liquefy gasses at atmospheric pressure. A compressor would not even be necessary in a gas liquefier using a cryocooler, but is helpful for removing water from the air and isolating nitrogen from air’s other component gasses. There are a few different basic types of cryocooler, but this article will highlight free-piston Stirling cycle cryocoolers. These devices are built with an internal piston that is driven by an electrical coil – a linear motor. The piston expands a working fluid (usually helium) in the device while a separate displacing piston moves the fluid to the tip. The piston then reverses direction, compressing the fluid as the displacing piston forces the fluid toward a heat-rejection area of the device. This process is repeated so that the working fluid is constantly being expanded at the tip, and compressed at the heat-rejection area. This causes heat to be pumped from the tip to the rejection area. The rejection area is cooled with atmospheric air, or other fluids that exchange heat with the environment.
Stirling cryocoolers are not relatively common devices, but they are used for RF filters that contain superconducting components. Such RF filters with their integrated cryocoolers can be found on eBay for under $300. One particular unit is the Superfilter built by Superconductor Technologies Inc. It contains a cryocooler that is rated at 140 watts of input power, and is extensively documented here (http://books.google.com/books?id=POLgG5mma6IC&pg=PA75).
I purchased the Superfilter on eBay and extracted the cryocooler. In order to test the device, I attached a small heatsink to the cooler’s cold tip, placed the tip into a household vacuum flask, and powered up the unit. After 30 minutes, I took the cryocooler out of the flask, and noticed a small amount of liquid air had collected at the bottom. Inspired by this success, I continued construction of a more complete liquid nitrogen generator. I already owned a 30-liter dewar (large vacuum flask) and fabricated an acrylic plate that would seal the top of the dewar while the cryocooler was also mounted to the plate with its heatsink hanging down into the neck of the flask. I also removed the cryocooler’s finned heatsink on its heat rejection area and replaced it with a liquid-cooling manifold. Liquid cooling lowered the heat rejection area temperature more effectively than forced air cooling, and this ultimately lead to higher system efficiency.
The liquid nitrogen generator has two basic sections, the dewar with cryocooler, and the air processing equipment that creates dry nitrogen from atmospheric air. The dry nitrogen is fed into the dewar at just above atmospheric pressure where the cryocooler chills the nitrogen until it liquefies and drips off the heatsink. Surprisingly, most of project’s time budget was spent designing and building the equipment to produce dry nitrogen from air. There are some companies who make dry nitrogen supply devices, but even small units are meant for much higher throughput than what is needed by this liquid nitrogen generator. Each liter of liquid nitrogen requires about 700 liters of room temperature nitrogen gas. 700 liters per day is only 0.5 l/min, a very modest flow rate. One popular, but unnecessary use for relatively low-purity nitrogen is filling car tires. I tried to purchase such a machine, but the cost and flow rate were much higher than anticipated. Instead, I found a very small nitrogen separation membrane on eBay. It’s original use was unknown. The separation membrane is the actual component inside commercial nitrogen generators that perform gas separation. The membrane is formed into a large bundle of hundreds of 2mm dia tubes. Air is fed under high pressure into one end of the bundle. The tube walls are semi-permeable and allow oxygen, water vapor, and other “fast” gasses to permeate relatively quickly. Nitrogen and slower (larger molecules) gasses do not permeate as quickly, so the concentration of nitrogen is much higher at the exit end of the tubes than it is at the input end. Higher purities of nitrogen can be achieved by restricting the flow rate through the tubes, thus allowing plenty of time for the unwanted gasses to permeate the tube walls and leave the system. The resulting nitrogen will contain trace amounts of argon and even smaller amounts of other noble gasses.
I also built a dessicator from aluminum cylinders filled with silica gel and plumbed this into to the system before the air reaches the separation membrane. These units are available commercially, and the one that I built is not particularly specialized. Separation membranes also exist for removing water, and this would be an improvement over silica gel dessicators, which require the gel to be dehydrated in an oven after it becomes saturated with water.
The liquid nitrogen generator has proved to be a reliable, but fairly slow method to produce small quantities of liquid nitrogen at home. The initial cool-down of the dewar takes about 12-18 hours, after which liquid nitrogen is produced at a net rate of 1 liter per day. The generator uses about 300 to 400 watts of electricity (includes the water chiller, which cycles on and off), so the energy cost for producing one liter of liquid nitrogen is about 8.5 KWh, or $1.10. This is substantially less expensive than having a thermos filled at a local welding supply store.
Here is a system overview.
This photo shows liquid nitrogen dewar (on the right) with the cryocooler mounted on top. The device on the left is a window air conditioner that was converted into a water chiller with liquid cooling lines running to the cryocooler. The small compressor on top of the air conditioner pulls air in from the atmosphere and sends it through the nitrogen separation equipment.
The cryocooler with custom heatsink. When the unit is running, the heatsink gets cold enough to condense nitrogen, and the newly formed liquid will drip off the heatsink.

This is the power supply for the cryocooler. The original control board from the STI Superfilter requires 27VDC, so I found a switching power supply from eBay and use that to power the control board.
This is a silica gel desiccator that I built from aluminum. It also contains coalescing filters and carbon filters to remove oil droplets and vapor from the compressed air stream.

This is the nitrogen storage tank with pressure sensor, valve and gauges.

Liquid nitrogen experiments:
Make Ice Cream
Mix a standard ice cream recipe in a large bowl.
4 cups half-and-half
½ cup heavy cream
¾ cup white sugar
2 teaspoons vanilla extract
pinch of salt
Add liquid nitrogen slowly while stirring the mixture. As the nitrogen boils, it will help froth the ice cream as it freezes the mixture very quickly. The rapid freezing produces small crystals and a fine texture in the ice cream.
Freeze a balloon
Inflate a standard latex balloon with air, then submerge in liquid nitrogen. The balloon will deflate dramatically as the internal gasses contract and even condense. After removing it from the nitrogen, it will reinflate as it warms. This process can be repeated many times.
Perform magnetic levitation on a superconductor
Certain high-temperature superconductors can be used at the boiling point of liquid nitrogen – 77 K. Once the material is cooled, it will exhibit “magnetic mirroring”, so that a permanent magnet can be levitated above the superconductor as its magnetic field is reflected. The best type of magnets for this are small (5mm dia or less, by 2mm long) neodymium-iron-boron magnets.
Make liquid oxygen
A variety of common gasses such as oxygen can be liquefied by passing them through a copper tube submerged in liquid nitrogen. Liquid oxygen can accelerate the combustion of common objects by creating a localized pure-oxygen environment.
Ping-pong ball spinner
Use a needle to puncture a ping-pong ball, then bend the needle to make the hole somewhat tangential to the ball. Repeat this on the other side of the ball with the hole “facing” the opposite direction as the first like a rotary garden sprinkler. Submerge the ball in liquid nitrogen for about 30 seconds, then remove it and place on a large flat surface. The ball will begin spinning as the captive nitrogen boils and streams out through the holes.
Effect on semiconductors
Connect various LEDs to a 9V battery with an appropriate current limiting resistor, eg 1Kohm. Submerge the LED in liquid nitrogen and note its color and brightness. As the semi-conductor cools, the band gap changes, causing a color shift. Some have also suggested the color shift comes from the spacing of the crystal lattice changing due to the very cold temperatures. Different LEDs will show varying degrees of color shift, so try a few from different manufacturers.

Old blog post:

You can generate liquid nitrogen (LN2) in the comfort of your own home with some parts found on eBay. I have proven that this is possible by purchasing surplus equipment and assembling it as described in this post. I spent over a year searching eBay, so these parts are not really easy to find, but the total bill for the whole system was under $500. The device consumes about 300 to 400 watts of electricity and needs no consumables (just atmospheric air). The LN2 is produced at a net rate of about 1 liter per day. This comes out to 9.6 kWh/liter or $1.15/liter, which is substantially cheaper than having the local welding store fill up a thermos (granted the thermos must be cooled as it is filled, thus requiring more than its capacity of LN2).

The most important part of this system is the cryocooler. This is a device that employs a thermodynamic gas cycle to pump heat through a very high temperature gradient. Many of these devices are self-contained and require only an electrical input to start pumping heat. The crycooler that I used was removed from a surplus RF filter which used the cryogenic temperatures to maintain a superconducting RF filter. http://www.suptech.com/home.htm
The crycooler itself has been fairly well documented:

I converted the cryocooler to be water-cooled on the hot end and attached a heatsink to its cold end. In operation, the cold end with the heatsink is inserted into the top of a large dewar. Eventually, the interior of the dewar gets so cold that the air will condense into a liquid and drip down to the bottom.

The second key part of this system is the nitrogen separation membrane. The is a device that accepts normal air, and produces relatively pure nitrogen. The waste products (mostly H2O, O2 and CO2) are vented into the air. Information regarding these membrane units is easy to find on the internet, but good luck buying one! They are nearly all produced for huge industrial installations, and those manufacturers will not even return phone calls from interested hobbyists. Asses! I spent a LONG time searching eBay, and eventually found a very compact unit, which was perfectly suited for this project. The nitrogen purity is dependent on the mass flow rate through the device. This means the flow must be carefully monitored and controlled. I will make another post that describes some fun stuff to do with LN2.


  1. Fascinating stuff Ben. Amazed you've managed to get all the bits to do this via Ebay...

  2. Thanks! Check out my latest stunt with cryogenic gasses:

  3. Ben,

    Do you have a bit more details on the parts and assemble of this system?

