Modulator plate current - major change in the circuit

[W4RFM]

So I am to the point where I can hear something and do some testing to my modulator under construction. (To Refresh: 6SJ7 pre, 6N7 buffer, 2x 6L6GC, driving thru a UTC S-9 transformer to the grids of 2x 811A's.) The 811's are not inserted, I put the appropriate load on the secondary of the driver transformer and
applied music to the second stage (6N7).  Sounds pretty good, a little high endish, but anyway, I inserted a 200 mil meter in the B+ to the 6L6's.  It rests at about 100 mA, and when I turn up the volume, the current goes down to about 80 mA! What am I missing?
 

[W1EUJ]

Any chance of getting a schematic? Too much to assume. When you debug, you must be careful of assumptions.

[AB2EZ]

I agree with David's comment... it would be very helpful to see the schematic.

Possible causes (in the order I would check them out)

1. Does the average (DC) screen voltage on the 6L6's go down when you increase the volume. If it does, since screen current can only flow in one direction... as you draw average screen current through whatever screen voltage supply resistors may be in the circuit, the average screen voltage will drop. Solutions: place a larger capacitor from the screen to ground on each 6L6; and/or use a regulated screen supply. [Using a larger capacitor from screen-to-ground on each 6L6 will reduce this problem for voice modulation (which has occasional voice peaks), but it won't reduce the effect if you apply sine wave modulation (which has a peak every half cycle)]

2. Is there a separate, negative grid bias supply (i.e. the 6L6 cathodes are grounded). If so, does the average (DC) grid-to cathode bias go more negative when you increase the volume? If it does, since grid current can only flow in one direction...  when the applied audio is large enough to cause grid current to flow in one 6L6 or the other (audio peaks), this will cause a voltage drop across the grid bias supply's source resistance (unless you are providing the bias through a center-tapped driver transformer, and using a grid bias supply with a low source resistance). In addition, if you are not using a driver transformer, the 6L6 grid current will charge up the coupling capacitors between the previous "phase splitter" stage and the grids of the 6L6's. In that case, the effect (reduced 6L6 plate current) will be much larger when you apply a sine wave (which has an audio peak every half cycle) versus when you are speaking (only occasional audio peaks... and therefore the charge on the coupling capacitors will bleed off through the grid resistors on the 6L6's fast enough to keep the coupling capacitors from accumulating too much charge).

3. Are you using cathode biasing for the 6L6's. If so, since the audio portion of the cathode current in each 6L6 flows through the cathode biasing resistor on one half cycle or the other (and always in the same direction)... the average cathode current increases when you turn up the volume. To fix this problem, replace the cathode biasing resistor with a Zener diode (its voltage drop will not change when the current through it increases). Use a resistor from your B+ (or low B+) supply to the place where the 6L6 cathodes and the Zener diode come together... to supply a small amount of current through the Zener diode (i.e., 1 mA). This small amount of current will keep the Zener diode at its full breakdown voltage even if you turn the modulator off during standby. That way, there won't be a surge of modulator current when you turn it back on. [i.e., since there is some capacitance, there would be no bias on the cathode for a brief instant, unless you provide this keep-alive current).

Good luck

Stu

[W7XXX]

First I would put the meter to read the cathode current. How are you biasing the 6L6's? What voltage is on the screens? I might suggest triode connecting those 6L6's and eliminating the need for grid bias and screen voltage. I use two 1614's (version of 6L6) in class B (triode connected) to drive the 813's also class B and the 1614's show no current at rest and the 813's rest at 50ma.

It sounds like something is not right in the biasing. Is the 6N7 driver (not buffer) transformer coupled to the 6L6 grids?

[w4bfs]

take a look at output waveforms ... are they distorted compared to input waveforms? .... if so, may be time to check out driver transformer ... resistance checks will not tell enough .... better to test each winding for self inductance ... serious differences indicate either a wiring error or internal problems ... let us know what you find ...73 ...John

[W4RFM]

I am trying to find a way to get the schmatic on the forum, in the mean time, I changed a few things last night after I posted, and today it seems to be working properly.  I do not have a separate negative bias supply.  I originally used 6B4G's here, (The 1956 Bill Orr General Purpose Modulator design), and I think my tubes are bad, so I plugged in the 6L6's and gave them a screen suplpy of 210 volts, and grounded the cathodes. I also moved my monitor pick up point to the "F" connection and one tap of the UTC S-9 driver transformer, ahhh hi-fi at last.  Today the 6L6GC's are running 301 volts on the plates, 210 on the screens, and drawing a combined plate current of 120mA at "normal" volume. According to what I can decipher from the N7JP tube data base, this should be about 12 watts which is exactly what I want, as I should need less to drive the 811's. I THINK!

