Returning to the am high level modulation

[AMLOVER]

Hi to all again,

I am thinking seriusly to return to my youth passion, the am high level modulation after a long journey to linears and PWMs.
There are remaining two main problems especially when big tetrodes are supposed to be used.
The first is the screen modulation which can be done from plate unmodulated voltage due to a huge resistor or from a separate dc supply and a reactor.
Did somebody here ever used 2 separate transformers or even one with two separate windings but connected in series in order to make the plate high voltage and then to supply the proper unmodulated screen voltage from the lower winding or the first transformer?
I provide a scheme to make it more clear.
This, if it works, could be a solution to avoid the huge energy consumptor resistor, to get the proper screen voltage in relation with the plate voltage swinging and in the same time to avoid the extra dc supply.

The question is if in such a case the screen is modulated efficiently and if using after that the screen reactor would add some help or it would be a meaningless addition?

The second main problem is how to avoid the back emf of the heising reactor in case we wish to have a very fast PTT and how to protect the modulator in case the transmitter load is lost for any reason. As far i have some thoughts on this problem i think i should open another topic later in order to share them with all of you.

Stefano

[w3jn]

You can see how Collins did it with the 30K.  Screen of the 4-125 is tickled by a tertiary winding on the modulation transformer.

[K9MB]

Seems like Gates also had a tertiary winding that they modulated the pair of 807s they used, in later models.

An alternative to using tje screen dropping resistor is to regulated the screen voltage with a shunt regulator capable of handing the screen current and feeding it through a 10-12H choke. The choke will allow the natural interaction of the screen  and plate to modulate the transmitter effectively.
Seems like Tom K1JJ did that on one 0f his transmitters, at least. Tom is an excellent source for this old school plate modulation system because he has built a number of them and has a ton of experience.

Edit:
Read this publication for all you need to know. See page 11.
I have captured diagram from page 11 below:
73, K9MB

https://worldradiohistory.com/Archive-Catalogs/EIMAC/Eimac-1951-Care-and-Feeding-of-Power-Tetrodes-1953.pdf

[AMLOVER]

Thank you both for the provided schemes.

My thinking is not to use complicated tertiary windings, not to use extra screen supply and screen reactor if possible and not to have a huge resistor from the  very high unmodulated voltage 4Kv+ to 500-600v.
I have never seen something like that I have described and I wished to know if somebody else has ever tried to get unmodulated plate voltage for the screen grid from a lower plate voltage level in the high voltage supply by using two transformers in series and in such a case if he managed to get sufficient screen modulation in relationship to the plate modulation.
The way I describe here (feeding the screen grid from a lower plate high voltage point) doesn't also need any screen protection sircuit as far the screen supply voltage comes from the plate supply voltage. If this way the screen modulates as it does when it is fed from the top of the plate high voltage then this could be a solution to many difficulties.
Let's see what others are thinking on the subject and of course Tom who has gone very deep in the high level am.
Thank you.

Stefano

[K9MB]

I am having difficulty seeing how that circuit can connect audio energy from the modulation transformer to the screen…
The tertiary mod tranny winding provides that audio and the dropping resistor does as well, since the resistor is connected to the plate voltage in the plate side of the mod reactor where the mod Tran output is applied via a capacitor.
The choke method isolates the screen so that audio is transferred internally via the swings in plate voltage during modulation because those variations cause instantaneous changes in the screen circuit.
Your series connection of transformers for adding together to form total voltage and tapping at the too if the lower voltage transformer- I do not see what it accomplishes.
Shutdown circuits to protect screens are theee in Eimacs publications, but they have nothing to do with modulation that I can see at present.
Maybe I am missing something?
73, Mike

[AMLOVER]

Mike,

you said 'since the resistor is connected to the plate voltage in the plate side of the mod reactor where the mod Tran output is applied via a capacitor'.
The most preferable way as is described in the books, is to connect the dropping resistor on the unmodulated side of the modulator or the reactor - where the hv supply is coming - and not on the plate side. There are of course few other comments on how to share modulated and unmodulated supply by connecting two resistors, one on the hv supply side of the modulator or reactor getting 60% of the voltage and the other on the plate side getting the rest 40%.
I am wondering if getting the screen supply from a lower plate voltage point the screen grid would be modulated in phase with plate as it happens when we get the screen voltage through a huge dropping resistor from the high voltage point.
This would serve to make very easy the procedure plate/screen high level am.
If this way is not however enough in order to help the screen grid to be modulated then why not to add the screen reactor.
At least this way we don't need a separate screen supply and the screen protection sircuit.
I was hoping that somebody had used it in the past and let us know the results.

