Pages:
Author

Topic: Pangolin Whatsminer M10 - 33TH at 2145W - page 4. (Read 22012 times)

legendary
Activity: 3612
Merit: 2506
Evil beware: We have waffles!
Those last few pics show the area that shorted as evidenced by the blast marks and residue left behind. Whatever caused the short seems to have been easily vaporized since the PSU still mostly worked nonetheless but at reduced power output.

If one pours through the API report as the miners start up you will see that there is massive amounts of testing going on and the results are there in the log. I do believe that PSU voltage is one of the things that is reported for each series of tests so maybe problems with it could show up there?

Aside from that: Damn they like using that white glue. Hopefully it is a hard epoxy to not hold too much heat vs a rubbery glue...
legendary
Activity: 4172
Merit: 8075
'The right to privacy matters'
Photos TIA
Photos
screen shots TIA
screen shots

Kudos to the post office $8.05  to ship psu NJ to VA In under 48 hours.

It is nice that it works better.

@ Pangolin  Please fix this for sarcheer   Grin

I am ordering the 72th to test for community so please continue to be helpful to us.
copper member
Activity: 330
Merit: 103
Well great news! There can be no doubt, the PSU was damaged and was limiting the freq achievable by the boards. Having swapped the PSU with the replacement Philipma1957 sent me, the freq are now:
491-524-478

Previously, it was the rarest of tunes that allowed for the middle board to scratch 500, with the others hovering in the 440 range. Hashrate is now 31.1TH/s.

Pics will be added later on today for those with more EE experience to dissect to see if there is any visible damage. I will say that the black sheeting within separating the top of the PSU from the bottom appeared to have been separated from the force of the pop, with only a tiny section of the adhesive holding it closed. The replacement from Phil, even from the exterior, was clearly closed.

Thank you so much Phil, your willingness to experiment for this community is a resource we all benefit from greatly.

Edit for Photos:



legendary
Activity: 4172
Merit: 8075
'The right to privacy matters'
I've got one board that's only running at 283 Freq, the others are at 550 or so. Means that effectively its almost in spec.

yeah of my eight units some boards are lower then others.

Once warrantees are done I will cherry pick and match the boards.

of my eight units one board of 24 is a bit flakey so 23/24 good boards and one that drops in and out.

@ sarcheer psu should arrive today
legendary
Activity: 3220
Merit: 1220
I've got one board that's only running at 283 Freq, the others are at 550 or so. Means that effectively its almost in spec.
copper member
Activity: 330
Merit: 103
When I remove the PSU to swap it for the replacement one Phil is sending me, I'll make sure to dismantle further than that and provide pics of everything I can.
legendary
Activity: 4172
Merit: 8075
'The right to privacy matters'
I agree. This might be anecdotal, but I'm running my M10 on 208 volts and it's been consistently hashing at the expected rate since I installed it a week ago.

Very nice of Phil to help you out. I hope it works.

well we get to do a good test.

if the psu I sent him  gives him 29.5 to 31th  it is not the 208 volts.

The solid buss bars neg and pos   are closely spaced so I think a  bug or a spec of something made the pop

and was fried. it may have damaged the 1 wire to 3 slot bridge as that is really close to the  buss bars.

this part

https://pangolinminer.com/product/miner-adapter-board-m10-m10s-d1/  

could be the issue  that is the part I had go bad on me I never did better then 30th and the boards dropped out.   when replaced I now do 30.6th

sarcheer lives close to me .

VA and NJ are same shipping zone for usps regional flat rate box b  so it was easy to ship the spare to him.

Us miners have to stick together  Grin
legendary
Activity: 1150
Merit: 1004
Combined with the stated specs for the PSU (below), I'm struggling to believe that the input voltage difference between 208V and 220V is what is hamstring the hash rate.

I agree. This might be anecdotal, but I'm running my M10 on 208 volts and it's been consistently hashing at the expected rate since I installed it a week ago.

Very nice of Phil to help you out. I hope it works.
legendary
Activity: 2030
Merit: 1563
CLEAN non GPL infringing code made in Rust lang
I have created a support ticket and will update going forward with how Pango resolves this issue.

From what I recall, the pop originated from the PSU and was loud enough to make me jump back. The PDU's are still being used right now without issue.

Edit: After giving it about 4 hours now, I just plugged it back in and everything is working... Very suspicious and will be sitting within view of this guy for the rest of the day with my fire extinguisher close by. Will still go through the support process with Pango to see what they have to say about what may have happened.

The only thing that crosses my mind is that a roach or bug crawling inside short-circuited something, but now that it popped and dried the unit would go back to work. Normally when something pops and trips a breaker or such i won't even try powering it again, at least without taking a look inside...

