New hash rate: 80.5TH
New wattage: 2266w
Initially, I would have to say that I only see about a 5% improvement in the overall efficiency. This is much less than Bitmain estimates, but there are a few factors that need to be considered. I am in the middle of the arctic blast and temps are really low. I need to let the miner stabilize and run for a full 24hrs.
The 24hr hashrate was about 78TH. W/TH improved slightly to 27.9. After 24 hrs I switched back to normal mode. I really don’t see much improvement in efficiency as claimed by Bitmain. Feel free to try it for yourself and post your thoughts/results.
I’m using firefox. I tried refreshing many times but still no low power option. Thanks to your genius suggestion, I opened up the miner’s ip in a firefox private window and magically the low power mode option appeared. Thank you
Now to test the 2 modes with a watt meter. The watt meter I used in the past craps out at 2000 watts. I can’t find one that will be able to withstand the 3000+ watts of the S19 in normal mode. Any ideas? I don’t want to buy a pdu with the meter built in because I would have to unplug everything else connected to the pdu to measure the watts of one miner.
For my S19J Pro 104
Normal Mode: 105.2 TH
Low Power Mode: 77.25 TH
I only tested the LP mode for 3 hours. The average hashrate was slowly going up. I suspect eventually it would be closer to 78TH eventually.
I’m no data scientist, but I grabbed a sample of the hashrate at the pools from yesterday to get an average and used my smart plug mesaure for the Wattage.
S19 Pro 82 on F2 Pool - 58.78 @ 1740 is 29.59 W/Th
S19J Pro 96 on F2 Pool - 69.14 @ 1740 is 25.16 W.Th
S19J Pro+ 117 on PegaPool - 86.3 @ 2000 is 23.17 W/Th
Pools are working out to be less than I see if I look at the miner, but it’s the pool rate I’m getting paid on so it’s the one that matters.
It’s very cold here right now. We’ll be on -1 Celsius overnight. I’m tracking KWh for taxes and both of the first 2 used 40KWh less in June than they did in April for the same run time with the same number of days in the month.
I got the 82Th as it was under 3KW and power supply was an issue at the time. Which I had got a bigger one.
I got 3 phase in from the street now so that’s less of a problem. I’m happy to get the most efficent boxes I can afford at the time, and run them in low power mode to get them more efficent. Might be a longer ROI but it’s more coin for the power input.
Also - my peak power rate is eye wateringly high, so I need to have them on smart switches that can turn them off during peak power pricing. The ones I have aren’t really safe over 2600 W. Waiting for some more Shelly 16A smart switches…
From my experience the cooler the miners the less power they consume. I think when I overclocked my S19 it runs 3500 watts when it’s 15 degrees Celsius or less and when it was 20+ it was running 3700 watts (including exhaust fans which is combined 150 watts) just my experience
Hello fellas, here is my numbers on normal mode and low hashrate mode.
Running 3x S19J-PRO 104T (1x Xinlin ctrl board , 2xAmlogic )
Test LH was made during last 4 days
Normal mode :
Avg 105T @ 3190W : 30.38 W/Th
Low hashrate mode :
Avg 81T @ 2227W : 27.49 W/Th
I see here that i win 3 efficiency points. great.
But we need to mention that this is during summer times, so i need more watts dedicated to the fans only to cool down a full powered hashboards, which is not the case in low power mode.
i think that this is great to switch to Low mode in summer times due to temperature , which is not the case when it’s freezin outside
hope that helps
I’ve been testing a profile from LuxOS on my S19 J pro 104TH which is 11.90 volts/200mhz frequency and I have it running at 930 watts/39.5TH, which is 23.5 watts/TH, I have 2 X exhaust fans which is included in that wall meter read, if I had fans like AC infinity which is much more power efficient it would be running 860 watts which is cool for a lot of people running off solar during off peak times