Thank you for your reply and input. The dry cooler fans are not variable, unfortunately. This would be great. It seems most of these systems that variable fans are on 3 phase only. I’ll be looking at this down the road (variable, not 3 phase unfortunately). I suppose if btc went to $1million then I could look at pulling in 3 phase. LOL.
I don’t think the foam would help as there isn’t much flow loss on the sides at all since the plate sort of boxes itself in. Inevitably there is a little, but with the oversized pump it doesn’t make much of a difference IMO.
I hear a lot of people running about 2.75-3 GPM through the miners. The oil pump in this thing would push about 6 GPM through each miner if the flow were directed well. There definitely is a happy medium, but at least with a number that high there is some margin for poor flow design. Inherently that is probably a part of the “design”
Hydro is amazing. I’ve always thought it would be amazing to live on property that has a large creek or small river running through it. Perhaps at the next property
@wired
I would like to add logic to the cooling system once I get heat reclamation up and running. This would involve powering one fan, both fans, or no fans. I will have active “radiators” and possibly a passive one to pull heat off in the winter and cooler months and then hopefully a water storage tank that will be applicable year round. With that the dry cooler fans will rarely need to work. I haven’t worked out if I’ll use a PLC or what to control everything. Any input on what would be a good direction for logic on a system like that would be great. It sounds like you have likely crossed similar bridges with your hydro setup. I basically need to control some valves to direct flow and turn fans on and off with a few temperature sensors to base these decisions off of.
I saw an interesting tank a while back that used the fans in the miner to circulate the fluid up through the miner and psu. You would have to have aftermarket firmware to control the fan speeds, but I wonder how well this does. You can obviously put as much or less fluid as you would like through it based on RPM of the fan. I thought this was kind of “hoaky” at first, but having this setup and seeing some of the challenges I wonder if it isn’t such a bad idea. MAYBE if I am able to get aftermarket firmware setup on a machine I could test this by building a little base for the miner to sit on and then run the fans at a low RPM to push the oil through. You’re talking a negligible amount of power to have them slowly turn.
When I get a little free time later this week or weekend, I’ll see if I can find that tank/system.
You could switch 5 solenoid valves 4 of which would have end limit switches.
On solenoid valves you have an actuator usually 24v on domestic heating, and you can get end limit switches built in. You can get SS actuated ball valves on ebay any size. The solenoids you want will close (usually spring loaded) if they are de-energised.
The end limit switches close when the valve is fully open, which stops dead ending of pumps etc.
I sketched a quick set of drawings on how I would go about it whilst I was having a lunch break.
They are rough and untidy, sorry for that.
It will prioritise the HWT and any heat left will either be returned to the dry cooler or go into the domestic heating radiators or heating coils. If the HWT is up to temp then it will bypass and send the coolant straight to the heaters. If neither are calling for heat (ie both bypass valves are energised) contactor coil 1 and 2 will also be energised and will open both contactors (the normally close ones next to the main) and that will turn de-energise the main solenoid.
You can pick up hydronic radiators and set them in series and or parallel to allow enough flow. There are tanks with built in coils available too. Again you can parallel them too. You just have to size the pipes big enough not to impede your flow rate too much.
A pressure relief valve will save your pump if things like solenoids failed.
The end limit switch wiring allows the main solenoid to close only when there is an open path back to the dry cooler.
With 2 x RTD switches you can control the fans by monitoring the outflow temp from the drycooler. You can do this with 1 smart controller if you want to monitor them or even 2 x Ranco’s.
Please check my workings as I did it in a hurry!
You can use any voltage you like - 24v is common for domestic thermostats, if you can not get a t’stat with the 2 poles then you can use a N.O. / N.C. relay.
Thank you very much for the information. I will be going through this very THOROUGHLY. I need some time when I don’t have a 2 yr old climbing on me while trying to digest info.
I have the solenoids he is talking about, but I use them on a remote controlled bushfire sprinkler system. I have them connected to a power point I can turn on/off via internet, as soon as there’s no power the solenoid closes. I have brass 2 inch and they are very reliable
Good to hear! I am excited for that aspect of the project. I will need to definitely get some of this done before winter so I can use the heat in buildings. If I don’t get the water storage tank setup before then that’s not the end of the world.
Today got to around 80F outside. The temps didn’t seem much different with 3 than 2. With that said, the last miner will go in this weekend and from there we will hope for the best!
The 4th miner is in the system. I’ll post some power consumption numbers on everything at some point over this coming week. I need to pull the electrical panel cover to put an “amp clamp” on the water pump and dry cooler fans.
I can say that the system is not more efficient than pure air cooling WITHOUT any additional fans to move air. How much so I will let you know soon I will also compare numbers of the components (no miners) as to what is advertised.
Interestingly, as SOON as the last miner went into the tank its hashrate did the exact same thing as the first miner. I wanted to give at least a day before saying anything.
edit: I will say the peace and quiet is well worth the 2-4 th/s I have lost. ha. Whether it will cause a premature failure on a board or not, I do not know. It is the same board on each machine. It’s the one with a frequency of 485.
