Morphing Air Conditioner into Autocascade System

Been pondering over this... what kind of impact would suction accumulator will have in autocascading system when load condition is not optimal per say. Would it start trapping r-123? Would R-123 be carried along with other gas state refrigerant? Would artificial boiling point of r-123 get much lower due to the pseudo refrigerant condtion? There is only so much volume of liquid r-123 we will have to begin with and I am starting to wonder if R-123 can pool itself in suction accumulator as liquid by itself effectively reducing amount of it flowing through rest of loop :)
Comments Michael? :)

Proper suction accumulator has a small orifice, no?
It would flow thru there with oil I'd imagine.

Quote:

---

There is only so much volume of liquid r-123 we will have to begin with and I am starting to wonder if R-123 can pool itself in suction accumulator as liquid by itself effectively reducing amount of it flowing through rest of loop

---

It shouldn't "pool" in the accumulator because a properly implemented suction accumulator shouldn't hold up much, if any liquid. Some accumulators (like the ones on small rotary compressors) are nothing more than a small receiver which adds additional volume to the returning suction line, and larger ones normally have a dip tube positioned slightly above the bottom. In either case the object is not to hold up liquid, but to allow it to flash off (expand) before entering the compressor. On a small rotary compressor (see image below) the accumulator is of the non-dip tube variety, and is normally plumbed with the inlet coming into the top, and the outlet going into the compressor. This type will not trap any liquid.

Edit: Yes Nol is also right, some of the dip tube style suction accumulators have what I would call metering holes drilled in the side of a dip tube riser which is open on the top. Any gas present would easily make its way out of the accumulator through the dip tube's top opening, and any accumulated liquid would be restricted by the metering holes along the side, thereby returning at a controlled rate.

Quote:

---

Would artificial boiling point of r-123 get much lower due to the pseudo refrigerant condtion?

---

Yes it will. In the CFC days, R-114 would have been the refrigerant of choice for the first component in an autocascade. When the non-CFC era first hit (and HCFC's were still considered to be ok) a combination of R-123 and R-22 (when in proper proportions) exhibited a similar boiling point as the formerly used R-114.

A question on HX length and size for you, plate HXs would work but are a bit costlier, for tube in tube setups is 1/4" in 1/2" good? Is 3/16" in 3/8" too restrictive for say a 3 stage autocascade running r11,r22,r23,r14 or similar?

Quote:

---

A question on HX length and size for you, plate HXs would work but are a bit costlier, for tube in tube setups is 1/4" in 1/2" good? Is 3/16" in 3/8" too restrictive for say a 3 stage autocascade running r11,r22,r23,r14 or similar?

---

Yes I agree, for small systems, especially autocascades, plate heat exchangers might get pretty pricey.

1/4" inside 1/2" will probably work, but I think you'll find that 2 each of 3/16" inside 1/2" will give you even better heat exchange (translates into shorter length required). It's a bit tight, so you'll have to roll out your tubing very straight and be sure to ream the 1/2" real good, otherwise you'll have a heck of a time stuffing the 3/16" inside the 1/2" tubing. For ideal performance using this design, a small 2 hole distributor should be used feeding from the liquid line filter-dryer into the dual 3/16" tubes. However a "pair of pants" style coupling would probably also work fine.

3/16" inside 3/8" will work just fine for the upper stages (cascade 2 and 3). In fact this is what I'll be using on my AC2 Project. Below I have included the heat exchanger designs I'll be using on this project. The trombone shape is to make it fit in the available space, you can wind them any way you want to. If you try to copy the shape I am going with, please double check the bend measurements are correct (I haven't verified this yet).

Explanation of Bend dimensions: The "0" mark corresponds to the zero starting point on a conventional tubing bender. Initially when the outer tubing has been rolled out, you would place a mark for every bend dimension shown, all starting from one end. After stuffing the center tube(s), you would create a 180 degree bend at every mark, moving down the tubing as you go. The best bender to use, is one that allows for the two halves to be separated. This makes it much easier as you are forming the coil, and when moving from one bend to the other (unfortunately my benders don't come apart, so it is a bit more difficult).

Yeah I've put triple 3/16" in 5/8" before and that wasn't fun ;)
Dual 3/16" work would great I think, though quiet a bit of pipe. Then I see you move to 3/16" in 3/8".
I've tried many pipe benders for 1/2" and none really work well sadly, bought two both are horrid. My 3/8" is easy tho on my lever pipe bender so my 3/16" in 3/8" HX's are great. I like the trombone style as well.
Best thing I've found for bending 1/2" is 3-4" PVC pipe or a cylinder of refrigerant.

And I know you're going to ask, so here is the complete piping diagram for my AC2 Project :D

Looks like exactly what I imagined! Though no subcooler or anything? Or think with only 200 watts of load you wont need it?

And yes I also thought you'd want to know what captubes I intend to use (funny how I can read your mind) ;)

These will be my first stab at it, might need to tweak it though. You'll notice that I am using all the same ID size on the captubes, this will keep it more economical.

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

Quote:

---

...no subcooler or anything? Or think with only 200 watts of load you wont need it?

---

Not for this temperature range. In fact I wont be using any Argon in the charge. Without the subcooler (and no Argon) I should get much better load capacity. In order for the subcooler to really work, it does steal about 1/3 of your final condensate, so ideally it is only used when you really need it.

Quote:

---

Best thing I've found for bending 1/2" is 3-4" PVC pipe or a cylinder of refrigerant

---

Keep in mind that bending around a cylinder will oval the tubing. This wont cause any problems with heat exchange, but will create a heat exchanger that is taller, then if a mandrel style bender were used.

What can you do with those flow factors? Ya it will oval the tubing sadly, but I havent a good tool otherwise for bending 1/2" properly. What do you use?

Quote:

---

What can you do with those flow factors?

---

You know to tell you the truth I forget what the flow factors in the chart actually equate to. So you might ask what good is it?

