3 stage cascade, need 3 compressors with 10:1 ratios. (R14 compressor)

[Cascadia]

Hey all,

My cascade design is requiring 3 compressors with compression ratios above 10. Running refrigerants 507, 170, and 14.

What compressors have you been working with for evaporator loads exceeding 300 watts? When I search manufacturer websites no one rates a compressor that is made for R14, but I've seen plenty of pics with people running R14.

Are you guys just working outside the manufacturer specs for refrigerants?

[[XC] gomeler]

Yes, we typically run compressors outside of their original spec.

[sdumper]

Whats your cascade targeted for? CPU cooling or something else.
Getting a 3 stage to hold 300 watts is a PITA. It can be done but the tuning takes a long time to get right.

I would go with r402, r170 and a blend of r1150 and r14 in the 3rd with .040 cap tube at around 10 feet 7 inches.

[Cascadia]

No CPU cooling for me... Actually I plan on desublimating CO2 from combustion exhaust gases. I'm here because every search for cascades comes to you guys! You must be good. I'm on a small team trying to build and demonstrate what we believe is the most energy efficient way to capture carbon dioxide and other pollutants from coal plants. (We can even build a coal fired power plant that has a negative carbon footprint)

But before we capture that CO2, we need to cool the flue gas down to -120 C to catch 90%. The full scale process cools using a turbine, but we cannot get one on this small scale. So hence a 3 stage cascade.

We've demonstrated the basics of our process already using liquid nitrogen but I want a mechanical refrigerator to give me a few hundred watts so I can do my experiments more smoothly. I also need practice! We have a huge project after this.

Here's what I got in mind:



People seem to suggest a cap tube instead of a TXV on the cold R14 loop. Why is that, sdumper? Thanks for the suggestions.

[Cascadia]

PH Diagrams for ya.

[sdumper]

Your r14 and r170 condensers are heat exchangers im assuming correct.
Wow that Flu gas is 500 watts...?

[Cascadia]

Your r14 and r170 condensers are heat exchangers im assuming correct.
Wow that Flu gas is 500 watts...?

The R14 condenser is the same heat exchanger as the R170 Evaporator.

The R170 condenser is the same heat exchanger as the R507 Evaporator.

We're just using small slip streams from the flue gas in our experiments, its only a few cubic feet per minute.

[DetroitAC]

Hi Cascadia,

Sounds like an excellent project, best of luck to you.

First to your original post, there are many good small hermetic compressors that operate well at high pressure ratios. I build math models of the compressor from mfg. spec. sheet data to evaluate performance at off-design conditions. Finding suitable compressors shouldn't be much of an issue, I recommend you start searching with Embraco and Danfoss. Both companies make good performing compressors, and also supply performance data.

If you will be able to use 240VAC, you will have a much easier time finding a suitable 1st and 2nd stage compressor. I haven't created a model of a three stage, but I'd guess you'll be looking a 1st stage too large for 110-120VAC.

You will also not be able to source compressors to match your model, there just aren't enough sizes available, so you'll need to rework your refrigeration modeling to suit your compressor options.

I also don't believe these systems should be operated in the simple cycles you have drawn. I think the very cold gas temperatures you have modeled will not be compatible with compressors designed for higher temperatures. In short...I think you need to superheat the suction gas to a much more reasonable temperature for compressor inlet. You can utilize the cold suction gas help condense and subcool higher stages.

[[XC] gomeler]

TXV can be used on the 3rd stage if you either buy a dedicated R14 TXV or you regas the bulb on a TXV with an R14 charge. There's a thread or two here on the procedure but I can't remember who did it.

I echo DetroitAC's sentiments on 240VAC. For a 500w load you are looking at a multiple kW load on the 1st stage which would be difficult to shift with the current limitations of 110VAC compressors.

Sounds like a fun project and I hope you guys keep us in the loop. Always fun to see non-overclocking cascades in action.

[Patrickclouds]

the efficience of oil seperation increase when you place the oil seperator after the desuperheater.

but be careful that parts of the refrigerant hasn´t already condensed in the r14 stage. critical temp of r14 is around -45°C. or use an air cooled desuperheater.

[mytekcontrols]

I think instead of building your own chiller for 500 watts, you might be better off buying one on the used market. Here's a spec list for some Polycold AutoC units. Based on what you are trying to do, I'd recommend something like the P-550ST or PFC-550ST (alternative models: PFC-550HC or PFC-552HC). It'll easily give you -130C at 500 watts. Of course the PFC-550LT gets even colder, but it maxes out at 500 watts which wouldn't give you any breathing room.

