You returned perfectly good polaris cards because you were not able to exceed 30mh on them?
What were you expecting exactly? 40mh per card? What made u have such absurd expectations? Do you have any mining experience?
Do you understand that a better memory strap can help you push the last drop of mining performance out of a card, but it wouldn't change the card itself?
Do you understand how graphics chips, or memory chips, or any kind of chips work? Do you know the term silicon lottery?

What memory? H5GC8H24MJ?

I have some of ASUS ROG STRIX RX580 8GB OC with Hynix H5GQ8H24MJ, but hashrate is terrible low 28-29
and unstable. Any ideas for good mode?

Does it mean there is no need to upgrade the drive for 290/390 owers?

So i have this weird problem.

My rig works fine with its current setup (Claymore 9.8, aug11 blockchain drivers, WIn10 7310) and has ran for 7 days nonstop, till I had to restart it.

First, GPUs 0 to 5 would be detected, then after 30 seconds GPU5 will show zero hashrate. Eventually 10 seconds later, it would hang. Then if it didn't froze on me, I would restart Claymore.

Sometimes Claymore restarts automatically when this happens, and I see a popup message on the lower right saying "AMD Radeon has been restored to default settings..."

Most of the time, after I see that error message and Claymore has restarted, All 6 GPUs would hash properly and run fine. Half the time, I had to do 3-4 restarts for it to work again.

But I just want to narrow down which is really causing the issue. I'm using new risers, tried changing usb cable, 6-pin sata power, but still the same issue.

Interesting note:
Of all 6 cards that GPU5 only is the single card that takes -150mV and still run fine. GPU-z shows .9375mV after the -150mV has been applied.

Cards are Asus Dual OC Rx 580 8GB 1260/2050/-125 to -150mV

Anyone?