Canaan Avalon 1246 Repair Suggestions Needed - 1 Hashboard not working

[RagingTalonS]

I bought I Canaan Avalon 1246 from Blokforge.

It hashed for 1-2 days then died. No hashboards working.

I contacted them and after a week or so got new firmware emailed and then installed it. Now at least 2 or the 3 boards are hashing (so not losing 100% $ and time).

They passed me off to Canaan directly to get warranty work. Spoke with Canaan rep directly in China through whatsapp and he said he would send a new hashboard to Blokforge and then Blokforge would send to me to install myself.

The issue is:

1: Im new to asic mining;
2. Ive never repaired or even tore down an asic miner;
3. There is NO videos on the 1246 or something even close I can find doing it on youtube or elsewhere that Ive searched.

Help please!


Updated 4/25/21:
I didnt get anywhere with Canaan. I got in touch with tech support via whatsapp but they said a hashboard was bad and they were sending a new one by end of March. It still hasnt come.

Yes I ordered to 90Th and Blokforge also bait switched me but did give me option to cancel order. I opted to keep the 85TH unit instead and it is earning .0005BTC avg per day now.

I tried to take my canaan 1246 apart. Pulled off fans and the cover on top and wiggled things around with my hands that I could and then put it all back and IT STARTED WORKING miraculously. Im not sure wtf I did but that little wiggle caused some connect with that bad hash board to connect with control board and it started working 3 boards now. Looked at a video of a bitmain s19 repaitr an I think it may be similar but I dont have the tech skills yet to do it. I made a video of what I did do though but it isnt uploaded anywhere yet.

[HagssFIN]

You can look my Troubleshooting guide for help, the link is in my signature space.

If you are totally new in electrical works, you could also take the pieces to some electrical appliance repair shor or to some professional electrician.

[RagingTalonS]

Thanks. Yes I did see that guide.

I tried to take it apart today. Took off fans and side of psu then couldnt figure out how to take off psu or controller unit on top without damaging so I just put it all back.

Now its hashing @ 80Th see if it dies again in 2 days like before. Supposed to be 87-90TH.

[spoonbandits]

There isn't any clear information out yet on how to service these so we will have to wait until more people crack them open. I had a similar experience, as opening/servicing wasn't as straightforward as other miners and I didn't want to risk damaging anything.

Some notes about my recent experience with Blokforge and the Avalon 1246:

I had a sort of "bait & switch" with Blokforge, not sure if they did it intentionally or not, but NOTE: There are various types of 1246s out there, and Blokforge was originally advertising the 90TH/s model, however they are shipping the 85TH/s model (1246-85). Also, Blokforge only supports for about 10 days after shipment is delivered, in which case you'll get pretty poor support from Canaan (w/big language barrier).

If you're running in "Normal Power" mode, and getting 80Th/s, and "High Performance Mode" gets you ~84TH/s, you most likely have the 1246-85. To summarize: When ordering the 1246 from Blokforge, please note that you may NOT get the 90 TH/s model. This being said, I have no issue with Blokforge, as they've always been helpful in procuring miners and always reply quickly to inquiries.

Now, technically - some issues I've found with my 1246:

1.) One of the ASIC chips misreads temperature way too high (~120-180C), unfortunately not worth repairing as it hashes fine. Yet this misreading will still affect the overall average temperature reading, and cause one of the fans to run unnecessarily higher.

2.) After about 2-3 weeks of mining, one of the fans is constantly at nearly 100% RPM, regardless of how cool the unit is running. I suspect this fan will die out sooner than it should.

3.) Factory reset did nothing to alleviate these issues.

Sending logs to Canaan simply gets a reply: "Everything is working fine with your miner". Additionally, there are no new firmware upgrades yet, and online resources are basically nil, so please be sure to post as much info as you can for those running/buying this unit.

For anyone who is interested, I would wait a little bit before buying the 1246 as it seems there are still some issues with it. Aside from the points mentioned, it is still pulling about ~.0005BTC/day depending on the pool.

[spoonbandits]

Hey. I have one miner that is not working and there is no support online and the email support is not responding.

Can you please provide me with the Firmware you got from Canaan and what is the best option to get in contact with Canaan support team

Thanks.

Did you use Blokforge or different distributor? I was able to get a somewhat newer firmware for the 1246 by contacting Blokforge through their support system (support.blokforge.com). Unfortunately Canaan hasn't updated their site with the link to the latest one.

[RagingTalonS]

Hey. I have one miner that is not working and there is no support online and the email support is not responding.

Can you please provide me with the Firmware you got from Canaan and what is the best option to get in contact with Canaan support team

Thanks.

