813 rig update and Modulation Transformer hookup

[KK7UV]

I’ve made some progress on my ‘Tesla’ 813 rig project.  As of my Oct. 2 post I was looking for some tips on getting my grid input tuning network working.  Ed K8DI spotted my error in a missing capacitor.  With that in place I got it working and delivering grid drive.  After playing around with turn ratios on the toroids and padder values, I now have decent grid drive.  With 2000v and 250mA on the plates, about 5 watts of drive from the exciter, and the grid tuned for an SWR dip on the exciter line, I get 380 – 390 watts output on all three bands 160, 75, and 40.  I am pleased with that efficiency!

Now I am ready to hook up the modulator.   The transformer is a Thordarson T11M77 ‘multi-match’.  The info sheet from the amfone files has the hookup info for various turn ratios,impedances,etc.  I found four lashups that I’d like to experiment with, having turn ratios of 1.15, 1.25. 1.32, or 1.49.
Now to my question, regarding the secondary side.  Using the 1.49 ratio by example below, does it matter which of terminals 2 or 6 connects to the capacitor and which connects to the finals end of the reactor?

[KC2ZFA]

250mA is the max secondary current on series hookups when the unit was new. Given the age of the transformer I would only consider parallel hookups.

[K9MB]

250mA is the max secondary current on series hookups when the unit was new. Given the age of the transformer I would only consider parallel hookups.

Agreed: nothing more useless than an open mod tranny winding and they are not making T11M77 any more.
Or consider getting a reactor and shunt feed it.
40 Henry choke would work.
Just need the extra space.

[WD5JKO]

Looking at the schematic, the DC current flows through an external reactor, and not the mod transformer. Therefore the 250ma current concern is more about what the reactor can handle.
As to the position of the capacitor, as shown will work, but the entire DC B+ voltage is across that capacitor. Putting the cap on the top side (Top Pin 2 to Cap, then to Finals, and low side pin 6 to HVB+), the capacitor mainly sees the AC voltage across it at the low frequency end of the audio when the reactance of that capacitor rises, hence, you might be able to use a lower rated DC rating on the capacitor? Comments?

Jim
Wd5JKO

[K8DI]

Looking at the schematic, the DC current flows through an external reactor, and not the mod transformer. Therefore the 250ma current concern is more about what the reactor can handle.
As to the position of the capacitor, as shown will work, but the entire DC B+ voltage is across that capacitor. Putting the cap on the top side (Top Pin 2 to Cap, then to Finals, and low side pin 6 to HVB+), the capacitor mainly sees the AC voltage across it at the low frequency end of the audio when the reactance of that capacitor rises, hence, you might be able to use a lower rated DC rating on the capacitor? Comments?

Jim
Wd5JKO

When not modulating, the voltage across the cap is the same in either position. You still need the same voltage rating, the mod transformer has virtually no DC drop across it. The question becomes more about cap contruction..is the case connected to one of the two terminals? if so, make it the ground and put it at the low end of the transformer...or you'll have to mount it on standoffs and have the case float at B+, which is less safe.

Ed

[WD5JKO]

When not modulating, the voltage across the cap is the same in either position. You still need the same voltage rating, the mod transformer has virtually no DC drop across it. The question becomes more about cap contruction..is the case connected to one of the two terminals? if so, make it the ground and put it at the low end of the transformer...or you'll have to mount it on standoffs and have the case float at B+, which is less safe.
Ed
[/quote]

Not true, as I stated, "Top Pin 2 to Cap, then cap to Finals and reactor, and low side pin 6 to HVB+"

You don't want the mod transformer secondary grounded, low side pin 2 needs to be tied to HVPS, not ground.

That means the HVPS filter capacitor needs to be large in MFD such that the modulator does not modulate the HVPS, say 25 MFD or more if you are going to modulate below 100hz.

The series Cap in the modified Heising circuit can be as low as 4uf, and if the cap is on the high side, it does not see the B+, so it can be de-rated in voltage rating.

Jim

[K9MB]

When not modulating, the voltage across the cap is the same in either position. You still need the same voltage rating, the mod transformer has virtually no DC drop across it. The question becomes more about cap contruction..is the case connected to one of the two terminals? if so, make it the ground and put it at the low end of the transformer...or you'll have to mount it on standoffs and have the case float at B+, which is less safe.
Ed

Not true, as I stated, "Top Pin 2 to Cap, then cap to Finals and reactor, and low side pin 6 to HVB+"

You don't want the mod transformer secondary grounded, low side pin 2 needs to be tied to HVPS, not ground.