  4. David, I am planning to bring the LN2 generator to the Maker Faire this year in San Mateo, CA. In preparation, I may be writing up some more explanations of how the device works. Currently, all information that I have is in this blog post. Is there any other specific question that you had?

  5. Well to be honest, my son (who actually happens to be a physics major at UT Austin) wanted to build something like this so we are just trying to get some clarity on the parts and assembly. I haven't tried to dig much deeper then what you generally wrote about in the blog.

  6. This comment has been removed by the author.

  7. JoeMattie, I am using the original drive circuit which accepts 27VDC and sends a sine wave to the cryocooler. The drive circuit interfaces with a computer via RS232 and reports that the output wave is about 11VDC at 60Hz.

    The drive circuit uses the temperature sensors to calculate the appropriate power level. When the cryocooler is warm, the piston will smack the housing if run at full power. As the device cools, power is gradually increased to prevent the piston from hitting its limits.

    Where did you get your cryocoolers?

  8. Yes, I would also like to know where you got your cryocoolers?

  9. I bought the SupTech Cyrogenic RF filter on eBay:

    I think it cost about $300 (used).

  10. Ben,
    I am an engineering student at Saint Louis University. I'm working on a thermal vacuum chamber and have been unable to find compact/cheap/diy LN2 generators other than yours. I have about a million questions for you about your creation (I'm sorry about the multitude of questions). Do you have an email address I could contact you with? If so, you can email me at will at sutton hyphen family dot org.

  11. Ben, William,

    I to am interested in understanding this thing more. If you exchange questions via email I would love to be the fly on the wall and listen in to this thread.

    At UT Austin, my son apparently has access to virtually unlimited amounts of liquid nitrogen but we would like to make one of these at home to play with.


  12. William, you can email me if you want, but I would encourage you to post your questions here on the blog. That way, everyone can benefit from reading our discussion.

  13. Very cool project! (pun intended of course) I've been looking off and on at those cryocoolers on ebay for use in cooling an amplifier for ham radio use, but they always go for more than i'm willing to pay (not much). I was also concerned about the disassembly of those unites since they were originally configured to cool a filter made from some really toxic materials. Would love to see more detail or pictures on the cryocooler itself.


  14. Hey Ben,

    I really enjoyed your demo at Makers Faire. I went again on Sunday and couldn't find your booth. Did you end up not going the second day? Also, I brought some half and half hoping to make some ice cream with your nitrogen.

  15. Sam, I was at the Maker Faire both days -- same place in Expo Hall. Thanks for the comment. I hope you enjoyed the Faire.

  16. Tony, I paid about $300 for the cryogenic RF filter unit on eBay. This device was made by Superconductor Technologies Inc. and contained the Stirling cycle cryocooler. The cryocooler was attached to the dewar containting the filter elements with a conflat flange. It was very easy to separate. The most difficult part was removing the cryocooler's finned heat sink, and attaching a rubber manifold that allows liquid cooling to be used. Good luck!

  17. So Ben, now that you have been to Maker Faire, did you or are you going to put some more detail into your notes as you suggested you might.

  18. David, sorry I have not had the time to write a detailed explanation. Please ask away if you have any questions.

  19. Can I use this to get high?

  20. Anonymous, perhaps you are thinking of nitrous oxide (N20). That compound contains nitrogen, but doesn't have any practical relation to liquid nitrogen.

  21. Wonderful project Ben!! Its odd that I came across your project on TecheBlog today. I've been reading into cryocoolers in the last week. I had been reading into Stirling engines (for my own enjoyment) and read about their use in cooling. Which I had no idea about before. I'm not seeing a personal need for liquid nitrogen, but I would like to learn more about the stirling engine as a cryocooler. However my web searches come up with mostly stirling engine links with little reference to cryocooler use. Do you have any web references you can suggest? Books I could buy?

  22. would it be possible to instead of using a nitrogen membrane create your own filter of some kind, such as passing normal air through a activated carbon filter to remove most impurities, than passing the resultant mixture through a basic solution to remove carbon dioxide, than another solution to remove oxygen, and finally pass the air through a column filled with an anhydrous salt to remove the water, thus resulting, in nitrogen? the nitrogen may not be completely pure but it may work for most purposes.

  23. i have a unit in my metal shop for producing oxygen to feed a radiator welding torch.

    it uses a ceramic filter to seperate the oxygen ie small molecules from the nitrogen ie large molecules and discharges the waste nitrogen back into the air

    maybe another way to get pieces

  24. Greetings Ben,
    I cant find your email address anywhere :-( so here goes....

    Would this be a suitable pump to use:

    According to the docs, Helium is used inside this pump. Is this the conversion where you used water to cool it instead ?


  25. Years ago I heard that Nitrogen boils off of liquid air leaving mostly Oxygen so wouldn't fractional distillation of liquid air serve to produce liquid Nitrogen without using a membrane filter?

  26. What if you could recycle the nitrogen from an oxygen producing concentrator like an Airsep Newlife or a DeVilbiss 505 or 515 model? There are plenty of used oxygen concentrator available. Oxygen concentration is 20.9% in the air, 78% nitrogen, we have only 1.1% of other gases, my question is what if we have a little of other gases in the liquid nitrogen?

  27. Why not make liquid air, then boil off the oxygen, and use it for your welding needs?

  28. Miroslaw said/asked" Oxygen concentration is 20.9% in the air, 78% nitrogen, we have only 1.1% of other gases, my question is what if we have a little of other gases in the liquid nitrogen "?

    Lefty Noonan answered - The 1.1% of other gases is mostly Argon, which, like Nitrogen, is completely inert - so no problem if it is in the Nitrogen.

  29. David Stanford said...

    "Why not make liquid air, then boil off the oxygen, and use it for your welding needs"?

    Lefty Noonan replied;

    To weld, the oxygen must be 99.999 pure.

    The oxygen boiled off in this case would only be approx 95% pure

  30. Look up "pulsed tube cryocooler", especially in academic literature. They are a type of cryocooler that has no moving parts, and could probably be built by someone with a bit of manufacturing know-how, for under 150 bucks, if one were so inclined.

  31. Any chance of an additional photo of the entire rig with the parts labeled like they are in the LN2 generator diagram. (e.g. I love the cooler box, but I can't workout to which part it corresponds. I suspect it is the plastic tub of coolant, but I could be wrong.)

  32. Looking at this, I think the odds of my finding a cryo-cooler are pretty small, but this is something I'd really be interested in making. I'd personally probably go for a more expensive option of buying in bottled nitrogen, which I could get hold of quite easily, to avoid a membrane hunt, but I suspect the only way I'd see a cryocooler is assembling one myself. That'd change the focus of such a project considerably. I suspect it'd be more a matter of building stirling heat pumps that got to lower temperatures as you got the hang of building them.

  33. @ Alex Wojtak

    Look up something called "Pulsed Tube Cryocooler". It's a Stirling cryocooler that has no moving parts (uses pressure waves instead), so it would be the easiest to build.

  34. I am looking for a start - what was the maker and part number for the nitrogen seperator and was it a PSA or membrane unit?

  35. Oh yeah and can we see either part or all of the N2 seperator in the pictures?

  36. Autonomy, check out my other blog post regarding the N2 generator:

    The actual separate membrane is the white in-line device at the right of the picture. It consists of a bundle of 1mm dia flexible tubes. The air is pumped into the bundle, and the flow is restricted at the outlet. As the air slowly passes through the tubes, the oxygen penetrates the tube wall, leaving nitrogen-enriched gas at the outlet.

    I bought the membrane off eBay. There was no manufacturer or model number info. As you may have found, nearly all nitrogen separation membranes are made for large industrial processes. It took a long time for me to find this one.

  37. I got my cryocoolers off of ebay.

    On the subject of pulse-tube coolers: I remember reading a paper a few months ago by a guy who built a working prototype out of a reciprocating jig saw and some tubing...

    And Ben, you're not kidding about it being a pain in the ass to remove the aluminum heat sink; I converted one of mine over to water cooling (using a coil of copper pipe in my case) and probably cut my hands upwards of 30 times in the process, heh.

  38. Just got my own cryocooler- Heatsink removes EASILY with GoofOff :D Given a good soak, they fell off on their own after 5 minutes!

  39. Anonymous, you must have a different model cryocooler. The cryocooler in the STI unit had its copper heatsink brazed or soldered to the stainless case of the cooler.

  40. Ah, I had assumed it was Twinbird, for no apparent reason. You should post your Youtube video on your site!!

  41. Dear Ben,
    When I contacted Superconductor Technologies Inc. regarding their Sapphire Cryocooler, I was surprised to see the cost of their cryocooler. It was 10000$ I am very eager to know how did you make your liquid nitrogen generator at less that 500$, it certainly is a remarkable achievement.

  42. niceboy, I bought everything on eBay. Check the first paragraph of the blog post.

  43. OK, I couldn't resist. As they say in the superconductor biz; "Resistance is useless". Or was that "futile"? I bought my STI Superfilter on eBay and just took the cryocooler out of it. Wow, that was fun! Ben, I take my hat off to you! You managed to figure out which connectors were required for the cryocooler only and which were for the filter control. Also how to connect the RS-232?!? My board is different from yours though. It looks like yours is from 2006 and mine is from 2001. I assume there is a temperature sensor in the filter can that is needed for the cooler circuit? It looks like there may be one sensor on the heat-reject side, and some kind of sensor on the modified KF vacuum fitting.