Did you use Blokforge or different distributor? I was able to get a somewhat newer firmware for the 1246 by contacting Blokforge through their support system (support.blokforge.com). Unfortunately Canaan hasn't updated their site with the link to the latest one.

I got some new firmware from blokforge.

[happysquid]

I lost a board, red hash2 on root menu and the H2 temp reads 32degrees f. I see the center board light is out so I removed it but chickened out on removing heat sink screws, don't know if there is anything I could do if I did find something....emailed and wtg response from Canaan.  

[HagssFIN]

First things first, can you post your log?
Please use the "code" function ( # ), while you post it.

[happysquid]

HagssFin,
I can copy the log but would need guidance on Code stuff...
Is new firmware utilized for batch management or will it fix stuff? like one board not hashing.
haha, honestly I did try to google how to #function, but thanks for the push in the right direction.

[happysquid]

Hagssfin,
  appreciate the assistance, I think I learned #function.. loosing btc is a good incentive...

Code:

{'STATUS':[{'STATUS':'S','When':554,'Code':11,'Msg':'Summary','Description':'cgminer 4.11.1'}],'SUMMARY':[{'Elapsed':648,'MHS av':53952122.14,'MHS 30s':59295536.61,'MHS 1m':58413199.35,'MHS 5m':50175096.20,'MHS 15m':28356242.80,'Found Blocks':0,'Getworks':26,'Accepted':119,'Rejected':0,'Hardware Errors':5,'Utility':11.01,'Discarded':3353921,'Stale':0,'Get Failures':0,'Local Work':15978,'Remote Failures':0,'Network Blocks':1,'Total MH':35109604396.0000,'Work Utility':756403.35,'Difficulty Accepted':7798784.00000000,'Difficulty Rejected':0.00000000,'Difficulty Stale':0.00000000,'Best Share':18284439,'Device Hardware%':0.0001,'Device Rejected%':0.0000,'Pool Rejected%':0.0000,'Pool Stale%':0.0000,'Last getwork':0}
{'STATUS':[{'STATUS':'S','When':554,'Code':70,'Msg':'CGMiner stats','Description':'cgminer 4.11.1'}],'STATS':[{'STATS':0,'ID':'AVA100','Elapsed':649,'Calls':0,'Wait':0.000000,'Max':0.000000,'Min':99999999.000000,'MM ID0':'Ver[1246-85-21022601_4ec6bb0_5f74eef] DNA[02010000c3adf89f] MEMFREE[1411984.0] NETFAIL[0 0 0 0 0 0 0 0] SYSTEMSTATU[Work: In Work, Hash Board: 2 ] Elapsed[649] BOOTBY[0x01.00000000] LW[411795] MH[3 2 0] HW[5] DH[29.661%] Temp[24] TMax[80] TAvg[65] Fan1[3598] Fan2[3604] Fan3[3572] Fan4[3572] FanR[44%] Vo[313] PS[0 1216 1252 176 2203 1252] PLL0[7901 2851 2135 1033] PLL1[7092 2818 2621 1389] PLL2[0 0 0 0] GHSspd[56921.95] DHspd[4.461%] GHSmm[59744.00] GHSavg[54055.58] WU[755147.76] Freq[357.66] Led[0] MGHS[27288.49 26767.09 0.00] MTmax[80 76 -273] MTavg[67 62 0] TA[240] Core[A3201] PING[28] POWS[0] HASHS[0 0 1] POOLS[0] SoftOFF[0] ECHU[0 0 131073] ECMM[0] SF0[520 540 560 580] SF1[520 540 560 580] SF2[520 540 560 580] PVT_T0[ 68 72 67 71 74 69 71 73 73 74 74 70 73 74 78 78 72 72 70 74 80 75 74 73 69 72 74 74 73 69 68 74 76 73 72 70 68 70 72 73 69 68 68 69 72 69 69 64 62 64 68 64 65 65 64 64 65 64 60 60 54 58 58 59 62 60 60 63 58 59 62 