Drifting plate capacitor setting for homebrew amp

[ka1tdq]

I built a homebrew grounded grid amplifier using a pair of 3-500's and it works/sounds good except that I keep needing to adjust the plate capacitor setting during a QSO.  It drifts so far that sometimes when I transmit the next time, output is zero watts. 

What's the cause of the drift?

All I can think of is that the B+ keeps dropping when I transmit and therefore the plate impedance does as well.  When the impedance changes, I need to tweek the plate.  During the next transmit cycle, B+ has recovered to no-load conditions and needs to be set back.

I'm using a 3000 volt / 1 amp transformer and 13H of inductance with about 42uF of capacitance.  Full wave bridge.

Jon KA1TDQ

[WA1GFZ]

Check the rotor shaft wipers that ground the rotor. I had problems with a cap like that in my big rig

[ka1tdq]

Oh!  What is it, like grease or something that causes a bad electrical connection?

[WD8BIL]

Do you have a  meter on the plate voltage? What's it show?
What is the plate current doing while all this is going on?
Is there a padding cap across this variable?
Do you smell anything getting hot?

[ka1tdq]

Actually, no... I placed no meters on this rig.  I don't know what plate or grid current are doing, or for that matter plate voltage.

Occassionally, something catches fire on the connection for the plate cap terminal.

I guess that's bad... :-)

[WD8BIL]

Yup........ fire IS a 4 letter word!

[WA1GFZ]

my crystal ball has three feet of snow over it so can't see the problem.
How can you build a rig without meters? You may not be tuned up properly.

[K1JJ]

Joe,

Long shot, but are you using any padding caps in the plate or loading? (C1 / C2)   Caps like door knobs can get warm and drift.

As Buddly said, take careful voltage and current measurements and watch what's changing for clues.

Another thang to do (CAREFULLY) is to spray some of that coolant spray onto suspected components once the drifting takes place. Even the grid tuning, if it has non-air dielectric caps, can drift and have an effect, though little in GG.  Do not spray the hot tube glass.

Also be sure the tank coil is not expanding from heat to cause the drift.  Heat seems to be my guess. Spray the coil too.  See if the drift swings back.

Last resort, replace the tubes in clase the plate structures are changing, though that is lower probability.

Be sure the plate tuning cap itself has enuff friction so that it, itself, is not slowly moving from fan vibration. That's happened to me before using air variables.

T

[W1RKW]

Squirt a small amount of Deoxit on the variable cap rotor brushes.  they may be oxidized and causing intermittent contact. Long shot but worth a try. Variable caps can become intermittent.

[KD6VXI]

Do your tubes change color during this phenomenon?

I had it happen once, turned out to be a ground loop.  The front panel wasn't making good contact, and when I'd reach up and twist the cap, it would 'get a ground' connection once again.

Also, Cblock known good?  If it's passin mo amperes than it's sposed to be doin, could be causing it to drift, changing the tuning of your tank.

--Shane
KD6VXI

[KA2DZT]

Actually, no... I placed no meters on this rig.  I don't know what plate or grid current are doing, or for that matter plate voltage.

Occassionally, something catches fire on the connection for the plate cap terminal.

I guess that's bad... :-)

WHAT!!

No Meters??  Better go back to the drawing board. 

Although, not having to cut out those PITA holes for the meters does save a lot of time and work.  But, then you spend more time and do more work trying to figure out why the rig doesn't work.

Fred

[k4kyv]

No Meters??  Better go back to the drawing board. 

And I thought it was bad enough to operate a transmitter without an oscilloscope!

[ka1tdq]

Thanks everyone for the tips.  I am going to pull out the RF deck and throw on a plate current meter (which is needed as a bare minimum).  Also, I tend to get a gutt feel and agree with K1JJ that heat seems to be the contributing factor.

I am using a big 30 KV doorknob capacitor as a Cblock which may also contribute. 

I never thought about expansion/contraction of the loading coil.  One side is held by a long section of solid copper wire and that may be causing it to expand/contract.

Jon

[KM1H]

A. You need to monitor plate AND grid current....period

B. That 30KV cap is most likely from a TV set and is not RF rated and WILL change value when heated.

C. Fire could be a sloppy connection or parasitics. You DO have parasitic suppressors or is that something else you left out.

D. What is in the input network and what is the filament choke?

E. What is the drive power on AM?

F. Any photos and schematic?

Carl

[ka1tdq]

Hi Carl

It's probably a sloppy connection on the plate cap that's causing the fire.  I did put in parasitic suppressors on both tubes.

