Example: right now my S9 is sitting at a set 750W. Temps are just shy of 100*C, cranking out 9.5TH/s at about 80W/THs.

If I set it to 300W, it tends to run at 110W/THs.

That's not what I mean, though. The fact that you can't dynamically change setpoints without the controls having to reset and "start from scratch." Basically it would be like setting the cruise control but the ECU can't adjust the throttle based on load. Stupid. Ignorant. Useless.

The dynamic scaling system based in board temp is laughably awful.

I want to be able to control the set power level at much more dynamic intervals so I can tie it in to my solar system and mine with calculated excess power after the battery bank reaches a certain of charge.

And, also based on ambient temperature (either air or liquid, as I want to use another So as an electric water heater).

As for dust, ocean let's you withdraw via boltz on lightning. I'm moving towards that. So, no dust.

Reply to this note

Please Login to reply.

Discussion

Yes I can confirm W/THs are pretty much the same for me.

I haven‘t tried to dynamically control my s9.

And even braiins pool supports lightning payouts (but I will also try oceans some day)

🎯I’ve moved to ocean for the same reason.

At 500W (499) I get about 81 watts per TH/s and at 300W (306) I get about 92 watts per TH/s.

You mean you want to be able to go from 750W to 500W to 300W without the boards dropping their Hash rates to zero before ramping up again?

Or something else?

“Start from scratch”

I just assumed this was a procedure to not burn up the ASICs by 1) stopping hashing, 2) run fans at full speed, 3) set fans according to configuration, 4) then ramp ASICs to Energy consumption configurations, 5) monitor temperature as the ASIC ramps.

If you mean they should be able to overclock and underclock without needing to “tune” to learn the stable set points I guess I don’t know enough to have an opinion. Once the set points are “learned” you can bounce between them without “tuning” with only the procedure above between the set points.

It is interesting to see the variability between what are supposed to be the sane boards.

Yes. That's exactly correct. Setting the power usage on the fly would be a much better control scheme for secondary uses.

Nope. On switching power targets, the fans drop out entirely... So, that's really poor controlling methodology.

The temperature monitoring drops off, too. Again, stupidly poor methodology.

May I ask you how you are controlling power consumption at your s9?

I have done it manual via the web interface so far.

Might save me some research time. :)

That's also how I'm doing it, which is another point of contention. I'd rather not have to do that.

Ah I see, yeah that’s not really practical