DIY Chiller Sizing Issue

Awesome work @Dannywarbucks! Nice to hear from you. I wish we could all collaborate and make some cool stuff! I’m soooo down.

2 Likes

Rated plate evaporators are better than coils, 2464 is a 1.5P compressor, I can provide you some info about the expansion valve, condenser etc if you are interested.

1 Like

I am very interested!

1 Like

2464 is R404, -40C compressor.
We use #2 Spool Danfoss expansion valve or 1.8m of Φ2.5 capillary.
Fill 450~500g R404, depends on room temp and the low temp you need

4 Likes

Condenser is 7.8 m2.
With regard to the rated plate evaporators, I am trying to get a drawing, will back to you soon

1 Like

Here is the drawing of the plate evaporators, Unit in mm. Heat exchange area is 0.53 m2.

4 Likes

Wow thanks for posting the design. I was going to build but trying to understand which diameter of capillaries tubing to use was not an easy answer to find.

2 Likes

I came across this resource the other day for capillary sizing: Capillary Tubing Info | Refrigerative Supply

Ive seen some pretty mean deals on modern expansion valve too.

Thanks @EffisonAndy

3 Likes

We use 2.5mm capillaries, 1.8m long.
But you have to test the machine to determine the length of the capillary.

1 Like

They are too good and too cheap to not use IMO. Other than the issues with hunting at VLT I mentioned they are just better than cap tubes.

2 Likes

I found the ZZ series from Emerson for VLT. But I just want to figure out -40 with either r404a or r290, as a start.

EEV’s seem pretty advanced, probably overkill for a first timer.

Just doing last minute searching on other comparable condensing units before I order the awa2464.

1 Like

Good luck @pangea. Love that u wanna use r290.

EEVs are like an easy button honestly. Very simple PID to control superheat. We use the Sporlan Kelvin II all the time but the temp transducer only works down to -60C. It wouldn’t be so bad except after that it commands the EEV full close and there’s not dry contact to command a compressor stop so yeah, guess what happens. Certainly makes you get your HPC logic down.

Other issue with the Kelvin II is that it’s pretty expensive. TBH it makes more sense to use a Siemens Logo or similar small PLC with a single PID loop and value chart to run the EEV. Next time I do one I’ll throw the program up on here and save everyone a couple hundred bucks.

2 Likes

Is it possible to pass compressed refrigerant directly though a brazed-plate heat exchanger?

For instance:

The red lines indicate suction lines to the pump inlet, blue lines are discharge lines from the pump outlet.

The orange lines indicate a chiller of some sort–if its possible to push compressed refrigerant through there without creating some kind of compliance nightmare, i’m all about it.

The plan is to transfer -10C ethanol from a larger holding tank into that small tank (the centrifuge is a CUP-15 for size reference), and further chill that ethanol to -40C using the brazed plate. This is primarily because -10 or -20C chillers are about 1/4th the price for double the amount of kilowatts.

I’d prefer not to get an ultra-low chiller–not because they aren’t awesome (because they totally are)–but because they’re super expensive.

Anywho, does anyone know if compressed refrigerant will work?

3 Likes

Plate heat exchangers work great for these applications. There are some issues with vapor holdup on the evaporator but nothing drastic. Absolutely no problem with that design (other than some nitpicky stuff like a liquid accumulator and expansion device). Also, you can get brazed plates in the 1000psi+ MAWP range.

2 Likes

Sounds like your in the right thread!

I posted this link http://s3.supplyhouse.com/product_files/FPN5X12-36-brochure.pdf for plate exchangers from GEA, lots of good spec and sizing info there for expansion or condensing.

The condensing unit Im looking at has the same capacity at -40 as a Julabo presto which retails at $40k

Seems like the EEV will be a future upgrade for me, I was eyeing the johnson control eev(cause they have a price listed), which looks like it’d work for -40. Probably opt for a sporlan or emerson txv.

I think this thread is a must read as well for anyone interested in this topic: Refrigerant For Jacketed Base? - #41 by Soxhlet

Also check out BHOgarts version: -40°C Single Plate Chiller — BHOGART

though no cooling capacity specs.

5 Likes

Duda diesel carries some big cheap BPHXs that I’ve used. I also troll eBay for SWEP and Alfa Laval units that are used/new-old stock. For the real badass applications Tranter maxchanger is an all welded stainless plate heat exchanger. We used one as an insterstage heat exchanger operating at like 1300psi once.

Only thing I will say about BPHXs is that they usually come with NPT threads which are fucking useless in a refrigeration system. I’d kill for someone who would make them with JIC ends instead. Usually we either weld the piping in directly or weld in a JIC adapter. In the past we brazed the threads directly but… I fucking hate brazing

3 Likes

Good call on the jic adapter welding and plate mfg recommendations, I will do some more searching there. I am also interested in dx cooling my processing recovery tubes as well.