Stefano

[W3SLK]

Probably the cheapest and hence mostly seen is the dropping resistor(s) from the plate to the screen. However, some tubes don't like to play well with this configuration. I was once considering making a 500 watt rig from a pair of 4CX250B's. I was advised to use an audio choke in lieu of the resistor for modulating the screen. I think there was someone on here that advocated the use of a secondary on an audio output transformer with the input leads open. Go with whatever works for your particular case. My HN-500, (a pair of 813's modulated by a pair of 810's) uses the resistors and I got good reports with it.

[K9MB]

Probably the cheapest and hence mostly seen is the dropping resistor(s) from the plate to the screen. However, some tubes don't like to play well with this configuration. I was once considering making a 500 watt rig from a pair of 4CX250B's. I was advised to use an audio choke in lieu of the resistor for modulating the screen. I think there was someone on here that advocated the use of a secondary on an audio output transformer with the input leads open. Go with whatever works for your particular case. My HN-500, (a pair of 813's modulated by a pair of 810's) uses the resistors and I got good reports with it.

I built a 4CX250B up for 2 meters and ran it class C but did not modulate it.
The thing about that tube and the 4cx300A and others is that it is not uncommon to see negative screen current in some operating modes like AB1 and only 20 mA for class C.
Also, the grid and screen structure is very fragile compared to those tough 813s, so regulated screen voltage is essential I think. I used two vr150 reg tubes for my screen regulators in my single 250B tube amp on 2meters.
That audio transformer primary is several henries, no doubt, so it makes a high quality low current audio choke and perfect in that application.
I had an 829B on 2meter AM modulated by an Eico 730 and of course used a resistor to power the screen and apply modulation as you did on the 813s.
Only downside is the extra heat and drain on the big B+ supply, but that seldom is a deal killer and it is a very simple solution. MB

[AMLOVER]

I could of course work all the usual ways to give screen grid what it needs but my question is if it could be done with a way which could solve lot of problems.
Some of them have to do with dropping resistor huge dissipation, others are the safety of the screen grid and the easiness of PTT.
Few of us who are using high level am have very late PTT and this is one of the problems we could solve with the way i propose.
My question is if with the proposed screen supply we can get the same percentage of audio as we get when we supply the screen voltage from the total plate high voltage supply by using a huge dissipation dropping resistor. In case by providing screen voltage as i propose in my schematic the screen grid will modulate as it should then this is the best way to supply voltage to the screen in relation with cost, easiness of PTT and safety of modulation parts including the tubes.

Stefano

 

[K9MB]

I could of course work all the usual ways to give screen grid what it needs but my question is if it could be done with a way which could solve lot of problems.
Some of them have to do with dropping resistor huge dissipation, others are the safety of the screen grid and the easiness of PTT.
Few of us who are using high level am have very late PTT and this is one of the problems we could solve with the way i propose.
My question is if with the proposed screen supply we can get the same percentage of audio as we get when we supply the screen voltage from the total plate high voltage supply by using a huge dissipation dropping resistor. In case by providing screen voltage as i propose in my schematic the screen grid will modulate as it should then this is the best way to supply voltage to the screen in relation with cost, easiness of PTT and safety of modulation parts including the tubes.

Stefano

 

I would love it if you could explain how you are going to get modulation on the screen grid?

[Opcom]

I could of course work all the usual ways to give screen grid what it needs but my question is if it could be done with a way which could solve lot of problems.
Some of them have to do with dropping resistor huge dissipation, others are the safety of the screen grid and the easiness of PTT.
Few of us who are using high level am have very late PTT and this is one of the problems we could solve with the way i propose.
My question is if with the proposed screen supply we can get the same percentage of audio as we get when we supply the screen voltage from the total plate high voltage supply by using a huge dissipation dropping resistor. In case by providing screen voltage as i propose in my schematic the screen grid will modulate as it should then this is the best way to supply voltage to the screen in relation with cost, easiness of PTT and safety of modulation parts including the tubes.

Stefano

What is meant by "very late PTT"? I probably know but don't call it that.