Well actually a lot of good:

1. Find all the flow rates for captubes in an existing "working" system
2. Compare existing system compressor displacement to new design's compressor
3. Re-calculate captube flow rates based on the percentage of displacement change
4. Find appropriate captubes in chart with same flow factor

So take a base working system as shown in Post #87, then decide how much displacement you intend to use, and get the percentage of change from the Tecumseh AJA4512-AXD as used, and apply it.

example:

• Captube #1 = 81"x0.055id = flow factor of 0.671 (extrapolated from chart)
• Let's say the compressor you intend to use is 75% as much displacement as the AJA4512-AXD.
• 0.75(75%)x0.671(original CT#1 flow) = new CT#1 flow factor = 0.503
• 90"x0.050 = a flow factor of 0.502 (from chart) close enough match

Quote:

---

Ya it will oval the tubing sadly, but I havent a good tool otherwise for bending 1/2" properly. What do you use?

---

I use something like this: Grainger 1/2" tubing bender w/1.5" bend radius

The condensing unit for my AC2 Project is based on a Goldstar GWHD6500R window air conditioner, which utilizes an LG QA090CBC Rotary Compressor with a displacement of 0.55 in3/rev (9.0 cc/rev).

Ideally I would have liked a "High Back Pressure" compressor instead of the "Medium Back Pressure" one that comes with the Goldstar. Or in other words I wanted a high starting torque unit. The compromise is that I may need either a larger expansion tank to keep the static pressure low, and/or a buffer valve that bypasses some of the discharge into the expansion tank for a few seconds on start-up. I'll know more when I get to the testing stage (which is still a ways down the road).

Here's a pic of the Goldstar unit. It comes with an electronic control unit with a remote. I haven't decided if any of the electronics will be usable in the final design. What I do like is the all metal cover, and the square shape, so I can probably fabricate a nice flat front panel for it, and get rid of the plastic grill.

Here is a look inside:

Now for some disassembly, where we get rid of the original evaporator, blower fan, and fan shaft. This area will be where the new autocascade stack will reside.

*Whoa*
So you think your only going to need a 9cc rotary to acheive those results?
Really?

Quote:

---

So you think your only going to need a 9cc rotary to acheive those results?

---

I was originally shooting for -100C at 200 watts, but realistically I would expect it to do 125-150 Watts at this temperature. This is based on a comparison of a similar (although not identical) product that Polycold used to produce called the PGC-100. It had a compressor with a displacement of about 2.0 in3/rev and achieved 450 watts at a temperature of -100C. This was about 3.6 times the displacement of the LG rotary compressor I am using (450/3.6 = 125 watts).

The PGC-100 had some other differences that are to my favor. First of all it was a primary+secondary system, with the secondary loop being used for chilling gas. My system is a primary system only that directly cools the evaporator. Without a secondary aspect, there should be less losses. The other disadvantage of the PGC-100, was that it was required to handle at least 1/2 the rated load at a temperature of -110C, this meant that the primary circuit had to be capable of at least -115C under 225 watts of heat load, factoring in the transfer losses between the primary and secondary loop. This also meant that Argon was required in the charge, as well as a subcooler to better utilize it under the high load conditions. Remember in order for the subcooler to work, it must steal about 1/4 - 1/3 of the final condensate.

Edit: My error, Argon was not used in this system, although I do believe a subcooler was still present (don't hold me to this).

So based on these differences, I would not be surprised if I can get -100C operation at 150 Watts, and possibly with a little luck, maybe even 200 Watts.

Sounds good thanks for the explanation :)
Hope you meet your goals!

Looking foward to seeing the end result. Out of interest, why do you use 2 pipes in some places in your design and 1 in others? Is it to keep mass flow high?

I can imagine that the point of that is better heat exchange early on when its really needed and theres alot more of a mix of gases, and later on when theres less of a mix its not as needed.

Quote:

---

Out of interest, why do you use 2 pipes in some places in your design and 1 in others? Is it to keep mass flow high?

---

It is to achieve the maximum surface area for heat exchange with standard sized copper tubing, while still having reasonable volume on both the suction and discharge paths. Basically 2x3/16" od tubes will give me greater outside surface area, then a single 1/4" od tube. An alternate choice would be a single 5/16" od tube which would leave very little annular space within the 1/2" outer tube, and be very difficult to slide it in. I do believe the use of dual 3/16" od tubes still has more surface area then the single 5/16" od tube (if I am not mistaken).

As we get farther downstream, and have gone through a few phase separators, the overall volumetric sizing of the heat exchangers can be reduced, since we have less refrigerant passing through (less mass flow). That is why the last 2 heat exchangers can be a single 3/16" od tube within a 3/8" od outer tube.

I fabricated the first heat exchanger today (Aux Condenser) see it below:

Sorry, you misundertood me, I meant on your actual piping diagram. From strainer 1 to phase sep 1; and phase sep 1 to strainer 2 there is 2 pipes the rest of the piping is on its own. Why?

Maybe we are both misunderstanding each other :shrug:

The first phase separator is made out of a 1 5/8" od copper tube with end caps. The bottom of this 1st phase separator feeds the first Bullet strainer/captube.

Phase separator #2 and #3 are made directly out of Bullet Strainers (they are one in same). Out of the bottom of each Bullet strainer we feed into a capilliary tube. Which in the case of captube #1 - captube #3 feed into the suction cross tube at the top of the next heat exchanger. Captube #4 feeds into the evaporator.

The first 2 heat exchangers (aux and 1st cascade) have 2 inner tubes which are part of the discharge (or high pressure circuit).

The last 2 heat exchangers (cascade #2 and #3) have only 1 inner tube associated with the discharge circuit.

The discharge circuit is = Red
The suction circuit is = Blue

Based on what I just stated; can you be more specific with your question?

Edit: Freddie123 I just realized you also misunderstood a refrigeration system schematic I posted earlier. Just to be clear, the AC2 Project diagram is not a schematic, but an actual piping diagram. In other words it is a cut-away view of an actual system so-to-speak. So a piece of tubing is represented as 2 parallel lines, unlike a schematic where it would be shown as a single line.

