Topic: Antminer D3 Blissz firmware (10/12 v1.12 update) - page 6. (Read 125923 times)

Yup, gave that a shot too, no luck. Oh well. Tried with different cables, different slots.

if if falls below 92% of the expected hashrate for a few minutes it will reboot.

sounds to me like a dead board that needs to be shipped back to bitmain, but you can try one more thing: try to connect it as first board and then disconnect the working ones. Try swapping the communication cables as well.

Seems I may have a dead board too 

Was running fine till last week, the last 6 chips decided to x, but the device seemed to be ok with the remaining 54, had it set to under 400 freq with no errors. (Tried to change voltages and feq and flash back to stock, still same last 6 died shortly after starting to mine.)

Few days ago the chain disappeared completely, but I could see it was still working by the total hashrate. Now its totally dead, hashing power nor the chain showing up at all.

Have tried resetting defaults, flashing back to stock, resetting again, flashing back to Blissz, underclock to 200 and set voltage to 13. Even opened her up to check, but nothing appears off.

Debug reports 0 ASIC for that chain, and the only thing I can see that is different is the temp sensor listing offset to 00 for that chain.

Jan 14 08:25:55 (none) local0.notice cgminer[376]: set_temperature_offset_value_chain_sensor: Chain2 Sensor0 temp offset : 00,
Jan 14 08:25:55 (none) local0.warn cgminer[376]: Chain 0 set_temperature_offset_value_chain
Jan 14 08:25:55 (none) local0.warn cgminer[376]: Chain 0 set_temperature_offset_value_chain_sensor
Jan 14 08:25:55 (none) local0.notice cgminer[376]: set_temperature_offset_value_chain_sensor: Chain0 Sensor0 temp offset : -2,
Jan 14 08:25:55 (none) local0.warn cgminer[376]: Chain 1 set_temperature_offset_value_chain
Jan 14 08:25:55 (none) local0.warn cgminer[376]: Chain 1 set_temperature_offset_value_chain_sensor
Jan 14 08:25:55 (none) local0.notice cgminer[376]: set_temperature_offset_value_chain_sensor: Chain1 Sensor0 temp offset : -2,
Jan 14 08:25:55 (none) local0.warn cgminer[376]: Chain 2 set_temperature_offset_value_chain
Jan 14 08:25:55 (none) local0.warn cgminer[376]: Chain 2 set_temperature_offset_value_chain_sensor
Jan 14 08:25:56 (none) local0.notice cgminer[376]: set_temperature_offset_value_chain_sensor: Chain2 Sensor0 temp offset : 00,

Any other things I can try before I unplug it completely and just run 2 @ 500watts at least

I would add this - low hashrate option - to my miner if i could, is hosted. I figure it drops to 12000GH/s and below and the host cannot reboot it each time, so it stays like this...

by the way, low hashrate option, at what degree of low hashrate it does reboot? would work for my example of 12000 gh/s?

Maximum Temperatures for the D3?

So, I've been confused about the maximum temps for the D3 the entire time and wanted to see if you guys know the correct max temps.

Just to recap, In the manual it states that the Maximum Ambient Operating Temperature is 40 degrees Celsius, and that the miner will shut down if the PCB Temp hits 85 degrees Celsius.

Now, I see most people referring to the Chip Temp, not the PCB Temp.  I've been trying to keep my Chip Temp below 85 by downclocking my D3s, but now I'm thinking that it's not necessary.  I've seen where the Chip temp on the S9s is commonly over 100C, and I saw a post which indicated it was from bitmain support that said the S9 chip temp should stay below 115C.

So, should when people refer to Max Temps, should we be referring to the Chip or PCB?

And if bitmain says in the manual to keep PCB under 85C what should we try to keep Chip temps at?

I built some sound reducing boxes (miners in garage), but they increase the temps a little.  Even in the boxes though the PCB temps are in the mid 60s on one D3 (smaller enclosure, 450 clock) with the Chip temps around 80, my other D3 (newer larger enclosure, 500 clock) has PCB in upper 60s, chips in low 70s.  I was worried the temps were too high, but now I'm thinking I should just crank them both up as long as they stay under 80 PCB temp?