I did get a new switch and will be reworking my network. I will likely setup a VPN to get the amlogic boards flashed. I haven’t had a whole lot of luck with getting an older xlinx board to work despite a couple people saying it “should”. @badgerlandcrypto provided a board to try for this. He has been of great assistance through this journey.
Yes, they can. You need to clean them with a special cleaner and then it is recommended to also put them in an another solution that coats the components to prevent any kind of corrosion. This can be done in an ultrasonic cleaner easily. I do not know if thermal paste should be replaced or not.
It’s definitely not a process you want to take lightly. It took me almost 2 hours a miner to prepare for immersion and it would take that or more to go the other way.
edit: I plan to keep these miners running for 2+ years, so I don’t mind. If you change miners frequently then definitely keep them air. There is probably more risk with immersion IMO with the systems available right now.
I think the ultimate immersion system for us home miners would be one that has interchangeable flow plates and a plate that is large enough to where you can add different miners, dimensionally speaking. Then the other aspect that would be amazing would be to have a baffle that could help increase or decrease flow to each miner flow plate.
The guys over at foghashing have hooked up with some handles for the miners. I liked these ones because they are minimal and currently the way I have the miners configured they would not block any flow. I’ll post a pic soon. I need to grab some hardware to get them installed. I think I’ll just install them while the units are in the tank.
edit: the ultimate miner setup is something I may end up designing for my system. I’ll most certainly need a good amount of time though and that’s hard to come by. I’ve got access to CNC, machinery, and can weld pretty much whatever.
thermal paste would probably need to be replaced. Amazing setup soccer! Thats interesting that the same FREQ boards all do that. Could be something with them interacting with the fluid maybe.
It’s only on two boards and my best guess is if the flow of the fluid was more evenly spread/better then it would not be an issue. The higher frequency boards may produce a little more heat.
That actually gives me an idea. I wonder if the position of them in the housing itself is different than the two that are functioning properly.
Nope, board position doesn’t matter. Just thought maybe if the 485 was in the middle or outside it would get better flow or more/less “heat soak” from the other boards.
It’s been a little busy around here with the 4th and such, but I took some measurements on the fan bolts/screws. They are an 8-36. I will stop by the hardware store today to see if they either have some black oxide or ss ones. If not then I’ll order and wait for snail mail.
The miners have been doing exactly what they’ve each been doing. The two hashing at degraded numbres for whatever reason (2% less maybe) and the other two working as if they didn’t know the difference.
It will be 93F ambient temps or so this weekend, last I checked. I don’t expect it to be an issue as it seems to hold up just perfectly fine in 90F or so. I set the shutdown temp right now to 50F outlet and the highest miner temps have been 68C, which I’m perfectly fine with as long as they’re getting cooled uniformly (unknown).
I’ll get to those power numbers soon for everyone.
I will put the numbers measured with an “amp clamp” for each component. I am taking the oil pump at face value since I don’t really have a good way to measure its consumption separately. Voltages were measured right at 240v for calculation purposes.
Complete plug and Play Immersion cooling System for mining
Miners Capacity : 6X Antminer S19 /S17/T17 and any Asic miner with silimer size
Power Load : 35 KW on 220-240V Single Phase circuit or 45KW on 380V Three Phase
- Fluid capacity : 180 L of dielectric fluid - Closer to 60 gallons 227L
Liquid Pool Material : Stainless Steal 304 - The flow “tray” is SS, not the entire pool
Outer dimenssion : 116 cm L X78 cm W X 58 cm H
inner dimenssion : 60 cm L X66 cm W X 55 cm H
all units comes with :
-Ethernet Wires
-Fan Spoofers
PDU and Power Cord
Water Pump 750 W 160 M3 / H (Draws 5.4A @ 240v = 1.3kW)
Oil pump 380W 160L/min capacity (Says 550W right on it, about 2.3A)
Dry Cooler
Oil Temperature sensor and Overheat Mechanism
Water Chiller optional for Hot Weather locations
-Total power consumption (excluding miners ) : 1650 W (Actual 1850W from above)
There has been NO mention of the dry cooler fan consumption which is: 4A x 240v = 960W
Actual consumption of the sytem without miners: 2810 W or 11.7A
The miners are pulling about 46.4A by pulling out the oil pump. That equals 11.136kW
Grand total = 13.95 kW
13.95 / 4 = 3.487kW per miner
3487 W / 134Th/s = 26.02 W / Th/s
Not THAT great, but not that bad either considering they’re silient and not heating my living and work spaces which I would have to help cool with hvac to make them livable. I also had many fans going as well to keep them cool by moving air. Each fan is rated for about 300W. There was probably a total of about 1.5kW in aux fans to move air.
At peak I would see right about 13A from the miners when they were “hot”.
13 x 4 = 52A
52A x 240v = 12.48kW
1.5kW + 12.48kW = 13.98kW (estimated total with air cooling and aux fans, no hvac)
Also take note that if I were to add two more miners to the immersion system then the efficiency as a whole would get better, assuming it would be ok and not over heat. I do know that as temps decrease so does the consumption of the miners themselves, excluding fan power.
All measurements were taken today with about an 85F ambient temp.
