Antminer D3 max temp

[justinjja]

I have been running L3+'s outside and they stay under 80.
I just moved a D3 outside and it is running in the low 90's is this too hot?

I can't find anything about the max temp on them.
I see people talking about running S9's at 100,
but then the L3+ has a check box to shut down at 80,
so I'm not sure where a D3 should be.

Thanks,
Justin

[bitcoinguy123]

I would be interested in this as well on the correct operating temperature range...

Would you mind posting a picture or two of your outside setup?  I am in process of getting things ready as I have my own little project to build out for.

You can reference what I am working on in this link.  phillipma gave me tons of great advice. 
https://bitcointalk.org/index.php?topic=2200357.0

[sil2222]

I have been running L3+'s outside and they stay under 80.
I just moved a D3 outside and it is running in the low 90's is this too hot?

I can't find anything about the max temp on them.
I see people talking about running S9's at 100,
but then the L3+ has a check box to shut down at 80,
so I'm not sure where a D3 should be.

Thanks,
Justin

80c is too high D3 becomes unstable at 70c+ which is when fan kicks in and you will get a ton of hard ware errors. You will be losing a lot of hashing also so its not good to run miner unstable. If miner is outside run fan at 100% this will bring your temp down and try down clocking to 456 frequency. 90c will cause miner to fail eventually. miner is stable at 60C-69C and hardware errors become almost 0. At 456 frequency miner hashes 16000 ,but you will get a lot less rejects. I was told by a tech at innosilicon once a chip hits 80c it will start to burn this is also true of video cards.

[justinjja]

So if I'm not getting hardware errors I'm fine?
Been running 8 hrs at 85c and 0 hw.
(but it's coolerat night)

Also this is a bitmain device not inosilicon, max temps could be completely different.

[justinjja]

Fan on auto doesn't hit 100% until about 90c,
So I would guess anything under 90c is ok.

[sil2222]

Fan on auto doesn't hit 100% until about 90c,
So I would guess anything under 90c is ok.

Ive never heard of a chip running at 90c being ok. Nvidea 1080 with Asus aftermarket heavy duty heat sinks would have thermal breakdown at 90c. You should contact Bitmain and ask them they usually reply within 3 days,but I would not risk 90c on any device. Weird thing is on D3 i tested the fans kick in 100% when chip reaches 71c,but you say on yours it kicks in at 90C ?

[Chan8]

Fan on auto doesn't hit 100% until about 90c,
So I would guess anything under 90c is ok.

Ive never heard of a chip running at 90c being ok. Nvidea 1080 with Asus aftermarket heavy duty heat sinks would have thermal breakdown at 90c. You should contact Bitmain and ask them they usually reply within 3 days,but I would not risk 90c on any device. Weird thing is on D3 i tested the fans kick in 100% when chip reaches 71c,but you say on yours it kicks in at 90C ?

I 100% agree, I don't think any chip should be running at 90c.. something will definitely happen. might not be now, but you will see something start failing soon.

[Mr_Snipes]

Fan on auto doesn't hit 100% until about 90c,
So I would guess anything under 90c is ok.

Ive never heard of a chip running at 90c being ok. Nvidea 1080 with Asus aftermarket heavy duty heat sinks would have thermal breakdown at 90c. You should contact Bitmain and ask them they usually reply within 3 days,but I would not risk 90c on any device. Weird thing is on D3 i tested the fans kick in 100% when chip reaches 71c,but you say on yours it kicks in at 90C ?

Friend of mine had Chips failing at temps over 70C° - after cranking up fanspeed its fine. The way the fan-control works now is bugged imo. They will most likely change it in a more recent firmware.

[noam46]

are you guys talking about 90 calculus? or Fahrenheit?

im running at 80-85, is this ok?

[Mr_Snipes]

C° = Celsius
°F= Farenheit
The Bitmain-gui always shows C°

[Crypto-Age]

85-90 deggree is really dangerous.
This machine works 50-70*C as optimum.
It is healty for your asic lifetime.

[flip909]

I have been running L3+'s outside and they stay under 80.
I just moved a D3 outside and it is running in the low 90's is this too hot?

I can't find anything about the max temp on them.
I see people talking about running S9's at 100,
but then the L3+ has a check box to shut down at 80,
so I'm not sure where a D3 should be.

Thanks,
Justin

80c is too high D3 becomes unstable at 70c+ which is when fan kicks in and you will get a ton of hard ware errors. You will be losing a lot of hashing also so its not good to run miner unstable. If miner is outside run fan at 100% this will bring your temp down and try down clocking to 456 frequency. 90c will cause miner to fail eventually. miner is stable at 60C-69C and hardware errors become almost 0. At 456 frequency miner hashes 16000 ,but you will get a lot less rejects. I was told by a tech at innosilicon once a chip hits 80c it will start to burn this is also true of video cards.