Excerpt from the DX thread previously linked:

So Im not exactly understanding how it wont be stealing much mass if you grossly oversize the evaporator. Is there downsides? The inside jacket surface area is much greater than say a 1/2 ton bphx.

I was reading this document from MIT on a compressor selection program they’ve written. http://web.mit.edu/parmstr/Public/hvac/compressor/Carlyle/CarwinDocumentation.pdf

It has some good info on learning what some of the performance data means and what determines or influences capacity of the compressor and how that relates to the evaporators refrigeration effect.

Im sure what Im not getting is right before me, but it hasnt fully clicked yet, I think I have to re read the mit doc several more times.

“As superheat increases, Hsh increases &
Evaporator Refrigeration Effect increases (as long
as the superheat does not become higher than the
desired cooling temperature for the unit)”

“as long as the superheat does not become higher than the desired cooling temperature for the unit”

if 5°F to 10°F is normal super heat, and the desired cooling temp for the unit is -40. I dont understand.

This also is not clicking yet:

"Lowering the return gas temperature will
increase the density entering the
compressor, thus increasing the evaporator
mass flow. This will increase the evaporator
refrigeration effect. Increasing the
evaporator mass flow will also increase
compressor capacity but the return gas
temperature enthalpy (Hrgt) will decrease
reducing the change in enthalpy. This leads to a lower
compressor capacity. In most instances,
the enthalpy decrease will be greater than the mass flow
increase which will result in lower compressor capacities. "

Bascially trying to understand what determines the enthalpy decrease vs mass flow increase, cause it sounds like an easy way to increase capacity is to lower the return gas temps from the standard 65f, which in my cold ass location should be easy.

2 Likes

So yes, oversizing the evaporator will require more refrigerant in the system and will have more holdup volume. Other than that though, no real issue with it. The way to look at it is how an EEV works (TXVs do the same mechanically, just less ideally): an EEV looks at the pressure and temperature of the refrigerant entering the compressor and calculates the superheat from that. If that is less than the setpoint, it closes the valve which reduces the saturation pressure and also consequentially the temperature at the inlet of the evaporator (because the refrigerant is now boiling at a lower temperature). BUT the density is lower because of the lower pressure and therefore the compressor is moving less gas.

When the system is running initially (let’s say cooling a massive amount of hot solvent) the evaporator is getting all the heat it needs it boil a shitload of refrigerant. Only limited by heat exchange, the EEV goes full open and still has more superheat than the setpoint. Design note here: if you way oversize the EEV you can build a system that just pumps refrigerant in a circle without building any pressure differential so… Don’t do that. Anyways, if the evaporator is way oversized here, it will just ensure that you max out your expansion valve but ultimately you are cooling faster because the EEV doesn’t have to close to ensure superheat.

Move along to where we’re at temperature. The process fluid is very cold and that makes the EEV close a lot to bring the saturation temperature sufficiently below the process fluid temp, otherwise no refrigerant will boil and no superheat will be developed. The larger the heat exchanger (actually the higher “theta”), the more easily the refrigerant can find the heat to boil and therefore will operate at a higher pressure ie higher density ie more chooch.

What the MIT thing is probably saying is that a less superheated gas is denser and therefore lends more capacity than a more superheated gas, which is correct but the more important factor is that the system achieves that lower superheat by allowing a higher evap pressure at a given heat load. That higher evap pressure = denser gas = more chooch.

Lastly, what I’m saying about not “stealing mass” is talking about fade out capacity. The standstill pressure is the same across the entire system. It is a function of the density of that refrigerant, the volume, and the temperature. Let’s say we cap that pressure at 700psi (normal max pressure of refrigeration components FYI). If you know your system volume, you can derive how much mass you are allowed. Now, when allocating your volume between suction and discharge (ie more sense and less sense), a higher ratio of suction volume (evaporator size) to discharge requires less refrigerant mass because the density in the evaporator is lower. If you minimize your condenser volume as much as possible and still require too much mass for your standstill allowance, the only solution is to add more volume on the suction side, commonly by putting a tank that literally only serves the purpose of taking up volume. Obviously if you have an enormous evaporator, the tank might not be necessary. Does that make sense?

4 Likes

Also, couple of post scripts:

None of this applies with a fixed orifice lol. I skipped that part of my self-education so don’t ask me how it all works when you just use a hole.

If you are looking to DX exchange with a jacketed column, it’s as simple as setting your standstill pressure/mass allowance to a bit below the working pressure of the jacket, but if that’s low it’s going to mean you need a massive amount of fade out capacity or another more sophisticated design.

And on your last point, what you’re really saying is that a lower suction superheat is more efficient as that is absolutely correct. It is standard practice to reduce that as much as you can without worrying about liquid slugging. Incorporating a suction accumulator is a very easy effective way to drop from 10 R to 3-4 R superheat. Honestly though if this is your first chiller build, don’t sweat it too much, it’s really not a major capacity factor.

3 Likes