Morphing Air Conditioner into Autocascade System

[Freddie123]

Mytek can't find cryo's post on the hot gas defrost feature. Am I right in thinking that it allows a continuous stream of hot gas into the evaporator as it runs? Why is a having a none icy evap useful surely it will just raise temps?

[boshuter]

Mytek can't find cryo's post on the hot gas defrost feature. Am I right in thinking that it allows a continuous stream of hot gas into the evaporator as it runs? Why is a having a none icy evap useful surely it will just raise temps?

I have a chiller with a hot gas bypass system, mine is controlled by a solenoid hooked to a temp controller; it opens and closes the solenoid to maintain the correct temp. There is a "T" right after the condenser on mine, one branch goes to the filter/cap tube and the other goes to the hot gas bypass regulator, the solenoid is after the "T" and right before the filter/cap tube; When the solenoid is closed the hot gas goes through the hot gas bypass regulator and into the evap. It's a very effective way to control evap temps without having to cycle the compressor on and off.

[jinu117]

In this case I think the original purpose was to... thaw evap off in hurry
Basically, something of this nature takes an hour or so to evaporator to get to safe temp to remove from CPU This can be done in mere seconds this way.
Also, I am thinking this could be useful for those situations where there is possible oil frosting with some sort of control logic or pressure sensors...

[mytekcontrols]

Jinu's answer is correct. It's just an option, that allows you to bring the evap (and refrigerant line) back up to room temperature very quickly.

Boshuter also brings up a very good alternative use, that being temperature control. Although most of you guys just want to get things as cold as possible, so temperature control is probably not an issue.

Edit: Happy Thanksgiving to all that celebrate this holiday!

[jinu117]

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?

[mytekcontrols]

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.

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.

[n00b 0f l337]

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

[n00b 0f l337]

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?

[mytekcontrols]

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).

[mytekcontrols]

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

[mytekcontrols]

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.

[mytekcontrols]

...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.

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.

[n00b 0f l337]

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.

[n00b 0f l337]

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

[n00b 0f l337]

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?

[mytekcontrols]

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

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

[mytekcontrols]

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.

[mytekcontrols]

Here is a look inside:

[mytekcontrols]

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.

[n00b 0f l337]

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

[mytekcontrols]

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.

[n00b 0f l337]

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

[Freddie123]

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?

[mytekcontrols]

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:

[n00b 0f l337]

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.