[ARC5] ARC-5 CHiiiiirrp!

[Dennis Monticelli]

Whatever the root cause turns out to be, I'll bet it's a thermally
induced variation somewhere in the signal path (from the 1626 all the
way to the antenna or dummy load itself and any poor connection in
between).  There is simply no normal electrical time constant in the
ARC-5 (or your power supply) that can produce such a slow time-varying
change.  I once had a one to two second chirp traced to a very flat
1625 (the other one was good).  The high grid current in the flat tube
was varying slowly as the internal tube elements heated up over those
seconds.  The slowly varying grid current slowly varied the 1626
loading which in turn caused the chirp.  Something in your rig or
power supply is changing slowly due to a thermal effect.  If you have
a scope, you can look at the circuit potentials and figure out which
node is drifting and then drill down from there.  The clue is in the
long time constant you are experiencing.

Dennis AE6C

On Sat, Nov 15, 2008 at 3:13 AM, David Stinson <arc5 at ix.netcom.com> wrote:
> (recent correspondence, copied to the list for info,
> comments and other suggestions)
>
> ----- Original Message ----- Subject: ARC-5 CHiiiiirrp!
>
> I've put a T-19 ARC-5 back on the air.  It has a long chirp on key-down.
> Drifts about 500 - 800 cycles in 5 or 6 seconds, then seems to settle down.
>
> 1625 plate voltage    =  490 key up
>                             =  450 key down and pulling 50 ma.
>       screen voltage =  140 volts regulated via 0A2 which stays lit key up
> or down
>
> 1626 plate voltage = 202 volts key down
>
> Cleaned the "decorative oxide coating" from the roller inductor & associated
> parts...........No change.
> Tomorrow I replace the 3 ea. .05 mfd caps in the can on the back wall.
> Could it be that my power xfmr is just too puny?  Key down pulls another 22
> watts out.  Perhaps key down causes the filament voltage sags a bit?
> Guess I could load down to 4 watts or so and see if there's any change.
>
> What other gremlins cause this type if fast drift/slow chirp?
> ------------------------------------
>
> How are you keying the rig?  The old "handbook" way was
> to tie the PA cathodes and osc. cathode together and key them
> to ground.  This is bad and chirp is inevitable, but not the freq.
> pulling you're seeing.  That's probably something different.
> If you're using the selector relay to key, make sure the
> osc. contact makes before the PA cathodes contact.
> You can gently bend them to accomplish this.
>
> How is your power supply constructed?
> Could you possibly send me a diagram?
> In the original design, all the power was derived
> from a simple resistor voltage divider off the
> 550 volt dynamotor.  Doesn't sound like something
> that would give you a stable rig, but it worked
> very well.  If you have a supply that can provide
> a solid 450-550 volts at 100 mils or better without croaking,
> it's still the best way to go.
>
> Are you testing the rig into an antenna, or a good dummy load?
> How are you feeding it to match 50-ohms?
> In any MOPA rig like this, changes in the antenna circuit
> will pull the VFO.  If you can find one, get yourself a great
> big *temperature-stable* 10-ohm resistor- wire-wound
> is OK for this test at 3.5 MC, if it's the kind that uses few turns
> of bigger wire- and tune up the rig into it using
> just the original output circuit.  Does it still pull?
> If not, there is a problem either in the things you're using
> to match 50 ohms (like a series cap that's heating and changing),
> or something crummy and heating up in your antenna,
> like a corroded coax connection.
> If it still pulls, the problem is the in the rig or the power supply.
> By the way- to determine if you resistor is temp stable,
> and it must be for this test, connect an Ohmmeter across it
> and heat it up with your soldering iron.
> If it drifts, it's no good for this app.
>
> Put some good contact cleaner, like De-Oxit, on every
> connection for the antenna circuit, especially the
> rotating contacts at the link coil and at the slip
> contact at the end of the roller coil.
>
> While the 3-banger .01 is often to blame, it's not always.
> Do you have a good cap tester
> that can test them under voltage? It's a good idea
> to replace them in any case, and you can get tiny little
