S9 with chips appear as XXXXX

[tbonetony]

Folks,

I have owned S1's, S3's, S5's, S7's, and now an S9. S9 is the first miner that I have so frequently seen chips turn xxxxx in some weird situations.

In my experience, it seemed to relate to either an unexpected connectivity disruption, or pool issues. However, all the previous miners could handle these situations without letting the chips fail to function. I am still unable to summarize the scenario that could make S9 chips become xxxx. Every time this happens, I have to reboot it to recover the chips.

What's even worse, is I have noticed it could start with a few chips become xxxx randomly on one or more boards, then it soon (within a minute or two) would spread to the entire row (board), then after a while, a second board could die, then another while, your last board.

The most annoying thing is S9 does not indicate such state to me. Like an S5, it beeped the hell out of me if network drops. this S9 would just quietly suicide without even lowering the fan speed.

Have you experience such situations and if so, what are your observations and thoughts?

Thanks for sharing.

[philipma1957]

Folks,

I have owned S1's, S3's, S5's, S7's, and now an S9. S9 is the first miner that I have so frequently seen chips turn xxxxx in some weird situations.

In my experience, it seemed to relate to either an unexpected connectivity disruption, or pool issues. However, all the previous miners could handle these situations without letting the chips fail to function. I am still unable to summarize the scenario that could make S9 chips become xxxx. Every time this happens, I have to reboot it to recover the chips.

What's even worse, is I have noticed it could start with a few chips become xxxx randomly on one or more boards, then it soon (within a minute or two) would spread to the entire row (board), then after a while, a second board could die, then another while, your last board.

The most annoying thing is S9 does not indicate such state to me. Like an S5, it beeped the hell out of me if network drops. this S9 would just quietly suicide without even lowering the fan speed.

Have you experience such situations and if so, what are your observations and thoughts?

Thanks for sharing.

first question.  your psu is?
second question your freq is ?

xxxx   many times is a show that the psu is running out of gas.

reduce freq a bit  say 600 to 550

More questions  do you set the fans to 80% or 85% or 90%?
or are you an auto guy?

another question  how how is it?
in the room and on the pcb and on the chips.

if the room is 90f or more  even if the pcb and the chips read 60c 90c

the psu may struggle.  here is why  most psus lose gas at 40 to 50c

so a 90 f room is more likely to get the psu too warm.

try to down clock freq from 600 to 550 or from 600 to 575

[tbonetony]

Folks,

I have owned S1's, S3's, S5's, S7's, and now an S9. S9 is the first miner that I have so frequently seen chips turn xxxxx in some weird situations.

In my experience, it seemed to relate to either an unexpected connectivity disruption, or pool issues. However, all the previous miners could handle these situations without letting the chips fail to function. I am still unable to summarize the scenario that could make S9 chips become xxxx. Every time this happens, I have to reboot it to recover the chips.

What's even worse, is I have noticed it could start with a few chips become xxxx randomly on one or more boards, then it soon (within a minute or two) would spread to the entire row (board), then after a while, a second board could die, then another while, your last board.

The most annoying thing is S9 does not indicate such state to me. Like an S5, it beeped the hell out of me if network drops. this S9 would just quietly suicide without even lowering the fan speed.

Have you experience such situations and if so, what are your observations and thoughts?

Thanks for sharing.

first question.  your psu is?
second question your freq is ?

xxxx   many times is a show that the psu is running out of gas.

reduce freq a bit  say 600 to 550

More questions  do you set the fans to 80% or 85% or 90%?
or are you an auto guy?

another question  how how is it?
in the room and on the pcb and on the chips.

if the room is 90f or more  even if the pcb and the chips read 60c 90c

the psu may struggle.  here is why  most psus lose gas at 40 to 50c

so a 90 f room is more likely to get the psu too warm.

try to down clock freq from 600 to 550 or from 600 to 575

I got the S9-B9 so it defaults to 550MHz.
Currently I am powering it with 2x HP-1000W which I used to power the S7.
I left the fans to auto and they run at 4600/4000 during the day and 4000/3600 at night.
Temperature usually is within 50~55 for PCB and 80-90 for chips, with ambient around 10~25 degC.
I will try your suggestions when I have some more time. Thank you, phil.

[stealthants76]

Hi guys.
Well you not really the only one. I had some xx but after restart most of the time it goes well. How ever IF i used the stock firmware. After update i got even more weird stuff going on.
My psu is antminer 1600w so no problems there. I even use another fan on it to keep it real cool.
But i just been talking to support about my last weird thing... After like two days of using the s9 like to get slower and slower... ? wtf

Temp is not a problem right now. As when the temp was at around 90C the s9 pushed on really well.. Untill now

Also one hash board gets pretty much hw unlike the other two. I changed the port on controller and got more better results but now again.. back where i was before :S

[philipma1957]

Hi guys.
Well you not really the only one. I had some xx but after restart most of the time it goes well. How ever IF i used the stock firmware. After update i got even more weird stuff going on.
My psu is antminer 1600w so no problems there. I even use another fan on it to keep it real cool.
But i just been talking to support about my last weird thing... After like two days of using the s9 like to get slower and slower... ? wtf

Temp is not a problem right now. As when the temp was at around 90C the s9 pushed on really well.. Untill now

Also one hash board gets pretty much hw unlike the other two. I changed the port on controller and got more better results but now again.. back where i was before :S

classic example of a hot tired psu.  drop your clock freq  say 550 is stock drop to 500  see if the hash rate drops after 2 days.

