Author

Topic: Antminer R4 trouble (Read 222 times)

legendary
Activity: 2030
Merit: 1573
CLEAN non GPL infringing code made in Rust lang
October 01, 2019, 08:12:09 AM
#11
Image is clear as day, however I have one small concern, since all those vents are intake , and by placing that gear this way you are blocking nearly half of the intake, this positioning does give the exhaust side a much better room , but also by limiting the intake space they boards can run hotter, have you actually measured temps in different positioning?

No, you only block a third, but that inevitably happens in any position anyway. There are 3 sides with grills to that fan. Usually scroll fans only use 1 wall with grill for intake, this has 3 walls and blocking 1 doesn't appear to harm. But you can measure it if you want. All 3 sides suck air, and its all pushed up thru the hashboards.

If you are still concerned, you could rise up the miner placing a couple of pieces of wood or something, but i found it unnecessary.
legendary
Activity: 2436
Merit: 6643
be constructive or S.T.F.U
September 30, 2019, 05:57:18 PM
#10
Thanks for the help, getting back you with results Smiley

Yes you can simply unplug board J7 and the miner should run just fine, but don't give up on it without trying everything at your disposal.

The gears i found from a chinese reseller in china of course, they have a group on telegram, pm me if you want the contact.
copper member
Activity: 97
Merit: 1
September 30, 2019, 04:14:33 PM
#9
[...]

Okay, there we go... so theoretically i could disconnect the board in port 7, and it should work if the other board is fine?
Please let me know where you found these machines? Still got some power to spare Wink

Thanks for the help, getting back you with results Smiley
legendary
Activity: 2436
Merit: 6643
be constructive or S.T.F.U
September 30, 2019, 03:06:16 PM
#8
Image is clear as day, however I have one small concern, since all those vents are intake , and by placing that gear this way you are blocking nearly half of the intake, this positioning does give the exhaust side a much better room , but also by limiting the intake space they boards can run hotter, have you actually measured temps in different positioning?
legendary
Activity: 2030
Merit: 1573
CLEAN non GPL infringing code made in Rust lang
September 30, 2019, 11:28:13 AM
#7
[...]

Yes they are very quiet, i forced mine to 100% fan speed and it is in a living room and nobody cares about it, it is that quiet, like a gentle breeze or a typical household fan.

The fan bearings probably last longer in the correct position. Your image shows the unit lying down, all you have to to is stand it up, but the scroll fan remains horizontal. I'll show you an image, unfortunately my internet is limited and searching for images is difficult for me at this moment.

You could also imagine a toaster, with each hashboard being one bread. The controller ends vertical.



Hoping that you can see the image, this is the only correct way to use the R4 (You can push the data ribbons to the side).

Most images on the internet disturbingly show them lying down flat, never do that, it damages the upper hashboard and overheats the unit because the fan is not that strong even at 100%. BTW, you can force the fan to 100% by disconnecting the blue wire (4), you can gently take it out of the plug without harming it by pushing a little dent on it (same as all computer fans). I think this fan lasts longer when forced at 100%, but YMMV. Some fans do 3240rpm and some do 3000rpm. If this fan fails i don't know where to get more, maybe alipay/alibaba but I'm don't know where Bitmain sourced those from.

Also there are a couple of variants, I'm not sure how many batches were (6?), but it seems the later batches used even cheaper fans, and its a bit smaller. Suffice to say, many people burned them due to bad positioning, AND Bitmain also did poor soldering of parts in some batches.

The R4 was a good design concept, handled badly by its own company. Besides, they don't care about home users, only large industrial miners, which is why they killed it. They work with the same PSUs as the S9, and even that weaker variant that Bitmain once did for them that could do 120v. 845W @ 8.5TH/s using Dec 2018 firmware according to official specs. Its pretty much a 2 hashboard S9, in fact the controllers can be swapped, but remember the S9 one expects 2 working fans with their firmware...

I think not all batches used the same Xilinx 1g ram controller, i remember a 512mb ram one being mentioned somewhere.
legendary
Activity: 3500
Merit: 6320
Crypto Swap Exchange
September 29, 2019, 06:18:10 PM
#6
[...]

All of mine are like that with no issues. The only ones I have problems with are the ones I got 2nd hand from people.
As with all things YMMV.
The biggest issue is the cheap ass bearings they have in the fan.

-Dave
legendary
Activity: 2436
Merit: 6643
be constructive or S.T.F.U
September 29, 2019, 06:08:51 PM
#5

  _Air exchaust_
 |  ^^^^^^^^^^  |
 |_Hash boards__|
[___Fan Intake___]


Do NOT lie flat, do NOT have scrolling fan in vertical position, EVER.