[w4bfs]

nice going Bob ... the 6L6 has too much internal resistance to make a good class B driver without using copious amounts of negative feedback for distortion reduction .... this can have its own share of problems  if there is enuff phase shift in the audio passband to change the fb to + .... do you have any 6080 or 6AS7 ?  ... I think your driver should have enuff swing to drive them with much less fb required ....73 .. John

[W4RFM]

Thanks Beefus, (by the way I usefd to have that WHO album, it was great.)

I think I have it under control now, I guess I need to see if in fact the grids are devising some kind of bias, but it sure sounds good, maybe a little distortion on the extreme lows, of course the Collins ART-13 Mod transformer will not pass that anyway, but I want to make it sound as good as possible.  I have about 30 henrys/250 mA to use as a mod reactor, and a 8 mfd at 3000 vdc oil filled on the way, so it's shaping up to be a decent rig.
It's kinda funny, I have took care of factory broadcast AM's for the last 39 years, but designing/building your own is a whole different deal, and it shows me how much I  don't  know...

[WA1GFZ]

You could zener regulate the 6l6 bias so it doesn't creap negative. A 10 watt stud zener or a string of diodes will hold the bais steady. Sounds like drive to the 6l6s is charging your bias more negative reducing plate current. A shunt zener regulator will bleed off any extra voltage. Also you could try loading the bias supply if it can handle the current. a zener will be the most stable.

[Opcom]

Please consider bypassing the cathode bias resistor with a cap with reactance at lowest expected frequency of Xc<(0.1 * (Ek/Ik))  or  the phool formula of R*C(uF)=25000. Many people do not bother to bypass zener diodes, only cathode resistors. An unbypassed cathode resistor is good only for class A push-pull work. In AB it will cause distortion.

[W4RFM]

So you can see what we are talking about. I am still trying to find a way to put the schematic on the forum.
Bob

[Ed/KB1HYS]

Very nice work om...

[W3RSW]

Yes it sure looks like a fine modulator.  Really neat, squared off wiring too. Say, you might experience some shorting under the driver transformer; that chassis cut-out looks a little close to the terminals. 

[W1EUJ]

Can you swap the small grey UTC transformer for the modulation transformer? Squaring off is a nice thing, but that doesn't look right. Or perhaps turn it 90 degrees counterclockwise?

[KB3DKS]

So you can see what we are talking about. I am still trying to find a way to put the schematic on the forum.
Bob

  Bob, just take a picture of the schematic from as close as the camera will focus. Macro setting, manual mode with the flash off and strong light from an angle, The other option is to scan the schematic and just post that image. Either way, common photo editing software will allow any contrast correction or conversion to grayscale. The pix posted are excellent. Looks like a nice circuit.
 A modulator is really only a normal audio amplifier with a mod transformer instead of a speaker matching output tranny. All other circuit considerations are the same unless audio feedback from the modded rf is desired.

 Bill,
KB3DKS in 1 Land

[w1vtp]

Sigh!

811's (not 811a's) and the ART 13 mod iron -- they bring back memories.  I had an Eldico TR-1 (not the TR-1TV) that had those parts.   I was 15 at the time.  Is that ART 13 mod iron going to give you the bottom end / high end freq resp you want?  Sure is purdy tho'  A work of art.  Where'd you get those tubes anyway?

GL, and thanks for the memories, Al VTP

[Opcom]

That is very nice work there Bob! You certainly are doing a fine job!

[W3RSW]

From looking at the under chassis pix it can't be too complex.
Find anything yet Bob?

Here's my old 811A lashup used to mod the singel 813 shoebox xmitter several years ago. It also had it's own power supply.  Three 500uf @ 450v Motorola electrolytics in series are underneath the chassis.

To reduce the complexity I just drove it with a RS 25 watt Pa amp into a backwards af output transformer (center tapped to the 811 grids, of couse.)

[W4RFM]

Please forgive the crudeness, but I found a way to post it based on your suggestions. Most of the changes have been noted, except the plan originally called for triodes in the driver position (6B4G's), the 6L6's shown actually have screens connected thru resistors to the B+ with 210 volts on them.The schem does not show that.
Bob.

[Opcom]

No reason I see that should not work. I notice no 811 grid winding center tap connection in the picture of the chassis, but I realize that has not been completed.

Differently than the original schematic, you are using:

301V on the 6L6 plates
210V on the 6L6 screens
Cathode bias resistor of ? ohms - -I notice the diagram says cathode ties to ground, so you must get bias from somewhere.

I see the resistor there in the chassis for the screen dropping. Does this also drop the 6N7 plate voltage to a low value?

The S-9 driver xfmr is rated 20W. Fixed bias or increased screen voltage on the 6L6's will get you closer. Can you raise the screen voltage (using its own resistor of necessary) and use fixed bias and still stay class A?  Push-pull into 5000 ohms should get 17.5 watts.

I notice in the RCA manual, that for cathode bias, 500K to GND is the maximum recommended G1 circuit resistance so that is OK in the circuit, and it would have to be 100K for fixed bias.

Are the grid drive voltages equal from one 6L6 to the other? Some adjustment of the 18K resistance value might be in order.

I did not find the S-9 connection/impedance diagram. It would be interesting to see that and understand more. From this, I suspect either the 6N7 plate voltage being too low or being affected by 6L6 screen current variations, the 6L6 biasing & screen voltage, or the 6L6 to 811 match.

There is no negative feedback from the 6L6 plates to the 6N7 plates or from the 811 grids to the 6L6 screens. Adding it might help the frequency response and improve the driver's behavior on the nonlinear loading of the 811's class B grids. Most 6L6 drive circuits for class B grids include some negative fedback right there at the 6L6 stage.

I found these two schematics which are very similar and might help.
 - radio_handbook_10th_ed_p315.jpg
I like this one best because it uses the driver screen grids to receive negative feedback which includes the 811 grid driver transformer and also provides some loading of the driver stage. If less loading is required, there could be put a resistor such as 5-10K between C15 and C16 to the screen grid connections. The loading and feedback level can be separately adjusted with these values and those of R20 and R21.

 - radio_handbook_10th_ed_p362.jpg
just shows one way to take negative feedback from the plates of the 6L6's and put it to the grids.
Since you are not using a transformer between the 6N7 and the 6L6's, it would be as well to try a 2 meg 1 watt resistor between each 6L6 plate and same-side 6N7 plate. Have to experiment with the value to make the best compromise. As low as 1M might be usable.

anyway just some thoughts.

[N2DTS]

I myself would not use capacitor input on the power supply, choke input gives much better regulation.
I dont EVER use capacitor input on any power supply.

On the mic input circuit, 1 meg will sound poor with an unamplified D104.

If you dont want a full size hole under the driver transformer, use some washers under the mounts to clear the lugs from the chassis.

I really hate iron that is built that way!

It looks very nice so far!

Brett

[W4RFM]

I agree Brett.  You'll notice in the chassis picture, I send the sum of the diodes to the 8HY choke (different from the drawing), I forgot to note that in my comments. 
The S-9 transformer actually has the solder lugs recessed about 1/4" or so, resulting in plenty of breathing room from the B+ to ground.

I only built the pre amp stage to keep as much of the original design as possible.  In operation, broadcast quality processed and E Q'd audio at line level will appear and connect thru that little UTC Ouncer transformer, and jump in at the gain control, the mike pre amp will probably never be used.

I think a lot has changed since 1957. Also the components are made much better in many cases.
Patrick I will take some reading and reply to your questions tonight. Thanks everyone.

The schematic has been updated, and is what is actually built, I think.
Bob

[Steve - WB3HUZ]

It doesn't matter that much if the cap is big enough. With the final amp current acting as a large bleeder/constant load, a cap input supply will work just fine on an plate modulated AM rig. Save the room from the big unneeded choke for a good mod tranny or mod reactor.

I myself would not use capacitor input on the power supply, choke input gives much better regulation.
I dont EVER use capacitor input on any power supply.

On the mic input circuit, 1 meg will sound poor with an unamplified D104.

If you dont want a full size hole under the driver transformer, use some washers under the mounts to clear the lugs from the chassis.

I really hate iron that is built that way!

It looks very nice so far!

Brett

[W4RFM]

Oh, I got an outstanding 8 MFD 3KV door stop from one of my broadcast engineer buddies, to go with the 34 Henries of Thordarsen chokes I scavenged off ebay. I am going to mount those on the power supply chassis.

[W4RFM]

See Guitar Amp Modulator, if interested. Thanks