has someone figures about max. chip temperature of Jupiter October and November batch. did Jupiters died because of to high chip temperature?
85C is top recommended max operating temp. I had some hit 95 and survive just fine(not recommended)Every board I had liked 75-80 for optimum performance.
Some say chilling works, but most of them have never seen 80C on their ASIC temps. If it were a GPU...
I'd say "cool it", but a Jupiter likes it toasty on the asic.
CLEAN power was most important..... meaning no spikes and fluctuations in the electrical feed.
also... tuning suite.... LOWER the voltages instead of raise to wake dead dies.
Wait aren't the sensors placed on VRMs or on boards? 85C for a chip isn't a bit low? I've read that some chips can go up to 120-130C.
but that's okay when the sweetspot is 75 to 80.
I have a very different experience with my ASICs.
On my ASIC I can gauranfuckintee lower temps result in higher Gh/s at the pool and everywhere else you care to measure.
Oct. 4 VRM Saturn stats after 1 day 2 hours.
330.323 Gh/s
000.422 %HW
Code:
die Mh/s HW% min max Volt Amp Watt W/Gh
0_0 849.30 0.107 ( 0.05 / 0.18 ) 0.80 52.19 41.80 1.025
0_1 849.65 0.104 ( 0.05 / 0.17 ) 0.80 51.19 41.16 1.009
0_2 847.85 1.119 ( 0.92 / 1.29 ) 0.81 55.94 45.20 1.111
0_3 782.99 0.095 ( 0.00 / 0.14 ) 0.79 48.06 37.73 1.004
55.0 C
1_0 872.48 0.003 ( 0.00 / 0.02 ) 0.82 53.44 43.71 1.044
1_1 874.42 0.003 ( 0.00 / 0.02 ) 0.81 54.12 43.73 1.042
1_2 894.81 1.361 ( 1.14 / 1.56 ) 0.82 57.88 47.63 1.109
1_3 899.04 0.590 ( 0.23 / 0.92 ) 0.78 57.88 45.08 1.045
58.0 C
The temp reported by the miner is NOT the ASIC or VRM temp!!!
The ASIC above @ 58C measures 73C on the bottom of the PCB directly below the ASIC.
Directly below the VRM's I get 90C.
Power consumption climbs with temp for equal output.
If I cool it more, efficiency climbs and errors lower.
I can adjust speed and W/Gh as needed.
Perhaps you define "optimum performance" differently.
YMMV
