triode neutralization

[KC2ZFA]

seeing the Bauer thread prompted me to sprint ahead of myself and try neutralizing the push-pull 812A transmitter I've been building the last year or so.

To do this (on 40M) I put a dummy load on the swinging output link and with no plate voltage on the 812's I proceeded to screw down on the neutralizing capacitors while watching the grid current variation as the PA plates were tuned through resonance. This current variation was gone when the caps got in the positions shown in the attached pic. Also, at this setting there was no current variation even with the swinging link all the way out of the PA coil.

Here's my question for those who have experience with such PA's:

Do the neutralization caps look like they're in the ball park ? I was afraid I would run out of capacitance before grid current variation was eliminated.

Thanks, Peter

[W7TFO]

Hi Pete,

Those look to have more than enough for that tube.  Might be a little tight (uuf wise) as in the photo.

PP is easier to neut, as you always have some out-of-phase energy close by in the usual cross-connection.

You ought to get a pronounced indication, maybe run a neon lamp around and see if you get RF variations whilst setting things.

A receiver tuned to your freq can be an audible tattletail as well.

Above all, don't be in a hurry...the 'sweet spot' goes by pretty fast.

73DG

[W8ACR]

Hi Peter,

You said that there was no plate voltage, but you didn't say whether the 812's were completely disconnected from the B+ supply. If I am not mistaken, the HV power supply should be physically disconnected from the final tubes. I have not neutralized a push pull final, but I have neutralized single and parallel triodes with pi network tanks. I simply put the tank at resonance and adjusted the neut cap for minimal reading on my homebrew neutralization meter. This meter is very simple to build and I think helps immensely. Look in any old handbook. It consists of a 1mA ammeter, a diode and a coil if I remember correctly. This seemed to work well. I have no experience with 812's, so I can't comment on the spacing of your neut caps. It might help to know the make and model of the neut caps.

Good luck, Ron

[KA2DZT]

Peter,

Caps look good,  xmtr looks good too.  You don't want to go any closer with plates of the neut caps or you may get some flash over.

Fred

[W3GMS]

Peter,

There are many ways to neutralize a final amplifier, but the procedure I use was taught to me by my mentor, Harry W3FDY (sk) many years ago when I was a kid!    It has worked flawlessly so its the only procedure I use. 

For triode's, kill the HV plate voltage to the final stage and for Tetrode's kill both the screen and plate voltage.   For the plate supply, do this at the base of the RF choke or further downstream towards the power supply rather than pulling the plate caps! 

Do the following procedure on the highest frequency band of operation

Put a dummy load on the output of the transmitter

Put a scope across the dummy load

Turn on the drive to the final amplifier

While watching the scope, peak both the tuning and loading capacitor for maximum feed-thru power as seen on the scope

Adjust the neutralizing capacitor(s) until you have minimum feed-thru of carrier on the scope

Go back and re-peak the grid and final tuning while watching for maximum deflection on the scope

Re-adjust the neutralizing capacitor(s) for minimum deflection on the scope

Keep doing this until no improvement is made

Have Fun!

Joe, W3GMS

   

[KC2ZFA]

thanks guys.

Ron, yes, the plate HV is completely disconnected, the cathode relay is closed, and the filaments are lit. The grid and cathode current meters agree.

For now my neutralization indicator is the (lack of) grid current variation as the plates are tuned through resonance.

I will have to make two nonconducting sticks with a little metal blade at one end to attach to the caps so that I won't have to reach back and turn their plates by hand.

[KA0HCP]

Peter,
I have a backwards method that may serve as a cross check.  In cleaning and preparing my Gross transmitter, I wanted some way to get the neutralizing caps in the ball park.   So, I looked up the reference internal capacitance of the 35T tube, then put my capacitance meter across the cap and adjusted it to the same number.  

I haven't fired up the radio yet so have no idea how far off I may be, but your picture looks comparable to how mine are set.

Bill

[k4kyv]

Hi Peter,

You said that there was no plate voltage, but you didn't say whether the 812's were completely disconnected from the B+ supply. If I am not mistaken, the HV power supply should be physically disconnected from the final tubes. I have not neutralized a push pull final, but I have neutralized single and parallel triodes with pi network tanks. I simply put the tank at resonance and adjusted the neut cap for minimal reading on my homebrew neutralization meter.

I discovered that when I built my single-ended plate neutralised final with the 304TL back in the 60's.  With the plate lead connected and the +HV simply turned off, I was getting a substantial DC plate current indication on the meter with certain grid and plate tuning settings. The neutralising indicator (rf meter, small pilot lamp connected to the output, or a neon lamp held close to the tank coil) would show the presence of RF and I couldn't get a clear null regardless of the neutralising capacitor setting. With the +HV lead physically disconnected, everything behaved just as it was supposed to.

Apparently, with grid drive and no plate voltage, but with DC continuity from the plate to ground, enough electrons are accelerated past the grid to strike the plate and create measurable plate current and cause a presence of rf in the tank circuit when it is tuned to resonance, regardless of the neutralising condenser setting.  I seem to recall that this wasn't a problem with the push-pull triode final; maybe the feed-through from each tube is of the proper polarity to cancel the other, but I always physically disconnect the +HV lead anyway.

Another way to use the grid current meter to check neutralisation is with the transmitter working into a normal load and the +HV lead connected but the plate supply turned off. Adjust grid tuning exactly to resonance as indicated by maximum grid current. Now turn on the +HV, load up the transmitter, and tune the final tank capacitor through resonance as you normally would for tune-up. Now carefully observe grid current as you rock the plate tuning slightly to one side of resonance and then to the other. If neutralisation is spot on, the grid current will peak just exactly as the plate current dips.  If it is not right on, the grid current peak will be offset from the plate current dip, and if the neutralisation is very far off, the grid current will not peak as the plate tuning is adjusted but simply increase on one side of the plate current dip and decrease on the other.  That's another method I use as a quick check from time to time or right after swapping out the final tubes, if I don't think a full-fledged neutralisation procedure is called for.

This meter is very simple to build and I think helps immensely. Look in any old handbook. It consists of a 1mA ammeter, a diode and a coil if I remember correctly. This seemed to work well.

You can also use a cheap Hammy Hambone SWR/wattmeter for neutralising indicator.  Just switch it to the lowest power range, and set the meter for maximum sensitivity in the forward power position. It seems to work more accurately with some kind of resistive load (a couple of 2w carbon composition resistors will do) with a total resistance in the vicinity of 50-100 ohms (not critical).  I have an old Mirage meter, and at maximum sensitivity I can get a good indication at less than 250 mW of power into a 50-ohm load. I start with a lower sensitivity setting and work down with the neutralisation null as the sensitivity is steadily increased. I then use the grid current meter tests to verify results.

With any final, don't expect an absolutely perfect null. Just get it as low as you can.

[KC2ZFA]

Another way to use the grid current meter to check neutralisation is with the transmitter working into a normal load and the +HV lead connected but the plate supply turned off. Adjust grid tuning exactly to resonance as indicated by maximum grid current. Now turn on the +HV, load up the transmitter, and tune the final tank capacitor through resonance as you normally would for tune-up. Now carefully observe grid current as you rock the plate tuning slightly to one side of resonance and then to the other. If neutralisation is spot on, the grid current will peak just exactly as the plate current dips.  If it is not right on, the grid current peak will be offset from the plate current dip, and if the neutralisation is very far off, the grid current will not peak as the plate tuning is adjusted but simply increase on one side of the plate current dip and decrease on the other.

This is great info. Thanks !

Now I have to suspend my stereoscopic vision so I can see each meter with each eye.

[WU2D]

Forward or backwards - it works - it is a bridge!

[KM1H]

For low power tubes such as the hybrid KW rigs I turn off the screens with the rear panel switch and then use an insulated tool on the cap for a minimum signal on another receiver. Ive checked that with a spectrum analyzer and its dead nuts on.

Not having bothered with PP since I was in HS (250TH's and 810's) the Class C single ended or parallel HF tube amps, triode or multigrid) are done with all voltages off, all PS still connected and feed in a sig gen and watch the spectrum analyzer; a scope would also be OK as well as an outboard receiver.

Since most of my triodes are 6M conversions of various GG 572B amps I use a wire between the tubes and trim until the feedthru is minimum and 35-40dB of isolation is easily attainable. Stock GG is about 15-20dB and the circuit is often squirrely.

Carl

[w1vtp]

Yup, Joe.  That's the way I remember it.

[KC2ZFA]

Thanks Bill, that's good to know.

I found some pics on the web of the Millen 90881 amplifier which uses the same tubes and layout and it looks like the neutralizing capacitors are also pretty tight there too.

Next I'll put 1KV on the plates and see what happens.

Peter,
I have a backwards method that may serve as a cross check.  In cleaning and preparing my Gross transmitter, I wanted some way to get the neutralizing caps in the ball park.   So, I looked up the reference internal capacitance of the 35T tube, then put my capacitance meter across the cap and adjusted it to the same number.  

I haven't fired up the radio yet so have no idea how far off I may be, but your picture looks comparable to how mine are set.

Bill