[Amps] Advice needed for SS amp

[Manfred Mornhinweg]

Alan,

> I am at exactly the newbie stage Manfred advocates: building a 50W 
> amplifier using a pair of one dollar switching FETs, in my case IRF640, 
> and a cheap 48V laptop PSU. I have a prototype more or less working for 
> use on CW on 60M.

Those FETs have massive internal capacitances. If you get those to work 
throughout HF, then you are ready to use better switching FETs up 
through UHF! ;-)

More likely you will have to keep that amp operating on the low bands only.

> I'm interested to know why a common mode choke type of balun is such a 
> bad idea, seeing as I am using one to couple the drains of my amplifier 
> to the harmonic filter.

With a common mode choke you get no DC insulation. In your amplifier, 
fed by a 48V power supply, that's absolutely no problem, and it's fine 
to use that common mode choke. The coupling capacitors will block the DC.
But when building an amplifier that uses the directly rectified and 
filtered 220V mains as power supply, safety regulations demand some more 
insulation than just a capacitor. Your power supply becomes 320V DC, and 
there is also a common-mode 220V AC on it. A capacitor would pass enough 
AC to be a hazard. It's here where you need an output transformer with 
galvanic insulation, ruling out a transmission line transformer and a 
common mode choke.

> I have 5 turns of RG316 on a core with OD about 
> 18mm, giving a choking inductance of around 17uH. Balance seems fine and 
> this configuration seems much less beset by problems with leakage 
> inductance compared to my attempts at making multi-winding conventional 
> transformers with bits of brass tube and so on.

Yes, you are right in that. But it can be done with conventional 
transformers too. It just takes some more effort!

I spent a few days until achieving a conventional transformer with a 1:4 
impedance ratio for a legal limit amp, that covers 160 to 10 meters. I 
had to use some tricks, and the end result is a transformer that does 
the job, but just barely. It would definitely NOT cover 6 meters.

> For "galvanic isolation" (I call it DC blocking) 

Its not the same thing. DC blocking passes AC. Galvanic isolation blocks 
both DC and AC. When powering an amplifier off-grid, the whole amplifier 
has the AC grid voltage on it, instead of being at ground potential. 
That's why in that case you need full galvanic isolation, and not just 
DC blocking.

Well, you MIGHT still do it just with capacitors, choosing their value 
so that they pass RF well enough, and block line frequency well enough. 
And then place an RF choke from the output to ground. That would satisfy 
basic functional requirements, but it doesn't satisfy safety 
regulations! And worse than that, it would couple a mighty amount of RF 
into the power grid!

So, you DO need galvanic insulation, via a conventional transformer, if 
you build an amplifier powered directly off-grid.

 > I am using physically
> moderately large 100nF multilayer ceramic capacitors rated 100V. Of 
> course it's the current rating that matters and I have no idea what it 
> is for my capacitors except it must be enough because they do not get hot.

So they should be fine. But try to get the current rating. Most 
datasheets are available online.

> OK, so the blocking capacitors in my baby amplifier don't have to handle 
> the 5 or so amps of RF in a 1KW+ amplifier, but such capacitors do seems 
> to be available, particularly if you use several in parallel.

Yes, they are available, even if they tend to become expensive.

What I miss, in terms of availability, are capacitors suitable for 
building good, safe, legal limit low pass filters, that don't blow up 
even at somewhat elevated SWR. What we need for that are capacitors in 
the range of about 68pF to 4.7nF, each rated at one kilovolt AT LEAST, 
and 6 to 12 amperes of RF current. And since we need about 18 to 45 of 
them, depending on what sort of filter bank we want to build for an amp, 
they shouldn't cost an arm and a leg each. That problem seems so hard to 
solve, that so far I have resorted to homemade copper/mica capacitors. 
But making these capacitors to the correct values is always 
hit-and-miss, because the mica sheets are of varying thickness.

Does anyone have a good hint about this?

> Or do really bad things happen at the higher voltages and currents in 
> serious amplifiers that I have overlooked?

I think that the only thing you overlooked was that Bill intends to 
build an amplifier powered directly from the grid, without a transformer 
of any sort in between. That's a different situation than using a 
galvanically insulated power supply and having the amplifier ground at 
true ground potential.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================