Hello Graham,
Friday, October 16, 2015
> I am wondering if this feedback or whatever sometimes happens has
> damaged a FET, can one or more of the 4 fail and still leave it
> operational, but with reduced power and more reflected power?
> Sorry to read that Chris ,
> I often thought wspr would of been better named 'ICARUS' , you
> where +10 dB and more in hand on your OP32 spots , I did see G3XKR
> showing -4 dB on the first day , but only the one time, later spots
> where lower , [the s/n is the average over the tx cycle and quite
> accurate],
> Could problems of struck after this ?
> The reflected power, is most likely to be the return of harmonic
> energy from the ATU, caused by waveform distortion, your class e/d amp is
> based on producing a pulse train, fo is recovered by the filter / ATU ,
> One of the £5 usb dongles in ds with sharp-sdr , will give you a
> usable spectrum plot of the relative levels , f0,f1,f2 f3 etc
> 73-Graham..
I have found the Zobel network resistor (22 Ohms, 25W) badly
overheated but not open. I also found one leg of one of the 4n7 caps
on the primary of T1 the output transformer broken. I don't have
another metal clad panel mount of the right resistance unless I
parallel two 47 Ohm 50 Watt ones, but there's not enough physical
space, so I have put the old one back until a new one arrives, and
renewed the 2kV 4n7 cap. I now see roughly square waves from both
sides of the output, looking before the filter network. Before I had
only one squarish wave and a funny wiggle showing on the scope.
I would like to know if the broken leg on this cap could have caused
the Zobel wirewound to overheat though. As I noticed the odour days
ago. Sadly my inexperience has made me build the thing with scant
reagrd to access for repairs or even to look for smoke. If I did it
again I would spread stuff about more in a single layer!
The meter still doesn't seem to read as high FWD, as it did, and shows
some reflected power, but it seems to be getting out OK again. As I am
learning a lot as I go it's no big deal. If it just worked I wouldn't
have learnt half as much about HOW it works as I do now :) It's TX'ing
OPERA 32 on 136 kHzas I type. Maybe now the output stage is correct
again the loading needs tweaking a bit. Scopematch looks OK though.
Thanks for the reply Graham.
--
Best regards,
Chris mailto:[email protected]
|
