Hi again!
Sorry for my very late reply... I had a few very bad weeks

But now, I have a few nice updates for you:
- 1st: I was getting all 6 GPUs up and running - stable with only one PSU!
- the next good thing is, I got now 7 GPUs hashing in one rig. Within the three starts with only one PSU, but than I decided to add another PSU.
-- I'm using a 1 to 4 PCI-e riser card and it works perfectly (almost... later more about the issue)
You where asking me about the brand and model of the "RMA-ing" GPU.
It was a MSI GTX 1070 Gaming X 8G.
Yesterday I got the information of the dealer that the graphic card was repaired at MSI and is now on the way back home.
So I guess that the gc could arrive within the next 4-5 days - so directly after easter.
As soon as it arrives I will add it of course to the rig

... and I hope that it still works!!
But now to another problem:
From time to time - so roundabout 3 to 24 hours - I got the error message "CUDA ..xyz... out of memory".
Unfortunately, I cannot really fix this issue.
After increasing the size of the win 10 page file to "minimum 17000 MB and maximum 57000 MB (amount of gpus (7) * 8 GB memory per card +1 GB)" it runs a bit better, but no smoothly as it should.
Do you have any idea maybe?
Thanks a lot in advance for your help!!!
...and here a few screenshots:
https://imgur.com/Dv1nXAfhttps://imgur.com/19e5Hv6https://imgur.com/DCS1ayO