Pangolin Whatsminer M10 - 33TH at 2145W

[philipma1957]

Who is the US distributor? .net?

I think so I will check with Offordscott

yes I just checked on facebook

[1714985711]

1714985711

[1714985711]

1714985711

[1714985711]

1714985711

[elokk]

Who is the US distributor? .net?

Yes

[ZimbaCardi]

Pango is saying I need to ship them back the damaged PSU at my own expense (don't like it, but no big deal and standard for the industry), but are also asking that I pay the shipping for them to send the new PSU to me.

Exactly what happened with the PSU on my Bitfily B1+. Decided to buy a new PSU from them and keep the old one old parts.

[fccs]

So odd question.  Has anyone noticed a hashrate bump since warmer temps?  I have not readjusted my setup for summer yet and the temps in location has risen about 10 degrees (It's about 80 degrees F right now).  My m10 miner has been averaging 35-37th for past 3 days now.  Do.. these things mine better in warmer environments? lol

[Steamtyme]

Well here's to hoping, I should have mine in the garage soon.  Summer does get warm and i only cool using ambient outside air, so at it's worst 35 celsius for a few days but high 20's for a good 6 weeks.

I can't say i watched it to closely, but i don't recall a jump in hashrate like that. I had a few times the basement was easily mid to high 20's .

Looks like you got the hardware lottery. Ill definitely keep an eye on mine especially the one that has only been in the cold garage.

[fccs]

The funny thing is this is the 31th model they were selling for like $700 some dollars.  I have a couple and all are averaging around 35th since it got warm out.  Very odd lol

[edonkey]

So I finally got around to swapping out the data bridge card. It took a while to get the replacements because I made the mistake of trying to communicate with them via their support system.

In case it wasn't clear, the only way you can get support is via Telegram. I should have paid attention to Phil's post about that, but I didn't have Telegram and didn't want to install it. Learn from my mistake and just use Telegram.

Anyway, I swapped out multiple new bridge data cards because they sent me several. But none of them made a difference.

But since I originally swapped hash boards and the problem stayed with the slot, not the hash board, I'm still convinced this is a data problem.

Perhaps my ribbon cable is bad? It could also be the controller, I suppose. But replacing the cable would be the first thing that I try.

Anyone know a source for replacement cable? I don't have one in my junk pile. Too bad those extra S9 data cables I have aren't 2 x 11.

Also, Phil you were asking me to snag you an extra data bridge card. I've got extras so I can do that, but it might be better to figure out where the fault really is before I send you a card that isn't tested.

[philipma1957]

well that sucks.  but  lets think it over.

controller

V
V  cable
V

slot 0
slot 1
slot 2

so  swapping the bridge did nothing.

this means

A) all bridges are good and some other thing is bad.

B) all bridges are bad and everything else is good.

What can we do on my end?  I have 8 working m10's with 1 bad board . Board is tested and does not work on any  of the 8 units.

but all 8 units can run all 3 boards.  as I have done lots of swapping.

If you send me 2 bridges and I swap them in and they work we then know the bridges are good.

Still leaves cable and or controller.

I have an un programmed controller which means if i send it to you  you need to go to telegram and have yang help you program it.

How about this send me 2 bridges I will drive up to clifton and test them.  I they work I will send one back and hold one as a spare.
I can also send you the unprogrammed controller yang will help you program it.

If it works  we know the old controller is bad.

If it does not work  we are stuck on cable or replacement controller.

maybe I can find an 11 x 2 pin cable    as it is easier to first test it.

[Steamtyme]

So I finally got around to swapping out the data bridge card. It took a while to get the replacements because I made the mistake of trying to communicate with them via their support system.
In case it wasn't clear, the only way you can get support is via Telegram. I should have paid attention to Phil's post about that, but I didn't have Telegram and didn't want to install it. Learn from my mistake and just use Telegram.

I never did try this route and honestly just forgot about it as my problem isn't so bad but is noticeable on LPM. I'll have to give this a go and see if I can get my underperformer up to snuff. Thanks for sharing, sorry I got no info for you on what you are looking for.

[edonkey]

well that sucks.  but  lets think it over...

Thanks so much for your offer to help Phil. But before you go to a lot of trouble, maybe I should do one more swap test.

If you remember what I did was swap 0 and 1. So as you look at the the miner from the front (i.e. the air intake side), I went from this:

S0  S1  S2
H0  H1  H2

To this:

S0  S1  S2
H1  H0  H2

This was based on the assumption that the slots are numbered from left to right. If I had any brains at the time, I would have rotated them entirely, moving every hash board like so:

S0  S1  S2
H2  H0  H1

This way all of the boards would have moved a slot and I wouldn't have to know which one is actually S0. If after that S0 is still bad, then clearly it's the data bridge, cable or (yikes) the controller.

Maybe I should still do this just to be on the safe side. At this stage I just have to swap the two outer hash boards to achieve the same effect as rotating.

Since the miner as running at a data center, I won't be able to try this until early this coming week.

In the meantime at least I'm running on two hash boards, which is better than nothing.

[philipma1957]

this is a dual row 20 pin cable

https://www.amazon.com/gp/product/B07FZWBP78/

pitch is 2.54 mm I think the dual row 22 pin is a tighter pitch

yeah  my bad slot was center  so swapping it left or right   works

I had a bad slot 1

It always stayed center.

when you go to check it  if it turns out to be a board let us know.

also see if you can measure pitch

11 holes x 2.54 = 27.94 mm  or just over 1 inch

this is close:

https://www.amazon.com/Sydien-2-54mm-Connector-Rainbow-Ribbon/dp/B07Q2SGFWM/

[edonkey]

Thanks Phil, as always.

