[HBR] That General Coverage HBR Project -- 2

[Walt Hutchens]

Jim said:
> C = .00000022 F or 0.22 uF
> 
> If there's a dozen or more bypass caps of .01 each, plus some other
> leakage - SNAP!!

You guys are terrific!  Don't ask me why, but it never occured to me to
try to calculate that.

As it turned out, the problem was a bad (cracked) feedthrough cap on the
filament line going into the 2nd mixer assembly.  Starting at the hot
end I went down the filament line pulling tubes in turn until it
switched from 'problem goes away' to 'problem is still there' meaning
that whatever component was the source, now had voltage on it.  Sure
enough, when I looked closely at the feedthrough going into that box, I 
could see a hairline crack.  Swapping it made the problem go away.

I really hate using those things -- they break at a careless look, or if 
you even wave a small wrench past them while still in the packaging. 
Mine are all salvage so I have several chances to break them -- 
attrition runs around 25%.  Then there's the risk of soldering ...

Anyway, now I know how to think about bypassing in the set.  The
filaments aren't quite so bad as they seem because the voltage declines
as you go along -- only the first tube has the full 120 VAC on its
filament bypass.  Other bypasses are on lines that are nominally DC-only
-- the currents would be just stray signals, probably mostly in the
microampere range.  The HV filter caps are huge and carry high currents
but they go hot line to neutral line -- there shouldn't be any current
in the ground wire.

I have, however, had trouble in the past with a bypass from the neutral
line to ground.  One factor is that there can be a significant voltage
between neutral and ground; I just measured 4 mV in this receiver with
the switch off and 0.75 VAC with it on.  Since that represents a current
flow of 0.3A (filament current), there's a roughly 2.5 ohm resistance in
the neutral line between the point where the pole pig is grounded and
the entrance to the set where I made the measurement.

There is considerable trash on the neutral line -- I have looked at it 
with a scope in the past and there's all manner of stuff.  Plenty of 
reason for heavy bypassing -- but you have to be careful.  If you use 
solid state rectifiers for HV supply then unless there's a series 
impedance you get very large current spikes in the neutral and if you 
observe that those VOLTAGE spikes (remember the 2.5 ohms) are a source 
of a buzz in your audio, then you might like to put a really big cap -- 
say 1 mfd or more -- from the neutral to ground.

You can help the buzz that way, but you are likely to get GFI problems
-- been there, done that, on previous sets.  It's better to add some
impedance in series with your rectifiers.  Either resistance -- using
vacuum rectifiers as in this set gives you that automatically -- or a 
suitable inductance.  The latter would be better for a transmitter or 
other high-current supply and I guess you'd want a reactance in the
ballpark of a hundred ohms at 60 or 120 cps, depending on the rectifier
circuit.  That would considerably flatten the current peaks and greatly
reduce the resulting noise pulses on the neutral line.

Not too many people build 'transformerless' transmitters these days, but
it can be done reasonably easily and with advantages similar to those
for receivers at modest power levels.  The killer at higher levels is 
the need for a voltage multiplier to get the necessary HV; at the 
currents typical of a transmitter the power supply filter (and series) 
caps begin to compare to a plate transformer in size.  I build an AM 
transmitter that ran around 40 watts output and worked okay, but I 
wouldn't want to try to go much higher.  I think that one was parallel 
22JF6's modulated by P-P 22JF6's with two or three more tubes as crystal 
oscillator and audio.  0.45A filaments.

The Rx audio stages are working okay for now and I'm trying to get the 
VXO BFO working correctly.  None of the circuits I have are exactly 
right and as usual I don't know exactly how to think about the theory ...

All things in good time.  It won't be possible to finish the VXO until 
the frequency range of the ladder filter is known.

Bob said:

> metric HW?
> mdmetric.com

Thanks -- bookmarked!

Walt
KJ4KV