[hpsdr] Unkeying Spike on Mercury/Penny revisited

[Tom Cathey]

Hi All,

I wonder if someone can shed some light?   It’s been over a year since I first mentioned an unkeying spike I get with my Mercury/Penny/Ozy combo.  

First, the system works great in all respects except for this unkeying spike problem.  I love its performance!

This week I changed to a new computer and put in the latest firmwares and W5WC software.  No change – still spikes on unkey.   On unkey, using either the MOX with mouse or straight PTT, I see the S meter slam to 60 db over and the spectrum scope transmit signal spike to the top of the screen. It doesn’t happen every time - maybe every third PTT.   BUT every sixth PTT cycle that the spike occurs, the Mercury looses its mind and goes into an oscillation. Penny them transmits with a pulsing sound. If I turn Atlas on and off and reboot, everything works fine until I go thru six PTT spike cycles and it happens again.

So far I’ve substituted in a rock solid 5 V supply and added large capacitors across each power lead. (-12V, +12V, 5V)        I’ve changed about every software parameter possible in the W5WC software.  I’ve even changed to other PWRSDR software versions with no change.   I don’t see any significant change in the supplies when looking at them on the scope – however, when the oscillation occurs, it seems to be everywhere, in the audio, power supply, etc.  Changing position of the boards has no effect on the spikes.

I can disconnect both ant leads from Penny and Mercury – or disconnect ALL external leads from the boards, except for the Ozy USB cable – and the spike still occurs.  The SM caps are all installed on the Atlas board as required.

Doug (W5WC) was nice enough to add some software keying delays in his software last year, but these have no effect on the problem. It seems the keying spike is being generated in a stage before the delays? I dunno.

I am convinced the problem lies in the hardware or in my layout and is something particular to what I am doing, since no one else seems to be seeing this. But I am almost out of ideas. I have a nice FT-1000D  200w solid state linear all set up and want to use the full HPSDR system on-line, but for now it’s just a glorified receiver.

One thing that may be a problem or not:  The power supply module is located about 8” away from the main connector and each supply lead is twisted with it's own ground lead. I know others running the supply off the board with no problems, but cannot be sure.  It’s a big job to relocate.  I added ferrite blocks onto various PS and control leads, but no effect.    

Any ideas on how to troubleshoot this spike? 


BTW, my thanks to Joe, K5SO for his patient help over the last year. Without his guidance, it would have been difficult getting the firmware and overall software set up.  Frank / WA1GFZ  has also been a great help. Thanks again, Joe and Frank!

73,

Tom, K1JJ
CT
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20121203/b3106ef0/attachment-0003.htm>