I just heard from Yang JYan. Turns out that the board order is reversed of what I thought. As you look at it from the front:

SM2  SM1  SM0

Makes sense when you think about it. English and Europeans read left to right. Chinese read right to left. Both want to count up from 0.

Too bad it's not clearly documented anywhere.

So I didn't move the problem board. It might be a bad hash board after all, not the slot.

I'll move it in the next day or so, then see what happens.

Phil, no matter what happens, I'll send you some spare data bridge boards when this is done. I've got plenty ;-)

I swapped SM0 and SM2 (the rightmost and the leftmost) and the problem moved to SM2. So unfortunately it's a hash board problem.

I've got a repair ticket from Pangolin. I'm going to send the bad hash board to be repaired.

Any recommendations for a shipping company to China?

[minefarmbuy]

DHL for quickness. We can setup a label for you if you request on website.

[philipma1957]

That lucky seven m10 went dead again.

A new bridge fixed it for a month but now it has gone doa.

I suspect a completely different issue this time.

The Clifton build uses Verizon modem to a switch to a router to four switches .

The network is only 192.168.1.1-256

We have about 70 miners. On that and no attempt was made to do any sorting just Dpch .

So 4 miners went down.

192.168.1.22. A t15
192.168.1.23. A s9
192.168.1.24. A t2turbo
192.168.1.25. A m10

Since the gear is scattered from 192.168.1.4.  To 192.168.1.100. I find it very odd that the four miners that dropped out were in consecutive order.

I got 3 to fire back up but that m10. Is like fu so I dragged it back to Howell and will see if it will work at my place.

No one is one pangolin for support unit Sunday night so I will play with this unit till Sunday then contact yang on telegram.

[fccs]

The chances of essentially 3 different branded miners in sequential order going down at the same time are so ridiculous it has to be related to something else.  Are these 4 units plugged into the same pdu?  Maybe a power issue or something happened with that strip?

[SaiyanPrynz26]

I am really liking my m10 miner. Just set up my s9s in a cooler to keep them silent and much much cooler. I’ll have to show my setup when I can figure out how to post pics on this site.

[Steamtyme]

They are nice, so far not to many issues reported. I think Phil's the only one I've heard of so far that died; that's the second issue on that 1 machine. Then another with a DOA Hashboard. I have one that seems to underperform in low power mode.  Glad to hear your is working out well.

To post images you will need to use a site like imgur.com, then you can paste the link here - They have an option to have it formatted for the forum which makes it pretty simple.

[philipma1957]

They are nice, so far not to many issues reported. I think Phil's the only one I've heard of so far that died; that's the second issue on that 1 machine. Then another with a DOA Hashboard...

I have 8 of them 7 are perfect.

The troubled one is back to working.  Still seems to have an issue with the bridge.  This where the controller cable is split and goes to the 3 boards.

I have a replacement bridge: https://pangolinminer.com/product/miner-adapter-board-m10-m10s-d1/

and I placed 3 little pieces of this tape to isolate it and help it to not vibrate

https://www.amazon.com/Scotch-Permanent-Mounting-Tape-Inches/dp/B00347A8GC/

this is working so far. Gear is in the garage  I will drive it up to clifton  when I get the m20.

[Steamtyme]

Not going to lie I'm a little irritated at the troubleshooting for Whatsminer's. Yang was helpful but they seem extremely limited in what they will try and options. It seems update the FW is the first place they start and sort of where it ends - we went through 4 I believe yesterday all newer, and performed worse. After that it was a guess that it's probably a PSU. This is all based on just the board temps showing up, not any additional digging in the logs. To top it off they say my original FW version never existed - but provided me with a close date stamp, so I'll try loading it tonight.

In the end I'll probably just never run that unit in LP mode - it hits 30.8 in normal mode. It all comes down to do I want to foot the bill and downtime to send the PSU back and hope that they were correct for warranty, or pay for 1 out of pocket - which I'm not interested in doing. I may order one of those bridges to see if that was the issue as it seems to have been a weak point in assembly, it's a 10$ hit and they could probably throw it in with my M20 demo. I might try some of those FW versions on the other models once my Power meter is constructed to see if there is any advantage  - I did note that they must be coming out with a liquid cooling option as it had a place on the miner status with 2 FW versions. I'll check the file names tonight and drop the FW versions I saw.

Edit:

So from the files left here are the FW versions they have out and will try. I honestly don't know where to get them from apart from asking them in Telegram. I never really looked into it until they figured that would fix the issue. - They may show in your overview with a .1 at the end when updated.

20190404.18   20190318.12   20190114.15

So for the 3 above it was late and I was just watching it all, I believe only 2 of these showed the "status column" for liquid cooling. There was also 3 errors on startup; I can't recall if this was on all 3 versions or the same 2. Finally they all hashed at 27-28 TH/S normal mode.

20181129.12 This one they gave me for rolling back. results are similar to the original so far 30.7 TH on normal.
20181127.17 My original 2 of my miners still have this version.

[NotFuzzyWarm]

Well that was fun.... Finally got around to changing pool #3 from Kano's de node to the Oregon one on the 3 M10 miners at home. Of course that meant restarting them and -- miner #2 did not want to fire up. Constantly reported 0 chips on slot-1 (they go from 0 to 2) which oddly enough is physically the board  on the left side and not the center board...

Looking into the air inlet end of the miner and going by the green LED's, the boards would fire up: middle on then off, right on then off, then both middle and right on and stay on - so far so good -- then the left board would come on for less than a second and take the other 2 down. Endlessly repeat.....

Fearing the dreaded bridge board problem I unplugged it, removed the 2 screws that hold the board in place for shipping (is a snug fit even w/o the screws) and reseated the board: Success! It lives again!

Bloody hell I hate it when a board that probably cost 50-cents to make takes down a whole machine...