    I should probably point out that if anyone Dremels apart the filter can don't touch or lick (eww) the exposed superconductor. The older filters are made of TBCCO where the T stands for thallium which is bad (goes thru the skin too). The newer ones like Ben's are probably made of YBCO which is OK but I still wouldn't lick it...

  44. Anonymous, my filter had an RS232 connector on the back. I connected my computer and realized that it required custom software. I called Superconducting Technologies and nicely asked if I could download the RS232 control program via their FTP server. It's a Java program. I then decompiled it, and found the maintenance password in the code. With the password, the program can be used to change the cooler control variables. It's pretty interesting.

    Also, the temperature sensors are used to calculate the appropriate cooler power. I physically moved the sensors from the original dewar to my custom heatsink and they worked for a couple years, then broke. I replaced them with just a plain resistor to trick the circuit into working, but I no longer have a cold-side temperature measurement.

  45. Hi Ben
    Ah, so you opened the can. I'm going to try that also as soon as I buy some more cut-off wheels...

    I'd ask you for your code but I'm sure it wouldn't work with my older unit and, I'm sorry to say, coding hasn't been in my skill set for too long. I'll figure something out.

    How often did you run the cooler in the last couple of years? The literature says these things are bullet proof. I think I'm going to leave the air heat exchanger on the heat-reject side and figure out some ducting for the fan. Unfortunately that will probably obscure the Stirling view...

    Also I think I'm just going to liquify air. I'm a lot lazier than you are, although, I suppose I'll have to dry the air first. Do you notice N2 condensation all the way down your condensor if you turn it sideways in a clear dewar?

    Again I want to state how impressive your project is. It's utterly amazing what you can do given a source of cheap doodads. eBay is wonderful!! This would make a great product if the components weren't so bloody expensive new.

  46. Hey Ben, I was just thinking, if you could set the temperature of the cooler to a little above the boiling point of N2, say 78K, then you should be able to condense just O2 from air for quick-start charcoal grilling .

    Also, if anyone is interested, the maximum amount of LN2 the STI cooler can generate given a perfectly insulated dewar is ~0.16 liters/hr. This is assuming 7W lift at 77K, 55080 J/mol, 28 g/mol, 807 g/l, and that I didn't goof up the calculation...

  47. Anonymous, I can send you the STI RS232 program if you want. email me at ben (at) magconcept (dot) com. I didn't need to do any coding myself. I've ramped my cooler up to 150W input power, which is the maximum the circuit hardware will allow. It's still running fine after a cumulative run time of about a month. You will definitely need to dry the input air. There is so much water in the air, your heatsink will become a huge icicle and the ice is actually a good insulator -- cutting your efficiency. Good luck!

  48. If you aren't worried about purity in the nitrogen why not just use an $80 home depot dryer
    and boil off the oxygen. This doesn't really help you but it probably will help decrease time and cost for someone else.

  49. To Anonymous 9/13: Thanks for the link!

    I finally got back around to playing with my old eBay Superfilter after I got my el cheapo eBay 24-27VDC power supply. The manual said I could communicate with Hyper Terminal, so I kluged a serial cable (it ain't pretty) and dusted off an old Windows ME laptop that I knew I was keeping for some reason... First time I fired it up it made an awful cyclic sound, but then it calmed down and went to 79K. Just turned it on again this morning and no bad sounds as it's cooling. I guess it had to train itself the hard way not to overdrive the Stirling!?! Anyway, not too shabby for a 9 year old system. The picture on the front of the manual has the cell phones of the day (~2000). Seems ancient...

  50. Anon, that's great! Yes, the piston inside the cooler will probably hit its internal limits if driven at full power while the cold end is not very cold. The stock driver circuitry monitors the cold end temperature and increases power as the temperature drops. Perhaps one of the temperature sensors is broken, or maybe the firmware has already been hacked (like I did) to increase total cooler power.

  51. Thanks for the tip on the pulse tube cryocooler! Must build one and impress friends with instant ice cream... -Dig your blog...sorry to not register to leave a comment but I've been enjoying reading about your experiments :) Cheers, Your Everyday Mad Scientist

  52. Another source for the Nitrogen Membrane is aircraft. My guess is this is for safety: as fuel is depleted they replace the void with Nitrogen from the air. Not all aircraft do this.

  53. I finally received my Superfilter in the mail yesterday! I fired it up off 27V and it seems to make a nice low frequency hum, and appears to be working, although there isn't much to indicate either way...

    I received just the unit, which is an 850B2RV20, an 850 MHz filter system. I opened it up and took a bunch of pictures which are on my flickr page here: http://www.flickr.com/photos/kc6qhp/sets/72157625382973371/

    I also inquired about the software with Suptech, but they replied that they are aware that people are getting these used, and that the cannot "sanction" this type of use. Clearly they want no liability with these things in the hands of anyone but their customers. While the likelyhood of someone suing Suptech because they hurt themselves with something they made is close to zero I'm sure they want to take no chances, epecially when it would not help their bottom line.

    So no software support, which means I'l have to figure ut other ways to determine if this thing is really cooling.

    Anon on 9/13 how did you determine that it was cooling through the serial port? I see nothing on the serial port coming out of my unit.

  54. Just a helpful update to other people attempting this. If you don't have working drive electronics, you can use a big variable autotransformer to ramp up the temperature by hand. Just go slowly and back it off if you start to hear it ping.

    I've had limited success with oxygen removal by "skimming" it from the surface of LN2 with a good rare-earth magnet (oxygen is slightly paramagnetic).

    Oh yeah I found the paper for the jigsaw-pulse tube cooler hack:


    I'll get around to trying this some day I'm sure :P

  55. Oh yes one more thing: you can simply use a good silicon rectifier diode as a temperature sensor. They have a fairly predictable temperature->voltage drop relationship. Just add a voltage source and a multimeter.

    At this point I've lost probably 15 of them into the bowels of my dewar. Solder doesn't hold up to well at that temperature it would seem.

    Good thing they're cheap :P

  56. Any ideas about efficiency of Joule-Thompson process? If I let 100bar nitrogen from cylinder at 90K (precooled) to the orifice and let it expand, how much LN2 do I get for 1000 normal-liters of nitrogen? What limits this? I'm thinking of once-through Linde-process - dump non-liquified nitrogen after heat-exchanger.. Thanks!

  57. Anonymous, check this out:

    I found a PDF on the site that says one of their JT refrigerators provides 250mW of cooling at 85K when starting with room temperature nitrogen. if you were able to pre-cool the cylinder, you should get substantially better cooling. The calculation is not trivial since the JT coefficient is heavily dependent on temperature.

  58. why hasn't anyone made a liquid nitrogen air conditioner?

  59. Anonymous, it is much easier to cool 50g of water by 1*C than 1g of water by 50*C, even though the amount of energy pumped is exactly the same. This is because insulation and high temperature differentials pose technical problems that make real-world cryocooling difficult. Air conditioning a room in a house or office requires lowering air flow temperature by only 12*C. Using liquid nitrogen at -196*C would not be an efficient use of energy because the energy used to produce -196*C nitrogen could have been used to simply cool air by 12*C directly via phase-change refrigerants ("freon") with much greater effect. Liquid nitrogen is only used where very low temperatures are required.

  60. First thank you all for the cool ideas and suggestions (pun intended).
    I bought a super filter on ebay and just got it up and running.
    Have not taken it apart yet, except took the cover of to look.
    I got mine from here:
    All works OK, used 2 old PC supplies with the 12 V in series.
    Make sure you remove the ground on at least one...

    Serial port seems to accept 19200 Bd, 8 data bits, no parity.
    Beware of hardware handshake, you do not need it, but if those pins are connected,
    the unit seems to toggle those lines, and may cause your communication program to get confused.

    0:00:31:03> t
    System Temperature Report:
    #1 #2 #3 #4 #5 #6
    BPF: --- --- --- --- --- ---
    BRF: --- --- --- --- --- ---
    Cold Stage: 79.76K 79.92K
    Cold Finger: 78.88K
    WR Cold Finger: 81K <--------- this is the actual main sensor

    Cooler Temp: 23.4C
    Ambient Temp: +23C

    0:00:31:05> l
    System Control-Loop Status Report:
    #1 #2 #3 #4 #5 #6
    BPF: --- --- --- --- --- ---
    BRF: --- --- --- --- --- ---
    System Error: +.06V <-------- this is the control loop voltage that drives the cryocooler
    it can go from +5 to -5 V I think.

    Cooler V (RMS): 63V <-------- system is in balance, this can go up to 117 V AC during startup.
    this voltage comes from a PWM amplifier and looks like 60 Hz.
    Supply Voltage: 21.0V <-------- 2 x 12 V PC suplly in series, some cable drops, fused with 10A.
    Back Off Count: 0
    DAC Back Off: 309
    (Int)LNA I(V): 3.66V
    (Ext)LNA I(mA): 688

    0:00:44:33> state
    System State Report:
    Cooler Controller/Driver: Normal
    Cooler Driver Status: Normal
    Coldstage TempSensor: 1
    Low-Noise Amplifier: On
    RF Bypass: Normal
    System Summary Alarm: Normal

    System SM State: 6:Baseline (5:Settle)

    It took more then an hour or so at too low supply voltage to get to 80 K.
    Everything in this unit seems 100% in order, I am going top give the seller a big + on ebay.
    I also ordered a better power supply just for this.