I was using a small T-network MFJ tuner for the input to the cathode.  I've temporarily bypassed it and just fed it with coax. 

I got the filament choke from RF parts... bifiliar wound and rated for the tubes.

Drive power is 20 watts using an Elmac A54H.

There's a good video on YouTube over the amp.  Just type in my callsign "KA1TDQ" and two videos should pop up.  The homebrew title shows much of the amps construction.

Thanks for the input too on the cap.  It's probably changing value with the heat and messing things up.

Jon KA1TDQ

[KD6VXI]

I've seen those TV doorknobs start to play havoc on 10 and 15 meters with as little as a couple hundred watts through them.

OK for bypassing the cold end of a plate choke, but otherwise it's too easy to order some russkie caps nowaways...  I've had great luck with them, at power levels higher than most would run...  Pretty much indestructable for the dough.


--Shane
KD6VXI

[Steve - K4HX]

Although it's somewhat risky operating w/o meters, it's not that bad if you know a few things.

First, he's driving the amp with an Elmac A54. The A54 is good for no more the 50-60 Watts. So, he cannot over drive the amp. Fifty watts of drive won't even get close to the plate current limit for a pair of 3-500z in GG.

If he has a wattmeter or relative power output indicator on the output, he can tune for max power out and be very close (if not right on) to the correct tuning.

Once you know the color of the plates at the normal/correct operating parameters, you can tell if things are OK by just checking the color of the tubes.

In other words, no meters is not as bad as it sounds in this case.

[W2PFY]

I thought the videos were great. I was a little worried when I saw the high heals and was afraid I would see some guy in drag, but alas, no bonny legs and knobby knees 

What receiver is that? Aside from using no meters,it looks like your in the spirit of radio and your building skills are up there. I can see you have a nice feel for layout. Maybe your the kind of guy who can build those super duper hi end audiophool amps as well

If you need meters, maybe someone on here has a stash that can help you out?

[Gito]

Hi

Sometimes the tuning changed,because the load changes
 in this case  the Antenna,
if there's "bad " connection or the heat that causes the Antenna change .
That can cause the load to change ,making the the transmitter untuned.


Gito

[ka1tdq]

Yeah, it does the same thing into a dummy load too.  It's definitely heat related. 

The receiver in the video is an IFR Communications set.  I was borrowing that from my company but now I'm using an Icom 718 as a receiver.

I've blown another plate choke on the Elmac (2nd in a week) so I'm slapping together a homebrew rig that's a little more husky.  It's a 4-65A modulated by a pair of 6L6's.  I should have it done in a couple weeks.

Thanks for the compliments on the video by the way.

Jon KA1TDQ

[KM1H]

Steve, an AF-54 is lucky to get 30-35W out but thats still enough to get 3-500 plates a bright orange as an AM linear especially with a healthy PS at 3000V out of the filter that doesnt sag like a SB-220.

I run about 25W into the LK-500 and a 350W carrier and they are a nice orange plus I have the fan in the high speed position.

[W2PFY]

It's a 4-65A modulated by a pair of 6L6's.  I should have it done in a couple weeks.

Please post some pictures when you get er done

[DMOD]

Also consider the condition of the Transit/Receive Relay contacts and the relay insulation material for carbon tracking.

I had this problem develop on my Henry one time, a 2K-4 with twin 3-500's.

Discharge HV system, and then Spray Deoxit on relay contacts. Then run a paper business card or strip of unprinted paper through the relay contacts to dry them off and to remove any solid particles. Wipe off relay block as well.

One more thing; if the cathode input circuit to the GG does not have pi matching-circuits, do use your low power mfj tuner for matching.

While the input resistance portion of the input impedance of twin 3-500's is
close to 55 ohms, you still need to tune out the reactance.

Phil - AC0OB

[Steve - K4HX]

That amount of drive is good for about 400 mA of plate current. You'll never burn up a pair of 3-500s that way.

If there's any problem with the setup it is that he cannot drive the amp with enough power to properly load it.

Steve, an AF-54 is lucky to get 30-35W out but thats still enough to get 3-500 plates a bright orange as an AM linear especially with a healthy PS at 3000V out of the filter that doesnt sag like a SB-220.

I run about 25W into the LK-500 and a 350W carrier and they are a nice orange plus I have the fan in the high speed position.

[KM1H]

Thats 100W PEP which will easly draw 800ma on voice peaks. If I drive the LK-500 at 30-35W the tubes really glow plus they are into splatter mode. And thats with only 2600V key down.

At 3000V and 400ma of carrier thats 800W doing nothing but helping the anode glow, there isnt much headroom left especially without a forced air chimney cooling system.