Can't help but thinking... finned tubing :P Of course, pressure drop will be there... but considering distance becoming shorter, might be moot issue :)

Quote:

---

Can't help but thinking... finned tubing

---

Something like Noranda makes. Yes it has excellent heat exchange, but sometimes the small passageways can hold on to oil in the wrong place, and next thing you know, you're blocked by frozen oil. I think it was about 21 years ago we tried building an entire stack with this tubing. It worked great for a few hours, and then suddenly started loosing capacity. It would freeze up about 2/3 up from the bottom.

Of course this was before we ever heard of a coalescing oil filter such as what Temprite makes. And we also didn't know it until much later, that the new compressor we were using was a major oil pumper. Had to get the manufacturer to add a 3rd ring for oil control.

Excellent work! If only 1/2" benders weren't $100!

wow, subscribed :yepp:

too bad we don't have ultra low temperature class at school, only "big/commercial" refrigeration.

Quote:

---

Originally Posted by mytekcontrols

Something like Noranda makes. Yes it has excellent heat exchange, but sometimes the small passageways can hold on to oil in the wrong place, and next thing you know, you're blocked by frozen oil. I think it was about 21 years ago we tried building an entire stack with this tubing. It worked great for a few hours, and then suddenly started loosing capacity. It would freeze up about 2/3 up from the bottom.

Of course this was before we ever heard of a coalescing oil filter such as what Temprite makes. And we also didn't know it until much later, that the new compressor we were using was a major oil pumper. Had to get the manufacturer to add a 3rd ring for oil control.

---

Yes I can see that but there are perpendicular as well as horizontal finned tube :) Perpendicular one I can see problem but horizontal one might not have that... :)

No doubt, time to get r23 me thinkies ;)
And watching with incredible rapt attention, do you think those capillary lengths are good though? Any reason to them or calculation? Seem sorta long after reading other lengths from Cryo-Tek.

Quote:

---

...do you think those capillary lengths are good though? ...Seem sorta long after reading other lengths from Cryo-Tek

---

Actually they have slightly more flow then what I calculated from Cryo-Tek's example. Remember that the system that Cryo-Tek's captubes were specified for has a compressor with 3.6 times the displacement of mine.

If you are concerned about the length, also remember that I am using all the same ID captube. Cryo-Tek's example was using several different ID's, which is fine if you have many different ones to chose from. I tried to keep it simple by only using a single ID and ordered a 100 foot roll. Also keep in mind that captubes work best when the length is kept somewhere between 5 and 15 feet. Less then 5 feet starts to act more like an orifice, and can produce greatly different flow rates with only a few inches of change.

Quote:

---

No doubt, time to get r23 me thinkies

---

Ethane works just as well, and can be used as direct substitute (might be cheaper).

Quote:

---

Originally Posted by mytekcontrols

The first 2 heat exchangers (aux and 1st cascade) have 2 inner tubes which are part of the discharge (or high pressure circuit).

---

That explains it. Cheers Michael, that HX is really nice looking too

Nah ethanes more expensive it would seems.
Well no actually hmm, just checked matheson...
$100 will get me 3lbs of ethane. Is it just as good as r23? I'd imagine since its lighter it might not be able to hold a load as well. Also, what about ethylene? would r170/r1150 work decently theroetically?

Quote:

---

That explains it. Cheers Michael, that HX is really nice looking too

---

Thanks Freddie I'm glad you brought it up. Often times I find it helps me do a better job at documenting a project, when people ask me to explain different aspects of it. Thanks also for your appreciation of the HX, because it was a bit of a struggle to build it that way. It sure would have been nice if the lever bender I was using, could have been separated into 2 parts after doing each bend. I'm almost wondering if I could make my own bender from something like a pulley.

Quote:

---

$100 will get me 3lbs of ethane. Is it just as good as r23? I'd imagine since its lighter it might not be able to hold a load as well. Also, what about ethylene? would r170/r1150 work decently theroetically?

---

I do believe ethane (R-170) works just fine as a substitute for R-23, boiling points are virtually the same, and on the positive side it has better solubility with the oil. I made up a charge last week for a large Polycold unit I was working on, and increased the ratio of ethane as compared to the R-23 (15 psi R-23, 65 psi ethane) it worked quite fine, and even exceeded specification. I suspect I could have just gone with 80 psi of ethane, and eliminated the R-23 all together (have to try that some day).

As for ethylene (R-1150) being used in combination with the ethane, yes this would also be a viable mixture, although it wont get you quite as cold as using R-14 which boils 15C colder (R-1150 = -104C, R-14 = -120C).

Edit: Correction, R14 boils at -128C.

Well what about r14 in combination.

Quote:

---

Well what about r14 in combination.

---

No benefit in using R-1150 (ethylene) and R-14 together.

Edit (10/10/2009): I have recently discovered that R1150 and R14 can play well together. Of course this depends upon what other refrigerants are being used in combination with, and how many stages are in the AutoC. Under the right conditions, the R1150 and R14 combo will yield evaporator temperatures with a smaller gradient from the the inlet to the outlet.

However R-170 (ethane) and R-14 would be fantastic :up:

Just trying to see what might be decent with what I have ;)
Hmmm r170 or r23....

I'm going to see what Airgas will do with me tomorrow for R1150/R170, just opened an account with them and crossing my fingers that they'll be able to supply me with decent prices.

Mathesons got good selection and great prices, but that hazmat fee is driving me nutters.
They have good prices on ethane, argon, butane hell just about everything.

I tried r1150 in an autocascade but didn't have good results with it. r170 works much better. I think the leap between r290/r507 etc and r1150 is too big.

I dont think its so much the leap as the lack of a gas like r11 or r123 that killed your setup thuogh Unknown.

How about a $4.00 3/8" od tubing bender? Only took a slight modification of an off-the-shelf 3" garage door pulley (used file to elongate holes in bracket). It works great for making a trombone style heat exchanger, clamp the pulley's shaft nut in a vice, and just walk the part you are bending around the pulley.

The v-shaped pulley helps to prevent the tubing from flattening out, and the bend radius = 1.5" which matches the one on my 1/2" bender. This ends up producing a better match for the 1/2" od heat exchangers, and allows it to sit perfectly on top of them (makes for a cleaner stack). So I'll be scrapping my original concept drawing for Cascade #2 and #3. I will provide details later.