    An other remark that MAY be important.
    IF you have neighbours and live in a flat, especially with wooden floors...
    They will come asking WTF you are running a diesel engine for in your house :-)
    Think of this as a 100W 60 Hz mechanical vibrator.
    This makes a LOT of noise during startup.
    Now that it is stabilised at 80 K, the fan makes more noise than the crycooler.
    Maybe it will be more quiet during startup if I run it from the correct 27V DC.

  61. I had the idea to make liquefied air by putting refrigerator or air conditioning compressors in series with their copper cooling radiators in between. On the high pressure end I would cool the last heat exchanger and have the out put filling the Dewar containing the last heat exchanger. I'm having difficulty figuring out if the pressure would become too great for my compressors? If not would the amount of cooling be enough that I would have a net gain in liquefied gasses? The most major drawback that I can see other than those described above is I would need to cool the output initially which would mean I couldn't turn the system off for very long unless i had a high pressure storage tank.

  62. Nathan, search youtube for tesla500. This user has begun making a liquid air generator using a method similar to what you describe. The key point is to circulate the cooled, compressed air back into the front of the system. Thus, you can extract more heat from the some parcel of air that continues to circulate. Simply having two compressors in serial will probably not extract enough heat to liquify air. Good luck

  63. Thanks I'll surly do that! I found this website which proves useful for my project. http://www.gizmology.net/liquid_air.htm

  64. I had an alternative idea of using vacuum refrigeration to freeze nitrogen so that I could condense air on the outside of my vacuum chamber. If I insulate the chamber in a Dewar or at least in Styrofoam I should get a net gain in liquefied gasses provided I cycle it back in the manor described above. Do you have any ideas on how I would make this work? I would also like to know how much better a Dewar is at insulating than a Styrofoam cooler.

  65. Nathan, yes the gizmology article sums up the process nicely. Of course, everything is easy in concept. The difficulty is making the design practical. In this project, the heat exchanger and maximum system pressure will be the major performance limiters.

    You are probably referring to "evaporative cooling." Even though the phrase sounds similar to "vacuum cooling", there is an important difference. Simply sucking all of the air out of a chamber will not lower the chamber's temperature. If the chamber has a liquid or solid in it, and the pressure falls below that material's boiling point, the material will boil and cool itself down until its boiling pressure matches the chamber pressure. You can make solid nitrogen from liquid nitrogen this way (same logic applied to the solid/liquid phase change).

    In order to build a normal heat pump cooling system, you need to compress a working fluid (eg helium), cool it, then move it into the insulated chamber, expand it, then move it back outside the chamber and compress it again. This completes a heat engine cycle and is how the Stirling Cryocooler works. Again, the problem is making the design practical. The working fluid must have its internal heat transferred to a temporary holding device while is is being moved into and out of the chamber. This device is called a regenerator and serves the same purpose as the heat exchangers shown in the gizmology article.

    In general, a dewar is the best-insulating container you can have. Styrofoam is good, but its insulation ability will depend on its thickness. Even a very thick styrofoam cup will not insulate as well as a normal coffee thermos. I'd estimate the difference at 1-2 orders of magnitude.

  66. Ah so that's why My google searches weren't turning up anything on vacuum cooling.
    With either system I'll likely be using several air conditioning compressors in series to do the mechanical work. The part tesla500 was missing in his system was an air dryer. Is there a cheep way to build one of these without using chemicals that will rust or otherwise damage my compressor?

  67. Nathan, I'd recommend silica gel for your air dryer, which is what I used in my LN2 generator. You can by a ready-made silica gel air dryer from Harbor Freight, or buy the raw silica gel pellets on eBay and build your own container.

  68. Come to think of it air conditioner units have air dryers built in. If those air dryers work properly for this purpose then I could build my whole system from used air conditioning parts. In that case I might be able to push the diy LN2/LO2
    price below $300.

  69. To measure these low temperatures, I build some simple stuff, and wrote some software, you can find out more here:
    This shows that measuring and logging very low temperatures does not have to cost much,
    provided you can use a soldering iron and run Linux.
    I made my own type T thermocouples for just a few Euro, those, together
    with the described simple hardware and not so simple software, work
    well from 70K to just below the melting point of 60/40 solder, about 190°C .

    And I want to thank Ben for the very inspiring projects!

  70. So the cooler had to be programmed first to get it to work properly? Is that because it needs to ramp up slowly as to not damage itself? It's not just something that can be plugged in to a power supply at full power, correct?

    I've been fascinated with LN2 for awhile now, but being able to make it yourself is just awesome.

  71. Anonymous, the cooler will work with the stock firmware, but I tweaked it to get more power. The firmware initially set the power limit at 130W (I think), but the hardware can handle more power, so I pushed it to 150W. The firmware takes care of the ramp-up to avoid damaging the cryocooler. Unfortunately, my temperature sensor broke, so the ramp-up routine ceased to work without knowing what the cold-tip temperature is. I now perform the ramp-up manually and don't measure the cold-tip temperature.

  72. Do you know what is the most efficient way to activate a superconductor
    and keep it activated do I need to produce liquid nitrogen on-site if so do you have any detailed instructions on how to make a liquid nitrogen generator.

  73. DefineEvil, many metals become superconducting at sufficiently low temperature (eg liquid helium temperatures), but special materials will superconduct at liquid nitrogen temperatures. These materials will not superconduct if the LN2 boils away, and the temperature raises above the nitrogen boiling point, so they must be kept cold.

    I am writing a book about my various shop projects, and will include the DIY LN2 generator. The book will be published next year by the No Starch Press.

  74. I'm interested in using LN2 to flash freeze some local fruits. Dipping them into the Dewar isn't practical. I want to do several pounds. I thought of pouring LN2 into an aluminum pot, then using a fryer basket to lower the fruit into the LN2. This creates a reaction somewhat like putting water into hot oil. It spits and hisses. Doesn't seem very safe. What would you recommend as the best way to take LN2 from a Dewar and apply it to the fruit to freeze it?

  75. Ted, it sounds like you are on the right track with using a dipping basket made from wire mesh. Maybe all you will need is a cover/lid for the pot with a slot to allow the basket handle through. I've seen fryers that are built this way. Maybe you can just buy a cheap fryer pot, and use it as-is.

  76. Thanks Ben. Thinking about it more, I think liquid ethanol chilled to just above freezing, might be even more effective. There would be no hissing or spitting.

    Any thoughts on the best way to chill 2 gallons of ethanol down to -60C? One friend said I should just take a whole bunch of pelletier elements and stack them.

  77. Peltier elements are good for controlling the temperature of tiny objects. You will need a huge number of them with a monster power supply to move appreciable amounts of heat.

    Maybe you could use dry ice to chill the alcohol bath. It's only $1 to $2 per pound at retail.

  78. Thank you for taking the time to explain this and even provide a Q&A.

    Having also seen some of your other machines - electron microscope etc - I am forcing myself to stick to one project question at a time.

    I am actively hunting a superfilter at the moment but wanted to spend the time waiting on preparing some of the other items. I have a query though about the stirling cooler which you might be able to help me follow.

    From what I understand, they usually are air cooled and you switched to water cooling for a better cooling effect. I can see the logic there.

    As an option to that, with various old compressors and two "portable" air conditioners going spare, I was wondering about using one of these to cool down a container of something like transformer oil and then, via a copper coil heat exchanger in the oil, circulate methanol or standard antifreeze through the system to cool the "hot end".

    I am fairly sure it would work ok but I was wondering if the likely result would be worth the effort.

    My logic was that if the "hot side" can be kept perhaps 20-30C cooler than by using water alone, it MIGHT mean that the condensation rate on the cold side was increased even if left at the default of 130 watts or so.

    Would greatly appreciate your thoughts before possibly embarking on a fools errand. You know VASTLY more about the cryo cooler than me and if I am making an obvious oversight, would really like to learn the easy way for once :)

    Kind regards - and thanks again for providing such feedback and sharing you fab gadgets and serious boys toys.


  79. Mark, take a look at pg.76 in the Google Books link in the blog post. The middle graph shows the cryocooler lift in watts vs the reject temperature. At 125W of input power, the cooler will lift 5.5W from 77K at a 75*C reject (heatsink) temp, but 9.5W at a 2*C reject temp. So, yes, reducing the heatsink temperature will allow the cooler to move more heat for the same input power. Lowering the reject temp by 20-30*C should buy an extra watt or two, which is 10% -- not bad.

    You could also chill the nitrogen gas that is being added to the vessel for liquefaction. This will reduce the total amount of heat that the cryocooler has to move.

    Good luck. Keep us posted on your progress.

    1. Hi Ben,
      There must be a temperature limit beyond which the Stirling Coolers hot end should not be lowered beyond. Any idea what that might be? Would circulating a -50*C coolant around the hot end be problematic? Alternatively, would it be simpler to just pre cool the nitrogen gas (to -50*C for example) to improve liquification?

      I've thought about using J-T cooling with a vacuum insulated counter current heat exchange in combination with the Stirling cooler to reduce the necessary high pressures required for a G-M liquifier. At a typical temperature drop of 0.25*C / bar drop in pressure, the starting gas temperature will need to be much less than ambient.


  80. Thank you for the reply Ben, very much appreciated and I shall go and read the link you suggest.

    I hadn't thought about cooling the gas - that should be relatively simple and as the AC unit should have some reserve from keeping the reject cool, it can also cool the nitrogen feed too. Good thinking - the nitrogen cooling part hadn't entered my mind :)

    I will let you know how I get on - I suspect the waiting game will be finding a superfilter in the UK or a US fellow happy to ship - I shall do more Ebay hunting.