That means the HVPS filter capacitor needs to be large in MFD such that the modulator does not modulate the HVPS, say 25 MFD or more if you are going to modulate below 100hz.

The series Cap in the modified Heising circuit can be as low as 4uf, and if the cap is on the high side, it does not see the B+, so it can be de-rated in voltage rating.

Jim


[/quote]

I agree, Jim.
 I confess that I did not examine the schematic, so failed to see the reactor 😳, but the cap should be on the top (pin 2).
Also, grounding the winding puts it at a high potential dc difference with the CT primary. Better if the windings see the same DC potential, so they are only stressed by audio swings.
Shorting to a case less likely, IMO anyway.
Some worry about which end of the secondary goes where because of capacitance I vaguely remember.
Also, with the reactor, spark gaps are a good idea. Collapsing fields yield very high transients.
K4KYV has posted a wealth of information on this and I think it is still in the AMfone Archives…
RCA Ham news can help with modulator designs to plot load lines.
The problem is that their are no data curves on 813s strapped as triodes, that I have seen.



http://www.amwindow.org/tech/pdf/art13shuntfed.pdf

Here is a short piece by Don Chester. He does ground the mod tranny secondary.

http://www.amwindow.org/tech/htm/modheising.htm

I love this simple method of designing a modulator. You do need the load curves, however.

http://www.amwindow.org/tech/pdf/rcamoddesign.pdf


Another classic byK5PRO

http://www.amwindow.org/tech/pdf/rcaclassB.pdf

http://amfone.net/Amforum/index.php?topic=26630.0

[w9jsw]

I’ve made some progress on my ‘Tesla’ 813 rig project.  As of my Oct. 2 post I was looking for some tips on getting my grid input tuning network working.  Ed K8DI spotted my error in a missing capacitor.  With that in place I got it working and delivering grid drive.  After playing around with turn ratios on the toroids and padder values, I now have decent grid drive.  With 2000v and 250mA on the plates, about 5 watts of drive from the exciter, and the grid tuned for an SWR dip on the exciter line, I get 380 – 390 watts output on all three bands 160, 75, and 40.  I am pleased with that efficiency!

Can you post your RF deck schematic? I would like to see what you arrived at for the grid circuitry. Did you add any neutralization?

[K8DI]

When not modulating, the voltage across the cap is the same in either position. You still need the same voltage rating, the mod transformer has virtually no DC drop across it. The question becomes more about cap contruction..is the case connected to one of the two terminals? if so, make it the ground and put it at the low end of the transformer...or you'll have to mount it on standoffs and have the case float at B+, which is less safe.
Ed

Not true, as I stated, "Top Pin 2 to Cap, then cap to Finals and reactor, and low side pin 6 to HVB+"

You don't want the mod transformer secondary grounded, low side pin 2 needs to be tied to HVPS, not ground.

That means the HVPS filter capacitor needs to be large in MFD such that the modulator does not modulate the HVPS, say 25 MFD or more if you are going to modulate below 100hz.

The series Cap in the modified Heising circuit can be as low as 4uf, and if the cap is on the high side, it does not see the B+, so it can be de-rated in voltage rating.

Jim

I'm sorry, I didn't draw out your proposed circuit before responding, so didn't note the difference from the OP's modulation scheme.

  What you are suggesting is a shut fed modulator, figure #2 in your first linked article. I am more familiar with the OP's posted schematic, which is the same arrangement used by Gates, Bauer, RCA, etc., in their broadcast transmitters, with the capacitor from one end of the secondary to ground. Doing it that way keeps the cap case safely grounded as well as keeping the dc differential between the mod tx primary and secondary down.  The cap's terminal to terminal voltage would not change if it were moved to the top of the mod tx, but it would increase differential voltages within the cap  to case and between the windings of the mod tx.

I don't see any advantage to the shunt fed modulator, rather, the opposite.  As the piece you linked to from Don K4KYV points out, the OP's/broadcast tx method has lower distortion, and with the cap at the bottom, less voltage stress.

Ed

[KK7UV]

See attachment John - the full schematic but I think you can see all the details in the grid circuit.  I may make some changes to the modulation transformer lash-up in the next revision - depending on the comments in this thread.

I did not use any neutralization since I did not think it might be needed for 160 thru 40 operation based on prior discussions on this topic.
I also did not use any grid swamping at this time.  I can add either/both if necessary.  Before I do any modulation I am going to run some tests with the spectrum analyzer and scope to see how pure the carrier is.

Further details around the modulator:  the dc blocking/audio coupling capacitor is a big G.E. 4uF, 4000v and the case is not common to either terminal.   The reactor is from a BC-1 (I think) and is 50H 700mA CCS.   PS Filter cap is 47uF 4500v.

