Morphing Air Conditioner into Autocascade System

[godmod]

Why reverse the logic? Wouldn't it work to take a SV that is curentless Opened?
That would also be a plus of security. If the HPCO fails (which is quite unlikeley) the SV would be opened all the time and HP would not rise.

Michael, do you know, did you try, or can you estimate at which HP the unit will settle if the SV is opened (thus the High side is directly connected with the Exp.tank)?
Maybe a solution with a second captube (not too restrictive) from SV to the Exp.tank would be nice to ensure that the presure drop is not too abrupt...
On the other hand the SV closes anyway at 300psig...

[mytekcontrols]

Heh I could get ones with opposite open and close for $25 ish

Opps!! Well I guess I should have waited a bit. Anyway what's done is done.

Adam can you post the source for these here for others to see?

Why reverse the logic? Wouldn't it work to take a SV that is curentless Opened?
That would also be a plus of security. If the HPCO fails (which is quite unlikeley) the SV would be opened all the time and HP would not rise.

Yes it could be done with a normally open solenoid valve, but the cost of this kind of valve is apparently much higher than a normally closed one (besides which, I have some N.C. valves that I got for free).

As for safety in the event of the valve failing to open; this is where the separate HPCO on the compressor power would protect the system. It would take a 2 point failure to have both safeties fail, which is extremely unlikely, and why European Directives and Semi-S8 only require 2 point redundancy in safety circuits for refrigeration systems.

As for cost; I think Chris proposed using a N.O. valve on his new project, and was shocked out of doing so by the price.

[mytekcontrols]

Michael, do you know, did you try, or can you estimate at which HP the unit will settle if the SV is opened (thus the High side is directly connected with the Exp.tank)?
Maybe a solution with a second captube (not too restrictive) from SV to the Exp.tank would be nice to ensure that the presure drop is not too abrupt...
On the other hand the SV closes anyway at 300psig...

I have to confirm this, and it varies as the system begins to cool, but I think if the Buffer Valve were to stay open, the pressure in the expansion tank would settle out at around 275 psig. Of course my tests are presently with the high pressure coming directly from the compressor, and not from the top of the 1st Phase Separator as proposed.

Interesting thought on the captube restriction. I don't really know if this would help much, if any. Kinda think that it is best to buffer abruptly and quickly for the best results (at least with my setup; based on the expansion tank size, and expansion tank captube -to- suction side flow rate).

[godmod]

Yes, those were hypothetical thoughts. I am curious how this solution will work and how it is going to affect the cool down phase. I hope you find time to work on the AC-2 soon! I love this project (gives me a lot of helpful information for my autocascade project).

[mytekcontrols]

Thanks as always for sharing all your ideas. And yes I too want to find the time to finish this project. I also look forward to seeing other peoples autocascade creations, whether or not they were inspired by what I am doing, and sharing here.

What size compressor, and how many stages will you be using on yours?

[godmod]

I am not sure yet whether i will take a rotary or a reciprocating compressor. I think rotary is the type of choice, and i will probalby take a 12ccm one with ~7200BTU. My biggest problem is how to make the pipe-in-pipe coaxial HXs. As for refrigerants i will probably take R600, R290, R23 (would prefer R170, but it is very expensive), r1150 and/or R14 (+R50?). But I also have some doubts, because i do not have any experiance with autocascading systems.
(i don't want to spam your nice thread with off topic posts, if you want we can discuss this via pm)

another questioning i were asking myself for days now: what do you think, will there be performance differences between your evaporator coil and a "normal" (direct die) evaporator? I think so. I thnik the performance would be worse, due to the reduced contact surface to the load and reduced surface where the refrigerant can evaporate. On the other hand, it should definately work with a "normal" evaporator, because your load values are in normal regions (100W-300W).

[n00b 0f l337]

We need different conditions, and really better and bigger evaporators. I've been wanting to get a much larger, almost LN2, pot evap for a while. 2.5" OD, 3-4" tall, stepper all the way up with very large center pole.

[mytekcontrols]

godmod --- Some things to take into consideration are how cold do you really need to go? If you can live with -85C to -90C, then leaving out a stage in what I have built, increasing the flows of the remaining 2 captubes (not counting captube #1), and going with a R-600, R-290, R-170, R-1150 (small amount) should get you in the range of 225-250 watts. Doing this and also increasing the compressor to 12cc would probably take you all the way up to 300 watts.

Just boosting the the compressor up to 12cc with what I've done, should easily yield -100C @ 200 watts.

As for how the test load I am using compares to something like a stepper evap, I'm not sure, but here are the specs for mine:

11 feet x 1/4" od Tubing
Internal volume = 3.53 cubic inches
external volume = 6.48 cubic inches
weight = 1.4 lbs of copper

[n00b 0f l337]

I'd drop the r1150 even in that mix, no need, r290 and r1150 will work out I think.

[mytekcontrols]

I'd drop the r1150 even in that mix, no need, r290 and r1150 will work out I think.

So leave out the R290 all together? Hmm... Yes, but based on what I've been seeing, you'll have to dumb-down the phase separation to make it work (in other words "very inefficient phase separation --- low volume").

In my early talks about using a combo like R-123 and R-23 (80C difference in boiling points), this was being linked for use with an autocascade with extremely poor phase separation (Polycold P-20). Apparently (and it should be no surprise), as the phase separation improves, so does the need for picking out refrigerants that don't take too much of leap in boiling points from one-to-the-other.

Of course this is all theory that I am spouting off the top of my head, but it does have an intuitive basis in what I have experienced.