    Have a great weekend.

  81. Hello Ben...

    This is amazing. What would I need to do to speed up the LN2 production instead of 1 litre/day to about 10 litres/day?

  82. Kirk, you would need to redesign the whole system using a different approach.

  83. Hi Ben;

    I want one! 8^)

    The most common practical use for LN2 is tho cryo treat heat treated metals. For instance knife makers do their own heat treat on some advanced metals and buy the LN2. One guy reports paying just over $1 a litre.

    Would it be possible to simply "auto fill" the dewar with some form of level detector turning the unit on and off as needed? Like a toilet bowl valve for LN2 making! [wink]

    Then the knife maker would not have to go buy the LN2 and treat the blades in relatively batches, instead treating 1 or 5 blades as needed on a daily basis, his dewar is always full!

  84. Shotgunner, the auto-fill part of the project would be relatively easy. The difficult part is removing the water and oxygen from the gaseous nitrogen stream being fed to the dewar.

  85. Thanks Ben.

    And the price of those circuit boards is sky high now! Everybody wants liquid nitrogen!

    Isn't the problem of scrubbing O2 and H2O already solved in your design?

  86. Shotgunner, I used a nitrogen purification membrane that removes a lot of the water and oxygen, but there are traces of both still in the stream because the membrane is partially broken and old. Finding a cheap small-scale nitrogen separation system is pretty difficult.

  87. Nathan:I posted last year asking about how to build a Hampson Linde cycle. I managed to obtain an ice maker from a scrap yard. The compressor and other vital components including an air dryer all work well. My problem is that I can only get a few degrees of cooling with an open loop at atmospheric pressure. So is it better idea to add more compressors to get a higher pressure differential,close the loop and re-pressurize the system,or both?

  88. Ben, is there anyway you could post a list of everything you purchased for this experiment, Im a student a Worcester State University and I would really like to make one of these. Also if one would want to make a mix of different gasses (in other words not pure N2) would I just remove the filtration setup? Also it would be great if we could email back and forth for this. I have a lot of questions about it. if you do want to email, my email is : gdervishaj(at)worcester(dot)edu

  89. Hello Ben!

    Is it possible to make liquid hydrogen (or O2) with the same setup, assuming that you have a pure hydrogen (or O2) source?

  90. Wyatt, oxygen boils at a slightly higher temperature than nitrogen, so it easily liquified by this setup. Hydrogen boils at a much lower temperature, and I don't think this particular cryocooler would be able to get cold enough.

  91. Any suggestions on a good cyrocooler or other cooling methods that may have a chance for liquid H2?

    I'm a inventor with a moderate understanding of physics and chemistry.
    I am trying to use liquid H2 and O2 as a storage medium in a brown gas (2H2 + O2 > 2H2O) combustion system. I could use simple compressed hydrogen and oxygen, but I believe that the liquid forms have more "storage potential". opinions welcomed.

  92. Hi Ben,
    There are Superconductor Technologies Superfilters on offer at eBay but they don't specifically say that they are cryogenic. Can you comment on this?

    Also, the price of these cryocooler-based systems are being offered for ~$2500/unit. Would be nice to see something in the $300 range but haven't for a long while.


  93. nitrous, as far as I know, all Superconductor Technologies Superfilters make use of cryogenically-cooled superconducting filters.

    Since I published my article, the market for used STI Superfilters has probably gotten a lot more competitive. $2500 seems exorbitant, though.

    Good luck!

  94. Hi Ben,
    Thanks for the reply.
    The unit number is 850A2RV20 but is a "Superfilter" (by description). They certainly are north of $500 for this one and the others, labeled as having a cryocooler are over $1500.

    Give us a shout when you're about to list yours! :)

  95. Could you write a procedure,which contains all the necessary parts and instructions to build the liquid nitrogen generator. I would love to build this generator if I could recive more simpler directions.

  96. Anonymous, I'm writing a book for the No Starch Press which will detail the LN2 generator in addition to many of my other workshop projects. The book should be out later this year.

  97. I will be wating for the book to read it.

  98. Hi Ben,
    I received a note back from Superconductor Technologies regarding Stirling Cryocoolers in their products. They advised that "all our Superfilter technologies include cryocoolers".

    That said, I wonder if you could give some specific insight into how the power supply drives the cooler. There are a couple units that are in the sub-$1000 range but a couple are listed as "as is". If the driver board is bad, will it be possible to drive the Stirling cooler with a simple DC power supply?

    Thanks Ben,

  99. nitrous, the cryocooler itself is just a linear motor, which requires a sinewave to drive it. The driver board has an inverter which approximates a sinewave at about 60Hz, 15V at 10A. The driver board also includes a current transformer to monitor the current going into the cryocooler, and also varies the frequency slightly (ie from 58Hz to 62Hz) -- presumably to achieve best energy transfer into the piston. I'll bet you could drive the cryocooler with just a large variac. You could use a low-cost watt meter to monitor power going into the variac at 120VAC, and then make the assumption that the variac is 90% efficient. Good luck!

  100. Thanks Ben,
    I'm not too worried about efficiency!
    I might just have to put a bid in on one of these Superfilters.
    Take care,

  101. Ben,

    Running the SF and it good to go. Just wondering why you chose to skip the air cooling for a water jacket. It seems to me that it was designed to cool with air cooling to 78K. Could you elaborate? Does the vacuum chamber just unbolt via the flange or does it need more work than that?

  102. Ok got the cooler off the dewar, real easy just unbolt the flange and pull. One question, where is the cold side temp probe? In the dewar? Looks like a bugger to open the dewar.

  103. Mike B, I used water cooling to get the heat-reject area of the cryocooler colder than I could with just air cooling. A lower temperature at the heat reject area will allow the cryocooler to pump more heat for a given power input.

    The cold-side temp probes (there are three) are located in the dewar, which is welded shut. I drilled a tiny hole to let air in, then cut the dewar apart with an abrasive wheel. Good luck!

  104. Hi Ben,

    I am very interested in your design as both a chemistry major and a DIY-er/tinkerer in such things.

    Only problem is, as you noted above, the market head for the cryocoolers has gone up, making someone working on a college kid (<$500 preferable) budget near impossible. Do you know of any other leads I could get on to find one for a *decent* price?

  105. Eli, sorry, I don't know of a source other than eBay for the cryocoolers. Good luck. -Ben

  106. How did you find the nitrogen separation membrane? What terms did you use? How could I find one?

  107. Also, what is the name of your book?

  108. Clark, they are difficult to find. Try searching for Nitrogen PSA, nitrogen membrane, nitrogen generator, nitrogen source, etc.

    My book project may be on ice for a while. It had the tentative title, "High Tech Hacks"

  109. Thanks so much! You have some fantastic stuff. I am building my own electron gun right now. I built one before but melted the filament in one blow. Can't wait to see what you do next.


  110. Ben,

    Got the Dewar open, I managed to pick the right end (tip end) to cutu open with a 4" grinder with a cutting wheel. cut just under the welds. Wow what a nice clean room fly to space kinda stuff. How much gold you think is in there?! :) SO I found 5 temp sensors. 1 on each filter module, and two on the copper spindle. But funny enough the board is only using two as the rest have their wires cutoff on the connector outside. Look like they build the canister with all three filter modules but configure the board based on how many the chassis is being sold as. The wires are there on the connector but are cutoff really short. What a bitch to un epoxy those thermal sensors. Managed to save three intact. Used a weller rework air unit and indirect heat and that finally got the epoxy soft enough. I really love the copper spindle but for the life of me I cannot figure out how its attached. I wanted to take it off. Epoxied? Heat fit? I dont know.

  111. Ben, I took your advice and yes the cryocooler head has 4 holes, and wow to my surprise they are already threaded. Boy are they small, really small, like not readily available small, but what's that pile of screws on the table from completely taking apart the dewar and its filters....hmm let me try them... Wow the screw fits! Again reinforcing the mantra to NEVER throw anything out!

  112. Ben,

    Any reason you chose to use a copper bar but aluminum heatsinks on the cold head? Any reason you did not use copper? how did you attach the aluminum to the copper bar? Epoxy or screws and heatsink grease?

  113. Mike, I just used the parts that I had available. Choice of metal was pretty arbitrary. I used screws and grease to attach the heatsink to the bar.

  114. Ben,

    Thanks for the info. Ok so I got it running gutted, it looks like it only uses one temp sensor in the dewar to run the cryocooler. And its required as the board will shutdown without it. I will submit the pinout details later when I get around to detailing it.

    Ben, I noticed that the power only starts running up the closer the dewar temp sensor is to operating temp, but I never saw it go over 115.7W and that is even after setting to 140W in config. Kinda cool watching the head smoke while operating.

    So how do you get the head to run full out all the time? Any thoughts on what is needed?

  115. Mike, you need to edit the power coefficients which are used by the device to calculate the appropriate power based on cold-side temperature. They are located somewhere in the maintenance software. However, the cryocooler will get too much power when the cold-side is warm if the coefficients are edited. You'll need to use the power limit to prevent too much power from being delivered while the unit cools down from room temperature. Read through the comments, I believe there may be more info up there.

  116. Ben, out of the dewar while running, I placed a temp sensor with thermal grease on the head and a fluke temp probe as well. Without understanding the way the program works in the SF. I could not reach a temp of below 77K that would liquefy Nitrogen, got to about 81K funny enough I thought open air would be the issue so I made a 1inch styrofoam shroud to cover the head and it actually was worse. like 84K to 86K. Can any help explain? This is still with the fan shroud and air cooling running.