Separate transformers for HV supply and screen supply, stacked? Yes something like that was done in the Collins 18S-4 transceiver. It used dynamotors for those DC supply voltages which amount to the same thing as stacked power supplies, and its modulation transformer has a plate and screen winding.

It's not stated what tubes you have in mind. I would like to have some example from which to make suggestions that would better answer the question. Even some random tube would help.

Looking at your schematic, R and R do not give values and some of that drawing is unconventional as to what is crossing over and what is connected. I assume the triangular items are bridge rectifiers and that connections do not use a dot. I'm not complaining but it's not quite clear to me.

R and R look connected in the middle and go through a 50 Ohm pot and relay to the screen. The problem with having audio voltage there is that there are also a lot of filter capacitors that would bypass the audio voltage to ground or to a filtered voltage.

It is not clear where the plate voltage is modulated from, or how the screen is supposed to receive audio voltage to modulate it.

It would be good to have a more detailed schematic if you can manage it. Attached is a PNG file that contains popular symbols that are easy to use in paint and other drawing programs. I used these a lot before using LTspice and creating lots of tubes and other symbols for it using its own editor. (no models, just for drawing)

As far as huge resistors, That power dissipation is all relative. What one person considers huge dissipation, other person may see as normal or trivial. How much power are you willing to burn or what is the space limitation for large resistor?

Protecting from reactor transients has a few options. spark gaps, or HV relays that close across the winding, or snubber, but these all depend on the voltage and current and rise times etc.

The PDF attachment in the next post is real examples of modulating a 4-1000 screen with a choke and resistor, an example of protecting 4CX250 screens with only two 25W resistors in an AB1 amp, which could be done similarly in a class C amp, and a three-4X150 example for modulating the screens of the tubes in parallel with simple adjustment for improving linearity. However, this is possibly not what answers your questions.

But forgetting all of that, the simplest thing is a screen dropping resistor from the modulated HV, next simple is a series screen choke (8H, but depends on screen current and voltage) supplied by a power supply appropriate to screen voltage, then next is the usual extra winding, and lastly, a cathode follower which is more complicated and driven by the modulated plate voltage and supplies a carefully regulated screen voltage.

All the common ways of doing this need some consideration for protection of the screen grids. The delicacy of the screen is different across many tubes and so are the power consumption requirements which are what drive the power ratings of any resistors in the circuit.

[Opcom]

screen voltage schemes

[K9MB]

Nice post, Patrick. I think you covered it well. MB

[W7TFO]

Push-pull plate mod or you're just kidding yourselves about power output...

73DG

[K9MB]

Push-pull plate mod or you're just kidding yourselves about power output...

73DG

And make that a pair of 833s- Triodes don’t need no stinkin screen supplies, but that is just my “neutralized” opinion…😉😂
Hey TFO, I bow to your AM knowledge always. 😎😁

[W7TFO]

Hi Mike, long time no hear!

833A's are a great tube, too bad they look like a skinned rabbit to me.

849's and 851's (aka hen's teeth) are lot more elegant.

But I digress...

73DG

[AMLOVER]

Patrick, Mike and forumers,

According to eimac and for Class C telephony (plate modulated) as is described here https://frank.pocnet.net/sheets/088/4/4-65A.pdf
'Modulation voltage for the screen is easily obtained by supplying voltage via a series dropping resistor from the unmodulated plate supply or .... or ....' ect ect ect.

What is meant by "very late PTT"? I probably know but don't call it that.

Very slow PTT means that in a high level plate modulated transmitter the sequencer normally has to deal for 3-5 steps with 100-500ms time difference from each step to the next one.

Separate transformers for HV supply and screen supply, stacked? Yes something like that was done in the
Collins 18S-4 transceiver. It used dynamotors for those DC supply voltages which amount to the same thing as stacked power supplies, and its modulation transformer has a plate and screen winding.
Stacking 2 transformers would help to get from the lower one the screen voltage but in the same time stacking the next one in order to 'build' the plate voltage from the same mains for safety reasons. This is a very safe way for screen grid in case there is a flash over or missing the plate HV because a rectifier diode blew up during transmitting.

It's not stated what tubes you have in mind. I would like to have some example from which to make suggestions that would better answer the question. Even some random tube would help.

The tubes can be any combination of parallel powerful tetrodes like 4-1000As or Qb5/1750s that are very demanding in screen grid current.

