You may remember a couple of weeks ago, I asked for comments on the amplifier
that was OK at 3.5MHz and went mad at 3.7MHz. At last, I'm at home for long
enough to actually look at it. Someone suggested that the 'non inductive' grid
shunt resistors wound Ayrton Perry style, might have had one winding go open,
and this was indeed the case. There were three 80 ohm ones in series: one was
bad. The bad one measured at 158 + j22, while a good one measured 80 + j7. Some
of the inductance was doubtless down to the leads - I estimate about +j2.
Interesting though, because we still have a Q of less than 1 for the whole grid
circuit, so it's somewhat surprising we can get an oscillation.
So I'm moving to a straight 75 ohm Corning glass - tin oxide 50 watt shunt
resistor. We'll see how that works out. Worst case SWR at 25 MHz is about
1.6:1, so with a tube driver, no problem.
The cathode circuit 1 watt carbon 18 ohm resistors had got hot - they're there
for NFB, with the filaments held above ground RF wise with a toroidally wound
choke. You don't need a lot of inductance to act as a choke at 1.8MHz when
there's 10 ohms of resistance across it. They are down for replacement with
some 'mini glo-bar' 5 watters I got at Dayton the other year
One interesting point about this amp (pair of 4-250A0 in parallel). It was
originally intended for military use from 1.6 to 22MHz, so you might expect
plate choke problems. Having it all apart, I had a look at what it uses. The
first choke is a solenoid, about 3 inches long, 1 inch diameter, 76
microhenries. Then there's a heavy duty high voltage, high kVA 680 pF ceramic
to ground. Then there's a single pie universal wound inductor, about 1 inch
diameter, and 650 microhenries, and another 680pF to ground across the HV. Now
I don't own a GDO (OK, Rich, you can say 'shame'!), so I tried an experiment.
Feed the tracking generator through 4k7 to the top of the choke where the plate
is connected, and put the FET probe from the plate to ground. The idea is that
a series resonance in the choke network should show up as a dip in the
transferred energy. The result was a nice series resonance at 700kHz, which is
where one would expect it - and no others. Calculation shows that the net
work only looks like +j728 at 1.8MHz, which seems awfully low, and will lead
to appreciable current in the solenoid - which is wound with about 20 gauge.
Yet it hasn't burnt up.....
Of course, after all the changes, we still have to see if the thing still goes
mad....It has parasitic suppressors in the grid, too. 5 watt carbons, 47 ohm,
pristine condition, with 3 turns, 3/4 inch diameter. 3/4 inch long across them.
The anode parasitic suppressors are similar to Glo-Bar, 47 ohms and 2 inches
long, 1/2 inch diameter, rated at 25 watts. I suspect the parasitic suppression
was put in like that 'because that's how you do parasitic suppression'. They
didn't have fancy network analysers in the late 1950's! Indeed, they wouldn't
have had a spectrum analsyer and tracking generator as good as the HP141.
73
Peter G3RZP
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|