Hello Chris,

I would be interested in the collective wisdom of the wise ones in
this group regarding replacement of FETs. Given they've all surely
had a strain when one or more blew, is it really wise to just
replace the ones that check bad? I'm leaning toward replacing all
four whenever one or more is dead. Am I being too paranoid?

I really wish we could socket the FETs.  When I am in there this
time I may see if I can find a not too messy / not too destructive
way to mount those copper strips the FET drain leads solder to so
they can't move around. When replacing FETs I find it oddly
aggravating holding that metal strip in place. Invariably the heavy
lead from the transformer then comes unsoldered from it and I have
to try to put that back without any drain leads or the snubber cap
coming unsoldered. It'd be nice if that copper strip were solidly
mounted and the transformer lead bolted to it (I think)...

I can still get FETs in the big package if I go 32N20 instead of 34N20.

73,
Paul


On 11/20/18 10:10 AM, Chris Wilson wrote:
> 
> 
> Hello N1BUG,
> 
> Probably  teaching  granny  to suck eggs but if you take off the HV to
> the  drains and just run the 12V supply and the exciter and `scope the
> gates  you  will quickly find the dead FET / FET's and also see if the
> driver IC is functional.  I think I have killed only one driver IC....
> :)  My  IC's  are  socketed, makes changing one a snap. I did consider
> socketing the FET's but I assume quite a bit of heat is disipated down
> the  source  lead  to  the ground plane of the PCB so I chickened out.
> Will  be  interested to see how the little FET's fair. I can still get
> the big packages but supply is intermittent.
> 
> Tuesday, November 20, 2018, 2:48:10 PM, you wrote:
> 
>> Hello Chris,
> 
>> Thanks for the information.
> 
>> I will order FETs and other parts today. It will have to be 32N20
>> this time, as you know the 34N20 in the large package is long
>> discontinued. While I am at it, I will get a supply of diodes and
>> resistors for present and future rebuild of that gate network when
>> changing FETs. I'd best order a couple driver chips too, as I had
>> problems with that before.
> 
>> I hope nothing else has been killed, but I haven't had time (or
>> energy) to pull it apart and go through it yet. I believe this
>> problem started with an oscillation in the exciter (converter) so
>> anything is possible.
> 
>> My LPF is the same as yours and Jay's. During initial testing with
>> the cover off, I found the center inductor was running warm. The two
>> on the ends did not seem to show much if any warming. In the
>> permanent form I put a fan on the center compartment, moving a good
>> bit of air across that inductor. I didn't do much testing after
>> this, but I think with the fan it should be running cool enough. I'm
>> running my amp at less than 50% power. Of course I have a fan
>> (rather larger than what Jay used) blowing air across the output
>> transformer in the PA as well.
> 
>> Perhaps the resistance change as things warm up is nothing I need to
>> worry about, but when I get the PA back working I will try to run
>> some better tests to see if I can find what is going on. Even if
>> it's not a problem, I would prefer to understand it. I'm odd that
>> way. I don't like mysteries! :-)
> 
>> 73,
>> Paul
> 
> 
>> On 11/20/18 9:04 AM, Chris Wilson wrote:
>>>
>>>
>>> Hello N1BUG,
>>>
>>> When I have blown a FET on Jay's amp design and replaced it / them I
>>> have sometimes seen less than optimal gate and drain waveforms from
>>> the replaced FET's. A simple meter "diode test" on the parallelled
>>> diodes / 10 Ohm resistors showed the diodes to be OK, but changing
>>> them restored things to a good pattern. i now change them as a matter
>>> of course if a FET blows. My matching transformer is a multi tapped
>>> primary / secondary pairing on three individually Kapton taped 77
>>> material 240 size cores stacked together. I see no heating t full
>>> power WSPR15 or Opera but they are in a sealed plasic box of more than
>>> generous size. I *HAVE* had issues with the ground terminals on the N
>>> type connector screws arcing although they appeared tight. I now
>>> solder direct to the socket flanges before mounting with the socket's
>>> flanges on the inside of the box. Hope you get it sorted soon, a box
>>> of spare FET's is a necessity, "things `appen" ;)
>>>
>>> The only thing that gets warm is the LPF, it's in three divorced
>>> compartments like Jay's and i think yours? I use a small PC processor
>>> fan on the middle box with an exit hole meshed over opposite the fan.
>>> Air flow is adequate to keep all three toroids near ambient with this.>>