As far as huge resistors, That power dissipation is all relative. What one person considers huge dissipation, other person may see as normal or trivial. How much power are you willing to burn or what is the space limitation for large resistor?

In the following attached scheme and using the way to obtain modulation to screen as per eimac, from the unmodulated part of plate supply, you may see that for screens demanding 600vdc/480mA the dropping resistor from plate HV 4kvdc to 600vdc should be 6.8K/1700w. This resistor will dissipate during transmitting almost more than a transmitter in full legal limit.... My propose is without the need of the dropping resistor just stacking one more transformer to get a much lower voltage for the screen but in relation with the one of the plate voltage in such a way that if something goes wrong all the voltages trip away in the same time. In case we need the extra screen reactor to feed audio the screen, we always can add it.

Looking at your schematic, R and R do not give values and some of that drawing is unconventional as to what is crossing over and what is connected. I assume the triangular items are bridge rectifiers and that connections do not use a dot. I'm not complaining but it's not quite clear to me.

The resistors R and R are the bleeders for discharging the filter capacitors and the diode sign is a rectifier bridge. Excuse me not to use LTSpice or similar programs as far I am an old fashion person. I will describe and paint as better as I can though.

It is not clear where the plate voltage is modulated from, or how the screen is supposed to receive audio voltage to modulate it.

I would ask exactly the same thing eimac when she proposes to feed the screen voltage from the unmodulated plate voltage through a dropping resistor.
If it is the dropping resistor itself that acts as a 'reactor' because of the long winding then I prefer to add in my screen feeding way the reactor choke as far I don't need to use a dropping resistor as far the voltage from the point I get it, is almost as I need.

Protecting from reactor transients has a few options. spark gaps, or HV relays that close across the winding, or snubber, but these all depend on the voltage and current and rise times etc.


Protection of irons especially during on and off is another beast that we'll talk about in future topic as far I have some more questions.

Triodes are for easy lovers .... the cow boys love rodeo ..... 

Stefano

[K9MB]

Hi Mike, long time no hear!

833A's are a great tube, too bad they look like a skinned rabbit to me.

849's and 851's (aka hen's teeth) are lot more elegant.

But I digress...

73DG

Hi Dennis, always great to see you post. Look for you when I read QRZ for wisdom and humor and “BS”😉😂.
Always entertaining and enlightening…
I always loved the candelabra elegance of a 304TL at full dissipation like a triple star!. The 833 is an ugly child, specially the new carbon anode 833C ones. No sun lamp there. 😉
I like that 851 envelope. If an 807 is named  for it’s shape, then this one must be a 2 quart Lager! 😎

[K9MB]

Patrick, Mike and forumers,

According to eimac and for Class C telephony (plate modulated) as is described here https://frank.pocnet.net/sheets/088/4/4-65A.pdf
'Modulation voltage for the screen is easily obtained by supplying voltage via a series dropping resistor from the unmodulated plate supply or .... or ....' ect ect ect.

What is meant by "very late PTT"? I probably know but don't call it that.

Very slow PTT means that in a high level plate modulated transmitter the sequencer normally has to deal for 3-5 steps with 100-500ms time difference from each step to the next one.

Separate transformers for HV supply and screen supply, stacked? Yes something like that was done in the
Collins 18S-4 transceiver. It used dynamotors for those DC supply voltages which amount to the same thing as stacked power supplies, and its modulation transformer has a plate and screen winding.
Stacking 2 transformers would help to get from the lower one the screen voltage but in the same time stacking the next one in order to 'build' the plate voltage from the same mains for safety reasons. This is a very safe way for screen grid in case there is a flash over or missing the plate HV because a rectifier diode blew up during transmitting.

It's not stated what tubes you have in mind. I would like to have some example from which to make suggestions that would better answer the question. Even some random tube would help.

The tubes can be any combination of parallel powerful tetrodes like 4-1000As or Qb5/1750s that are very demanding in screen grid current.

As far as huge resistors, That power dissipation is all relative. What one person considers huge dissipation, other person may see as normal or trivial. How much power are you willing to burn or what is the space limitation for large resistor?