Well here it is:

Here's a close-up of the completed pulley bender. The slot allows you to adjust the bracket for a tight fit on the tubing, this helps prevent the tube from kinking. The secret to using it, is to try and keep the part you are bending underneath the bracket, and maintain tension on both sides of the tubing (It can be a little tricky at first).

And here's what the first HXC produced with the pulley bender looked like (I miscalculated on the length, so ended up adding a swaged extension tube to make it right).

Edit: You'll notice that I reduced the number of turns by one from my original drawing. I needed the extra room, and felt that a little less length on the last 2 cascades was acceptable.

Reminds me of Tim Tayler on "Tool Time" :)

Great work! I look forward every morning to looking on this thread!

Regards,

Philipp

PS: I am planning on an autocascade in a very similar way like your drawing (I studied all the old threads by you and Cryo-Tek). Now that I have a starting point about capillary length (thanks to you), I think about the charge. How do you charge a unit like this? Do you mix the gasses up previously in an empty bottle and then charge from there or do you charge one refrigerant at a time up to a certain pressure? Or do you charge a unit while running up to certain points, eg. Ref#1 until HX#1 reaches -XX degrees, then Ref#2 until HX2 reaches -xx deg. and so on? I would appreciate the info very much.

Quote:

---

Reminds me of Tim Tayler on "Tool Time"

---

I wonder if Binford ever made an autocascade? If they did, you know it would be powerful and get very very cold :sonic:

Quote:

---

Great work! I look forward every morning to looking on this thread!

---

I'm glad you are enjoying it :)

Quote:

---

How do you charge a unit like this?

---

If you have an electronic scale, and your unit is not too heavy:

• Place the unit on the scale
• Zero the reading
• Charge in your 1st, 2nd, 3rd... liquid refrigerant (R123,R22,ect.) into discharge side only
• When all liquids have been charged, open all valves on manifold, allow pressure to balance, take pressure reading
• Add gases by pressure (allow 5 minutes for soak-in before going on to next gas, top off if needed, wait another 5 minutes, and repeat top off if needed --- repeat until pressure no longer drops)

If your unit is too heavy for a scale, then use a cylinder as a transfer tank. Same procedure as above, but when adding gases shake the cylinder to speed up soak-in.
Do the following when the entire charge has previously been put into the transfer cylinder:

• Connect hose from cylinder to unit
• Open valve, purge line with refrigerant
• Invert cylinder (for liquid), and open discharge valve only on manifold
• When flow slows down, open all valves on manifold and allow pressure to balance
• If compressor has suction valve, front seat it and close off discharge valve on manifold
• Run compressor and allow it to suck in remaining gases from cylinder

If you have a charging glass instead of an electronic scale, you can pre-charge the glass with the liquids, add them first, and then proceed with the gases, same as was done in the first example.

If your system has an expansion tank, you'll need direct access to it when adding your gases (a way to bypass the restrictive captube).

Of course this all assumes that you know what your total charge is going to be. When first developing the charge, you'll want to start out with about 2/3 of what you think you'll probably need. Later when the system is running, and based on temperatures seen at the exit of each heat exchanger, you can test adding small amounts of the different refrigerants. If the compressor discharge is running hot (0ver +125C), then very slowly add in additional liquids an ounce at a time through the suction port.

When adding gases, it is good to add these from a standard size cylinder with a known volumetric ratio as compared to the system volume.
For instance; before charging your refrigeration system, fill a small cylinder with nitrogen or argon to 100 psi. Put a gage on the cylinder, and a gage on the refrigeration system. Allow gas to flow from the cylinder into the refrigeration system until the system pressure has gone up 5 psi, and then stop the flow. Take note of how far the pressure has dropped in the cylinder (i.e.; dropped from 100 to 85, or 15 psi). Now you have a way to measure in a specific amount of gas (every 15 psi taken out of the cylinder will equal a 5 psi increase in the refrigeration system's static balance pressure). Of course this would be added into the suction side of a running unit.

Later as I get into the charging phase of my unit, I'll give you the refrigerants and amounts I will be starting out with (haven't gotten that far yet).

:confused: Have I forgotten anything? It's late, so my mind is getting a bit fuzzy, and this is a lot of stuff to cover.

Picked up a 4" garage pulley and damn your right, just what I've been looking for!

Isnt the internal volume of those bullet strainers too low? Flow rates through a low volume tube like that would be too high to allow for phase seperation, wouldnt they?

Or are these bullet strainers larger than we are used to seeing?

Regards

John.

Edit: Meant refrigerant stream velocity, not flow rate.

I think we've been overestimating the need for this kind of separation. If you have a liquid seal at the bottom, gas will leave gas passage ;)

Quote:

---

Originally Posted by n00b 0f l337

I think we've been overestimating the need for this kind of separation. If you have a liquid seal at the bottom, gas will leave gas passage ;)

---

But liquid droplets can leave at gas outlet if the velocity isnt slowed or the droplets hit an impingement screen. And gas disolved in liquid will escape the liquid seal if not given enough time to come out of suspension;)

Regards

John.

Which is true, but the way his seps are setup the mix comes in and heads to the bottom, a liquid seal is then created.

Quote:

---

Isnt the internal volume of those bullet strainers too low? Flow rates through a low volume tube like that would be too high to allow for phase seperation, wouldnt they?

---

Actually the volume is quite large as compared to the ID of a 3/16" od tube that feeds it, in fact my Bullet Strainer Phase Separators are probably 4-5 times the volume of what Polycold used on their P-100 and PGC-100 product lines for the upper phase separators. So what did they use? Believe it or not, a 5/16" copper tee, angled downward with the side port passing gas, and the bottom port passing the liquid. When you get into the upper separation points, the volume of refrigerant is considerably less than what you see in the first stage (hence the larger size of the 1st separator as depicted on the piping diagram, and the larger flow of the 1st captube).

Believe me, the bullet strainers will do just fine as phase separators :up:

I know I have been showing great progress on this project this last week, but the holidays are over until Christmas, and now I've got to get back to my real job :shakes:

So for the next couple of weeks there wont be much new stuff to show (sorry PhilippF). However there is a very real possibility that the new HXC stack will be completed by the end of this year :cheer:

As it comes together I am getting more, and more excited. I think it will be very clean build, and something quite amazing if it meets my expectations.

...must...complete...unit...must...complete...unit ...

lol PhillippF. Maybe that'll work :up:

Hey Michael!
http://www.extremecooling.net/index/...okaskade1_.jpg
http://www.extremecooling.net/index/...tokaskade_.jpg
Look at my autocascade. My phasesep is quite huge with steelwool inside - but don't seperate as i want. I have two sightglasses, in each liquid-line.
It is charged with R290/R23. I get much R290 in the liquid-line from the r23-stage.

Any advice?

Thank you very much!

lol i already put in my subconstions that i have to reach -80c this year!

@ mytekcontrols: so the r22 and r123 they are mixing and create a ''new freon'' kind off .why not use just 1 r*** like lets say r134? is there any reason, could u explain a bit pls(i saw low presures and high temps for condensor so there is something ;))
would using a real extra stage just for r123 improove things?

second question: is it ok to use 0,7mm(0,028''?) cap lines?

thx

I am no expert on auto-c (obviously) but there was some patent related to using the HFC only for autocascade. R-134a was definitely mentioned in there.
Thing would be imperical data will largely have to be rebuilt around new gas, and time and money spent getting to it while R-22 is still widely available.
Also, R-134a's efficiency isn't quite that good and will change balance of current design (cap tube, charge amount, etc) most likely (as well as PT... which if you noticed is shared among stages...)

PS) I also remembered something, some of the gases could fractionate (how much does it matter? not sure) and cause unwanted behavior too... don't think it will matter much in grand scheme of things but it surely would not be easiest thing to develop without good idea on what to look for.

Beleive that mix is r123, r134a, r508a/b, r14.

Quote:

---

Originally Posted by n00b 0f l337

Beleive that mix is r123, r134a, r508a/b, r14.

---

what the........ 123 and 134a? no r123 and r22?
are u using those types mytek?

thx,
q

Thats for a more environmentally friendly mix I beleive.

Hi Moc thanks for joining the discussion :)

As for your autocascade: it's not the fault of your phase separator that is causing the problem, but the lack of a HXC before it (Aux Condenser). Coming in at room temperature is the limiting factor for good separation. If you put another plate HXC before it (exchanging with the suction return from the existing cascade plate HXC), you should see somewhere around -30C at the phase separator. Give it a try and let me know what happens.

Quote:

---

mytekcontrols: so the r22 and r123 they are mixing and create a ''new freon'' kind off .why not use just 1 r*** like lets say r134?

---

If I could get R-114 (boiling point = 3.6C), then this would be what I'd substitute for the R-123/R-22 mix. However I don't believe there is a good alternative available, at least not in a freon. R-134A might work, although it does boil off at a colder temp, and probably wont work as well as either R-123 or R-114 in transferring heat at the air cooled condenser.

Edit: HFC R236fa is a good replacement for R114, although it will necessitate the use of POE oil in the system.

Quote:

---

is it ok to use 0,7mm(0,028''?) cap lines?

---

Well it all depends on what displacement of compressor you intend to use, and what flow rates you need. Go back a few posts in this thread and you'll see a chart and a discussion on captube sizing.

Quote:

---

what the........ 123 and 134a? no r123 and r22?
are u using those types mytek?

---

I'll be using R123, R22, Ethane,R14 as my first mix. And to improve the oil return, I'll be using POE since it is quite miscible with all my choices. The R123/R22 liquids will be a 50/50 ratio, and probably also the Ethane/R14 gases as well. This will be my starting point, and dependent on temperatures seen across the HXC's, additions will most likely be needed.

Speaking of temperatures across HXC's...
I said this before, but I'll say it again, when developing a new autocascade design it is very informative to thermocouple the In & Out of everything within the refrigeration circuit.

So I'll end up with the following TC assignments:

1. Compressor Discharge
2. Compressor Suction
3. Air Cooled Condensor-OUT (Liquid Line)
4. Auxiliary Condenser Suction-IN
5. Phase Separator#1
6. Cascade Condenser#1 Suction-IN
7. Phase Separator#2
8. Cascade Condenser#2 Suction-IN
9. Phase Separator#3
10. Cascade Condenser#3 Suction-IN
11. Cascade Condenser#3 Discharge-OUT (Final Strainer)
12. Evaporator-IN
13. Evaporator-OUT

This will give me a very clear picture of how efficient each stage is performing, and will help reveal the effects of refrigerant additions.

If I get it right: You dont charge every refrigerant in a row but start with 2 mixtures (R123/22 and R170/14) and change the ratio of these two mixtures in the tuning process?

EDIT: I wanted to charge the following way (please correct me if I tell nonsense):

Vacuuming and then breaking with R123 up to 1 barg (2 bara). Adding R22 up to 3 barg (4 bara) and then adding R23 up to 5 barg (6 bara). I imagine this will be a nice starting mix. Then turning the unit on and watching the first HX (where R22 mainly should evaporate) and perhaps add R22 until I have no significant temperature differential along the HX (I picture then being enough R22 to be there to cool the whole HX). Then repeating the same thing with the Evap (for R23). Do you think this will work?

Regards and thanks,

Philipp

Going by pressure would be really hard to control to be quite honest with you unless you are sure of charging condition (R-123 would most likely not be issue as you will end up liquid charging most likely, R-22 would be as it will be from liquid -> gas, R-23 most likely not since it will be gas -> gas typically).
Go by weight... and pressure :) At least that's how I played on last mixed refrigerant build... (nearly a year and half ago or something...) Kind of failure back than as goal was to lower temp under load but I ended up gaining capacity -_-;
PS) Hey Mytek, how do you dispose of mixed gas? Link to place that are nationwide chain (or something like that) disposes of these mixed gas specifically? Only places I found were pretty damned remote and I am tired of having this mixed gas of recovery tank for.... damned long taking space.