  117. I think the white heastink thermal grease may be insulating at such cold temps? Any ideas anyone?

  118. My name andy from colombia anchasom@gmail.com could you send for me one equipment of nitrogen? I would apprecciatte, Thanks

  119. Ben,

    OK been monkeying with the unit for many hours and still cannot get the head to reach the 76.5K target temp. I even got a small 6oz dewar and used foam to seal it onto the head. I bolted the temp sensor right to the head. The system is showing 78.345K.

    I even monkeyed with the power coefficients and noticed one thing, despite setting the upper limit from 120w to 140w and futzing with the coefficient setting I got it to demand 140W but actual never went over 120W. I did set the max limit to 145W Going to the status page shows the PWM drive at 100% and cooler power error of 20w. So it looks like no mattter what you do the boards drive will only deliver 120W to the sterling.

    Back to the cooling issue, wondering whether the thermal sensors I removed from the dewar are contaminated / compromised. I tried 2 different ones and both are the same.

    I am 100% that the system was working and reached temp before taking the whole thing apart. I thought I had it cooling low running the head in free air once, wonder whether the head is not working properly.

    Any thoughts / feedback is much appreciated.

  120. Mike B, if the PWM drive is %100 and the power error is 20W, check the crycooler driver input voltage. It must be 27 or 28V at the terminals. If you are using a bench supply, there could be substantial voltage drop between the supply and the input terminals on the cryocooler.

  121. Ben,
    Checked the voltage drop and the largest one was their fuse, fixed that and still no change.

    I did manage to make liquid nitrogen! FINALLY. A few things to note. My tem sensor never would go below 78.3K with normal air. My dewar seal is kind of crude foam and most likely leaky so I started to blow nitrogen around the seal area where it had ice and immediately I noticed the thermal temp go down quickly almost 1 deg K interesting. I guess it take dry nitrogen to make the temp go lower and that the water vapor in normal air prevents the dewar from getting colder? I just ended up bathing the outside with nitrogen and in the morning there was a bit of liquid nitrogen in the dewar. Now on to further the plans and get a proper flange etc... bigger dewar, bigger heatsink......

  122. Great! Yes, CO2 and water vapor might be presenting too much of a thermal load on the cryocooler.

  123. Hi, thanks for sharing nice information with us.

  124. Ben,
    I would advise your readers to be sure to remove the oxygen from the air prior to condensing the gas to liquid as liquid oxygen is very dangerous and can cause an explosion of great magnitude if mixed with or allowed to permeate conbustible materials. With pure oxygen combustible material is just about everything.

  125. Ben, I have been following your projects for quite awhile and was most impressed with your LN2 generation system. I lucked out and found a local service guy that had 4 fully refurbished superfilters. They literally look brand new without even as much a a scratch. He included the setup/testing manuals as well. Not much in them as far as diagnostics. Is the java based control program really just a GUI for serial control?

    I want to interface either an arduino or raspberry pi to provide a control system via serial. Don't have the cryocooler fired up yet. I'll report back when I get my 27v PSU in a day or so.


  126. Unknown, that's great! Yes, the Java program appears to just be a serial communication front-end. It's possible that the cryocooler provides console access via a serial port for sites that do not have the Java program installed, but I never used it that way. There are a lot of settings and data that can be viewed in the GUI, so the serial protocol could be fairly involved. Let me know what you find. -Ben

  127. (I'm Unknown in the last post above)

    When I first got the unit, I set it up and followed the cool-down process in the operation manual. The system uses a state machine to control the process and has eight working states and one fault state:

    0) Initialize - Initial power up state
    1) Idle - Cold stage is warm, dewar not cooling
    2) Coarse Cool-Down - Cooling, cold stage is above 85 K
    3) Fine Cool-Down - Cooling, cold stage is below 85 K
    4) Over shoot - Cold stage is cooler than set point (nominally 78 K) and the control algorithm integrator is settling
    5) Settle - Cold stage temperature is settling, circuits are switched to normal
    6) Baseline - Cold stage temperature has settled and baseline data is being collected
    7) Operating - System is operating normally and being checked for deviations from the baseline

    8) Fault - Alarm relay is switched on

    The system took about 2 hours to go through stages 0-6 and displays the current state of the unit with blink codes. At stage 7 the bypass light goes out and the ready light goes green. At this last state the unit is supposed to be at 78 K. Haven't got around to hooking up the serial port so can't query the control board and see exactly what it is doing.

    After this initial test for functionality I decided it was time to open up the unit and start converting it to liquify air/nitrogen. I removed the dewar from the cold side. The dewar emitted a constant stream of water vapor from the cold port for at least 4 hours. Shows how good the dewar is and just how cold 78 K is!

    I tried measuring the voltage driving the cryocooler while it was operating in bypass/startup mode from the control board. It looked like 105-116 VAC which blew my mind since the whole unit runs off of 27 VDC. It cools very well but doesn't get under -100 F in bypass mode. I have not opened the dewar to try and harvest the temperature sensor. Without the dewar connected and in place the unit doesn't cool down beyond this. I really don't want to destroy the dewar to just get a thermistor/thermocouple or whatever is in there for temperature feedback.

    Another guy has just run his off a Variac at 16-20 VAC, a video of his unit is here http://www.youtube.com/watch?v=4_bGkztd7t0 I tried doing that but did not have as much success as him. It gets very cold but not as cold as it seemed from the control board. Not sure what control board does as far as frequency or voltage. Got to head to the local MakerSpace and put it on the scope to see. I was not brave enough to push the voltage past 24 VAC off my Variac.

    Do you have an idea what voltage & current the linear motor is driven at from the original control board? The board you show is different than mineI would really be interesting if the cryocooler can just be run from 110 VAC.

    I have a few photos for anyone that is interested: https://www.dropbox.com/sh/3uvk2gd5yke1b4l/j7-9eiwFc7

    I'll try to get the manual scanned and linked somewhere.

  128. Ken, there are two models of cryocooler. The one that I have runs at about 27VAC at full power. However, I have seen photos of a similar-looking unit that apparently runs at about 120VAC. Good work! Let us know if you start making LN2!

  129. Ben, successfully made about a liter of liquified air! Since the unit I have is an older 2001 model, it is one that runs at the 120 VAC. I set a time lapse camera on the unit when it was powered by the control board and saw the voltage go as high as 132 VAC. I hooked it up to a Variac and slowly ramped up to full 120 VAC. Not sure how far I could push it but I'll stay at 120 for now.

    I would still like to get the real control board working and have opened the dewar to get at the temperature sensor. Grinding off the weld is a pretty tedious process. After opening, I realized it probably have been better to open at the weld on the top. Crazy stuff in there. Added additional pictures to the link above since I haven't seen any on the web before. Now to plan on how to get the temp sensors out on one piece. Are they just thermistors or thermocouples, rtd, etc. Just to cover my butt what was the value for the resistors you used to mimic the output of the sensor. ;-)

  130. Ken, that's great! I believe the temperature sensors are just two-wire devices -- probably thermistors. I seem to remember 5-10Kohm being the resistance when at LN2 temperatures.

  131. Hello Ken and Ben, I saw a couple of these superfilters on ebay now. My question is, how do you capture the LN2? Did you remove all the material, and just put a regular thermos flask under the cooler?? Or is there the flask already built in and has clean hoses with LN2 coming out? I would like to buy this unit off ebay to get some LN2 to experiment with food to freeze them. I would be happy if you can guide me, thank you

  132. Admin, there is a bit of work involved with converting the Superfilter into an LN2 generator. Did you read my blog post? The cryocooler gets cold, and the LN2 drips off into the collection dewar. The Superfilter's existing dewar is not really that helpful since it's built for RF filters.

  133. Air seperation can be done using Zeolites and Pressure swings. You fill a pressure tank with Zeolite. Then add compressed air. When you release the pressurized air the first ~20% of air release will be 95% oxygen. After that the remain air will by 95% nitogen. The Zeolite will grab the nitrogen first leaving just the Oxygen as a free gas.

    This may be a more practical method of seperating Nitrogen or Oxygen.

  134. read the complete blog. very interesting!
    I have a dewar question.
    built a stainless 304L 170 gallon dhwt. now am thinking of vacuum insulating it. could these cryomachines be used to assist in vacuum pumping? say precooling before the pump for higher density ? any thoughts on materials to make the envelope?
    radiant shield? am thinking of using argon to flush and dry the chamber. is the model 850A2RV20 superfilter a good candidate for cooling?

  135. I just purchased one of these and...oh no! The fault lamp is blinking. I disconnected the RF equipment from the front but I'd be surprised this is a problem. Any chance it cares that the "AUX" plug is disconnected? I'll see if I can get serial diagnostics up but could really use some input here if there are any ideas as to whats wrong. I'm using a brand new (verified working before plugged in) 27V SMPS.

  136. Operation/install guide: http://siliconpr0n.org/uv/sti_superfilter/SuperFilter%20Operation_Install%20Guide.pdf

    Datasheet: http://siliconpr0n.org/uv/sti_superfilter/SuperFilter%20Datasheet.pdf

  137. So, read through the manual and all it is that it reads alarm initially during startup. I'm good to go!

    0:02:42:05> Temp
    System Temperature Report:
    Cold Stage: 79.52K
    NR Cold Finger: 79.08K
    WR Cold Finger: 77K
    Motor Temp: 19.7C
    Ambient Temp: +19C

    I'll post back when I do something interesting

  138. John, that'a great! I'll send you the program via email.

  139. Hello where can ı get cryocooler as cheap?

  140. v.ramanathan@outlook.comMay 19, 2013 at 12:19 AM

    Hi Ben, Can a cryocooler be used to make ice - and how practical would that be? Thanks for your time.

