Ponyprogging ATMEGA328 U3S

No more 6m signals from Japan  time  to finish some projects

After faulty applying  12V on the QRP-LABS U3S smoke came from the BS170
Display back-light ceased out U3S was dead did not boot the ATMEGA328 was blown.
Placed the working U3S in its case
Applied  power on it but forgot to check the Voltage  major mistake.

The ATMEGA and the BS170 where smoked but no other damage
The  Si5351a Synthesizer board has its own 3V3V regulator that protected it
The QLG1 GPS rx was not yet connected besides its chip has it own 3V3

So i had to program a new  ATMGA328   with U3S V312.hex

The ATMEGA328 was programmed  with an other device verified okay
But it would not boot. I had to use my own ISP Ponyprog and SI-IO interface

Ponyprog V2.08d beta has the ATMEGA328 in the device list
Earlier version not.
With SI-IO to the U3 SPI socket it first failed to see the 328.
2 things where wrong:
1.The U3 Reset needs R3  100k  (91k)   not the jumper to +5
2. Had to remove C1 on CLK. 
Then it could read and write to the 328
Reading the security/configuration bits   only  BODLEVEL0 was checked
un checking this bit and the U3S was booting again

But much lower output level 50mW and the other BS170 already heats up !
No way reaching 500 800mW  like my first U3S does with a cool BS170
It seems every U3S behaves different
Could it be that the BS170 also is counterfeit ?
The circuit is simple not many parts involved the Si5351A drives direct over 100nF the BS170
It is 5V powered over a bifilar wound choke. The output is token over 100nF that's about it.
It is as simple as can be .

Removed the 5W-PA since it never could never bring the 28MHz  power to the wanted 5W level
A mistery to me.
It seems an easy job for  the RD15HVF1  to amplify  500mW to about 5W ??