LF: W1VD amp help - MY FINAL POST

[N1BUG <paul@n1bug.com>]

For sake of completeness, as there were some following this hoping
to learn from my experience...

After a break I went back to work on the amp.

I swapped out the IR2110. No significant change.

I swapped out the 74F74. No significant change.

I replaced all four 1N4148 diodes on the gates with diodes from a
different lot/source. No change.

I experimented with the resistors on the gates, from 4.7 ohms to 47
ohms. Lower value resistors make the cross conduction problem and
drain ringing worse. Higher value resistors do the opposite but with
some penalty on power output.

I swapped out all four FETs. This returned drain circuit ringing on
both sides to equal (as it had been originally) but did not fix the
problem.

I built two more versions of the driver board, same circuit but
different physical construction. This definitely makes a difference
in gate rise/fall times, cross conduction and ringing but none of my
variations made it good enough. In fact the first version was the best.

I tried moving the driver board around, mounting it closer to and
farther from the FETs with varying gate twisted pair lead lengths.
This made a difference but seemingly not in any predictable way.

I believe there is some issue with my particular build wherein for
whatever reason the driver cannot turn the FETs on and off quickly
enough. It definitely appears the slow gate rise and fall times are
causing or at least worsening my problems.

Jay's design does work. I hear several of them on the air every
night and they are reliable. For whatever reason my particular build
does not work. Please don't judge a book by its reader. :-)

I wish to commend Jay on his tireless assistance in the effort to
figure this out.

In the end I feel I am not learning anything from this and I have
decided to move on to other things. I am also looking forward to a
summer break from building.



Since I am considering an upscaled version of my current class E
amplifier, I put that back on the bench and was taking some
measurements. It uses a single 34N20 FET. In particular I looked at
the gate waveform with the same scope, probes, and settings I've
been using with the W1VD deck. I was shocked at how steep the rise
and fall times are and how it seems to 'power through' the FET
turn-on anomaly. This stands in stark contrast to the W1VD deck with
its slow gate rise/fall and long hiccup at FET turn on. At least I
know the scope is capable of displaying a 'crisp' gate waveform. ;-)

Paul N1BUG