Quote:

---

If I get it right: You dont charge every refrigerant in a row but start with 2 mixtures (R123/22 and R170/14) and change the ratio of these two mixtures in the tuning process?

---

PhilippF look what time it is! Why am I still awake? I keep hearing a message in my head saying to complete unit... complete unit... I'm seeing spinning disks... ;)

Seriously though, I talk about the ratio between the two mixtures, but I was just comparing what I call the liquids (R123/R22), and the gases (R170/14) as two separate groups, and applying a ratio of proportion. The liquids will be charged into the system by weight, and the gases by pressure. Each component is charged into the system (or transfer tank) one after the other, or in your words "in a row".

Quote:

---

I wanted to charge the following way (please correct me if I tell nonsense):

Vacuuming and then breaking with R123 up to 1 barg (2 bara). Adding R22 up to 3 barg (4 bara) and then adding R23 up to 5 barg (6 bara). I imagine this will be a nice starting mix. Then turning the unit on and watching the first HX (where R22 mainly should evaporate) and perhaps add R22 until I have no significant temperature differential along the HX (I picture then being enough R22 to be there to cool the whole HX). Then repeating the same thing with the Evap (for R23). Do you think this will work?

---

As I said, the liquids are charged by weight. Then adding to the static pressure of the liquids in the system, you would add the R23 (in your case).

However if you do build a system with 3 cascades, you'll need something like R14 as well to get any appreciable cooling at the end (-110C to -120C system). If you don't want to use R14, then eliminate cascade#3 and phase separator#3. Of course this would be a -75C to 80C system.

Edit: Although having one less cascade wouldn't preclude the use of R14, it just wouldn't be quite as efficient.

When tuning the system by adding refrigerants the object is to minimize the temperature differential between the Suction-IN vs. the Discharge-OUT of each HXC, while still having a high differential from one end to the other. See example below:

Quote:

---

Hey Mytek, how do you dispose of mixed gas? Link to place that are nationwide chain (or something like that) disposes of these mixed gas specifically? Only places I found were pretty damned remote and I am tired of having this mixed gas of recovery tank for.... damned long taking space.

---

A lot of times I can re-use it if it's not too bunged up, either as a top-off to a low charge, or sometimes by using a cold trap to separate out the liquids. Whatever can't be recycled internally, yes does unfortunately get disposed of by a refrigerant recovery place (I don't recall the name of the firm that my place of work uses). I am lucky, since I can dispose of my experiments along with their unusable stuff free-of-charge.

Time to get a charging scale ;)

Quote:

---

Originally Posted by mytekcontrols

I am lucky, since I can dispose of my experiments along with their unusable stuff free-of-charge.

---

I'll stick with R290/R170/R1150. Yay hydrocarbons :D

mytek, you use r123, so if i intend to use hydrocarbons (290/170/1150), just like gomeler, would you recommend using R-600a or even R-600 as the refrigerant with the highest boiling point?

PS. I am following your thread eagerly, I like the high quality information you provide :up:

Thanks for sharing all your knowledge!

I'd say from my limited understanding, R-600 or R-600a would be ideal choices for the auxiliary cooler for the initial separation. Might have the term incorrect for the aux cooler, but it was a pre-stage HX meant to max out the condenser and to help increase efficiency. Feel free to correct me if I am incorrect, been a while since I've thought about this stuff.

Yep I'd say r600 as its a higher boiling point.

Quote:

---

I'd say from my limited understanding, R-600 or R-600a would be ideal choices for the auxiliary cooler for the initial separation.

---

Quote:

---

would you recommend using R-600a or even R-600 as the refrigerant with the highest boiling point?

---

Quote:

---

i intend to use hydrocarbons (290/170/1150), just like gomeler

---

Hydrocarbon Refrigerants that were mentioned:
R600 n-Butane -0.5C
R600a IsoButane -11.7C
R290 Propane -42C
R170 ethane -88.6C
R1150 ethylene -103.7C

I think between R600 and R600a, I'd go with the R600 (remember earlier I mentioned that R114 would be an excellent warmer boiling candidate to replace the R123/R22 combo).

In the CFC days, a polycold charge consisted of:
R114 +3.6C
R13 -81.5C
R14 -129.9C (hmm this looks colder then I remember -- maybe check other MSDS's)
Argon -185.7C (no it wasn't condensed, just absorbed)

I think a charge made from R600, R170, R1150 comes the closest to the first 3 CFC choices, and would be the simplest combination that would work for -100C operation. I would go with this first, and see what you get. Using only 3 components will make it a lot easier to tune. After you have gotten all that you can get from this combination, go ahead and try adding a small amount of R290 to see if it buys you anything.

Don't expect too much load capacity at -100C, since you will be operating close to 0 gage pressure, and not have a tremendous amount of mass flow. On the other hand you should see very good performance at around -90C to -95C.

Quote:

---

Might have the term incorrect for the aux cooler, but it was a pre-stage HX meant to max out the condenser and to help increase efficiency. Feel free to correct me if I am incorrect, been a while since I've thought about this stuff.

---

I call the first HXC (before the 1st phase sep) an Auxiliary Condenser since it is essentially just an economizer, or what I believe some of you also call an SLHX. All the following HXC's are considered to be Cascade Condensers. Hey gomeler I read your article on autocascades... very nice :up:

you said, that you recommenc a temperature difference of approximately 50°C between the refrigerants. Isn't the difference between R-600 and R-170 a bit high? According to a P/T-chart, ethane has a pressure of 27bar at 0°C :eek:
What about a mix of R-600/R-290/R-1150 or R-600/R-507a/R-1150? That would be closest to the "50°C-Rule"...
How much does Ethane in the US or in other countries than Austria cost? Here it is about 280€ for 3kg. A bit expensive :(

In the r114,r13,r14 mix, would r12 in between not help a little? It seems the jump is rather high, even if r114 and r13 might make a mix in the r12 range. You seem to skip that stage as well, r600, r170, r1150. Would r290 not help that much? Or give better load capacity?

[offtopic]

Which program do you use to make your designs?

Paint?

