Again sorry for the delay
It is simple problem bad batch of components.
All production is checked and bad components exchanged.
The problem persist only @ full load , so assembly then test , then disassemble ,exchange the component , assembly then test.
At the end disassemble of the test coolers and stacking for shipment.
Again I do not want to make a saga out of this problem and not going to make update posts with :" new "xxxxx" on the way, ETA tommorow "
Just a bad batch of MCU's, that is all.
The bad part is it took us a week to find out what it is , then we have to  get MCU's from different batch and check every single board.
I will not make statements like "next week" or similar.
We are working and as soon we are ready the boards will go out.
I understand that 2 weeks delay in BTC world means a lot and not happy with the situation and with the 20 hour work day lately.
Again remember that with a few exception most of these orders are paid with chips and we take all the costs for production including all this aditional labor for test and exchange . Minesr source are covering all the transport costs for the most of North american orders
Supplier  just delivers us new MCU's, but tons of additional work is on us.