Do we know what the result is if you power the controller before the hash boards?
Do the hash boards not get detected?
Can you fix it with a soft reboot?
Yes you can fix it.
I've simply had the board not boot because it is not detected at boot.
I would probably avoid doing this as i'm unaware if there might be electrical issue related to this that would cause hardware damage.
Anyways, a soft reboot worked for me, but i didn't repeat the experience. Also a power cycling would definitively fix it if no damage was caused.
It's a great pity that we are having to guess why the instructions are to power up the Controller after the Hash boards. It would be much more useful if we were told what the problem(s) are if this is not done....
So to continue the
guessing. In practice I suspect that most users are doing what they have always done and are powering up Controller & Hash Boards simultaneously, and are getting away with this?
There are then some reports of S7 that seem to have a lower Hash Rate & Higher HW errors either at switch on or after a power outage, that can be fixed by a Software Reboot, which could be interpreted as being the same as powering up the Hash Boards first, so perhaps this is the reason?
My latest best
guess on the reason is nothing to do with the above problems but all to do with the fact that the controller now has a separate supply whereas in the past it was powered from the hash board and hence was powered simultaneously, with no risk that the Controller could be powered & the Hash Board not.
I think what Bitmain is guarding against is a situation where only the Controller is powered at the point that the Hash boards start taking power. The reason being that I suspect there is then a risk that the unpowered Hash Board attempts to take current from the Controller?
Unfortunately just another guess, but there you go.
Rich
This scenario happens with s5 controllers and hash boards also.
The s5 controller gets power from ONE hashboard so they would both go on at the same time.
The other hashboard however may not get powered at the same time.
I've had to soft reboot when the second hashboard chips don't show up live after a power on
or when nicehash does a pool switch and the hashrate goes to ZERO on the miner.
Sometimes it never comes back and locks up, other times one hashboard comes back and no amount of soft reboot
will fix it. At that point a power cycle is needed.
Glad I have the remote control PDU's so I can remotely power cycle instead of going the the PSU
and toggling the switch.