My original post as to which modulation transformer terminals (2 and 6) go to the capacitor and which to the modul. B+ end of the reactor was based on some notes I had from an old K4KYV comment in which he said the polarity of those secondaries in a multi-match transformer MUST be correctly wired.  I do not want to get the polarity backwards and blow something up.

[K1JJ]

Hi Steve,

A beautiful looking rig as well as a nice, organized schematic. This build brings the old 813s X 813s rig into the next phase of evolution.

Did you build and test the GFZ MOSFET audio driver yet?

It will be interesting to see if the lack of neutralization and no-grid-loading is stable enough on 40M.  My bet is you can get away without the neutralization, but may need some light grid loading, like 3K or so on 40M and above .  But ya never know...  your RF layout may be real FB.  :-)

Keep posing pics and results!

Tom, K1JJ

[KK7UV]

Hi Tom.  Thanks!   and yes, with some help from Frank I got the GFZ amp working last winter - on the bench anyhow with an audio generator and simulated loads.

With regard to my modulation transformer polarity question, I found a post of yours from a while back on how to determine that by shorting the primary leads, and driving the secondary with a little audio thru a resistor approximating the plate Z.  I've done similar with smaller unknown 'universal' audio transformers but didn't know why it didn't occur to me on a big mod. tranny.  duh.

Steve

[w9jsw]

I will be interested to see how you hook up the numbered pins on the mod trans. I have the same one waiting to be hooked up. Currently my 813 is running on BC-1T Iron. I am building a 4-400 modded by 3-500s that will be transplanted on the BC-1T Iron. I then have a 11m77 that i will combine with a BC-610 HV trans and choke to run the 813 rig.

I really like your innovative grid circuit. Very nice.

John

[KK7UV]

Here are my T11M77 test results on the bench this morning.  I used the hook-up from the info sheet for a 1.49 turns ratio (step-down).

In the first test (left side of the diagram), terminals 6 and 12 were at common ground with the generator.  This produced the expected voltage step-down, but with a 180 degree voltage phase shift between primary and secondary.

In the second test (right side of the diagram), terminals 2 and 12 were at common ground with the generator.  This also produced the expected voltage step-down, but no voltage phase shift between primary and secondary.

Does this result tell me where to connect terminals 2 and 6 in the diagram in my original post?  Is it just a matter of selecting in-phase or out-of-phase?   

[K8DI]

Is it just a matter of selecting in-phase or out-of-phase?   

There's only two reasons to choose:

1) you design negative feedback from the secondary side into your audio chain, because the number of inversions in the speech amp, modulator and transformer matters to make it negative vs. positive feedback.

2) your audio chain includes asymmetrical limiting so you need to keep the larger positive audio peaks modulating in the positive direction vs. negative direction.

Otherwise, I am of the opinion that it can't possibly matter. Keep in mind, your hookup for testing isn't how it runs -- the primary is push-pull, so by definition, the two ends are out of phase with each other. the polarity reference would be the center tap.

Ed

[w9jsw]

Note that the GFZ driver board will use negative feedback and his schematic shows a feedback tap marked NFB from the plate of one of the mod tubes back to the board.

[KK7UV]

I need some more guidance here.

I now have the GFZ audio driver hooked up to the modulator and ready to try some tone tests.  But I'd like to understand the following....

1.  Where should I set the bias level in the GFZ for an initial test to make sure the idling plate current in the modulator tubes is at a safe low level?
2.  Does a more positive bias increase or decrease the idling plate current?  (John - I notice you used +8v to get 20mA idling current.  Is that per tube or for both tubes?)
3.  Where do I measure this bias?  Is it from the MOSFET gates to ground?   Or at the tube grids to ground?   

[w9jsw]

I put an ammeter in the grid line to set my bias. Setting by voltage was not accurate. I tried to set it with the mod plate current meter but that was also not accurate. Set each tube to 20ma.

Be sure to hook up the feedback circuit or use a temporary feedback or you will run into trouble with gain.

Tom may have better advice.

[w9jsw]

Steve,

You got that rig on the air yet?

John

[KK7UV]

Not on the air just yet John - only the dummy load.

Everything seemed to behave OK with the 'local' feedback in the GFZ amp and bias at zero yielding 24mA of idling modulator plate current.  Ran a few tones up to full modulation and they looked good.

Today I removed the local feedback and installed the 2Meg NFB string. 

Next step is to use your method of biasing one side to 20mA, then biasing the other side until 40mA total.

Then I'll attempt to get the 6dB NFB dialed in and do some thorough tone tests.

I hope that's the logical order anyway.