WSPR today U3S deluxe project continued

last night no TA decodes i.e. WD2XSH/17  not seen  anymore

heard by SV8RV/1  and EA7HPM just over 2000km

but surprize this morning 28MHz open to VK
2017-03-05 09:30  VK3KCX  28.126060  +3  0  QF22qd  10  PG0DX  JO33ce  16471  312 
 2017-03-05 09:30  VK3KCX  28.126030  -8  1  QF22qd  10  PG0DX/1  JO33ce  16471  312 
 2017-03-05 09:16  PG0DX  28.126136  -12  1  JO33ce  200  VK3KCX  QF22qd  16471  78 
 2017-03-05 09:20  PG0DX  28.126135  -4  -1  JO33ce  200  VK3KCX  QF22qd  16471  78 
 2017-03-05 09:18  VK3KCX  28.126062  -8  0  QF22qd  10  PG0DX  JO33ce  16471  312 
 2017-03-05 09:30  VK3KCX  28.126058  -13  -2  QF22qd  10  PA0O  JO33hg  16442  312 
 2017-03-05 09:28  PD0OHW  28.126174  -27  0  JO33lc  5  VK3KCX  QF22qd  16424  79 
 2017-03-05 09:24  VK2KRR  28.126164  -23  0  QF34mr  10  PE1MPX  JO32gc  16392  313 
 2017-03-05 09:24  VK2KRR  28.126110  -11  4  QF34mr  10  PG0DX/1  JO33ce  16380  316 
 2017-03-05 09:20  PG0DX  28.126119  -1  0  JO33ce  200  VK2KRR  QF34mr  16380  73 
 2017-03-05 09:24  VK2KRR  28.126150  0  -1  QF34mr  10  PG0DX  JO33ce  16380  316 
 2017-03-05 09:24  VK2KRR  28.126138  -21  0  QF34mr  10  PI4THT  JO32kf  16367  314 
 2017-03-05 09:24  VK2KRR  28.126147  -16  0  QF34mr  10  PA0O  JO33hg  16351  316

U3S continued

Meanwhile  4 spare Si5351A arrived
First attempt to replace it on the older 2015 version Si-board that came with the progrock.
Sudden it had no more RF on any of the 3 ports and no I2C init. Possible ESD damage?

B.t.w  i gave up to use the Progrock  as  96.000MHz l.o. for the 23cm XVTR The result was even worse then with the original  96 Xtal  96MHz was random jumping and not on freq.

Managed to replace the Si chip  it was not easy, this thing is really small.
Normaly only handled by pick and place robots and not by an om.
After  replacing the Si chip and checking the 10 pin chip conections with  improvised needle test pins
and yes 1 pin needed reflow..
But.. testing on Progrock it still does not finish initial I2C communication check with the Si5351a
The LED stays red, it does not flicker at boot time and goes off as indication of completed I2C connection.
The other Si-unit a newer V4 with  TCXO option pads does communicate with the Progrock I2C
But when this one is plugged in the U3S, the display say Error: Si5351A  No Clk
I do not know exactly what "clk" is missing  it can not be the I2C that is present and the 27MHz clock also it must refer to the U3S output signal then  but   RF is on all its 3 ports  when in the Progrock.
The ports work.
What is wrong with the 2 units is still misty for me

Measuring the I2C  SDA SCL levels i see a constant communication data stream  on the not working unit.

I2C bit time about 3.6us  # 180kb/s
The Progrock does not recognize this data?

The other unit shows short data activity just after boot.
Then the RED led flickers and goes off like that is how it should be.
The level is perfect 5V ttl shape on both units.  the pattern seems not much different in timing.
An I2C  data analyser it would be nice to have one
So the Si5351A  seems to work normal,  the level shift BS170  okay 2-way data.
All is normal..  but the unit fails  no clue yet

The 3rd Si5351a unit  QRP-LABS send a free replace for the progrock.
 2015 Si-unit version without the TCXO option.
It it the only left that works with the U3S  after I added the LM317  needed for use with U3S
but Progrock has its own special low noise 3V3 circuit. 
However  this Si works on both boards with the LM317.

It all looks nice blue LCD in the U3S case the QLG1 extern over a 4wire extention cable on a shelf.
But the WSPR signal from the U3S on 12m 10m and 6m does freeze the WSJT-x  and also WSPR 2.21!
Many IMD products  and drifting like a free running VFO..
I did not yet dare to go on air with this signals.
Band filters can not filter the close IMD products of course.
The BS170 "PA" is in class A  but with switched off BS170 5V  
IMD is present on the  Si5351A ports

Now I wonder why do i have this trouble?  am I the only one?
Not seen other hams  with this U3 experience
Only LA3JJ Jon Ove describes how to stabilize the drift on 28MHz
but i could not fix that.

And PA3FYM  Remco noticed Progrock output glitches during I2C comunication.
All other hams  using the U3s  have perfect working stable devices...

This U3S still is an nice project
Reading my experience could give an other impression.
Until the breadboard 5V went  wild all was going fine.
With the spare Si5351a unit it worked.
I still hope to get a reliable working Beacon with clean reasonable 5W output