[ARC5] When a "Good" Tube, Ain't- AN/ARC-2 Success (long)

[Glen Zook]

I have always said that the equipment itself is the best tube checker in the world!

Over the decades, I have found numerous tubes that checked "good" in a tube tester that would absolutely not work properly in a unit and I have found numerous tubes that checked "bad" in a tube tester and would work fine in the unit.

I do have several tube testers but seldom use any of them.
 
Glen, K9STH


Website:  http://k9sth.com



On Friday, March 21, 2014 6:56 AM, David Stinson <arc5 at ix.netcom.com> wrote:
 
To recap:
My Collins AN/ARC-2 PTO would drift up about
400 cycles on key-down, then back again on key-up.
This seemed to follow the dynamotor's change in RPM
from receive to transmit.   Since the PTO is multiplied
in this rig, the result was an unacceptable "swing"
on 40 meters.

A couple of people said: "It's WWII; you'd expect some drift."
but I don't believe that Art Collins would have let it go like 
that.
The Collins-designed TCS oscillator is very stable
and this one is a later development.

I checked the B+ buss feeding the PTO and determined
the buss was at 260V during receive, 235V on transmit.
That's a 10% drop and would be expected from a
dynamotor source.  It looked suspect, but not in the
light of the paragraph above.  Still, I investigated it.
You can get to the 260V buss at the terminals of the
5.5 uFd buss filter cap on the right side.  One terminal
of the cap goes to ground and the other has three wires.
One of the wires is the 260V source.  I pulled this wire
and regulated the buss to 200V using a string of 5-Watt
Zener diodes with a 4.7K dropping resistor.  The buss was
solid during RX and TX and the 200V level made little
difference in power out.   This had no effect on the drift,
so I removed the regulator.

I listened to the drift again.  On transmit, the drift closely
followed the spooling-down of the dynamotor under load.
On receive, it recovered more slowly than the dyno's
spooling-up.   This followed the 2-volt sag on the A+ buss
on transmit, but not on receive (the buss recovered much
quicker than the oscillator).
I started scratching my head again.

About this time, The Almighty reached down and
turned-on one of those tiny little LED lights in the
cob-webby cavern between my headphones.
If the oscillator tube, a 12SJ7, had low cathode emission,
this is just what one would see-  emission dropping off
quickly as the A+ drooped, then recovering more slowly.
I started twisting away at the bunch of PTO screw.
Thank you Mr. Collins for making them "captive" screws;
I'd have dropped half of them inside and had to
man-handle that heavy beast around to shake them
back out.

Got the 12SJ7 out and onto the TV-7 for testing.
Minimum "good" indication was "40" and this tube
tested "45."   Hmmmmmm.....
Not great, but not "bad" either.
This is where it gets interesting:
I thought about what this tube was actually seeing.
Reached over and turned the "Line Adjust" pot to
lower the filament voltage a few percent....
and the tube's emission swooped down below 40!
Cranked it back up to the proper setting and the
12SJ7's emission slowly crept back up to 45 in
just the way I was hearing in the PTO.

Of course, we've all seen tube emission drop
with lowered filament voltage, but I've never seen
one drop like a stone over just a few percent change.
Went spelunking in "The Cave" and dug... and dug...
and DUG into my "12 V Octal" box until I managed
to find one NOS 12SJ7.
The ARC-2 PTO is now quite steady on all bands.
There is a shift (not drift) of about 100 cycles between
RX and TX due to the big difference in load
on the oscillator, but that's minor and acceptable.

Next:  Since I have the beast on the bench, I'm going
to investigate some of the non-destructive ways
to tighten the IF bandwidth.

GL OM ES 73 DE Dave AB5S




______________________________________________________________
ARC5 mailing list
Home: http://mailman.qth.net/mailman/listinfo/arc5
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:ARC5 at mailman.qth.net

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html