> poly film caps that can solder right across tube sockets
> to replace them.  Don't remove the wiring to the old caps.
> Just cut it loose and follow them to their connection points,
> which are at tube socket, IIRC.
> Solder the caps there to ground.
>
> Check all the solder connections.
> Some of them have gotten flaky over the years.
> Reheat any that are suspect.
>
> Has the rig been in "ham" use before?  If so,
> you will want to check your 1625s for grid emission.
> These rigs were often abused by hams.
> Substitution is the best test, and ditto for the 1626.
> Also check all resistors for discoloration.
> Hams often "over-volted" the rigs with
> the results you'd expect.
>
> Here's a symptom that's been showing up alot lately-
> If you look at the end of the big variable caps,
> including the one under the VFO cover, you'll see
> that the connection point for the stators is a press-fit
> terminal to which a wire is soldered.  After all these years,
> these press-fit connections are beginning to get
> intermittent and Hi-Z.  Use a soldering gun to
> solder the place where the stator has been press-fit
> onto the connection strip.  Solder the pressed end
> nearest the wire terminal connection, and do that on
> every variable cap, including the one under the VFO can.
> Be careful not to over-heat and melt the little insulation balls
> that hold the cap together.  I've never had this happen
> but I guess it could.
>
> How much does your Osc B+ change from key-up
> to key-down?  Can you put a good meter on it,
> one that has enough resolution to follow the voltage
> during a key-down, to see if the voltage drifts in
> time with the VFO drift?  A scope would be better
> for this test if you have one that can handle the voltage.
> You can use a hi-Ohm voltage divider as isolation if you
> scope won't handle 250 volts.  Connect a 10-Meg resistor
> to the voltage source.  Connect the other end to your probe,
> the from the probe to ground with 1 Meg.  Use good
> resistors that aren't going to drift around on their own.
> If the voltage does drift with the VFO pull, you need to isolate
> which side of the equation is the "variable."
> Measure the Osc stage current, then
> do the Ohm's law calculation and substitute
> a big resistor for the VFO stage.
> Connect it between the power supply output
> and ground through something to "key" it.
> Does the voltage drift the same way as it does
> with the VFO connected?  If so, the problem is
> in the power supply- what kind of filtering is in that
> part of the supply?  If it drops to your "key down"
> level of about 202 volts and stays there without
> drifting, the problem is in the rig.
> Note that this isn't 100% certain; there could be
> some complex reactance interactions,
> but that's not likely.
> Also note- If your Osc voltage is derived from the
> same transformer as your PA supplies,
> you'll need to load the PA plate and screen supply
> points with honking big resistors to simulate the transmitter
> load in order to make the test accurate.
> If you keep the test short, or if you simulate a lower
> power-output tuning condition,
> you can get away with smaller resistors.
> If the VFO still pulls when you're tuned for lower
> power out, that will make it easier to find load
> resistors for subbing the PA stage.
> I know- this is a pain.  But such is our poor lot. ;-)
>
> While I've never seen it, I've heard that
> RF can get back into the power supply and
> cause drift like this.  Are your power supply
> leads properly filtered and bypassed at the
> supply to suppress RF backfeed?
> Again- I've not seen it, but it does make some sense.
>
> In the rig, you could have either a bad cap
> at the plate of the 1626 (one of the 3-bangers),
> grid emission in any of the tubes (test by subbing)
> or the hi-Z connection at one of the variable caps
> we talked about earlier.
> Of course, there could be a problem with other
> components in the oscillator circuit, like the
> grid leak parts or the mica bypasses or
> the plate resistor, but I've never seen any of
> those go bad.  They could; I've just not seen it.
>
> This should get you started.
> Let me know what you find.
>
> 73 Dave S.
>
>
> _______________________________________________
> ARC5 mailing list
> ARC5 at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/arc5
>