Edit: This stuff just keeps on evolving.

[n00b 0f l337]

Not the r290, but the r170

[godmod]

Can you say to what extent reciprocating compressors would also be suitable?
I have a bunch of big compressors, e.g. an Maneurop MT-28 (48cc/rev) and many others. although i would rather make a autocascading system with a small (12cc) rotary because of size and energy consumption.

[mytekcontrols]

Not the r290, but the r170

Got you

There would be a 61C gap between the R-290 and the R-1150... might be wise to implement a Partial Phase Separator on the last stage prior to the final cascade condenser (see below). This way some liquid will be passed on to the next stage along with what ever gas is present, there by insuring that a disconnection between stages doesn't occur (as what happened to me on my R-600/R-170 charge test).

Some experimentation will be needed to find the optimum height for the exit tube.

Edit: Notice that this is a modification of my existing upper stage phase separator design.

[mytekcontrols]

Can you say to what extent reciprocating compressors would also be suitable?

No problem using reciprocals. just so long as motor size vs. displacement is reasonable (medium to high back pressure).

[n00b 0f l337]

Yeah I think that would work well, what filters do you use? I think I asked before but I cant seem to find something similar to what you use.

[godmod]

No problem using reciprocals. just so long as motor size vs. displacement is reasonable (medium to high back pressure).

Thanks for the Info.
edit: ok, i already thought this. and of course R404a comps would be better, because of stronger motor (high HP)

This way some liquid will be passed on to the next stage along with what ever gas is present, there by insuring that a disconnection between stages doesn't occur (as what happened to me on my R-600/R-170 charge test).

Some experimentation will be needed to find the optimum height for the exit tube.

The right "length-setting" might be really hard to find. Aren't the liquid levels loaded different than in idle? The problem i see here is that "the gas out pipe sucks" either liquid or gas. There will be an alternating oscillating gas-liquid compositon. I visualize that that way: a constand volume flow out of the strainer through the gas-discharge. the composition will be, depending on the refrigerant level in one moment liquid, until level drops, then gas until liquid is high enough. maybe sometime it will equilibrate, but I think varying load could cause a oscillating system (keyword: feedback). Just thoughts, might also be delusions

[n00b 0f l337]

No I think thats valid, however by charging till you always have spill over, you can fix said problem

[mytekcontrols]

The problem i see here is that "the gas out pipe sucks" either liquid or gas. There will be an alternating oscillating gas-liquid compositon.

No I think thats valid, however by charging till you always have spill over, you can fix said problem

You are both correct

Yes the idea is to lower the threshold for flood-over within the phase separator, but mainly during low-to-moderate heat loads. I believe it is during this time that the system can get in trouble when refrigerants with wide differentials in boiling points are used. It goes into what my original mentor would have called "self refrigeration". I call it inter stage disconnection. There is another technique that has been used, which combines a distributor with a phase separator, although in a system with a single tube-in-a-tube HXC it isn't required.

Basically you want to keep the first stage separator from getting too cold and stealing all the lower boiling refrigerants through absorption into the warmer boiling refrigerants (what I call the liquids, due to the fact that they are charged into the system with either a charging glass or a scale in their liquid state).

The simplest partial separator is to use an inclined tee, with the gas exiting the side connection. Although I'm not sure how well this would work in a very tiny unit (the strainer that follows would still hold a lot of liquid).

Yeah I think that would work well, what filters do you use? I think I asked before but I cant seem to find something similar to what you use.

I get them from Baker Distributing which is a local refrigeration parts wholesaler. I think they are part of a national franchise.

[n00b 0f l337]

Ah I was wondering about no load, freeze back through all the stages I guess...

[tim-]

very nice as already said

i'm actually thinking, since i don't get anything less then 14-17k btu rotarys and up.

i'm thinking of using one of those which is 3-3½times larger then your compressor (27-34ccm). how would you suggest to dimension the captubes? make them 3-3½times less restrictive or is it something else to remember?

condenser and hxes need to be larger as well due to the larger compressor and flows.

regards
Tim

best regards
Tim

[n00b 0f l337]

Figure out how many times your displacement is, take his mass flow figures and multiply by that and then find new captubes to fit your new massflows.

[mytekcontrols]

i'm thinking of using one of those which is 3-3½times larger then your compressor (27-34ccm). how would you suggest to dimension the captubes?

This will link you back to the discussion on captube sizing:

...I also included a captube sizing chart in case you wish to use a different ID.

More on Partial Phase Separators

Here is another idea out of the past, that was used to create a partial phase separator that would meter out liquid as the liquid level rose within the separator. The higher the level, the more liquid would be passed on to the next stage along with the gas.

Metered Partial Phase Separator:

[n00b 0f l337]

Very cool idea indeed. I like it. Sending you an email Mytek.

EDIT: Actually just had an idea since I can't seem to find those strainers...

[mytekcontrols]

Although I have presented several schemes along this theme, my answer would be NO!

The reason being; you'll be sacrificing thermodynamic efficiency in the process. It is always better to design with the best possible phase separation, heat transfer, and minimal pressure drop. If you have to resort to partial phase separators to get your system to work properly, then you're probably sacrificing efficiency in doing so.

Of course sometimes due to the refrigerants that are available (or not available), you are left with no choice. So because of this, I think it is still important to cover the topic of partial phase separators, but it is also just as important to try to design without relying them.

[n00b 0f l337]

So in your opinion, propane to ethylene. Thats -42C to -104C. Or a 62C boiling point delta. Do you think a partial phase separator would be necesary? R600 to R170 is a 89C boiling point delta.