Anyone have the official Chip temp max?

Yes, open a support ticket with Bitmain and ship that hashboard back for repair.

it seems like its realy dead. mine first chain is dead aswell. i tried all cables etc.
i will send it back 4 warranty repair on monday.
the warranty procedure of bitmain is the biggest shit ive ever saw on the internet! But i managed to get it done.
hopefully they dont send back used crap. as they did to others *fingers crossed*

@blissz
great work with the firmware! i like the fact that my miner dont sound like a F1 car anymore cause i can manage fanspeed properly!
is it possible that you make a shutdown button at the GUI? or is the hardware not made 4 such a feature?

Blissz, no go. chain 1, 200mhz, 13. Stepped up to 14, 15, 16, 17. Nothing. The other two cards run fine when connected. The machine is out of warranty. Please advise. 

Yep. In a "I like to know" way i guess. It gives me an idea of the steps and what to expect. I've messed quite a bit with input voltage with the S'es and pen/resistor mods on S1/S3. Anyhow, thanks for the work, per board settings is top notch.

I can't really tell the real voltage as I hardmodded my unit at first. I can't tell it per ASIC chip as I haven't measured, but for the total chain the voltage is something like this:

lowest: 7.3v
default: 8.2v
highest: 9.1v

I hope this helps

I rolled the dice and flashed my D3, i have to say, this firmware has a lot more proper-like features than Bitmain's stock crap. Kudos there. I was wondering about the voltage, are you using numbers because you can't actually read/tell the real values?

So I started to try to work on this... and here is my opinion on the matter AAAAARRRRRRGGGGHHHH!! stupid @#$$%#@^#^# thing!!!
There was much wailing and gnashing of teeth leading up to that conclusion.
I'm a little ocd about stupid things though so I'll go from "I never want to see that thing again until there is a baseball bat in my hand"
to "welllll maybe I could try this one other thing...."

Thank you for reply!
I tried to use method described in the topic start at first and... it didn't work. The thing is that miner doesn't get connected to network at all. It doesn't get IP, so I cannot get to the web interface.
The green light on miner starts blinking, blinks for some time and then goes off. Then, miner tries to start again (check fans etc).
Before adding files from custom firmware there was no even check procedure. So, I guess some of the necessary files disappeared...
How can I get full set of files? ...

The topic start shows you how to make a correct sd card for recovery, but in your case I don;t think it's needed: just connect one chain instead of them all and see what happens. It can also be the case that you set your clocks too high: so after it started up: go the advanced screen directly and set the voltage to lowest and frequency to 400Mhz as a starting point. I am quite sure that will help you out

Glad it helped you out

Please try to run it with only that board connected to the first chain connector. Try to swap PSU / chain cables to see if it gets detected. Try to start at default voltage (13) and set a low frequency to see if you can get it working (200Mhz).
If that works: decrease voltage / increase clockspeed slowly until you found the sweet spot

It would be great if it was possible to separate the different algorithms to have the option to run them on their own.

The best thing about this firmware for me is the fact that my miner runs despite the front fan sensor being faulty.  Previously, it would shut down because the fan is not sending the correct signal to the control board.  Now it runs and ignores that (which is perfect).  The fan works 100%, speeds up, slows down, but doesn't send it's speed to the controller.  I don't have to replace this expensive $80 fan any more now, it just runs.

Unlikely.  **

**Which from me translates to no way in hell really but I don't like to make absolute statements...
there is always something you don't know like :
*IF* (also unlikely)  bitmain has secretly developed the asics to do the missing algos and has some sort of virtual secret club handshake to turn it on
X11 to x13 to x15 is simply adding a few algos and with an FPGA "technically" easy-peasy but with a hardcoded chip? The ability would have to already be there before you could do it.. now that doesn't mean bitmain at this very moment isn't getting ramped up to run x13 and x15 and anything else really .. 'cause if you can do those then "almost" any algo can be done with asics with little extra R&D costs.

See what I'm sayin'?