My D3 19.3 GH from factory is running at 78 Degrees on 1 board. others range 70 - 74. I wanted to overclock but don't know about temps of chips. how can >70 be an issue if mine form factory is like this as standard.

[krzy]

https://imgur.com/a/5Kcy3

Am I fine with these? I don't get it, I have it outside in cold winter and it's still overheating? How the fuck am I supposed to cool it down then?

I get 0 HWs though, thoughts?

[dlezama]

Fan on auto doesn't hit 100% until about 90c,
So I would guess anything under 90c is ok.

Ive never heard of a chip running at 90c being ok. Nvidea 1080 with Asus aftermarket heavy duty heat sinks would have thermal breakdown at 90c. You should contact Bitmain and ask them they usually reply within 3 days,but I would not risk 90c on any device. Weird thing is on D3 i tested the fans kick in 100% when chip reaches 71c,but you say on yours it kicks in at 90C ?

It's common for example for automotive grade chips to support up to 125C. As someone else said, it depends on the chip. Lots of opinions in this thread, but I wonder what are their sources, definitely not a data sheet This is the only thing I've found in the manual, note that it refers to PCB temp, not the chip:
Note: The D3 miner will stop running when the Temp(PCB) exceeds to 85°C, there will be an error message “Fatal Error: Temperature is too high!” shown in the bottom of kernel log page.

In my case I'll worry if they get close to 90 or if I see more than a few HW errors per hour (the D3 is a useless heater anyway, so why worry?  )

[wonseok4840]

My D3 is the 19.3GH/s batch, and it is running very hot. Like, VERY, VERY hot. The ambient temperature is not that hot, and it is in my basement. The weird thing is, the PCB temperature is very good (High 50's), but the chip temperature is boiling hot. It is around 80 degrees, and the one of the ASIC boards are always an outcast in terms of the temperature. One time, one ASIC board was 120 degrees, and the rest was on the low 80's. Another time, it was on the 50's, and the rest was on the 80's. What kind of problem would that be? Any help is appreciated.

[wonseok4840]

Also, when I looked at the kernel log, it said that there was a temperature offset. That was the case with all 3 of my D3's, and the offset value was random as well. Is there a way to fix this?

[flip909]

My D3 is the 19.3GH/s batch, and it is running very hot. Like, VERY, VERY hot. The ambient temperature is not that hot, and it is in my basement. The weird thing is, the PCB temperature is very good (High 50's), but the chip temperature is boiling hot. It is around 80 degrees, and the one of the ASIC boards are always an outcast in terms of the temperature. One time, one ASIC board was 120 degrees, and the rest was on the low 80's. Another time, it was on the 50's, and the rest was on the 80's. What kind of problem would that be? Any help is appreciated.

wow

I have 1 19.3 D3 i dont want to overclock since its running hot anyway but yours is far too hot surely?

Contact Bitmain and send them logs. they will advise.

[clyjr]

My D3 is the 19.3GH/s batch, and it is running very hot. Like, VERY, VERY hot. The ambient temperature is not that hot, and it is in my basement. The weird thing is, the PCB temperature is very good (High 50's), but the chip temperature is boiling hot. It is around 80 degrees, and the one of the ASIC boards are always an outcast in terms of the temperature. One time, one ASIC board was 120 degrees, and the rest was on the low 80's. Another time, it was on the 50's, and the rest was on the 80's. What kind of problem would that be? Any help is appreciated.

When one is reporting really high, then really low, and seems to alternate like that, it's not a valid reading.   It's a bug, most of the time you can unplug the miner for a few minutes, plug it back in and the temps are all correct (within a few degrees of each other).   One of mine does this when I change settings sometimes but it always gets fixed with a hard reset (unplug/wait/plug)

[nguyenthong]

Great

[wonseok4840]

Now, it is doing weirder stuff. I tried to flash the latest firmware, but even after that, there were other errors. I got rid of the heat problem by running it in my garage at winter time, but the problem is that the chips don't get found. There are supposed to be 60 chips on one hashing board, but for me, it always changes from time to time. Once, it only detected 54, (Not X's on the chip status, the chips were just plain out missing), and once, it detected all 60, but 3 were dead (X). To make matters worse, it got 2000 HW errors in less than a minute! I will be sure contact bitmain about this, but honestly, I don't know......