Quote:

---

Originally Posted by godmod

you said, that you recommenc a temperature difference of approximately 50°C between the refrigerants. Isn't the difference between R-600 and R-170 a bit high? According to a P/T-chart, ethane has a pressure of 27bar at 0°C :eek:
What about a mix of R-600/R-290/R-1150 or R-600/R-507a/R-1150? That would be closest to the "50°C-Rule"...
How much does Ethane in the US or in other countries than Austria cost? Here it is about 280€ for 3kg. A bit expensive :(

---

I recommend r600/r290/r170. maybe add some r1150 to make a r170/r1150 mixture but I would add another stage for it.

Quote:

---

Originally Posted by mytekcontrols

I call the first HXC (before the 1st phase sep) an Auxiliary Condenser since it is essentially just an economizer, or what I believe some of you also call an SLHX. All the following HXC's are considered to be Cascade Condensers. Hey gomeler I read your article on autocascades... very nice :up:

---

That's the term I was looking for, auxiliary condenser! Like I said, it's been a long time since I've bent my mind around this material. I'm actually in the process of revising all my articles as I feel they were written when I had a much smaller understanding of the principals behind this stuff. I still do, but it's a slighter wider understanding and actually includes some experience now :D

Quote:

---

ou said, that you recommenc a temperature difference of approximately 50°C between the refrigerants. Isn't the difference between R-600 and R-170 a bit high?

---

You guys are getting good at this :)

Actually although I did say 50C spread is a good starting point, I also think I said something about not all designs adhere to this. Another way to put this is that if you are starting with a blank slate, with no past experience in autocascades, then keeping the spread between refrigerants to approximately 50C will give you a high likely hood of success. However it is not written in stone, and mixtures with far different spreads have been proven to work quite well (R114, R13, R-14, Argon is one of them). I was simply suggesting a hydrocarbon version of the highly successful CFC mixture that Polycold used to use. No guarantee that it will work the same, but it certainly seems likely that it can. However I also introduced the possibility of adding an additional component to help bridge the large gap between R600 and R150. Not that this will necessarily be needed, but something if needed will save the charge.

Quote:

---

Which program do you use to make your designs?

---

I use DeltaCad for my diagrams.

are there any gas mixture rules of advice and/or any captube flow rate procedures you would take with single stage designs?

Quote:

---

are there any gas mixture rules of advice and/or any captube flow rate procedures you would take with single stage designs?

---

I'll have to defer this to some of the other people on this forum, who have much more experience with single stage systems.

Quote:

---

What about a mix of R-600/R-290/R-1150 or R-600/R-507a/R-1150? That would be closest to the "50°C-Rule"...

---

This is referring to:

Quote:

---

The best initial approach I can suggest (and could perhaps be considered a rule of thumb), is to try to pick refrigerants with a difference in boiling points of approximately 50C (+/- 10C). Of course an autocascade will work even if these conditions are not adhered to, but the idea is that this should give you the best initial results. The rest comes down to intuition based upon a lot of experimentation (or should we say a "gut" feeling). Some people also call it black magic.

---

Which was from Project Whop Post #139 it was my mistake to call this a "rule of thumb", which has been reinterpreted into a "Rule". In the future I will have to be more careful how I word things.

What I really meant to say, was that this is a good "Starting Point", but unfortunately bunged it up by adding the additional qualifier of : "and could perhaps be considered a rule of thumb". Sorry guys :(

Anyway moving on with my tail between my legs, what has been observed in the past, is that refrigerants with widely different boiling points can still have great success at working together in an autocascade. I talked about this earlier in Post #8 where I discuss the possibility of using only R123 and R23 to build a -80C system (a CFC version of this would be R11 and R13). Below is a schematic of how little is needed to make such a charge work (single-stage autocascade --- as simple as it gets).

10 seconds q' :D can u pls tell me fast i mean u got experience enought for this small 1: applying your last schematic, the above 1, using r11, r22 and r23 and a heat load of ~200watt, what size compressor do i need?

thx for all the good info's that u put here for us;)

mytekcontrols thanks for posting such usefull informations and tips. You explained many things that i had in my mind with simple words. Thanks a lot.

Quote:

---

Originally Posted by mytekcontrols

What I really meant to say, was that this is a good "Starting Point", but unfortunately bunged it up by adding the additional qualifier of : "and could perhaps be considered a rule of thumb". Sorry guys :(

---

No, No. You did make it clear that it is just a rough way to go with the 50K difference. I just wanted you to tell me if you know or think if one or both of these mixed could work :)
Thanks for the drawing, thats exactly how i will build it. Hopefully i won't fail, but who knows, autocascading is a science :shrug:

Yes Mytek we really cannot thank you enough for bringing your expertise here.
I am going to order r170 I think, and attempt to find a r600 source. Hydrocarbons are nice and friendly, and flammability isnt really a concern.

Quote:

---

10 seconds q' can u pls tell me fast i mean u got experience enought for this small 1: applying your last schematic, the above 1, using r11, r22 and r23 and a heat load of ~200watt, what size compressor do i need?

---

I'll address the part about "what size compressor do i need?" with it all depends on what temperature you require while under a 200 watt load, and of course somewhat dependent on what refrigerants you'll be working with, and how much heat exchange + pressure drop will exist. Let me just say that it gets complicated real fast, if you need a specific answer to that question, and to be quite honest, I'm not the guy with that kind of answer.

Edit: Check out the project described at the very beginning of this thread, it may help to steer you in the right direction with this.

I'm more like the guy that works on cars in his garage, in his spare time. Although he may have done it for years, he is not in the same league as the people that actually designed the engine itself. Sure he can maybe take an engine, re-build it, couple it with a transmission, drive train, perhaps build a frame, and figure out how to get it all to work fairly well together. But bottom line is, he probably does so more by feel, and a bit of luck.

However I do realize unlike some (or many) of you, I also have the luxury of having easy access to a vast array of refrigerants that can be quite expensive, and hard to get on an individual basis. So of course you are going to want to get it right the first time. So that is why with this build, and the one before it, I will show you at least 2 working versions of autocascades. This comes inclusive with everything needed to replicate my results: Compressor specs, Heat Exchanger details, and Refrigerant Charge. The first one which is of the same design as the diagram I just showed you, was designed for a specific application, and is locked into that use. So it can no longer be used by me to test alternative charges. However this second version starting out with 3 cascades, is for my own personal satisfaction, and is not being built for anything specific. Also unlike the first build, this one will be load testable.

