i don't understand why in my test braiins result worse than bitmain......
i've tested an s9i 13,5 on bitmain selected default hashrate with low voltage and result Gh/s 13560 W 1153 without error stable.
but with braiins Gh/s 13470 W 1170 with 1 error .... with 625 mhz and undervolt to 8.75.
You can take a look at the logs from bitmain's if you are curious, as it sometimes shows the voltage and speed and try to match. But don't be too reliant on it, discover your best settings by yourself.
Do not pay much attention to errors (unless there are way too many, but wait like an hour first), just at the actual and ideal hashrates, they should match as close as possible. I never used the second decimal, just things like 8.7, 8.6, 8.5... And speeds like 650, 625, 600, etc.
So you are going to do this on each hashboard. Pick a desired speed and start lowering the voltage, until you see the aforementioned values differ too much, then go back a notch in that voltage (OR) lower the speed and start again. Various people here managed to do about 80w per TH, things like 8v 500mhz or so so about 10 TH @ 810W.
Every single hashboard behaves different, you can find the absolute best your hashboard can do with bOS.
It might be useful to do tests against reliable pools such as ckpool which shows your miner stats without registering. The best efficiency is not necessarily close to "stock" speeds (and usually aren't).