  141. Had a question? I just got my cryocooler and I notice in your pictures that you drilled 4 holes in to the copper clad on the cold end of your unit. how deep can you drill those holes? Also: How did you fool the control board into thinking that the the S.C.filter was plugged in to it? I saw in the picture of the power supply that you put a resister somewhere on the end of that plug.

  142. Joyangel, my model already had four tapped holes in the copper tip of the cryocooler. They were unused in the Superfilter design. They are about 4mm deep.

    Originally, I used the stock temperature sensors mounted on the cryocooler to monitor its temperature with the circuit as intended. Eventually, the sensors broke and so I replaced them with a plain resistor (at room temperature) across the sense leads to make the circuit think the tip is always about 80K

  143. Could you possibly produce liquid helium by cooling the hot side with niquid N?

  144. Garbled, the Stirling cryocooler uses helium as its working fluid, which must remain a gas for the cooler to work as designed. Also, the thermal leakage from the tip to the body might be too great for liquid helium temperatures. Search for Gifford-McMahon cryocoolers. There are 2-stage units that can produce LHe.

  145. Ben You can write a book to share your DIYER spirit do? Too many problems to be solved

  146. Hi Ben,
    I've been looking for a cryocooler and the Superfilters are pretty rare and pricey on Ebay. I did find however a bunch of APD Cryogenics Helium Cryocooler Compressors, and didn't know if they would have the same necessary component inside. Do you think these would accomplish the same thing? There isn't a ton of literature on them as the company has been bought out.

    Below is the link for the ones I was thinking about.

    Thanks for any information

  147. Mike, those helium compressors don't provide any cooling directly. They must be attached to a cryocooler or cryopump. The high pressure helium is expanded in the Gifford-McMahon cycle to remove heat from the target.

    I've wanted to experiment with those compressors and cryopumps for a while. It will definitely pump enough heat to make LN2 -- probably at a much higher rate than the system I have.

  148. Hi Ben
    I have been looking for LN2 generator to perform lab experiments in undergraduate program at our institute. I found you article of internet "http://citizensciencequarterly.com/2011/05/liquid-nitrogen-generator/". I think this was the thing which i was looking for my lab. Please send me some details of equipments and how to connect them so that i can build it in our lab. i searched "super filter 2" to extract its cryocooler but didn't find it. Please send me some details on yasiralikhan_25@yahoo.com

  149. Yasir, I'm afraid that I don't have any additional documentation. Feel free to ask me specific questions about the project.

  150. Hi Ben
    In your article i am confused with a little point. I searched in literature that working fluid is compressed and its heat is rejected to environment (heat rejection area which is heat sink in literature) and then it is expanded so that it can extract heat from a chamber (tip in this article) to be cooled. And i noticed you have mentioned it in explaining cryocooler.

    But in explaining the design you mentioned that a heatsink is attached to tip. I think cryocooler doesn't have heatsink so that it is attached to tip?
    I am confusing. Please give me some explanation.

  151. The heatsink provides a way for heat in the dewar to be conducted into the cryocooler tip so that it can be pumped out of the system. If the interface were very small, the cryocooler tip could be very cold, while the N2 in the dewar would still be relatively warm.

  152. Hi Ben!
    I am facing a serious problem in finding reliable and cheap cryocooler. I searched much but i only found new one which are very costly. I also searched on eBay.com and other similar but due to insufficient experience and knowledge i am not sure which product will contain a suitable crycooler. For example i have searched this product

    Please guide me. These products contain cryocoolre or not. If not please give some link or product name which is available.

  153. Yasir, those cryopumps require a separate helium compressor in order to function. Search for "helium compressor" on eBay. The high pressure helium is expanded in the crypump in order to remove heat from the system, then is returned to the compressor for recirculation.

  154. Then my problem is not solved. These will not work for me because i need crycooler so that i can build LN2 generator at our laboratory. Please give me some information i.e., name of some products which already have cryocooler and i can extract easily it from those products

  155. Yasir, you could buy the helium compressor and the cryopump. Or you could buy the Superfilter. Or you could buy the Elan2, which is an already-made LN2 generator. These are the only options that I know without doing additional research. Good luck.

  156. Thanks for your kindness Mr. Ben. Actually i am very new in this field as i completed my studies recently and joined my institute who is planning to arrange cryogenics lab experiments for undergraduate students. Moreover I lack knowledge about engineering so i feel more problems. But from your article I got a lot of ind=formations.
    My question is any superfilter contain cryocooler? for example this one


    Thanks for your kindness again

  157. Yasir, I believe all STI Superfilters have a cryocooler.

  158. This comment has been removed by the author.

  159. This comment has been removed by the author.

  160. Great article, but you have this one detail wrong:

    "The tube walls are semi-permeable and allow oxygen, water vapor, carbon dioxide and other “fast” gasses to permeate relatively quickly. Nitrogen and heavier gasses do not permeate as quickly, so the concentration of nitrogen is much higher at the exit end of the tubes than it is at the input end... The resulting nitrogen will contain trace amounts of argon and even smaller amounts of other noble gasses."

    N2 is both lighter and smaller in volume than all of these gases except for water vapor and Oxygen. O2 is actually heavier, but smaller in volume. So, how then does the device work? When I hear permeation, I think volume--not mass, so most likely it is only filtering out things which are larger in volume than N2: Oxygen and water vapor (plus Helium and Hydrogen which are obviously very rare).

  161. Unknown, yes, I meant "slower" instead of heavier, and I believe you are correct that the membrane only filters based on molecule size.

  162. Amazing article Mr.Ben!. I have a question that might sound stupid to u?.
    Can I start a small scale business in LN2 Production? what machinery
    would I require ? What is the budget required?
    I have very little knowledge about engineering.

  163. Ben, sup-tech has a device called a Sapphire Cryocooler. I am going to build one of these liquid nitrogen generators and i wanted to know if this sapphire cooler will work?

  164. oh and also how long could this device work for?

  165. IN reply to Sud, Ln2 production Businesses Use machines that cold MILLIONS of dollars. This experiment is for small scale things for like making ice cream :). The élan 2 could get you ctarted but if you are thinking of selling it to bass shops you are getting into the millions

  166. Ben,
    STI's cryocoolers are 10,000$ now :(

  167. how can I buy your nitrogen liquid generator ben..

  168. Hey Ben; Long time reader; Big fan. I think I've been following your exploits ever since Hackaday did a article on your LN02 Generator a few years ago, I've subscribed to your Youtube channel and love seeing the stuff you get up to, one word about your machine shop. AWESOME. The next project I’m embarking on after this one is a supercritical dry ice setup like what you did with the acrylic and machined aluminium, just need to get my hands on a CNC mill first.
    Anyway; back to the subject at hand, I've recently acquired a STI Cryocooler, the only issue is it didn't come with the associated power supply/control board. From previous comments it seems I can use a Variac in regards to powering the Cryocooler, however this leaves limited control over the cooler and as well as the fact I live in Australia and the mains frequency is 50hz whereas the sticker on the Cryocooler denotes 60hz, while I am unsure if this will effect the Cryocooler one can only assume working at a lesser frequency would result in a smaller cooling ability.
    You suggested in one of your posts, The control board may shift the frequency as well as voltage to the Cryocooler, dependent upon tip temperature; I was wondering if you have any other information in regards to the control board, i.e. voltage/frequency Vs time/temperature graphs of Cryocooler / scope (I know you just got a new Tektronix one) screenshots so I can make a control board myself;
    Also, I was wondering if you could approximate the quantity of CMS? In your nitrogen generator, I know you picked up a ready made device, I am not so lucky but am thinking about building a PSA large enough for approx. 700 liters per day, yours does not seem too large, the only other person I can find on the internet who has attempted such a project built his PSA with 40kg of CMS, but I think it is for a much larger volume of Nitrogen then is required for 1L a day of LN02. There must be an approx. Pressure/Volume to nitrogen ratio somewhere for building PSA’s;

    Any help you could give me in regards to this would be appreciated; I have a few experiments of my own I would like to try with Cryocooling and would really appreciate it.
    Also if you could shoot me a email with the program for the STI Cryocooler it may have some data on the Cryocooler that could be useful.
    Regards. Christopher.

  169. Hi Ben,
    Well I finally got my super filter up and running. Interestingly, the first time I ran it, there was a fairly dense vapor cloud of what I thought might be silicone oil. After 1.5 hrs, the fault bypass light went out. This is consistent with reaching temperature, if I read the manual correctly. What I'm wondering is the following.

    Do you think that the dewar could be maintained intact, using some of the electronic lead vias to pipe the gas in and the LN2 out?

    The other question relates to high temp super conducting materials contained in the filter. Any application you can think of for this?

    As far as I can tell, all SCT super filter units use the same Sapphire stirling cooler. On the other hand, Coleman briefly produced a Stirling chest cooler for under $500. No longer available, but used a different companies cooler engine.


  170. Nitrous, the wires leading into the dewar cannot really conduct gas in or liquid out. I'm not sure how it could work.

    The superconducting material in the dewar can be removed, and you levitate a small neodymium magnet. It's a neat and classic demo.

    Good luck!

  171. I was wondering if the Dewar vac could be maintained while expanding one of the vias. It depends on how the dewar is designed.