If you choose to go off on your own, and build something different, and use a different charge. I wish you great success, and hope that whatever I have shared on this forum will have helped you in some small way. But please do understand that I am not an "expert", and have never, nor will ever claim to be so. I do this strictly for the fun of it, and also to share what I have learned with others.

When it comes to using hydrocarbons in autocascades I am the newbie on this one (other then R170). There are other people on this forum that have a much better feel for this sort of thing, and I think some have showed great results on actual working builds. LukeXE comes to mind, although I forget what refrigerants he has been using. Please tap into these other valuable minds, and don't just take my word for it.

I thank you all for the following acknowledgments (I hope I didn't leave anyone out, sorry if I did) :up:

Quote:

---

thx for all the good info's that u put here for us

---

Quote:

---

Yes Mytek we really cannot thank you enough for bringing your expertise here.

---

Quote:

---

Thanks for the drawing, thats exactly how i will build it. Hopefully i won't fail, but who knows, autocascading is a science

---

Quote:

---

mytekcontrols thanks for posting such usefull informations and tips. You explained many things that i had in my mind with simple words. Thanks a lot.

---

Stay tuned for my progress on my AC2 Project :)

You are being too modest Michael :)
The experience and gathered set of data can't be denied.
As you know, lot of this heat flux and flow, etc are being calculated based on experimental data (unfortunately, most of equations available are close but no cigar as complexity of behavior under gradient temperature, heat flux, and surface tensions are not something that has been fully analyzed yet).
This comes as even more of complication I bet when it comes to auto-cascade. There might be equations that people use that are pretty close but I bet last 20% of it comes from years of experience. (in fact, neural network AI is quite often used in industry it seems in field of cryogenic... which is nothing more than emulating what human mind is capable of).
When such new variables comes in as different gases, etc... of course, it might be some larger amount of change until you get it under control in your mind :P
We really do appreciate your contribution even if there are some things overlooked as it is still TONS better than what is available to use easily (nope... after reading 5th cryogenic book... lol... you still have a lot of imperical data that is useful to us than these books do... haha)
Many thanks to people like you and Detroit AC who are in field and spend your valuable time to enrich the community again.

To me, it seems that Mytek is valuable because hes got a lot of first hand experience with exotic refrigerants and knows the in's and out's of cryogenics in a way that users here dont. Not saying the regular phasers here dont know anything, Mytek just has a lot of first hand knowledge that us here cant get from books. :yepp:

If only Mytek had a "search" function so he didnt get bombarded by questions. ;) :rolleyes:

I love watching projects like these and reading posts about theory and planning of refrigeration like this... I learn a lot more than just reading threads about "Why doesnt my MachII like my QX6700" or "My Vapo is broken." Its almost a privilege to stand back and watch great minds converse like this.

:up: to you guys.

Quote:

---

Originally Posted by jinu117

Many thanks to people like you and Detroit AC who are in field and spend your valuable time to enrich the community again.

---

:yepp:

On a slightly off topic note, is it just me or butane nearly impossible to get here at a decent price. In UK it seems it'll cost you $30 or so for a big cylinder, like we do in the USA with propane, but butane here I've gotten quotes for $300+ for 99% purity n-butane.

Aren't there cylinders of butane used for camper stoves or are they a dirty blend of various hydrocarbons? Not talking about the green camper fuel cylinders as those are nearly pure propane.

Here in Europe Butane is quite common for camping stoves, but its a wild mix of n-butane, iso-butane, propane and some unidentified hydrocarbons, so I would not recommend using it. Also lighter gas (for lighters, you know what I mean) is mostly butane, but I wouldnt recommend it either.

well.. I can get 100% pure propane, n-butane, iso-butane from gas supplier here ... since it's used in different kind of heaters.

Yeah well I cant seem to find that ;)

Quote:

---

You are being too modest Michael
The experience and gathered set of data can't be denied.

---

OK perhaps I am, but It's always made me a bit uncomfortable when too much is riding on what I say. And although within this forum I may sound quite confident in my answers, some times I am shooting from the hip, maybe saying something that I hadn't let my mind fully wrap around around yet. Or another way to put it, would be to say I am thinking out loud.

Quote:

---

...in fact, neural network AI is quite often used in industry it seems in field of cryogenic... which is nothing more than emulating what human mind is capable of

---

Have you guys put one of these together yet, and tapped it into these forums? I think it would be a great idea, and while you are at it, add Polizei's search function request to it :D

By the way, thanks jinu and Polizei for your appreciative comments as well (it means a lot to me) :)

Back to the AC2 Project:
I will be updating the piping diagram soon, to include a combination buffer/expansion aspect which I believe will be needed to have enough gas present to realize the full load rating of the unit, while at the same time accommodate the compressor's start-up requirements.

I will also be adding in all 13 temperature points I described on an earlier post, so everybody can have a much better idea of the locations. This will be very important when I get into the testing stage, since I will be sharing test data that will reference these same points.

Hopefully the first tests will commence in late January of 2008 :up:

13 temperature points? Does that mean youre going to be measuring gas temps throughout the entire system?

Excellent idea there... tons of info awaits!

More importantly he's now given me enough insight as to how to charge. Charge to get as little a delta at outputs of an HX.

Delta between what?

Output high side of HX and input temp.

Im confused... low stage or high stage?

I have too much to learn. :shrug:

Quote:

---

More importantly he's now given me enough insight as to how to charge. Charge to get as little a delta at outputs of an HX.

---

Quote:

---

Delta between what?

---

Quote:

---

Im confused... low stage or high stage?

---

Polizei check out Post #165 for more info on what Nol is talking about.

Quote:

---

13 temperature points? Does that mean youre going to be measuring gas temps throughout the entire system?

Excellent idea there... tons of info awaits!

---

Yes to all your questions. It will give us some very valuable data that up until now, has been sadly missing with the autocascade projects.