[anakinisme]

Folks,

I have owned S1's, S3's, S5's, S7's, and now an S9. S9 is the first miner that I have so frequently seen chips turn xxxxx in some weird situations.

In my experience, it seemed to relate to either an unexpected connectivity disruption, or pool issues. However, all the previous miners could handle these situations without letting the chips fail to function. I am still unable to summarize the scenario that could make S9 chips become xxxx. Every time this happens, I have to reboot it to recover the chips.

What's even worse, is I have noticed it could start with a few chips become xxxx randomly on one or more boards, then it soon (within a minute or two) would spread to the entire row (board), then after a while, a second board could die, then another while, your last board.

The most annoying thing is S9 does not indicate such state to me. Like an S5, it beeped the hell out of me if network drops. this S9 would just quietly suicide without even lowering the fan speed.

Have you experience such situations and if so, what are your observations and thoughts?

Thanks for sharing.

first question.  your psu is?
second question your freq is ?

xxxx   many times is a show that the psu is running out of gas.

reduce freq a bit  say 600 to 550

More questions  do you set the fans to 80% or 85% or 90%?
or are you an auto guy?

another question  how how is it?
in the room and on the pcb and on the chips.

if the room is 90f or more  even if the pcb and the chips read 60c 90c

the psu may struggle.  here is why  most psus lose gas at 40 to 50c

so a 90 f room is more likely to get the psu too warm.

try to down clock freq from 600 to 550 or from 600 to 575

I got the S9-B9 so it defaults to 550MHz.
Currently I am powering it with 2x HP-1000W which I used to power the S7.
I left the fans to auto and they run at 4600/4000 during the day and 4000/3600 at night.
Temperature usually is within 50~55 for PCB and 80-90 for chips, with ambient around 10~25 degC.
I will try your suggestions when I have some more time. Thank you, phil.

Here to answer your problem, the temperature for chips must not reach 90 if reach it will die off and casue the chips to became XXXXXX. It has become a dead chip. See which blade it is and send it back to claim warranty from bitmain.

[stealthants76]

Hi guys.
Well you not really the only one. I had some xx but after restart most of the time it goes well. How ever IF i used the stock firmware. After update i got even more weird stuff going on.
My psu is antminer 1600w so no problems there. I even use another fan on it to keep it real cool.
But i just been talking to support about my last weird thing... After like two days of using the s9 like to get slower and slower... ? wtf

Temp is not a problem right now. As when the temp was at around 90C the s9 pushed on really well.. Untill now

Also one hash board gets pretty much hw unlike the other two. I changed the port on controller and got more better results but now again.. back where i was before :S

classic example of a hot tired psu.  drop your clock freq  say 550 is stock drop to 500  see if the hash rate drops after 2 days.

But while i tested the psu on S7 and it worked well??? I have a different fan on the psu also so it is alot more cooler than before.

[bilabonic]

Hi guys.
Well you not really the only one. I had some xx but after restart most of the time it goes well. How ever IF i used the stock firmware. After update i got even more weird stuff going on.
My psu is antminer 1600w so no problems there. I even use another fan on it to keep it real cool.
But i just been talking to support about my last weird thing... After like two days of using the s9 like to get slower and slower... ? wtf

Temp is not a problem right now. As when the temp was at around 90C the s9 pushed on really well.. Untill now

Also one hash board gets pretty much hw unlike the other two. I changed the port on controller and got more better results but now again.. back where i was before :S

classic example of a hot tired psu.  drop your clock freq  say 550 is stock drop to 500  see if the hash rate drops after 2 days.

But while i tested the psu on S7 and it worked well??? I have a different fan on the psu also so it is alot more cooler than before.

BUT they are still hitting 90 !!!!

[philipma1957]

Hi guys.
Well you not really the only one. I had some xx but after restart most of the time it goes well. How ever IF i used the stock firmware. After update i got even more weird stuff going on.
My psu is antminer 1600w so no problems there. I even use another fan on it to keep it real cool.
But i just been talking to support about my last weird thing... After like two days of using the s9 like to get slower and slower... ? wtf

Temp is not a problem right now. As when the temp was at around 90C the s9 pushed on really well.. Untill now

Also one hash board gets pretty much hw unlike the other two. I changed the port on controller and got more better results but now again.. back where i was before :S

classic example of a hot tired psu.  drop your clock freq  say 550 is stock drop to 500  see if the hash rate drops after 2 days.

But while i tested the psu on S7 and it worked well??? I have a different fan on the psu also so it is alot more cooler than before.

BUT they are still hitting 90 !!!!

60  90 is fine 
65  95 is fine
70 100 is pushing your luck

[tbonetony]

I think my S9 is good now. The issues I was seeing seemed to only happen when changing a pool via MRR. Some renters use less stable pools that can cause frequent disruptions. When that happened, my S9 got screwed up, maybe due to the tight PSU it paired with.

[ANTZ25]

I'm having similar issues, I just posted the detailed problems with caps, https://bitcointalk.org/index.php?topic=1643027.0 is this similar to what happen to you?