A metering device i belive is what is used. Went to hvac school and did a year of work but honestly dont remember much so take that with a grain of salt.
2 Likes
The iced tech is nowhere near expensive. Maybe 3k for the condensing unit, 500 or so in 404b and the lineset
It’s just a emerson condenser hooked up to a tube and shell
Emerson FFAP-050Z-TFC-071
3 Likes
I think the primary difference is this won’t technically be a single unit, which I believe would describe a chiller. It’s a design that people can alter to fit their particular system and needs.
So, for example, let’s say a company had 4 separate extraction units that they want to use CO2 to chill, one could position them in such a way that having the lines plumbed correctly would allow them to simply attach a line from a source tank. Whether it’s a giant tank or 70lb tank, and chill a single unit, or all 4 units, and have manifolds leading to the pump and compressor constantly recovering into tanks to be rolled over and used again, or fed back into the giant source tank.
This could function almost exactly like an extraction machine that uses CO2 to extract, but, with significantly lower operating temperatures and pressures. All the while, having the scale of it only be restricted by number of pump/condenser recovery units, or the size of the pump/condenser.
The condenser for this idea is relatively simple, a jacketed cylinder containing lots of tiny tubes for the CO2 to pass through and condense to liquid state. The pump does most of the work via compression so the jacket doesn’t have to be super cold.
So, I still think the system is going to be comparable in complexity and cost to an actual chiller with a refrigerant distributor/load switching station, but with a bunch of downsides. The biggest problem I see is choosing an appropriate pump: any HVAC compressor (scroll/hermetic) is going to be out because you will pump all of the oil out of it. That leaves you a selection of waaaayyy more expensive slower pumps (piston recovery pumps or diaphragm pumps). In order to get adequate cooling, you need a lot of pressure drop so the evaporator side will have to stay at a very low pressure; that means either a very big vessel (not practical for CO2 for a whole shitload of reasons) or a quick pump.
Shell and tube is an expensive option for a condenser, but there also aren’t a lot of plate exchangers that can handle CO2 pressure. Honestly, an HVAC condenser rated for r410 is your best hacky option.
I can drop some components and such for an “open source” cascade chiller using 410 and 744 but it’s not very “DIY” because to work with CO2 pressures you need to braze a ton of stainless to copper joints which is absolute hell.
1 Like
I know @Cannagas makes they’re dry ice. Maybe they can share some specs on the machine they use and we could extrapolate some info from that…
Ok so here goes with another long post. I got an okay to share some details about a process chiller we built for condensing CO2 at -24F; I guess we all realized that we have no interest in selling chillers outright and hopefully someone rips off the design so we can just buy one lol.
Capacity is approx 80k btu at -24F, but it obviously can run colder. We’ve ran it down to -60F and then it started freezing the glycol.
The low stage:
We used a Copeland ZO88KCE to pump r744 (CO2). I can’t tell you our cost but I think retail is around $3k. Unfortunately, despite being a freaking awesome refrigerant, they’re not real common so the Americans at least won’t be able to find them used/new old stock.
The evaporator heat exchanger is a duda brazed plate, I think it’s a 60 plate 4x24" but I don’t remember exactly. They’re so cheap, just oversize it. We use the exact same HX for the interstage heat exchanger too but again, only because it’s oversized.
Expansion valve is a sporlan SER-C; sporlan will tell you that you only need an SER-B because of the pressure drop allowed but that is a crock of shit, ate a whole rev number for us lol. Depending on who’s building it you can run it off their Kelvin II superheat controller or you can run it off your own PLC. A little note here: the stock temp transmitter only goes to -60F and below that the controller will close the valve, blowing all your refrigerant and oil out the relief valve if you don’t address your logic accordingly; you’ve been warned. For lower temp operation, a mechanical TXV might be better but it will probably hunt like a mofo because there is not a product designed for these temperatures at this size.
We did not need a fade out tank because of how much evaporator piping we had, but there’s little downside to having one. I HIGHLY recommend wiring an unloader valve to run between the discharge (before the condenser) and suction of the evaporator to keep it from spiking pressure on startup because you will obviously want to keep the condenser-side volume as low as possible (this will mitigate the standstill pressure issue). You will also want a HP and LP cutout obviously, but you also need a pressure switch to ensure the low stage is sufficiently cold before it starts; this is much more effective than using a temp transmitter to do the same.
Now, this low stage hasn’t been “ratified” by Copeland but with our parameters it was well within their operating envelope. However, as you try to run this colder, oil management may become an issue which I have yet to address. Builder beware.
Pretty much all condenser piping is .035 wall 5/8 OD 304l welded, all evaporator piping is 1-1/8 OD of the same. The mueller HP copper can also do the pressure but it was waaaayyy more expensive than the stainless. Relief valve is set for 750 psi, service valves are where they always are.