62 61 65 61 64 65 63 62 65 62 65 69 67 68 67 62 64 68 67 69 67 65 66 69 68 69 68 65 67 69 70 70 68 67 65 68 68 69 70 65 66 69 69 70 67 65 64 67 68] PVT_T1[ 61 61 68 65 62 61 61 65 64 67 66 63 62 64 67 68 69 63 65 64 69 76 66 65 65 68 69 70 67 68 61 65 66 66 65 67 64 65 68 68 65 64 59 64 69 63 64 61 59 62 64 60 67 58 58 61 65 65 57 57 54 52 50 55 60 55 54 54 55 59 59 56 57 59 55 57 58 61 62 59 59 59 61 63 62 60 59 60 62 64 63 61 61 62 65 64 63 63 61 65 68 65 60 66 63 63 66 66 61 64 63 63 64 63 64 64 64 64 61 65] PVT_T2[-273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273 -273] PVT_V0[294 298 299 293 292 293 294 292 287 292 291 289 287 287 283 283 288 286 287 285 287 289 289 290 291 285 285 283 288 288 287 289 291 285 290 289 289 287 289 287 292 292 289 289 291 292 294 293 294 292 293 294 295 293 295 297 296 299 296 301 298 295 298 292 298 300 292 295 298 292 292 295 295 298 298 291 292 292 295 299 298 298 298 294 294 292 293 297 295 297 292 294 291 287 287 291 292 292 294 293 289 290 297 295 293 295 294 293 294 295 294 291 291 292 293 290 293 293 297 299] PVT_V1[293 292 302 289 293 289 290 291 295 291 292 292 290 286 287 287 286 287 290 289 284 293 289 290 286 286 285 292 294 292 286 285 286 290 291 292 286 287 283 281 284 282 291 291 292 294 294 294 288 287 289 289 294 293 292 294 299 295 295 302 301 299 295 297 297 294 295 293 292 295 298 297 293 297 296 295 297 298 290 291 294 296 294 298 290 294 289 291 291 291 293 298 294 294 295 292 290 290 291 292 289 291 289 292 294 291 290 294 290 292 292 296 293 293 295 298 295 293 292 299] PVT_V2[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] MW[208455 208491 0] MW0[42 35 35 31 32 23 35 29 31 33 52 34 31 29 27 31 35 26 27 32 41 34 33 30 38 27 34 21 32 33 34 50 25 25 48 30 31 37 34 29 30 28 37 29 32 26 36 39 27 36 41 40 23 48 21 24 38 33 36 41 43 41 36 29 38 33 50 36 30 31 51 25 37 35 36 30 27 38 44 27 38 28 31 35 31 26 39 38 38 31 37 36 38 36 30 36 32 37 31 25 31 31 29 31 39 40 34 45 31 31 34 40 34 28 31 32 28 33 27 32] MW1[22 35 34 45 29 28 19 40 33 52 34 31 29 35 27 44 35 35 26 31 28 22 41 29 31 25 29 21 31 48 26 28 31 40 25 28 28 22 38 35 34 20 23 33 33 37 37 36 37 31 43 38 36 37 36 28 26 25 29 36 35 43 38 42 33 44 32 39 28 35 26 39 35 37 44 36 34 50 30 39 32 29 26 32 39 31 37 20 35 25 26 30 39 26 47 44 31 36 30 29 35 27 34 37 29 26 24 36 27 22 31 33 37 46 42 26 40 29 38 27] MW2[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] ASICCRC0[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] ASICCRC1[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] ASICCRC2[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] CRC[0 0 0] POW_I2C[OK] FACOPTS0[] FACOPTS1[] ATAOPTS0[--avalon10-freq 504:524:544:564 --avalon10-voltage-level 42 ] ATAOPTS1[--avalon10-freq 520:540:560:580 --avalon10-voltage-level 38 ] ADJ[1] COP[0 0 0] MPO[3200] MVL[87] ATABD0[520 540 560 580] ATABD1[520 540 560 580] ATABD2[520 540 560 580] WORKMODE[1]','MM Count':1,'Smart Speed':1,'Voltage Level Offset':0,'Nonce Mask':25}

[HagssFIN]

Based on all the value readings, and ECHU, your third hash board is broken indeed.

ECHU (error code hashing unit) for it is 131073.

To figure out the ECHU sum meaning,
we need to do "reverse sum math".
131073 = 1 + 130172
, in which:
1 = CODE_IDLE
130172 = CODE_HUFAILED

See more here, in Table 4. -> https://bitcointalk.org/index.php?topic=3224256

[happysquid]

HagssFin,
  I've referred to table 4 and I would assume from your reply I am at the mercy of Canaan. when I opened unit, I noticed the center board did not have red light on, so I assumed it was board #2. The temp that showed zero celcius is labeled H2. Of H0/H1/H2. Your review mentioned 3rd board? so I am assuming board with light out in the middle is bad. The temps # labels have me confused as to physically which one is bad? should I go by the light out or persue understanding the configuration of HU labels? in another post a fella mentioned Canaan would send blokforge (US seller) the replacement for him to install.
Thank you! your guides are most impressive. I have alot to learn. I thought it would be easy.

[spoonbandits]

HagssFin,
  I've referred to table 4 and I would assume from your reply I am at the mercy of Canaan. when I opened unit, I noticed the center board did not have red light on, so I assumed it was board #2. The temp that showed zero celcius is labeled H2. Of H0/H1/H2. Your review mentioned 3rd board? so I am assuming board with light out in the middle is bad. The temps # labels have me confused as to physically which one is bad? should I go by the light out or persue understanding the configuration of HU labels? in another post a fella mentioned Canaan would send blokforge (US seller) the replacement for him to install.
Thank you! your guides are most impressive. I have alot to learn. I thought it would be easy.

Seems there are a lot of problems with the Avalon 1246, and people are waiting on replacement hashboards. One of the hashboards on mine isn't working either, and sending the logs to Canaan they verified that it's indeed faulty. The board failed after about 8 weeks.

Canaan said they have multiple sources for their ASICs, so hopefully the turnaround time won't be too long, but who knows.

[HagssFIN]

The turnaround time could be minimized if Canaan agrees to send a new hash board in advance, and the user or the user's electrician replaces the old hashboard with the new one.

This was the idea that they had, at least in 10 series time, I think.

Not sure how much Nasdaq listing has changed things.
One thing I have noticed to change, is that I can't seem to buy a 1246 from them, for a review.
The last reply I got from them is from February, even though I have spammed them with emails also after that.

When Steven was in the house, he gave me a miner for review.

Now I can't even buy one for review.

I guess they don't care about reviews anymore...

Truth to be told, I'm losing my interest in ASIC mining, because it looks like its becoming impossible for a small guy like me to do.

Thus, few more miner generations, and I will know nothing useful to help other people here at the forum.

[happysquid]

HagssFin,
I greatly appreciated your help. I was educated by you on the log and had the forethought or afterthought to send it as a reply to my initial email. I wont throw names around as I am at their mercy, lol. The fella at Avalon replied and agreed that I need a hash board. And to work with blokforge supplier to fullfill this need. so I am happy to report that they are supporting the customer. I would think that if the market wasn't as strong at the time that you would have gotten one to review? maybe not as they have Voskcon review video already on their page, next to the manual in support page. I like blokforge but was a bit dissapointed that during my chat to resolve issue I wasnt assured that it will be taken care of. To him it was standard reply, but to me the machine I spent a handfull of bitcoin on was huge to me.  When it broke my lil world fell apart, kinda.

[HagssFIN]

Yeah I agree you on that one. If manufacturers ask as huge prices as they currently do, the gear better be bullet proof.

If the user is left with some probability risk of gear breaking,
then that down time risk should be seen in the initial price being lower than it now is.

[happysquid]

i cannot say it was anyones fault that the chip shortage or issue, my March batch came out 85th instead of 90, and the deal was accept or cancel. that was brutal ontop of premium cost.
 

[kano]

...
Not sure how much Nasdaq listing has changed things.
One thing I have noticed to change, is that I can't seem to buy a 1246 from them, for a review.
The last reply I got from them is from February, even though I have spammed them with emails also after that.

When Steven was in the house, he gave me a miner for review.
...

All the original devs I knew left.
I was in regular contact with a few of them and the last contact with one of them I was told they'd all left.

Steven appears to be no longer be there now either.
I tried to contact him recently about something we had discussed a year ago, but his email says it no longer exists.
I instead contacted sales directly, and they never bothered to reply (that was 1.5 months ago)

Xiangfu is no longer CEO - though I didn't ask him that directly, but it would appear Ng has replaced him.

[HagssFIN]

That's weird, I would've thought they at least reply you,
since they have made co-operation with you long time (eg. burn-in testing Avalons in Kanopool, they use cgminer, etc.).

[spoonbandits]

Yeah I agree you on that one. If manufacturers ask as huge prices as they currently do, the gear better be bullet proof.

If the user is left with some probability risk of gear breaking,
then that down time risk should be seen in the initial price being lower than it now is.

Canaan reported in a press release that they have received a massive order "Canaan Announces Significant Customer Order of 11,760 Bitcoin Mining Machines" https://canaan.io/1081.html, so for their own sake they ought to have their chip sourcing strategized. Such large orders would naturally mean they should have the ability to provide replacement parts, as well. But I also wouldn't be surprised if their priorities shift from smaller/home mining operations to large scale miners. Still, they're going to need to be able to source replacement components for even bigger operations if this is the case.

Under the contract, Canaan will ship and install its A1246 ASIC AvalonMiners throughout 2021, adding a total of 1.05EH to Mawson’s global mining operations.

Considering that the 1246 has their proprietary onboard controller, and I'm guessing that a large mining operation is going to be using their own home-cooked controllers and software, this order is probably for a stripped down version?

Regardless, hopefully these components will be available quickly, as they've most likely ramped up production.