Nice ascii artwork, but I don't seem to understand how is it possible to position the miner this way, if the description in the image below is accurate



source: https://www.eastshore.xyz/tutorial-antminer-r4-bitcoin-miner-16nm-quiet-asic-miner/

I found a few brand new R4s in China, they go for about 270$ each , I might get me 1 or 2 of them to set-up at home for the winter since they are really quite, so I get to use the free limited power at home to earn some extra BTC and keep the house warm, I never had any of those gears before and I am eager to test them, also I would like to know if they are really quite enough to sit in the next room.

but i dont what to look for?

This is what to look for

Code:
main.c:9626: retry check chain[6]: asicNum = 63

Chain 6 has 63 asics, so all good.

Code:
main.c:9626: retry check chain[7]: asicNum = 0
driver-bitmain.c:1825: set_reset_hashboard = 0x00000080
driver-bitmain.c:1825: set_reset_hashboard = 0x00000080
driver-bitmain.c:1825: set_reset_hashboard = 0x00000000

chain 7 however is showing 0 , this does not mean all chips on J7 are dead, it usually mean 1 chip is dead and that is good enough to kill the board, a lose cable or any of the other reasons that have been already mentioned may cause this issue, so try to troubleshoot it , but be prepared for the worst and the most likely scenario which is a dead hash board.
legendary
Activity: 2030
Merit: 1573
CLEAN non GPL infringing code made in Rust lang
September 29, 2019, 02:58:13 PM
#4
1) Do you know which board went bad? Every R4 I have had issues with the top board went bad. Heat rises....who knew...

2) A trick that sometimes worked for me was update to the latest firmware (2018-11-19) then go back to the 2017-04-28 firmware and if that did not work go back to the 2018-11-19 firmware. That has allowed me to revive some dead boards now and then.

-Dave

The only reason that would happen is if they place the miner wrong. The R4 has only and only one valid position, exhaust air going UP. Any other position and you are damaging it.

Of course, Bitmain was very negligent and even their pictures often show them wrongly placed. The R4 can only be used with the exhaust air pointing up like a chimney, with the scrolling fan lying flat precisely because of what you said.

Incidentally this makes it very unsuitable for the typical mining farm arrangement.

Lets do some ascii artwork:

  _Air exchaust_
 |  ^^^^^^^^^^  |
 |_Hash boards__|
[___Fan Intake___]


Do NOT lie flat, do NOT have scrolling fan in vertical position, EVER.

Check your wiring connections. Mine is with 2018 firmware and it works fine. As with the S9s, test each hashboard separately, swap power and data cables.
legendary
Activity: 3500
Merit: 6320
Crypto Swap Exchange
September 29, 2019, 10:10:14 AM
#3
1) Do you know which board went bad? Every R4 I have had issues with the top board went bad. Heat rises....who knew...

2) A trick that sometimes worked for me was update to the latest firmware (2018-11-19) then go back to the 2017-04-28 firmware and if that did not work go back to the 2018-11-19 firmware. That has allowed me to revive some dead boards now and then.

-Dave
legendary
Activity: 2506
Merit: 1714
Electrical engineer. Mining since 2014.
September 29, 2019, 09:59:00 AM
#2
The problem is with the hash board connected to the control board connector J7.

I have you tried these tricks? -->

1. Try username RadekG's method for testing one hash board per time.

I found solution by unplugging one board and run miner only with ONE board until it warm-up and start hashing. After that I did the same with second board. Both boards worked, so I plugget them together and miner started with both boards running as new. Check kernel log for ASIC=63 during warming-up phase.

2. Try to swap the data cable in to another connector in the control board.
See if the problem is the same when using another connector.



3. Try to reset the miner using the IP Report button reset method.
https://support.bitmain.com/hc/en-us/articles/115003595387-How-to-reset-miner-to-factory-settings- (see chapter 3 !)

If your hash board still does not wake up, I think it is safe to say that it is a dead one.
copper member
Activity: 97
Merit: 1
September 28, 2019, 08:22:10 PM
#1
Hey!

some days ago, one of my Antminer R4s halfway died, one board was hashing nicely, and the other wasnt available. All this was seen on pool-side stats.
I tried rebooting the faulty one, without any good results. so i tried re-flashing the currently most up to date software for the Antminer R4 found on Bitmains support.
this night i took it all apart, cleaning the hashboards and cleaning the controller aswell as the fan and entire case, all clean!
still, no results. therefore im asking the community, which i know have several good people at finding errors in these Antminer-machines.
i have a copy of the current Kernel-log, but i dont what to look for?

Hope to get some help from you professionals! Cheesy

Link to the Kernel-log:  https://pastebin.com/JiyCnc5b

Kind regards emilio932
Jump to: