-196C and colder?

[[XC] gomeler]

I just spent a happy hour jumping around the web reading about the founding discovers underwent to produce liquid helium for the first time. This naturally got me thinking and raised a few questions. I highly doubt this is the forum for these answers but perhaps someone can point me in the right direction.

In a similar fashion to using a sub-cooler with R14 and Argon, would it be possible to use a blend of nitrogen and neon to lower the effective boiling point into a region between nitrogen and neon's boiling point? In what clear reading I could find, it seems like they are making the jump from nitrogen straight into helium which to me sounds insane as that's nearly a 80K jump in boiling points. Then I remembered the R14/Argon trick which is about a 60K jump, and a proven mix.

From my understanding, liquid helium is such a small atom that it will literally pass through cryogenic cooled copper, so neon sounds like a great choice as it's much larger with an atomic mass of 10, just 4 less than an LN2 molecule.

At this time I have no intention of building anything in this temperature range but after seeing the AMD Phenom II overclocks (6.4GHz CPU-Z shot on LHe) it got my brain thinking. Less so in making a functional cascade of this scale and more so as a thought experiment, I'm curious what it would take. Feel free to chime in, if you know of any relevant patents, detailed schematics(I love that word), or brains I can pick, let me know.

I crave knowledge

[alexio]

Neon is expensive. I'd say if you really need lower than -196C you need to evaporate the LN2 under sub-atmospheric pressure. It's like how a rotary compressor can pull a vacuum to achieve lower temps with a specific gas than is possible under atmospheric pressure. At least -210 should be doable.

[[XC] gomeler]

I've thought of that also, not really ideal. Limits you to either low capacity or using monstrous compressors to have enough capacity at an extreme vacuum to be useful.

[mytekcontrols]

Chris --- If you want lower then -196C; a helium cryopump like CTI's Cryo-Torr is the most efficient way to go. Phase change just gets too inefficient at these low temperatures, even with a good autocascade design (4 phase seps) and a subcooler.

Although if you insist, it can be done as you suggest, and neon will be required, as well as nitrogen, and argon. But it will require a considerable amount of horsepower.

A Helium Cryopump utilizes compressed helium gas, which is sent through consecutive expanders in order to obtain cooling. Loosely based on the Stirling Cycle Heat Engines, only done in reverse.

[alexio]

I've thought of that also, not really ideal. Limits you to either low capacity or using monstrous compressors to have enough capacity at an extreme vacuum to be useful.

I'm not talking about a closed loop system. I think you don't need such a powerful compressor at all. You could have a big fan on top of an LN2 tube (with an adapter) and the LN2 comming in trough a small tube on the side. I mean a 30W+ fan, so it's not that small either. Everything should be "airtight" obviously.

[[XC] gomeler]

Alexio,

I've considered brazing my f1ee shut and inserting a ln2 feed tube and a second port to connect to a compressor. Using a pressurized dewar, feed in ln2 and use the compressor and a solenoid on the ln2 feed to boil the ln2 at slight vacuums. I figure a large stepper evap would be better though as otherwise the nitrogen will not spend much time in the pot.

Michael,

I imagine that pump uses the process of expanding helium to drop the temp, linde cycle I believe? When I am back at my pc I'll have to check those out. I just thought of neon and such because compressors are easy to find second hand compared to special pumps.

I wonder, a normal compressor would probably not be too thrilled with a -200c suctin temp. I guess special cryogenic compressors would be needed?

[alexio]

Alexio,

I've considered brazing my f1ee shut and inserting a ln2 feed tube and a second port to connect to a compressor. Using a pressurized dewar, feed in ln2 and use the compressor and a solenoid on the ln2 feed to boil the ln2 at slight vacuums. I figure a large stepper evap would be better though as otherwise the nitrogen will not spend much time in the pot.

That's a good idea. But, as long if you're not pumping liquid it shouldn't matter what design the evap has other than the usual considerations for efficiency. In any case the LN2 will only boil off following the law off physics. What I mean with that is that any LN2 not in contact with the copper won't get "lost" by inefficiencies. In fact it will "cool" the LN2 that is in contact will the copper.

[[XC] gomeler]

That is very true, I keep thinking of it as an open system. It would indeed keep the gas layer between evaporation and the base. I want to use capillary tubing to spray liquid nitrogen at the base and then suction from the top. I might still go through with this, just make plate that'll form up with the top lip of the F1EE and then use something like the benching station that Team Japan uses with an anchored rail and a large screw to apply pressure downwards on the pot. This way it would at least be a non-permanent solution.

I think I'll actually try this whenever I have the chance. I'm just afraid that a normal compressor's internals won't stand up to the thermal stresses up a -19xC. Anyone have thoughts on this? Guess I could always try with a cheap compressor

[alexio]

What is the specific heat of gaseous nitrogen? I thinks it's pretty low and since compressors generate a fair amount of heat themselves...

It could be an idea to change the standard oil in the compressor with something specified for lower temps. Maybe put some extra grease on o-rings, etc. I think you may be just fine without these measures though.

[[XC] gomeler]

The oil was something I was considering, alkybenzene would be a good option as it stays non-viscous at low temperatures. I am just more worried about what the 250 kelvin delta will do when frigid gas hits warm metal. Seeing as it won't be under any pressure, the worst thing I could imagine would be the internals of the compressor cracking and it violently destroying itself within it's hermetic shell.

Ok this is starting to sound like a fun idea, might try this on an upcoming weekend. I really need a large shop to work in and retirement to set in at 22, I have too many projects I want to undertake.

[alexio]

To prevent rapid cooling (compressor) and heating (N2 gas) you can consider a cold start. With that I mean not to let the compressor run before you start injecting the LN2, so the deltaT remains as small as possible.

[mytekcontrols]

I wonder, a normal compressor would probably not be too thrilled with a -200c suction temp. I guess special cryogenic compressors would be needed?

Both the IN and OUT connections running to and from the Expander (Cold Head) are at room temperature and uninsulated (typically stainless steel flexlines with self sealing disconnects). The actually pump part of the system is just a refrigeration condensing unit, with an oil separator, and what they call an Absorber to catch any remaining oil from getting into the Expander.

These units are typically built to trap water vapor and other non-condensible gases in a vacuum system. Essentially they can be considered as vacuum pumps in their own right. Hi-Vac pumps that is, since they still need a mechanical vacuum pump to take it down to a reasonable pressure. Typically 200 microns is the cross over point.

Here's a good demo: Helix CTI Flash Tutorial

And here is a very good explanation: Closed-Loop Gaseous Helium Cryopumps

As you can imagine these are usually not cheap, but sometimes they can be had on the surplus market for a bargain.

[[XC] gomeler]

From what I gather, that cryopump tutorial is a high performance vacuum pump that operates by condensing the "suction" gases upon the various heat exchanger arrays and causing a vacuum to form. The pump has a limited volume it can absorb before it must be regenerated by warming up and evaporating off the gases from the fins, collected I imagine for further use/filtration.

From what I gather, the cryopump is fed a high pressure helium supply that is gradually expanded to utilize the Joule-Thomson effect to chill the various stages of the pump down to 10 Kelvin. So, I guess I need to read up on this Joule-Thomson effect a bit. From what I understand it sounds like they are starting with very high pressure helium, and by using a series of valves and insulated expansion chambers, cause the helium to liquify. Having trouble wrapping my brain around this concept, seems to wrong. Tomorrow when I'm clear headed I'll think about it again.

Once again thanks Michael, you've been incredibly helpful and I appreciate it.

[mytekcontrols]

From what I gather, the cryopump is fed a high pressure helium supply that is gradually expanded to utilize the Joule-Thomson effect to chill the various stages of the pump down to 10 Kelvin. So, I guess I need to read up on this Joule-Thomson effect a bit. From what I understand it sounds like they are starting with very high pressure helium, and by using a series of valves and insulated expansion chambers, cause the helium to liquify.

Actually not really Joule-Thomsom, which is more of a passive system (no moving parts), but more like Stirling Cycle which uses a displacer assembly with a regenerator (stacked copper screens). The Gifford-McMahon cycle refrigerator which is used in the CTI or APD Helium Cryopumps, is an improvement on the Stirling Cycle systems, and is used through out the high vacuum industry.

Having trouble wrapping my brain around this concept, seems to wrong. Tomorrow when I'm clear headed I'll think about it again.

Well tomorrow has come, and I have some more material for you to read when you get a chance

This covers the history of refrigeration, and shows all the different approaches that were taken to achieve cold. It also has a section that gives a good break down on the Gifford-McMahon cycle refrigerator.
Introduction to Cryogenic Engineering

And this covers a bunch of Pulse Tube refrigerator designs, which are the newer kids on the block.
Pulse Tube Cryocoolers

Enjoy

[[XC] gomeler]

Just got home to a proper computer screen. I had been trying to read those documents on my iPhone, can't read anything at 320x240 Going to spend my breaks today between school work reading these documents, that's 200 pages of slides Will get back to you after I absorb all this, might be a day or two

[[XC] gomeler]

Ok I skimmed a lot of that, understood some of it, recognized some of it, and started researching the stuff I didn't understand. Stopped for the night and I had nearly 30 tabs open just because of that CERN PDF. I'm going to continue reading into this over the next few days, I'm sure some questions will pop up.

[mytekcontrols]

Ok I skimmed a lot of that, understood some of it, recognized some of it, and started researching the stuff I didn't understand. Stopped for the night and I had nearly 30 tabs open just because of that CERN PDF. I'm going to continue reading into this over the next few days, I'm sure some questions will pop up.

Yes it's pretty interesting stuff. To tell you the truth, I haven't done much more than skim through it myself.