High Stage:
We ran a ZPT166KCE tandem r410a compressor for the high stage. This allowed us additional capacity that we actually used as a dryer for the system air compressor. This is a much more common compressor with plenty of cross options.
Evaporator has been discussed already. For the condenser we used two 4-5 ton rated goodman slab coils. TXV sized out as a ERZE-6. Same deal with all the piping as the low stage, stainless is cheaper than the HP copper so, yeah, we did that.
Unloader circuit is helpful for the high stage as well if you don’t want to run a liquid accumulator.
We charged for the interstage to be operating around 32F so that you could visually see which stage was “winning”. If you bumped that up a bit, you could eliminate the entire high stage by using the facility chilled water (if it’s an option) which is usually around 44F. The next thing to do would be an R23 or N2 “even lower” stage but I haven’t had any time to select components for that.
If anyone has specific questions let me know. I can draw up a PID if y’all like but it’s probably not necessary for anyone that actually plans to build one. Total parts cost will probably come out in the $6-7k range, but could be quite a bit lower for a smart shopper/scrounger. It’s enough to condense 4+ kg/min of CO2 at -24F so it’s a pretty big unit.
12 Likes
Awesome write up thanks for taking the time. Gets me excited.
3 Likes
Of course there’s already a thread about this, lol.
We’re currently looking at picking up a bizzybee system and it would be the first rig I’ve ever run using LCO2. We also happen to have an Apex 25L CO2 machine just sitting around collecting dust. I see a lot of talk on this thread about utilizing the CO2 in a more traditional chiller method, but would you not be able to just inject the LCO2 from one bottle into the jackets and hook up the Apex diaphragm pump to the exhaust from the jackets to pump into an empty bottle? Or are there pressure issues that I’m missing here with how fast the CO2 could be pumped out of the jacket before it over-pressurized?
1 Like
As pointed out above, using a valve to create pressure drop and then using a pump to compress it so that it condenses is just a regular vapor compression cycle. You would need cooling to actually condense the CO2 with liquid and bing bam boom, you’ve created a chiller
2 Likes
Hmm. Well, the Apex is already set up with a jacketed column, but I’m guessing that wouldn’t provide the most efficient surface area for cooling. I’ll do some homework this week and talk to a friend of mine who’s well versed in HVAC and this Apex machine and report back, because I feel like I’m not well enough versed in this area to even ask intelligent questions yet. Good to have a thread to put me on the right track though. Been missing this community lately. 
3 Likes
We’re happy to see you back around here. I didn’t mean to say that you can’t use you apex as a chiller because it’s pretty nearly that already. It’s just obviously not the most economical approach to making a chiller that uses CO2. The Copeland ZO series compressors exchanging to a regular split system AC or a building chilled water system can be pretty badass for several tens of thousands of dollars less than even the cheap second hand apex units.
1 Like
Copeland ZP series?
1 Like
ZP is rated for I think 600 psig for r410a, which is just a little on the low side for CO2. The ZO series is built for transcritical CO2 applications so they’ll do 1000 psig or so. They’re “the tits” for low stage on cascade systems.
4 Likes
I wish I could just pay you to come out and build one for me. Your understanding on this hard to find
5 Likes
I’ve seriously considered giving up “the big bucks” to be a roving refrigeration dude because I find it wicked fun lol. I haven’t made much progress on my mega chiller thread but I promise to share the chiller for the people when I get back to that project.
I think understanding of refrigeration systems is going to be one of these things like running SPD or a closed loop that this community will make pretty standard in the near future. Especially since the principles are very similar to running hydrocarbon systems already.
8 Likes
I understand the basics but without enough experience I think I’ll just waste my time trying to build a custom chiller without someone smarter than me moving the project forward. It’s like trying to do the digital work for packaging stickers and then somehow figure how to print batch information in the exact right spot.
I can do a lot of things, but I definitely know when to stay in my lane.
4 Likes
It’s kind of one of those things that you have to fuck up once or twice and then you know lol. I always love teaching people things, maybe one of these days I’ll grab a pile of parts and throw a little chiller building shindig.
Also, once I put that BOM and PID together, it’s going to be like paint by numbers. Somewhere I put up some details on the cascade 410/744 chillers we built. Not the same level of detail but the design is mostly there already, it’s basically just putting the parts together in the right order
6 Likes
I’ve made a couple chillers for clients. It’s alot easier than you think. You can buy the complete compressor section or build it. I used a simple metering txv and a stainless coil in a bath how cold you want to go?
4 Likes
I want to have around 18-20kw available at -60c
4 Likes
Injection and recovery I assume?
2 Likes