[ANN] US/North American Bitfury sales NEW STOCK ***NOW SHIPPING***

[klondike_bar]

Code:

bad      off     per chip        good cores
0:      848     35.892  37.522  2507    76      4       0       0       16     00       (2.243/chip)    34%
1:      848     36.307  37.491  2536    15      0       0       0       16     00       (2.269/chip)    35%
2:      848     27.431  27.946  1916    134     16      5       2       16     00       (1.714/chip)    30%
4:      848     35.720  36.708  2495    68      2       0       0       16     00       (2.232/chip)    34%
5:      848     37.180  37.152  2597    14      1       0       0       16     00       (2.324/chip)    35%
6:      848     38.898  38.917  2717    8       2       0       0       16     00       (2.431/chip)    36%
8:      848     33.000  34.309  2305    82      4       0       0       16     00       (2.062/chip)    32%
9:      848     33.529  36.698  2342    82      1       0       0       16     00       (2.096/chip)    32%
A:      848     0.000   0.000   0       0       0       0       0       16     00       (0.000/chip)    9%      speed down
C:      848     35.648  36.793  2490    16      1       0       0       16     00       (2.228/chip)    34%

So, yeah, I have 1 bad H-board it appears.  It throttles down for about 15 minutes and then just goes to 0 GH/s

Uh, are these still covered under warranty if I heatsinked the back, Dave?

Edit: This is interesting, a second board just did the same thing. :|

   Your Hboard maybe overheat that why Hboard shutdown and you need better cooling. 

this. When i pencil mod my 2-card rig, the maximum i can run stable is about 38.5-39GH per card. If i run higher than that, they become sensitive to the changes in ambient temperature (my apartment always gets hot in the evening) and would do what youve seen, where chips drop to zero and either restart again a few minutes later or require me to stop/start miner.

you either need better cooling to the card, or slightly tune back its performance. whats with your card 2 btw? 27Gh and lots of errors - is that also cycling between 36->27->36? if so it needs better cooling also and it will be stable

[1714648126]

1714648126

[1714648126]

1714648126

[1714648126]

1714648126

[1714648126]

1714648126

[1714648126]

1714648126

[1714648126]

1714648126

[tacotime]

Code:

bad      off     per chip        good cores
0:      848     35.892  37.522  2507    76      4       0       0       16     00       (2.243/chip)    34%
1:      848     36.307  37.491  2536    15      0       0       0       16     00       (2.269/chip)    35%
2:      848     27.431  27.946  1916    134     16      5       2       16     00       (1.714/chip)    30%
4:      848     35.720  36.708  2495    68      2       0       0       16     00       (2.232/chip)    34%
5:      848     37.180  37.152  2597    14      1       0       0       16     00       (2.324/chip)    35%
6:      848     38.898  38.917  2717    8       2       0       0       16     00       (2.431/chip)    36%
8:      848     33.000  34.309  2305    82      4       0       0       16     00       (2.062/chip)    32%
9:      848     33.529  36.698  2342    82      1       0       0       16     00       (2.096/chip)    32%
A:      848     0.000   0.000   0       0       0       0       0       16     00       (0.000/chip)    9%      speed down
C:      848     35.648  36.793  2490    16      1       0       0       16     00       (2.228/chip)    34%

So, yeah, I have 1 bad H-board it appears.  It throttles down for about 15 minutes and then just goes to 0 GH/s

Uh, are these still covered under warranty if I heatsinked the back, Dave?

Edit: This is interesting, a second board just did the same thing. :|

   Your Hboard maybe overheat that why Hboard shutdown and you need better cooling.  

Thermal diode reports chips are 51 C VRM is 54 C :|

I swapped the cards into different slots and now they're both no longer throwing lots of errors or shutting down

Clearly H-boards are delicate flowers

Most recent stats:

Code:

board-2 speed   nrate   hrate   good    errors  spi-err miso-er duplic  good   bad      off     per chip        good cores
0:      848     36.851  37.195  2574    75      0       1       0       16     00       (2.303/chip)    19%
1:      848     35.419  37.290  2474    9       1       0       0       16     00       (2.214/chip)    18%
2:      848     37.452  38.685  2616    18      1       0       0       16     00       (2.341/chip)    19%
4:      848     35.262  36.434  2463    60      2       0       0       16     00       (2.204/chip)    18%
5:      848     36.035  36.846  2517    18      1       0       0       16     00       (2.252/chip)    19%
6:      848     37.982  38.600  2653    8       0       0       0       16     00       (2.374/chip)    20%
8:      848     31.668  34.045  2212    77      1       0       0       16     00       (1.979/chip)    17%
9:      848     34.847  36.550  2434    69      0       0       0       16     00       (2.178/chip)    18%
C:      848     35.777  36.518  2499    30      2       0       0       16     00       (2.236/chip)    19%
D:      848     32.885  31.941  2297    13      1       2       0       16     00       (2.055/chip)    17%

[tacotime]

Yeah so that one board is still bad.  Keeps shutting itself off even if I downclock it so the onboard temps are only 30-40C.  I guess I'll have to talk to dave about RMAing it.  This happens no matter what slot it's in.

[klondike_bar]

Yeah so that one board is still bad.  Keeps shutting itself off even if I downclock it so the onboard temps are only 30-40C.  I guess I'll have to talk to dave about RMAing it.  This happens no matter what slot it's in.

if you are mixing capacitor and no-capacitor boards, try putting ones with caps behind those without. In theory, this may have a stabilizing effect on the bank depending just how the m-board circuit is laid out. Of course a bad board is a possibility too, but they are very simply designed so there are not many failure modes available (bad chip, broken component, bad solder, etc)

make sure heatsinks are on the backside of the PCB, and put a heatsink on the thermal vias opposite the small volatge regulator. putting a heatsink on the inductor may help as well, as these components are both near thier capable limits at 30A across cards doing >35GH. keeping thier temp down may help

[Keefe]

I just powered up my new full rig, v3.0 M-board, 16 v2.2 H-boards. No heatsinks, 3 fans side by side (120mm, 148 cfm each). Getting 575 GH/s without tweaking. Using 622W at the wall, plus 30W for the fans (PSU is 86% efficient). Very nice!

The v3.0 M-board has two 6-pin PCIe power sockets and also the screw terminals like the v1.0 M-boards. Using an extra adapter from an older kit, I'm able to plug 4 cables from my PSU into the M-board. This should solve my concerns about pulling alot of current through just two cables.

[achimsmile]

Yeah so that one board is still bad.  Keeps shutting itself off even if I downclock it so the onboard temps are only 30-40C.  I guess I'll have to talk to dave about RMAing it.  This happens no matter what slot it's in.

Same thing happened to me, chips were cool enough. Turned out that the power regulator was too hot, despite heatsinks (I measured 90°C with IR-Thermometer). I moved the 3 fans from the top to the side of the regulator, which helped a lot. Now my 10 Boards are stable at 345-355 GH/s.

[Soros Shorts]

Yeah so that one board is still bad.  Keeps shutting itself off even if I downclock it so the onboard temps are only 30-40C.  I guess I'll have to talk to dave about RMAing it.  This happens no matter what slot it's in.

Same thing happened to me, chips were cool enough. Turned out that the power regulator was too hot, despite heatsinks (I measured 90°C with IR-Thermometer). I moved the 3 fans from the top to the side of the regulator, which helped a lot. Now my 10 Boards are stable at 345-355 GH/s.

Aren't these new H-cards supposed to have trim pots that would allow you to reduce the voltage output of the regulator? The voltage preset at the factory may just be a little too high for your operating environment. This symptom was common with the August H-cards if you "over pencil mod" them. I have yet to receive my Oct kits, but if this were to happen to mine I would try reducing the voltage a little on those cards that were shutting down.

[Doff]

I bought a Multimeter to be able to check the voltage, however I do not know where on the board you would test that.  If someone could make a screenshot of that I would really appreciate it, or maybe it already exists?

I have tried all sorts of fans and none seem to keep the regulator cool enough to keep them from cycling. Mine just turn off as well, and do not come back up.

[goxed]

If you donot add heatsinks to the regulator area, cards will cycle. The real point of adding heatsinks to the ASIC is to keep the regulator cool. Add sinks to the regulator and atleast the bottom 2 rows of ASICS and reduce STRESS in your life .
BTW, I also did another mod, but I will not advise anyone to do that unless they know what they are doing. It's obvious if you look at my H-card very carefully.

[goxed]

I bought a Multimeter to be able to check the voltage, however I do not know where on the board you would test that.  If someone could make a screenshot of that I would really appreciate it, or maybe it already exists?

I have tried all sorts of fans and none seem to keep the regulator cool enough to keep them from cycling. Mine just turn off as well, and do not come back up.

Check voltage on top of the inductor (grey box with PULSE printed on it and any of the GND pin from the PCIE power supply)

[Doff]

I did that actually, but I am still having issues with it. It made them stay up a tad longer before they started dropping off.

[Doff]

I bought a Multimeter to be able to check the voltage, however I do not know where on the board you would test that.  If someone could make a screenshot of that I would really appreciate it, or maybe it already exists?

I have tried all sorts of fans and none seem to keep the regulator cool enough to keep them from cycling. Mine just turn off as well, and do not come back up.

Check voltage on top of the inductor (grey box with PULSE printed on it and any of the GND pin from the PCIE power supply)

Ahh, that's how you do it. Thank you.

[Soros Shorts]

Add sinks ... and reduce STRESS in your life .

I don't know about that ... sticking on $400 worth of tiny heatsinks per rig was a pretty stressful and time consuming experience for me. I wish somebody would make custom heatsinks for these.

[goxed]

Add sinks ... and reduce STRESS in your life .

I don't know about that ... sticking on $400 worth of tiny heatsinks per rig was a pretty stressful and time consuming experience for me. I wish somebody would make custom heatsinks for these.

It's a one time exercise, and then peace of mind.

$400 :O They usually sell for $40 good for 4 cards. Actually, I realized after sometime that heatsinking only the bottom 2 rows is sufficient to keep the regulator cool.
I see that heat-sinks you used are a different that the VGA RAM sinks.  

[tacotime]

More updates on my bad card:

- No matter what I do it will shut off within 5 minutes of beginning mining, although it does not throw a lot of errors while it is mining.
- The shutdown is a slow taper in noncerate to about 10 GH/s, after which is hits 0 GH/s.  After this the board does not restart.
- Heatsinking the board like crazy has no effect and thermal diode readings all over the voltage regulators max out at 45C when you start it up, so it's not any kind of heat issue
- Manually setting the clock speed has no effect
- Board position has no effect

At this point I'm just going to say it's a bad board and needs to be RMA'd unfortunately.  Dave, will you accept RMAs with heatsinks on them?

[goxed]

More updates on my bad card:

- No matter what I do it will shut off within 5 minutes of beginning mining, although it does not throw a lot of errors while it is mining.
- The shutdown is a slow taper in noncerate to about 10 GH/s, after which is hits 0 GH/s.  After this the board does not restart.
- Heatsinking the board like crazy has no effect and thermal diode readings all over the voltage regulators max out at 45C when you start it up, so it's not any kind of heat issue
- Manually setting the clock speed has no effect
- Board position has no effect

At this point I'm just going to say it's a bad board and needs to be RMA'd unfortunately.  Dave, will you accept RMAs with heatsinks on them?

Can you please post a picture, front and back side? I had a card performing similarly, I carefully monitored the power consumed by the card. It was ~70 watts. Had to reduce the voltage a bit to make it reliable.

[tom99]

Yeah so that one board is still bad.  Keeps shutting itself off even if I downclock it so the onboard temps are only 30-40C.  I guess I'll have to talk to dave about RMAing it.  This happens no matter what slot it's in.

Same thing happened to me, chips were cool enough. Turned out that the power regulator was too hot, despite heatsinks (I measured 90°C with IR-Thermometer). I moved the 3 fans from the top to the side of the regulator, which helped a lot. Now my 10 Boards are stable at 345-355 GH/s.

Aren't these new H-cards supposed to have trim pots that would allow you to reduce the voltage output of the regulator? The voltage preset at the factory may just be a little too high for your operating environment. This symptom was common with the August H-cards if you "over pencil mod" them. I have yet to receive my Oct kits, but if this were to happen to mine I would try reducing the voltage a little on those cards that were shutting down.

   Yes, newer Hboard got SMD trim pot on it.

[Keefe]

I bought a Multimeter to be able to check the voltage, however I do not know where on the board you would test that.  If someone could make a screenshot of that I would really appreciate it, or maybe it already exists?

I have tried all sorts of fans and none seem to keep the regulator cool enough to keep them from cycling. Mine just turn off as well, and do not come back up.

Check voltage on top of the inductor (grey box with PULSE printed on it and any of the GND pin from the PCIE power supply)

Just FYI... There is a small problem with measuring voltage like that, but it's still the easiest way in a crowded rig. There's a significant and varying voltage drop on the ground line between the power supply and a card. Just between the M-board's GND terminal and a slot's ground pins, there's a drop of between 0.023v (slot 0) and 0.067v (slot F) on my v1 rig. So while it may look like you're running a card at 0.900v when measured the easy way, it could actually be as low as 0.833v. And I didn't even account for any further drop from the slot to the chips' ground plane.

[Keefe]

If you donot add heatsinks to the regulator area, cards will cycle. The real point of adding heatsinks to the ASIC is to keep the regulator cool. Add sinks to the regulator and atleast the bottom 2 rows of ASICS and reduce STRESS in your life .
BTW, I also did another mod, but I will not advise anyone to do that unless they know what they are doing. It's obvious if you look at my H-card very carefully.

Ah, I think I see what you did. The thin wire over the fuse? Is that supposed to break if the fuse trips, so you can see that it happened?

[tacotime]

More updates on my bad card:

- No matter what I do it will shut off within 5 minutes of beginning mining, although it does not throw a lot of errors while it is mining.
- The shutdown is a slow taper in noncerate to about 10 GH/s, after which is hits 0 GH/s.  After this the board does not restart.
- Heatsinking the board like crazy has no effect and thermal diode readings all over the voltage regulators max out at 45C when you start it up, so it's not any kind of heat issue
- Manually setting the clock speed has no effect
- Board position has no effect

At this point I'm just going to say it's a bad board and needs to be RMA'd unfortunately.  Dave, will you accept RMAs with heatsinks on them?

Can you please post a picture, front and back side? I had a card performing similarly, I carefully monitored the power consumed by the card. It was ~70 watts. Had to reduce the voltage a bit to make it reliable.

Sure, just a minute.  I checked the voltage (Inductor +, GRND -), 0.862 V for this card versus and average of 0.860 V for the other cards, so the voltage is apparently not the issue.