I reworded my post above as I thought it could come across wrong. Tried to smooth out tone, not sure if it worked or not. But I just don't want to see guesses that are anecdotal as fact. And it is zero against you on this just anecdotal is very hard to take as fact as before when we talked some that were assumed as APW3's were DPS1200's so anecdotal really can bring false positives -
https://bitcointalk.org/index.php?topic=1493601.msg15323232#msg15323232I can say one has been tested in low 90's F with APW3 and it worked fine with S9. So that is a true test with one, and I realize my subset of 2 with one here is small. But my actual test's don't show this.
don't worry we are friends not an issue. I put in my biggest two reasons after you posted.
the graphs on bitmaintech are right above that drop in amps is hidden by the scale the 10 c it takes to drop is 2 spaces and should be ⅓ of 1 space . so my guess is mostly based on those graphs and the fact that the s-2 had a 1000 watt psu and it needs a 1200 watt psu.
I have never owned or tested the antminer psu.---Fact
You own and tested it and it works ------------------------Fact
No long term tests have been done on a batch 1 s-9 at 650 freq --------- Fact
Long term is 30 days - not a fact but what I use to define long term as I have had psu's start to fail at the 30 day marked when pushed hard 24/7
The graphs show if it gets hot it drops. ------------------------------------------------------ Fact
The graphs are scaled to make you think the drop off is 6x less steep then it is. --------- Fact
The s-2 had a 1000 watt psu that failed often ---------------------------------------------- Fact
These are what we have as facts.
What it means is ?
What I would do is keep my eyes open if I had antminer psus. powering batch 1 ,2,3
if indeed the batch 4 scales back to freq 550 the issue may be a mute point.