  172. Folks might be interested in the Sunpower cryocooler manuals at the bottom of the following page. Sunpower owns the technology for the Sapphire, and it looks like the Sapphire is very similar to Sunpower's 5W model. They emphasize the importance of cooling the hot end, and say their controller begins by centering the piston, so just any source of 60Hz AC might not work (or perform a reliable startup at least).


  173. Trying to find a Controller for a CryoCooler if anyone has one; Don't mind if its broken, happy to repair, will pay.
    have a Cooler without a controller.
    Also, How did you make the water jacket for the hot end Ben?
    regards Christopher

  174. Hi Christopher, I might be willing to sell a spare controller. You can contact me at (obfuscated against spammers - remove letter Q) jeQtmonQk at gQmail dot cQom. - apologies for replying in the thread but I can't find out how to reply individually.

  175. ST is willing to give people a manual for the Sapphire. Here's the important page on supplying power. It looks like a variac is fine, but the voltage has to be kept low, and the power ramp-up curve has to be obeyed. The fins need to be kept below 80C.


  176. where does one get the ST manual from?

  177. Christopher, please check your email that you gave above (I emailed you a day ago) for an email with subject "cryocooler controller" or email the jetmonk email a couple of posts above.

  178. Thanks to jetmonk for sending me some very helpful tools and docs. In return I'd like to offer the list pointers to some very informative patents

    General search for all STI patents (There are a lot!)

    Core Cryocooler design (Exactly how it works)

    DSP temp Control (Including the full source code for an older version of the firmware)

    Analog Temp control (Explains the loop calculations included in the DSP control code)

    Dewar sealing (Included just because I love the cold welding of clean Indium surfaces!)

  179. Before cutting it open, could one of those people who have ground apart the dewar confirm my thoughts?

    The dewar is heavy duty and single skinned, with the electronics mounted inside exposed to the vacuum directly (for simplicity and to avoid condensation/contamination).

    The heatsink mounting the electronics attach to the inside of the end of a pocket in the dewar, with the cryocooler cold head being inserted into the pocket and pressing on the bottom with thermal material. The NW40 flange is used to stop moisture condensing on the cold finger.

    From the look of it, the electronics are all assembled to the bottom of the dewar which is then welded into place (the rest having being prefabbed), with the whole lot then evacuated and baked, before being sealed off.

    The large unused connector at the bottom may be a getter to preserve the vacuum long term.

    Does that make sense?

  180. I couldn't wait and cut the unit open. Pretty much as I expected, except there was also a pocket in the middle of the heatsink with zeolite beads to improve the vacuum once cool. You can tell how good it was because I didn't open till 6hrs after switch off, yet the moment air went in the unit started to collect dew and then froze solid!

    Two questions for those who have already taken these units fully apart.

    * How is the copper block attached to the cold pocket? I'd like to reuse the top of the dewar including the cold pocket and block, so don't want to damage them.

    * Does anyone have a pin assignment from the umbilical to the thin wires inside the dewar, as some of these came off when I opened the unit.

    thanks in advance, Colin.

  181. Is the copper block hard soldered on to the end of the cold finger tube, meaning I need to grind/turn off the sides of the dewar if I want to reuse the block and pocket?

  182. As requested by several people, photos of disassembly so far


    Dewar opening was via a bench grinder (to remove weld and get a small access under lid) combined with chisel and hammer (not subtle, but the lid is a very tight fit even once the weld is removed).

  183. Three minutes with a narrow kerf, fine grit cutting disc in my angle grinder and the rest of the dewar was easily removed. Knowing what I know now, this is definitely the way I'd open one of these dewars in the future.

    Album above updated with extra photos, including breakdown of the filter/LNA assemblies.

    Could someone who has already done this, confirm that each of the three blocks has two twisted pairs feeding the ceramic assembly at the bottom of the unit, and that from the look of it, one provides power to the LNAs mounted inside the block, whilst the other connects to what I assume is a diode temperature sensor, just visible one the ceramic hybrid?

    Have people just pried the sensor out and used as-is, or have people tried using an alternative diode as a sensor?

  184. Voltage drop reading seem to backup the inner pair of wires connecting to a silicon diode on the ceramic carrier, and with some care I've managed to separate the carrier, diode and wiring from the filter module. Given that the diode junction is exposed with extremely fine bond wires, I've very carefully protected the junction with a small drop of epoxy resin, because otherwise just the slightest touch could damage the sensor.

    I can already see the voltage varying with temperature, and I'm going to try and reconnect to the original cryocooler controller to verify it still works. Typically silicon diode temperature sensors have a bias current of either 10uA or 100uA, so I'm going to check pairs on the connector with a 100K resistor to see of I can separate the sensor connections from the LNA power feed.

    The other sensor is the one on the lid, which I'm guessing from appearances is a thermocouple for sensing ambient. Does anyone know what type it is?

    (More photos uploaded)

  185. It looks like only two of the three diode sensors are actually used. I'd thought the third had come loose from the feedthrough when I was taking the lid off, but I now suspect it was simply never connected, and just included as a standard part of the filter module.

    The sensor circuits are the two pairs of contacts on the side of the umbilical connector which has the wires covered in black sleeving. Bias current is 100uA. Once I get going I'm going to see how closely the response curve matches that of the BAS16 diode which I've used in the past for temperature sensing (and is detailed here http://www.sciencedirect.com/science/article/pii/S0011227505001591). If it is close enough it would remove the need to strip down the filter assembly, and give a ready supply of replacements.

    I'm still unsure what the metal "wire" welded to the lid was all about, but it isn't the ambient sensor as this appears to be on the control board. Given the spot weld, perhaps just a ground/shield connection to the dewar? Anyone know for definite?

    My goal is to produce a cryogenic purifier for He-Ne (by condensing out N2, O2, etc) but I'm going to start by building a air liquifier first. I'll update the list on how that goes.

  186. For a first test I've adapted a 1.2L Thermos ThermoCafe vacuum flask to fit the end of the cold finger. I chose the flash because it is for food as well as drink so has a very wide mouth, but conveniently the inner liquid stopper has just the right dimensions to be altered to mate with a NW40 flange on one side and a snug fit on the cold finger in the middle. The cold finger has a bent bit of 28mm copper tube attached with two of the small screws from the filter module and some Apiezon N grease for thermal transfer. I'm using one of the rescued sensor diodes for measurement, again with a little grease for thermal contact.

    The whole lot is mounted in place of the standard dewar, with the original case cut so I still get good airflow over the cryocooler heatsink.

    Photos of all this (along with breakdown of one of the filter modules) now uploaded to album.

    I've also upload the cooldown graph which shows the cold finger hitting 80K in just over two hours. It is stable at this level, hopefully because I'm now condensing the equilibrium mix of 50:50 nitrogen/air with a boiling point of 81.6K http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/liqair.html

    To get maximum power, I tweaked the control loop constants once I got down to around 81K (at 140min into the run), so that I've got around 135W going to the cooler (temp sharply drop almost another 1K in graph). The way I've currently done this I'll need to reset before the next cooldown (to avoid overdriving) but looking at the code it may be possible to alter in a way which can just be left "as-is".

    I will leave it running for a couple more hours and see what I've got. Fingers crossed!

  187. Again I can't wait.

    Turned off the cooler and removed the dewar. Encouragingly there was very little ice on the cold finder (so just enough air getting in to be liquidized) and the characteristic sound of liquid gas swilling in the flask. Emptying out I would guess I had around 50ml of liquid air. :-)

    More photos and a video uploaded

    Proof of concept done, so now on to improvements to allow a) more liquid air to be made b) purification of my He-Ne mix. The key part of the latter is a heat exchanger to allow gas flow through the system without a huge heating effect.

  188. Not many diy project projects on the net about liquid nitrogen generators. Found another one were they used a scuba compressor to cool through regenerative cooling. Problem is, that 3000 psi compressors are all super expensive. I would think that if you build a super efficient system, thought regenerative cooling you should be able to reach 70K with a much lower pressure.

  189. What about psa for initial nitrogen "filtering"- med O2 equipment usually keep the oxygen and discard the nitrogen; I'm sure that can be rectified!

  190. Quick question for those wh have the superfilter. Can the Stirling cooler be separated from the dewar without destroying the dewar? When I removed the collar that seems to couple the cooler to the dewar, the cooler can only be slid out a cm or so.

    Any advice?

    1. I remember it being a very tight fit, and required a fair bit of force to pull the cryocooler out of the dewar. Besides the clamp, I don't think there is anything holding it other than some thermal grease and friction. Be sure to let the device warm up over several days to allow the thermal grease to thaw.

  191. Thanks Ben,
    I was hoping to not have to pull the dewar apart just to get the cooler out.

    1. Oops
      Didn't see the charging pipe on the backside of the Stirling engine.
      With the dewar restraining band loosened but still attached to the frame, I lifted the cooler out without trouble.

      The sensor to the hot area appears to be necessary to allow the driver board to continue to power the engine after initial start up. Did you substitute a resistor for that sensor?

      Thanks again Ben

    2. I don't remember spoofing the hot-side sensor. I either altered the firmware to not read that value, or my unit didn't care what the hot-side temperature was.

  192. Could you do a youtube video on this project?

    1. Perhaps. The LN2 generator is mostly in pieces distributed around my shop, and there are always new projects that keep me busy. I agree it would be fun and interesting to take another look at it.

    2. Is there any way i could somehow contact you other than on the blog?