What a curious gamble you made by powering it again, wanted to see if flames would come out or something?
copper member
Activity: 330
Merit: 103
So Phil is going out on a limb and is sending me his replacement PSU to determine whether or not my PSU is damaged. This is a very generous gesture from him and ill do whatever is necessary to make sure he is made whole when this is settled.

I'm quite convinced its the PSU given the events that precipitated the reduction of hash rate, namely the pop, flash of sparks, and puff of smoke from the PSU. Combined with the stated specs for the PSU (below), I'm struggling to believe that the input voltage difference between 208V and 220V is what is hamstring the hash rate. I will add that the autotune function is not raising the avg board freq to what it was prior the PSU incident. Current board freq: 445-499-434 resulting in 28.5TH/s. I can only hope the boards were not damaged.

Model:
P10-12.5-2250-V1
Input:
176-264VAC 50/60Hz 25A Max
Output:
12.5V 180A
12V 10A
legendary
Activity: 4172
Merit: 8075
'The right to privacy matters'
It was provided directly from Pangolin support.

Update on my M10:

Unfortunately, it would appear I am at an impasse with Pangolin support after hours of communication on telegram. They are denying anything is wrong with the PSU on my M10 (despite not asking me to send them pics or remove it) and are now claiming two things:

1.) The input voltage of 208V is low and that 220V is ideal.

2.) The hash rate is within normal variance as relayed by their engineers of +/- 5%, despite being 8.1% off of spec at 28.5TH/s.

At a loss of where to go from here.

I have pm'd you

here are my m10 numbers

My oldest m10  drops a board on and off it is out of warranty

m10one16 is the oldest one

mten7 is the one  I needed a repair part

they all are a little better then yours  but I have 227 volts input

my newest is in the garage it gets 239 volts and is doing 31,387  with really good temps of 66-67-72

copper member
Activity: 330
Merit: 103
Where is this FW available?

It was provided directly from Pangolin support.

Update on my M10:

Unfortunately, it would appear I am at an impasse with Pangolin support after hours of communication on telegram. They are denying anything is wrong with the PSU on my M10 (despite not asking me to send them pics or remove it) and are now claiming two things:

1.) The input voltage of 208V is low and that 220V is ideal.

2.) The hash rate is within normal variance as relayed by their engineers of +/- 5%, despite being 8.1% off of spec at 28.5TH/s.

At a loss of where to go from here.
legendary
Activity: 3220
Merit: 1220
Quick update: Telegram support recommended trying downgrading to January FW to see if that would fix the low hashrate problem that appeared after the PSU problem. As my M10 is not run at home I had to wait to come in this morning to give the downgrade a try and, unfortunately, it has not helped. No change in hashrate, still only ~28.5 TH/s.

Will get back in contact with support over telegram tonight to see about next steps.

Where is this FW available?
full member
Activity: 1022
Merit: 221
We are not retail.
We've had zero warranty/returns even with the m10s.
legendary
Activity: 4172
Merit: 8075
'The right to privacy matters'
My stats are 8 m10s 7 were good and have mined with no issues .

1 needed a 1 wire to 3 slot bridge replaced.
copper member
Activity: 330
Merit: 103
Quick update: Telegram support recommended trying downgrading to January FW to see if that would fix the low hashrate problem that appeared after the PSU problem. As my M10 is not run at home I had to wait to come in this morning to give the downgrade a try and, unfortunately, it has not helped. No change in hashrate, still only ~28.5 TH/s.

Will get back in contact with support over telegram tonight to see about next steps.
member
Activity: 123
Merit: 21
The one I ordered arrived at my hosting facility with dead fans and 1 bad board. Grrrrrr.....
legendary
Activity: 4172
Merit: 8075
'The right to privacy matters'
I can only hope! Hopefully that what they want me to do as part of troubleshooting.

Unfortunately, I still cannot get this M10 to hash at 31TH/s. Last night I was seeing 31.1TH/s, now its hovering at 28.8TH/s.

Freq:
451-505-442

105 eff. chips per chain.

Oh the telegram people should show in 4 hours to help about psu.

yeah this one on monday  I got is good in fact it is the coolest running one I have ever owned

legendary
Activity: 3612
Merit: 2506
Evil beware: We have waffles!
Try a CGMiner restart. After waiting for a 1/2 hour between restarts my sorta gimpy #4 miner usually 2x restarts, sometimes 3 to get up to full speed. For me SM0 is the problem child in Normal mode, Low-power mode is a no-go as SM0 and SM1 fail on the final reboot after tuning is done..
copper member
Activity: 330
Merit: 103
I can only hope! Hopefully that what they want me to do as part of troubleshooting.

Unfortunately, I still cannot get this M10 to hash at 31TH/s. Last night I was seeing 31.1TH/s, now its hovering at 28.8TH/s.

Freq:
451-505-442

105 eff. chips per chain.
Pages:
Jump to: