Author

Topic: Share your experience with Bitmain T21s. (Read 117 times)

full member
Activity: 219
Merit: 426
November 01, 2024, 11:05:39 AM
#7
at danieleither

my four are fine.

but they use epic control boards and are set to 3000 watts.

They all use 240 volt psu's the one that the s19xp's use.

 these psu's just version f spec to 277 watts.




https://altairtech.io/product/apw121215-bitmain-power-supply-psu/

must be version f



and my units are doing about 17 watt a th

I have spare oem t21 high volt psu's and wires I could send them to you as I do not need them.

are you in usa? if I recall you are not in usa

Thanks Phil, however I'm not the OP. No problems here (I always carry spare PSU's, control boards and fans) Smiley
legendary
Activity: 4326
Merit: 8950
'The right to privacy matters'
November 01, 2024, 11:01:50 AM
#6
at danieleither

my four are fine.

but they use epic control boards and are set to 3000 watts.

They all use 240 volt psu's the one that the s19xp's use.

 these psu's just version f spec to 277 watts.




https://altairtech.io/product/apw121215-bitmain-power-supply-psu/

must be version f



and my units are doing about 17 watt a th

I have spare oem t21 high volt psu's and wires I could send them to you as I do not need them.

are you in usa? if I recall you are not in usa
full member
Activity: 219
Merit: 426
November 01, 2024, 10:28:49 AM
#5
I'm only running 24 x T21's total (across two sites), however out of 5 or so failures, 4 have been down to the PDU connection issues, the other was a dead PSU
full member
Activity: 219
Merit: 426
November 01, 2024, 10:27:37 AM
#4
Are you using the Bitmain PDU's? I've had a LOT of problems with my T21's, but most have been attributable to poor termination of cables to the rear of the PDU sockets. These sockets are TERRIBLE in design, often the pins come out on the rear of the PDU socket (therefore dropping a phase). The miners still power up but display all sorts of problems and error messages (not related to power) such as 0 Asics detected, messages about checking connections to hashboards etc

Just noticed you ARE using Bitmain PDU's.

For the miners that are problematic, try pulling the power socket out of the PDU (they just pull out) and check the connections on the back - if they're like mine, you'll find cables that either fall out (and therefore weren't making a proper connection in the first place) or are on the verge of falling out.
full member
Activity: 219
Merit: 426
November 01, 2024, 10:09:15 AM
#3
Are you using the Bitmain PDU's? I've had a LOT of problems with my T21's, but most have been attributable to poor termination of cables to the rear of the PDU sockets. These sockets are TERRIBLE in design, often the pins come out on the rear of the PDU socket (therefore dropping a phase). The miners still power up but display all sorts of problems and error messages (not related to power) such as 0 Asics detected, messages about checking connections to hashboards etc
legendary
Activity: 2436
Merit: 6643
be constructive or S.T.F.U
August 28, 2024, 05:58:40 PM
#2
T21 is generally a "bad" miner, while I have not personally operated any, I did help some folks acquire them "against my advice", I believe the first batch was the worst one, there seems to be a flaw in the design where the chips in the corners get too hot and eventually fail, many people chose to change thermal paste because they believe that's the main reason why these T21s get very hot.

Obviously, you would still see people who love the T21s, again, it's all about luck and which batch is screwed and which one is not, Bitmain is known for making terrible batches, heck, even an entire product line can be bad like the 17 series especially the T17s, so the reality is, you will get mixed views, very different answers, but someone saying they run T21 just fine doesn't mean everyone will get to say the same thing.

I have had thousands of mining gears run under my hand, be it miners I ran, or bought for clients, I still up to this point don't understand why anyone would choose Antminer over Whatsminer, the little efficiency improvement Bitmain has can never make up for the high failure rate, I had a Whatsminer M21s running alongside dozen Antminer 17 series, in a very hot summer with a single fan and full of dust, the worst it would do is show "over temp error", reboot and start mining, while nearly all of the Antminer experienced issues and some died, generally, you shouldn't run Antminers in hot weather, underclock the shit of them and always make sure the environment temp is no more than 30c or so, if you have Whatsminers, 40c will do just fine.

I suggest you find a good local repair shop and get them to fix your gears (they are usually fixable and it's most of the time just a few chips that go bad, and once you fix them they go on for a good period), sending gears back to Bitmain is usually a waste of time.

If you would share your container design, some of U.S might be able to help you improve cooling, although, if it's freaking hot outside, your best bet would be to underclock for the 2-3 hot summer months.
newbie
Activity: 3
Merit: 0
August 15, 2024, 03:27:34 PM
#1
We bought 162 T21 190T from an authorized bitmain dealer and deployed the miners this May at our farm in North Georgia. We have experienced a high failure rate this summer and are struggling to parse out what are miner hardware issues and what are environmental overheating issues. It has been a hot dry summer for us and we experienced a good amount of overheating in this container but many of the T21s did not recover. As it stands ~30 of the 162 have experienced major issues.  I suspect the issue could be related to the way our farm goes about cooling in the summer but am not fully convinced. We run S19s in the same way but if they overheat, they generally come right back up at night.

We are running these T21s in an Antbox originally designed for S9s and similar legacy models. We have 10 containers like this, mostly S19s and have had no real issues running newer machines in older containers until now.
In order to run these 3 phases miners, we upgraded all the electrical in the container (transformer, wiring, Bitmain PDUs) but did not add any auxiliary fans. These containers use passive cooling with a hot side and a cold side and no auxiliary fans; it relies on the miner fans themself for airflow. This has been working for us with Avalon 1246 and S19s but I'm afraid we might have found the power density limit with the T21s

The two main problems I see is either on the miner page, an error code that reads "y for repair check connection to hashboard x" and occasionally, if I catch the log mid overheat, I notice that the first chip on each chain is 20*c hotter than the other chips. From past experience this jumps out to me as a hardware issue since all the other chips are in the normal range. It seems the 30 bad miners all have 1 chip temp way hotter than others if you catch the log at the right time. If I look at the log after, it generally is stuck on find 0/108 asic chips on each board or sometimes 50/108.Also, I notice get-power-type-fail on all the T21s on boot up, even the ones that hash perfectly fine - I think this is unrelated

We sent 15 seemingly defective miners to warranty repair and received them back about 2 weeks ago. Of those 15, only 7 are hashing consistency, 7 are failing again and 1 was declared scrap. The technicians assured me they were tested for 4 hours before declared good, but many were seemingly DOA when returned. Same errors as before.

I am confounded. I guess I just want to hear other experience with T21s

What is everyone's failure rate with T21s so far?
How are you cooling your T21s?
Has anyone experienced these issues described?
Experience with T21 repairs, did they work?
 
Jump to: