Seeking assistance with 1965 ARRL SSB Tube transmitter project?

[KC9KEP]

Hello all,

I really appreciate everyone’s help on these projects!  Part of the issue for me is “I don’t know what I don’t know”.   Sometimes it’s a real challenge to even determine what part of the circuit has an issue ..

This project is the 1965 ARRL filter sideband transmitter.  It features a McCoy type filter and sideband crystals.

http://www.bignick.net/ssb/281_.pdf
http://www.bignick.net/ssb/282_.pdf
http://www.bignick.net/ssb/283_.pdf
http://www.bignick.net/ssb/284_.pdf
http://www.bignick.net/ssb/285_.pdf

Pictures of my project:

http://www.bignick.net/ssb/top_front.jpg
http://www.bignick.net/ssb/top_rear.jpg

I had much trouble just getting as far as I have.

First, I couldn’t get *any* drive to the output.  After days of researching the circuits and tracing wires, I discovered that I had installed a 47K Ohm resistor rather than the required 47 Ohm!  DOH!

Then, I had major instability in the output section.  Part of the problem was solved by shielding various power leads.  Shielding the 6DQ5 didn’t seem to help much.  (Note, white-wire is the neutralization lead.)

http://www.bignick.net/ssb/shield.pg.jpg

Then, I read that VHF oscillations (I was getting about 145 MHz) can be caused by poor wiring of the output tank.  Fat, direct wires fixed that.

http://www.bignick.net/ssb/fat_wires.jpg
http://www.bignick.net/ssb/fat-wire.jpg

Then, I could see that the 8.9985 Hz was getting into the audio pre-amp, so I moved the crystals to a more remote location.

http://www.bignick.net/ssb/crystals_moved.jpg

One problem I have not fixed is the neutralization.  When I remove the 700VDC, 27 Ohm and 5.1K Ohm cathode resistors from the 6DQ5 and adjust the neutralization capacitor for minimum output, the null occurs when the capacitor is full open.  I think the output is more stable with the neutralization capacitor disconnected.

http://www.bignick.net/ssb/neut_cap.jpg

And .. I am able to transmit to my ICOM IC-730 while connected to a dummy load.  But, the audio sounds as if one were talking into a ventilating fan .. Any ideas what this symptom is telling me?

And, probably most troubling is that the transmitter is generating very little power output.  It’s barely moving the needle on my cross needle meter.

I went back to peak the L1 and L2 coils in front and after the SSB filter.  When they are peaked, I get more carrier output (maybe 20 watts), but I should not be seeing any carrier in the output for SSB.  The ARRL book says nothing about adjusting these coils.

And .. the 6CL6 amp runs VERY hot.  I can see the tube shield blistering.  Is this normal?  The voltages look good to me?

http://www.bignick.net/ssb/hot_6CG7.jpg

Incidentally, my test set up is:

Power to transmitter via isolation transformer and Variac. 
Transmitter output to current tap, power output continues to Cantenna load.  Current tap goes to 50 Ohm load, 10 dB attenuator to scope and 10 dB attenuator to Rigol; DSA 815-TG.

http://www.bignick.net/ssb/curremt_tap.jpg
http://www.bignick.net/ssb/rigol.jpg

I know this is a lot of info but any help is appreciated as I correct one issue at a time.

73

--KC9KEP

[AC2EU]

And .. the 6CL6 amp runs VERY hot.  I can see the tube shield blistering.  Is this normal?  The voltages look good to me?

Be sure that the tube bias voltages are correct ( especially G1) and that there is no parasitic oscillation.

[KB1GMX]

Several things I'd suggest...

Use ceramic caps, lot of them in the .005uF range on all the filament pins to ground, ALL the
DC points to ground.  Use short leads on those ceramics.

I've built a lot of multi-tube gear and the filament and HV lines tend to carry stuff around.
I've also found many of the handbook designs didn't work unless I added a lot more bypassing.

That Neutralization cap looks to be to many plates (too much C).

Make sure the filament 6.3V line really is, high is bad.

Insure all the screen bias lines are both bypassed with very short leaded .001 or .005uf caps
to ground.

Make sure the screen bias on all tubes really is the correct values.

Looks like that driver got the plates red, pull the shield and see if it does when powered.

Note insure all the caps used have adequate voltage margins.  the coupling cap from the
driver to the final should be at least 1000V (driver plate is around 350 and the grid is minus
35 or more volts), a 500V part may not take that for long.

Oh, the right angle bends and all that is terminal bad at HF RF.  SHORT DIRECT LEADS.

Thats enough for the first round.

Allison

[KB1GMX]

One more thing that fat read wire in the picture with the shield..  Is that HV for the final?
If so I could not determine from the pictures, is there a cap on the topside from the feed point on the plate choke to ground.  There should be a 1000V, .005uf right there and use very short leads.  If not that will
bring RF under the deck to all the low level stuff.

Allison

[N2EY]

Everything KB1GMX says is good advice. I'll add/reinforce this:

1) The neutralizing cap is way too big (too much C). Rather than yanking plates, put a fixed capacitor of about 10 pF in series with the lead to the 6DQ5 plate circuit.

2) There should be a .005 to .01 bypass cap at the cold end of the final plate RF choke, on top of the chassis.

3) The 6CL6 should not run so hot. Check voltages and resistor values.

4) Shielding and bypassing are essential. Ground leads must be short.

73 de Jim, N2EY

[N3QE]

And .. I am able to transmit to my ICOM IC-730 while connected to a dummy load.  But, the audio sounds as if one were talking into a ventilating fan .. Any ideas what this symptom is telling me?

Choppy AF modulation brings to mind two common things: Motorboating in audio stages? Squegging in a RF stage? Scope will help a lot. Unless you have a very wide bandwidth scope, you might not see the UHF parasitic when "Squegging" but you will see the choppy audio-frequency relaxation cycle quite clearly.

6CL6's used as class A drivers will run hot. They don't last forever either, but being a TV-type they are quite common and affordable. Might check plate current with a measurement, some small shifts in screen voltage can cause large departures in plate current.

Tim.

[KC9KEP]

Thank you everyone for your help on this project.  I’m starting to feel very dumb about now .. I didn’t think that I’d put so much time into this project and not get satisfactory results by now :-(

So, here’s what I’ve done to date:

-   Added six (6) 0.005uFD caps .. one to each filament connection.
-   Added 0.004 to the feed end of the plate choke to keep RF above deck.
-   Removed plates from the neutralizing capacitor.  (I did this before reading N2EY’s message .. )

Here are some photo's .. (Please let me know if I'm doing this correctly ..)

Here’s the choke bypass .. Hope I did this right.  I only had 0.002’s so I paralleled them.
http://www.bignick.net/ssb/choke_bypass.jpg

I don’t know if this was in my last photo’s, but I used braided leads to the Neutralization cap.
http://www.bignick.net/ssb/braid_to_neut.jpg

Here’s the 6DQ5 w/added filament cap.  There already are bypass caps on pins 3 and 6 as indicated in the schematic.  Are these leads too long?
Voltages as follows:
Pin-1 Grid (-107VDC) key up, (-81VDC) Key down.
Pin-8 Screen 148VDC
Pin-3 Cathode 0.0VDC Key up; 0.288 VDC, key down; (across a 27 Ohm Cathode resistor = ~11 mA)
Here's something that I didn't know - Plate Voltage - Key up 1,066VDC, Key down; 966VDC.
Could that be causing too much gain?  I'm using a Hammond 278X transformer; 800V CT 200MADC.
http://www.bignick.net/ssb/6DQ5.jpg

Here’s the 6CL6 w/added filament cap.  The plate voltage is ~355VDC which seems high.  Cathode is about 5.56VDC.  Grid to ground is 0.012VDC so I suppose the grid is negative relative to the cathode?  I didn’t see any parasitics.  The tube manual says that pin-7 goes to cathode but ARRL has it indicated as connected to ground?  It already has a 0.01uFD bypassing its screen.
Pin-8 (screen) is 245VDC.
http://www.bignick.net/ssb/6CL6.jpg

Here’s the 6BA7 w/filament cap.  It already has a 0.001uFD cap bypassing its screen.
http://www.bignick.net/ssb/6BA7.jpg

Here’s the 6AU6 w/filament cap.  0.001uFD bypasses pin-6.
http://www.bignick.net/ssb/6AU6.jpg

Here’s the 6CG7 w/filament  bypass cap.  Not much to bypass ‘cause it’s a triode.
http://www.bignick.net/ssb/6CG7.jpg

Here’s the driver cap between 6CL6 and 6DQ5.  It’s rated at 630VDC.  I’ll need to find a larger voltage replacement.
http://www.bignick.net/ssb/driver_cap.jpg

Here’s an under the chaassis view.
http://www.bignick.net/ssb/under.jpg

Filament voltage measures 6.7 VAC with 110VAC line voltage.  Is that too high?  If so, how could I correct it?

Do you suppose that I should re-dress those fat RF leads?
The driver tube runs very hot, but the plate is not red.
Some of the tubes have redundant pins.  Should these be connected?

Note:  Instability happens with the 80 meter coils but not the 20 meter coils.  I’m not sure if this is a clue.  Also, last night I had connected my Rigol DSA815-TG to the output section with the transmitter powered down.  I injected the tracking generator at pin-6 of the 6CL6 and the input to the analyzer to the coax output.

I could easily see that I could peak and move the peak resonance for 80m.  But, there is a second strong peak at 1.5MHz as well as the desired 3.5MHz peak.
I don’t know what is causing the 1.5MHz peak or if it’s an issue or not.

Bottom line – unfortunately, no discernible change.  Looking at the output with the +700VDC, 27 Ohm and 5.1K resistors disconnected, the output is minimum when the Neutralization capacitor plates are full open.  I have yet to try and additional 10 p in series.   I’m not certain that I could get the neutralization leads any shorter or more direct that they are.  I’ve replace the old “White-Wire” with a heavy gage single conductor.

I’m out of time for now but will pick up on this again soon.

Thanks again!  I’m stuck!

73

--KC9KEP

[N2EY]

Couple of things:

1) Others have told me that the use of iron-core RF chokes in tube projects sometimes causes all sorts of odd behaviors that do not occur with air-core RFCs. In one case of a 6AG7-6DQ5 MOPA, there was absolutely no drive until the 6AG7 plate choke was changed from an iron-core to an air-core.

2) The 300 uuf cap from the cold end of the grid coil to ground is part of the neutralizing circuit. It may need to be bigger.

3) 6.7 volts heater voltage won't hurt anything.

73 de Jim, N2EY

[KB1GMX]

The driver tube runs very hot, but the plate is not red.

Measure current and check against design or tube manual.
The screen voltage may be a tad nigh or the cathode resistor is a bit too low.

Some of the tubes have redundant pins.  Should these be connected?

YES!  They are there to help lower the inductance from that element to the circuit.

Also I see those yellow (in the pictures they are yellow) caps are they
real RF rated or what?

the 6CL6 to 6DQ5 cap is adequate at 630V. 

If your seeing multiple resonances that means a choke or other circuit component is
not up to the task.  For example the plate choke may be the correct inductance but
due to construction may have unwanted resonances.  Same for other parts used.
Any of the coils in the driver and final if they deviate from the original design for
inductance or capacitance used can impact that as well.

The usual trouble shooting for something like that is sequential from input (mic)
to output getting each state working as you go.  The input to the driver (no
6cl6 or 6dq5 plugged in) should be audible at the expected output frequency
and clean. If not stop and fix that.  and by clean I mean all unwanted mixing
products suppressed and no spurious stuff moving down. then get the 6cl6
going then 6dq5.


Allison

[KC9KEP]

Thank you everyone for the help!

Last night, I discovered that removing the canisters that house the tank coils, greatly change the tuning.
This is no surprise, since these canisters are NOT aluminum.

(I knew this when I created the canisters but I had used what I could find.)  So, I'm running with
the canisters removed for the moment.

I also replaced three RF chokes.  Two of them appeared to be sintered cores.  I also replaced the
plate-choke with a NOS National 2.5 mH choke.

When disconnecting the plate voltage and two cathode resistors from the 6DQ5 output tube, I am
now able to easily adjust and null the output with the neutralization capacitor.  So, the is one BIG win for me.

At this point, the handbook says that I should check the low-level output with a receiver.

My FT-243 crystal is 5.240 MHz.  My SSB crystal is 9.0015 MHz.  (I read that when you subtract the
two frequencies, you get "sideband inversion" hence the 9.0015 MHz xtal for 80 meter operation.)

My target frequency is 3.7615 MHz.  My ICOM IC-730 is tuned to 3.7620 MHz.  (The IC-730
is not necessarily calibrate either.)

Here is a recoding of what is sounds like.  Can someone tell me is this is correct?
(This is what I mean about not knowing what I don't know :-)

http://www.bignick.net/ssb/KC9KEP_test.mp3

Thanks again!

73

--KC9KEP

[AC2EU]

Here is a recoding of what is sounds like.  Can someone tell me is this is correct?
(This is what I mean about not knowing what I don't know :-)

http://www.bignick.net/ssb/KC9KEP_test.mp3

Hard to say without doing some "divide and conquer".  Try putting a clean sine wave into the mic input and make sure that there isn't some distortion that originates in the audio stages by tracing with a scope. Work your way out from there.

[N2EY]

I also replaced three RF chokes.  Two of them appeared to be sintered cores.  I also replaced the
plate-choke with a NOS National 2.5 mH choke.

I think the 6DQ5 plate RFC was OK - it's air-core in the pictures.

(I read that when you subtract the
two frequencies, you get "sideband inversion" hence the 9.0015 MHz xtal for 80 meter operation.)

No. Here's how it works:

In any mixing system, there are three frequencies of primary concern: the input frequency, the output frequency, and the oscillator frequency.

The ONLY time you get sideband inversion is when the oscillator frequency is higher than BOTH the input and output frequencies. In this design, that doesn't happen with a 9 MHz SSB generator and 5.2 MHz oscillator, regardless of whether you choose 20 or 75 meter output.

The filter is centered on 9 MHz. Using a 9.0015 MHz xtal means the carrier is above the sideband - so the result is lower sideband. Using an 8.9985 xtal means the carrier is below the sideband - so the result is upper sideband.

There is a persistent myth that the use of a 9 MHz SSB generator and 5 MHz LO will get you LSB on 75 and USB on 20. But it's not true, because sideband inversion doesn't work that way.

73 de Jim, N2EY

[GW3OQK]

I think the 300 pf at the bottom of the grid coil should be more like 1000 to 1500 pf. Then the neutralising cap can become effective.
Best of luck.
Andrew