In fact my direct experience with this kind of refrigeration is very limited, although for a while, I used to come in near contact with it almost every month when I was doing field service work for Polycold. A lot of the same industry that uses Polycold's, also use the Gifford-McMahon cycle refrigerators. Because the capacity of these style units for pumping water vapor is limited, Polycolds were often times used for this aspect, and the Gifford-McMahon units would be left to trap out the remaining gases (Nitrogen, Argon, Oxygen).

I don't know if you guys realize it; but much of the technology that you use every day (incorporating ICs, CPUs, Hard Disk Drives, CDs, DVDs, USB Memory Sticks, Digital Cameras, Cell Phones, and almost any computerized device you can imagine), owes its existence to Vacuum Coating technology, Polycolds, Gifford-McMahon cycle refrigerators, Diffusion and Turbomolecular Pumps (2 other Hi-Vac pumping methods). And in each and every case, water vapor is the killer. Thus the need for a water vapor cryopump such as a Polycold.

Anyway, I hope I haven't gotten you too side tracked with all this reading

[ultralo1]

As always I am amazed at the tech knowledge.

I have only been exposed to 3 cold heads, two were used on MRI systems and one on a vacuum desposition system. They all were water cooled. I always wondered how they worked and i am going to try to got the manufacturer school on them in the next year or two.

The two for the MRI systems are used as a cold Sheild (finger) setup with in the insulated helium cavity. They are used to help condense the vapor helium back into liquid and this also increase the vacuum level which aids in the efficency of the insulation. When the cold head is not operating the rate of helium evaproation more than doubles. Helium refills for the big system (human size) are $30000. Without the cold head this was being done once per month.

The other one is used in conjuction with a turbo molecular pump to achieve a vacuum of 10X-12. They were gold plating DNA strands for backscatter electron microscopy.

[tiborrr]

It's not really a closed loop system, but still a fun project. Only backdraw is that condensing medium, which is LN2, runs out pretty quickly.

http://www.extremecooling.net/index/...opic.php?44612

[quintus]

So we need: a big compresor, or several in tandem, good separation of nitrogen and oil, and a bigg a$$ LN2 plant!!!

LE: It is a closed loop
But if we do it unclosed? like cap line(bigger diameter than the usual 0,031) taking directly from dewar the liquid N2...Then wtf do we do with the compressor... will it freeze? Discharge line goes directly outside dumping the N2 back to where it came from Maybe completing the compresor at X times benchez with oil...
...
regards,
q

[[XC] gomeler]

Used babelfish and my eyeballs to translate. Looks like they took a single-stage, used LN2 for the condensing medium, and nitrogen in a heavy vacuum? Very interesting but very expensive Been busy with class the last two days but this upcoming weekend will be chock full of research afterwhich I'll continue to probe the idea of a mechanical system to hit -200C

[Moc]

Looks like they took a single-stage, used LN2 for the condensing medium, and nitrogen in a heavy vacuum? Very interesting but very expensive

This is exactly what they did. Unrockstar is a cracy person !

[ilkkahy]

I have thought about possibility of using venturi effect with Ln2 to create the underpressure. It would be pretty easy to try out too.

http://en.wikipedia.org/wiki/Venturi_effect

[[XC] gomeler]

Found out my college library(Georgia Institute of Technology) has a rather nice section on cryogenics, alternative refrigerants/designs, and even a book or two on "mixed refrigerant" systems. Also found 5(Five!!!) professors or PHD/Masters students on campus that have done research into using single and multi-stage refrigeration systems to enhance performance of microprocessors AKA overclock. Just glancing over their work, they built a small R404a/R508b cascade with a Danfoss BD80F and NF11FX. I kind of giggled at their maximum heatload of 40w @ -78C but by contacting them I'm hoping to get access to their advisors and pump them for information regarding cryogenics.

They also have done extensive work on evaporator design with evaporators that are 3.1mm thick and feature incredibly small channels or pin matrices. Skimming their work, which was quite extensive, an alternating pin matrix was most efficient.

Will chase these people down on Tuesday, no school on Monday and no hope finding someone on Friday before a 3 day weekend

[Sgrios]

Found out my college library(Georgia Institute of Technology) has a rather nice section on cryogenics, alternative refrigerants/designs, and even a book or two on "mixed refrigerant" systems. Also found 5(Five!!!) professors or PHD/Masters students on campus that have done research into using single and multi-stage refrigeration systems to enhance performance of microprocessors AKA overclock. Just glancing over their work, they built a small R404a/R508b cascade with a Danfoss BD80F and NF11FX. I kind of giggled at their maximum heatload of 40w @ -78C but by contacting them I'm hoping to get access to their advisors and pump them for information regarding cryogenics.

They also have done extensive work on evaporator design with evaporators that are 3.1mm thick and feature incredibly small channels or pin matrices. Skimming their work, which was quite extensive, an alternating pin matrix was most efficient.

Will chase these people down on Tuesday, no school on Monday and no hope finding someone on Friday before a 3 day weekend

Keep us posted Gomeler. Sounds very interesting. Than again, I crave knowledge so.