I know $10,000 might be a bit steep for prototyping. But sometimes they can be gotten for as little as $1,000-$2000 from a vacuum coating company that is downsizing, and selling off unneeded equipment. Of course this is As-Is without a warranty.

[Cascadia]

Thank you for the help gents! Excellent advice.

Detroit, when searching for compressors and they don't have one that technically fits your refrigerant, do you have a method for which refrigerant compressor you choose? I've noticed a few R22 compressors and R404 compressors most frequently. Ultimately I need to justify why I'm putting a refrigerant through a compressor that was made for a different refrigerant. My plan is to find the flow rate and temperatures it was designed for and then try to match up as best I can. I added a recuperator to warm the R14 suction while cooling the discharge gas below atmospheric temperatures.

Also, are -40 and -20 C are decent temps for compressor suction? It likely depends if the gas density is too great for the compressor to handle..
Patrick - Thanks, I moved the oil separators after the air coolers. As for the R14 I'll be operating well below the critical pressure so unless my pressure creaps above 20 bar the R14 shouldn't condense until -65 C.

And to both, we do have access to 240. Thats a good call.

Updates from your comments:

[Cascadia]

I think instead of building your own chiller for 500 watts, you might be better off buying one on the used market. Here's a spec list for some Polycold AutoC units. Based on what you are trying to do, I'd recommend something like the PFC-550ST (alternative models: PFC-550HC or PFC-552HC). It'll easily give you -130C at 500 watts. Of course the PFC-550LT gets even colder, but it maxes out at 500 watts which wouldn't give you any breathing room.

Oh you're the perfect guy to ask about that! I talked with Polycold a few months back about buying a unit. But when I asked them if the unit could handle a continuous load for hours on end they didn't give me a definite answer because they said 'its not designed for that'. My boss wouldn't allow me to spend $30k on something that they wouldn't confirm could deliver the load we wanted.

In your experience, have you seen those units take their max rated load for hours continuously?

[Cascadia]

Oh one more though, Michael. What if I downgraded my load to something more like 200 watts. Would it make more sense to build the unit in that case?

[mytekcontrols]

Oh you're the perfect guy to ask about that! I talked with Polycold a few months back about buying a unit. But when I asked them if the unit could handle a continuous load for hours on end they didn't give me a definite answer because they said 'its not designed for that'. My boss wouldn't allow me to spend $30k on something that they wouldn't confirm could deliver the load we wanted.

In your experience, have you seen those units take their max rated load for hours continuously?

They are put into many applications running 24/7 without any problems. The biggest service issue with these units, is that the Parker or Cajon (VCR) cryo-fittings connecting to the customer coil or evap have a tendency to leak if not held rigid (i.e.; prevented from twisting). Some customers opt to eliminate this problem and braze or weld their connections instead.

Edit: Many of these applications cycle between Cool-Defrost-Standby and then repeat. Cycle times in minutes are often 15-3-10 for the respective operational loads I listed. This constant cycling is much harder on the unit then just using it in constant cooling mode at full load. The defrost is achieved by hot gas derived from the compressor's discharge after being scrubbed of oil by two coalescing oil separators in series. The return path is always through the HXC stack and back to the compressor's suction. So an extended defrost will tend to heat the stack back up, thus causing high pressure build-up after defrost terminates due to the HXC stack recooling that is necessary.

The "P" units are without a defrost option, which is probably better suited for your application.

[mytekcontrols]

Oh one more though, Michael. What if I downgraded my load to something more like 200 watts. Would it make more sense to build the unit in that case?

Yes and no. If you are looking at building quite a few, then I would say yes. But if only one or two, then probably not worth the R&D.

If you only need one unit, then also check out the P-400 specs. This will give you -120C at 300 watts, and is about 1/2 the price of the other unit I mentioned. Although these are only available used, because Polycold no longer produces it.

BTW; if you decide to buy used and contact that company that I linked the spec sheet from. I do believe they might have one P-400 laying around gathering dust.

[Cascadia]

Those are great options.

We just need one unit on this small scale, but we will need a much larger, slightly different unit for a project right after this. The biggest pro for building this would be learning lessons on a smaller unit before tackling the bigger and more expensive one.

[mytekcontrols]

Probably both projects can be handled with used Polycold chillers. This would certainly save you money over new, and at least let you prove out your technology without breaking the bank. I don't know what the spec will be for your larger system, but a Polycold PFC-1100HC unit will handle around 2000 watts at something approaching -120C.

