Antminer R4 - Wrong chip index

[GWhisper]

I've got an R4 which will not mine, the system boots and it seems that when the auto frequency script runs it hits an error and can't proceed.

Kernel log-

Miner fix freq ...
Fatal Error: DownOneChipFreqOneStep Chain[-1] has Wrong chip index=7893959

Here's the code-

if(max_freq_chipIndex<0)
   {
      sprintf(logstr,"Fatal Error: DownOneChipFreqOneStep Chain[%d] has Wrong chip index=%d\n",max_freq_chipIndex);
      writeInitLogFile(logstr);
      while(1)sleep(1);




Any ideas?

I've rebooted and power cycled a number of times, re-flashed the firmware - tested with another know good PSU. I've logged a ticket with Bitmain but thought perhaps someone on here had seem this issue before and may have a fix. I thought it would be unlikely for both hash boards to be faulty.


***edit - I think https://bitcointalk.org/index.php?topic=1741088.0 is the same problem.

My testing has concluded I have a dead hash board - miner is now functional with just a single board connected.

[smracer]

I just got 10 R4's in today and 1 of them has this same problem.  I really don't want to have to send another miner back to Bitmain.  Anyone have any ideas?  I updated to the latest firmware on Bitmain's site and still have the same error.

Chain[J7] has 63 asic
Chain[J8] has 0 asic
retry Chain[J8] has 0 asic
retry Chain[J8] has 0 asic
retry Chain[J8] has 0 asic

Miner fix freq ...
Fatal Error: DownOneChipFreqOneStep Chain[-1] has Wrong chip index=7893903

[fanatic26]

A bad chip is a bad chip. The way the boards are laid out it only takes one bad chip to take a board out. There isnt much you can do except warranty it.

[qctechno]

I was missing a board on my R4B1 , cold solder on the big inductor. Re-soldered it, need a lot of heat to do the job
I redo all connector too. Cleaned with alcohol , dryed with compress air. Back in business. 

[GWhisper]

A bad chip is a bad chip. The way the boards are laid out it only takes one bad chip to take a board out. There isnt much you can do except warranty it.

I've just disconnected the faulty board and I'm mining with the remaining board. Bitmain want the entire R4 back not just the hash board - I'm waiting until after CNY to ship it off, no need to make a bad situation worse.

[NotFuzzyWarm]

A bad chip is a bad chip. The way the boards are laid out it only takes one bad chip to take a board out. There isnt much you can do except warranty it.

Iv'e just disconnected the faulty board and I'm mining with the remaining board. Bitmain want the entire R4 back not just the hash board - I'm waiting until after CNY to ship it off, no need to make a bad situation worse.

I just joined the sick R4  Club... ref https://bitcointalk.org/index.php?topic=1597639.msg17559940#msg17559940
For now at least the bad board is still somewhat working with 55 chips vs the normal 63 and giving 1-1.4THs vs the usual 4.4THs .
Lovely timing....

[MyRig]

Hash PCB is probably damaged. 

Any mod or self attempt repair may result in warranty denial tho.  Just be careful.  To be safe for warranty, when the unit does not recover with simple power off power on, or testing with already announced trouble shooting method then start warranty process or repair process be the safest way.

We do see many "eye opening" hash pcb and control module send in for repair that were previously self attempted and that process caused more damages than it fixed the problems.

I've got an R4 which will not mine, the system boots and it seems that when the auto frequency script runs it hits an error and can't proceed.

Kernel log-

Miner fix freq ...
Fatal Error: DownOneChipFreqOneStep Chain[-1] has Wrong chip index=7893959

Here's the code-

if(max_freq_chipIndex<0)
   {
      sprintf(logstr,"Fatal Error: DownOneChipFreqOneStep Chain[%d] has Wrong chip index=%d\n",max_freq_chipIndex);
      writeInitLogFile(logstr);
      while(1)sleep(1);




Any ideas?

I've rebooted and power cycled a number of times, re-flashed the firmware - tested with another know good PSU. I've logged a ticket with Bitmain but thought perhaps someone on here had seem this issue before and may have a fix. I thought it would be unlikely for both hash boards to be faulty.


***edit - I think https://bitcointalk.org/index.php?topic=1741088.0 is the same problem.

My testing has concluded I have a dead hash board - miner is now functional with just a single board connected.

[GWhisper]

Now my second R4 from batch 6 has now lost a hash board. Reliability isn't looking so good. That's 2 dead boards out of 2 miners.

http://imgur.com/a/H07m7

[fanatic26]

I am pretty sure the reliability issues are going to be the same across all the .10j/gh chips regardless of them being in an S9 or R4. They are trying to run them on the jagged edge of the power required to run the chip properly hence the T9 being slightly less efficient but most likely much more reliable.

[GWhisper]

We'll see, I should have some T9's early next week. Fingers crossed, they fair better.

[boomin]

Chalk up another R4 problem.  Same shit - day after I got it.  I have 2 one has been going for 3 days *got here 1 day earlier for some reason.  Now I didn't even get 48 hrs out of this one.  Not cool. 

[beltsniffer]

Chalk up another R4 problem.  Same shit - day after I got it.  I have 2 one has been going for 3 days *got here 1 day earlier for some reason.  Now I didn't even get 48 hrs out of this one.  Not cool.  

Same problem here. Though mine only hashed on both boards for about an hour. Now it's just one board. Hopefully, that one keeps working until after their holiday.  I was thinking of sending to Bitmain Warranty and eating the repair cost, since it saves on shipping and downtime.  Anyone know what they are charging for an R4 repair?

[boomin]

Charging for a repair?  That's ridiculous!  ITS under warranty.  The real problem is with trying to run the unit with only one board is the airflow.  The air takes the path of least resistance and will bypass the heatsinks if you try to take out the other board.  Mine originally worked with both hash boards, but I noticed that it was only hashing on one board.  So the natural thing to do is restart it....DONT!  Then you have to troubleshoot the bad board and disconnect it from the hash plug to make just the one board work.  They should advanced exchange a mining board on something less than 30 days old.  I don't mind paying for the board and waiting for a credit, but to buy something "brand new" and have it not make it 48 hrs is ridiculous.  In any other industry the company would be run in to the ground and no one would support them EVER AGAIN!  Now we know why United States has laws governing too much control of an industry.  Makes me sick I have to work on anything I just paid money for.  Can you imagine buying a new Ferrari and the tire falls off and they say oh we are sorry, please send back your $ 300k car for a month while the value of your new car is going way down so we can do what we should have done in the first place!?

I am not happy. 

Bitmain needs to address this issue!  I will pay the .25 btc for a hash board to be sent NOW and then when you get my defective one I ALREADY PAID FOR back then return my .25 BTC deposit.  The difficulty is going thru the roof since they just disconnected all these machines and hooked up their "new in 2 months" to us machines when we pay for those machines. 

Does anybody else think this sounds unreasonable?  any other company would advanced exchange the entire unit for a new one *but I understand BTC is a different game.  They would need a deposit....

Please chime in...

Boomin

[beltsniffer]

Charging for a repair?  That's ridiculous!  ITS under warranty.  The real problem is with trying to run the unit with only one board is the airflow.  The air takes the path of least resistance and will bypass the heatsinks if you try to take out the other board.  Mine originally worked with both hash boards, but I noticed that it was only hashing on one board.  So the natural thing to do is restart it....DONT!  Then you have to troubleshoot the bad board and disconnect it from the hash plug to make just the one board work.  They should advanced exchange a mining board on something less than 30 days old.  I don't mind paying for the board and waiting for a credit, but to buy something "brand new" and have it not make it 48 hrs is ridiculous.  In any other industry the company would be run in to the ground and no one would support them EVER AGAIN!  Now we know why United States has laws governing too much control of an industry.  Makes me sick I have to work on anything I just paid money for.  Can you imagine buying a new Ferrari and the tire falls off and they say oh we are sorry, please send back your $ 300k car for a month while the value of your new car is going way down so we can do what we should have done in the first place!?

I am not happy. 

Bitmain needs to address this issue!  I will pay the .25 btc for a hash board to be sent NOW and then when you get my defective one I ALREADY PAID FOR back then return my .25 BTC deposit.  The difficulty is going thru the roof since they just disconnected all these machines and hooked up their "new in 2 months" to us machines when we pay for those machines. 

Does anybody else think this sounds unreasonable?  any other company would advanced exchange the entire unit for a new one *but I understand BTC is a different game.  They would need a deposit....

Please chime in...

Boomin

How about not even 4 hours. Well, I had good luck with Bitmain's stuff. I've been buying their products since the 335mh stick miners. So when I got the first R4 and it worked well, I figured I'd gamble on a second.

Bitmain Warranty is and isn't Bitmain. I personally think they should be honoring Bitmain's 180 day warranty, especially if they are gonna call themselves Bitmain Warranty. The fact that I even have to pay anything to ship it back to them irks me, but I knew their terms when I gambled on it.

Paying for shipping to Colorado vs Hong Kong plus the shorter downtime, might be worth it if the repair cost was cheap enough.

[sidehack]

Far as I know, Bitmain Warranty is not Bitmain at all but a separate entity that got good at fixing Bitmain's stuff on their own and decided to make a business of it.

Additionally, Bitmain never sold a 335MH stick miner.

[beltsniffer]

Far as I know, Bitmain Warranty is not Bitmain at all but a separate entity that got good at fixing Bitmain's stuff on their own and decided to make a business of it.

Additionally, Bitmain never sold a 335MH stick miner.

It's quite probable that I have the mh/s incorrect. It's been a while since I fired them up. They were USB miners.

kinda funny though, they made me almost as much money as my r4 does today at quite a bit less cost.

As for Bitmain Warranty. I just assumed they were affiliated, since they were using the same font/logo as Bitmain albeit a different color ant.

[Andrkondr]

Most strange thing is that if you will disconnect power supply cables from the non-working board and keep data cable connected "Wrong chip index" error still trying to find working asic on that board and shows 0 asic. In the end just block miner hash possibility.

How it can be possible?
How controll board receive info from non powered board?
Or maybe Hash board was not power plugged at all after mulfunction?

[NotFuzzyWarm]

Most strange thing is that if you will disconnect power supply cables from the non-working board and keep data cable connected "Wrong chip index" error still trying to find working asic on that board and shows 0 asic. In the end just block miner hash possibility.

How it can be possible?
How controll board receive info from non powered board?
Or maybe Hash board was not power plugged at all after mulfunction?

The I/O signals (typically 3.3v) between controller and hash boards is fed from the controller board via the data cable. At least on s7's and I have to assume also with the S9's and R4's that cable also feeds the PIC. The 12V only feeds the ASIC cores.

[philipma1957]

yeah my 50 day old unit has  1  board like the op's.   I purchased it on Dec 08  I have tired a few ideas none work.

My 11 day old has good boards and fan is terrible nasty sound..


 I am going to try to get someone to authorize a fan swap on my end  so that I only send in 1 machine.

[boomin]

Has anyone had any luck trying to fix it?  I see a couple weird posts that dont make any sense but its frustrating as hell.  I see the pic is actually verifying the hash board SN with the controller now tho.

boomin