Keep in mind that most of these systems require cooling water, and are not usually air cooled. However Polycold 400 series units were made in air cooled and water cooled configurations. But in all cases they require 3 phase power.

[DetroitAC]

Detroit, when searching for compressors and they don't have one that technically fits your refrigerant, do you have a method for which refrigerant compressor you choose? I've noticed a few R22 compressors and R404 compressors most frequently. Ultimately I need to justify why I'm putting a refrigerant through a compressor that was made for a different refrigerant. My plan is to find the flow rate and temperatures it was designed for and then try to match up as best I can. I added a recuperator to warm the R14 suction while cooling the discharge gas below atmospheric temperatures.

Also, are -40 and -20 C are decent temps for compressor suction? It likely depends if the gas density is too great for the compressor to handle..

You basically have to do the best job you can with what's available and what information you have. Personally, I like to design very conservative and do everything I can to insure things won't break. Once you have a set of design conditions for your refrigerant gas, you can evaluate:
Suction temperature
Power required by the compressor
Discharge temperature

In my opinion these should be in the same ballpark as the hermetic compressor you choose, as the compressor mfg develops their compressor under these conditions. Especially important and often overlooked (on these forums) is compressor motor power. R410a compressors have enormous compressor motors for their displacement, R134a compressors have small motors for their displacement. Refrigeration Utilities will allow to compare the power required easily.

I think you might warm the suction gas only until you get within the design range for the compressor. i.e. if you choose a LBP compressor for your R14 stage, warm the suction to -40C, maybe even colder if you are willing to take a little risk, since there is a huge compressor displacement penalty for warming the suction gas.

Your 3 cycles have reasonable temperatures to me except the R14, I'd run the suction gas into the compressor @-40Cish

I think Mytek's suggestion of a used polycold unit is good one. If you need to develop the capability to build a 3 stage cascade, you should build it yourself. I'd buy a $30k polycold before I'd build a single 3 stage cascade myself, since there are dozens of issues that must be learned by failure and redesign. Polycold has already paid for those failures, and is building workable units. Their reluctance to sign off on your specific use conditions wouldn't bother me.

[Cascadia]

Michael, I called Pat and M&T and I think we're going to pick up a 550ST. And likely get the larger 1100 when the time comes. I was excited about building my own unit but this is far more practical!

Thanks for your help guys!

[Cascadia]

One more thing, Michael.

I just looked at the polycold energy requirements. For the 1100 it requires 18 kW for 1.8 kW of cooling. I'm pretty sure I could get better energy efficiency out of a 3 stage cascade, right?

I had a model for a 2 stage cascade and it required a lot less energy than the polycold. I'm not sure I'll be able to use it due to the high energy requirements. It kills our process efficiency.

[mytekcontrols]

I just looked at the polycold energy requirements. For the 1100 it requires 18 kW for 1.8 kW of cooling. I'm pretty sure I could get better energy efficiency out of a 3 stage cascade, right?

I don't think so. An autocascade only uses one compressor, and no isolated intermediary HXC's like on a cascade system. To me it's kinda like comparing an autotransformer (single winding with taps) to a standard transformer having two isolated windings. There are inefficiencies when trying to cross the threshold, so in a step-up or step-down application, the autotransformer has better efficiency, but you do loose the isolation. Pretty much the same thing applies in a auotcascade as compared to a cascade. The cascade gives you isolation, which makes it easier to figure out the refrigerant charge, but you sacrifice efficiency in the process. An autocascade is very tricky to develop a charge for, but once you have, you reap the rewards in higher energy efficiency.

But we must also factor in how cold do you really want to get? Because overshooting your target by a large degree, probably means that you are not optimized for the particular application. And not being optimized also eats up additional energy. So for instance a Polycold normally runs as cold as -140 to -145C under light loads. If in your case you really want -120C at a moderate load, a different choice of refrigerants may yield better efficiency by giving up the ability to run colder. This would be true for either a cascade or autocascade.

So perhaps in your final application a more specific charge might be in order (i.e. reduce the amount of Argon).

I'm glad you and Pat made the connection, and they had something to suit your needs. I still occasionally do contract work for them, and will most likely be involved with testing whatever unit you guys decide to go with.

One last thing...
18 kw would be at maximum rated heat load for a PFC-1100ST, with less heat load, you'll draw less energy. Also doing a good job of matching your needs to a given Polycold will help keep the energy requirements down. The 1100's use a 10 hp compressor and the next lower model uses a 6.5 hp compressor, with the hp ratings getting incrementally smaller as the model number also gets smaller.