In the following attached scheme and using the way to obtain modulation to screen as per eimac, from the unmodulated part of plate supply, you may see that for screens demanding 600vdc/480mA the dropping resistor from plate HV 4kvdc to 600vdc should be 6.8K/1700w. This resistor will dissipate during transmitting almost more than a transmitter in full legal limit.... My propose is without the need of the dropping resistor just stacking one more transformer to get a much lower voltage for the screen but in relation with the one of the plate voltage in such a way that if something goes wrong all the voltages trip away in the same time. In case we need the extra screen reactor to feed audio the screen, we always can add it.

Looking at your schematic, R and R do not give values and some of that drawing is unconventional as to what is crossing over and what is connected. I assume the triangular items are bridge rectifiers and that connections do not use a dot. I'm not complaining but it's not quite clear to me.

The resistors R and R are the bleeders for discharging the filter capacitors and the diode sign is a rectifier bridge. Excuse me not to use LTSpice or similar programs as far I am an old fashion person. I will describe and paint as better as I can though.

It is not clear where the plate voltage is modulated from, or how the screen is supposed to receive audio voltage to modulate it.

I would ask exactly the same thing eimac when she proposes to feed the screen voltage from the unmodulated plate voltage through a dropping resistor.
If it is the dropping resistor itself that acts as a 'reactor' because of the long winding then I prefer to add in my screen feeding way the reactor choke as far I don't need to use a dropping resistor as far the voltage from the point I get it, is almost as I need.

Protecting from reactor transients has a few options. spark gaps, or HV relays that close across the winding, or snubber, but these all depend on the voltage and current and rise times etc.


Protection of irons especially during on and off is another beast that we'll talk about in future topic as far I have some more questions.

Triodes are for easy lovers .... the cow boys love rodeo .....  

Stefano

The Eimac note specifies a resistor that is high enough in value to drop the plate voltage down to a value compatible with the screen circuit. That high resistance will automatically make the screen see a high impedance for audio. After that, the interaction of plate voltage swings during modulation will be mirrored in the screen circuit because the screen is coupled to the plate, so the audio swings are there, just like they are for the regulated screen voltage and screen reactor circuit.
OpCom answered the issues of flux collapse discharge by spark gaps and the screen control circuits he posted and protection circuits abound. i favor fixed cutoff biasing for both the modulators and amplifier tubes and sequencing is essential.
I can see no advantage in the series transformer supply and some disadvantages.
Maybe I am missing something, and there are multiple ways to modulate a transmitter, and DG knows 100 times more about them than I do, but to me, it looks like riding this horse may get you nowhere.😉
MB

[W7TFO]

Screen mod attracts those that have a hard time finding mod iron today.  It's a thing...

73DG

[K9MB]

Screen mod attracts those that have a hard time finding mod iron today.  It's a thing...

73DG

Similar to metaphorically to liposuction to get rid of that beer gut we all despise, but it does do something, I guess because it takes something away and gives part of it back and they call it 100% modulation. Reminds me of the plans that politicians offer every election year…😉😂😂

[KD1SH]

I'm an admitted newbie here compared to some of the AMfone luminaries, but if I might humbly make a few points:

   With a tetrode PA, modulating the screen voltage isn't optional; it's absolutely required, and the audio waveform applied to the screen must be faithfully in phase with that applied to the plate, not just close. Taking the screen supply from the secondary of the modulation transformer, or from a tertiary winding of the modulation transformer, or by "self modulating" the screen, insures this proper phase fidelity. Any other method which implies  deriving this modulation from some point prior to the modulation transformer - if that is your intent - would undoubtedly result in modulating the screen out of phase with the plate.
   If your circuit supplies rectified, filtered, and regulated DC screen voltage, at the proper amplitude, a choke of the appropriate inductance would allow the screen to self-modulate, which is the only way in which I can see that your circuit would work. The modulation applied to the plate causes changes in the screen current, thus imposing the audio waveform on the screen, while the choke acts as a barrier to prevent that audio waveform from being shunted to ground in the screen DC supply. In effect, the choke allows the screen to be tied to a "stiff" DC level, while allowing it to "float" at audio frequencies to wherever the plate wants to pull it, perfectly in phase with the plate.
   We very often find ourselves inclined to "reinvent the wheel", but often need to remind ourselves that the wheel has been around a very long time simply because it works so well. I think you'll find that the traditional plate modulation "wheels" will roll just fine for you.

I could of course work all the usual ways to give screen grid what it needs but my question is if it could be done with a way which could solve lot of problems.
Some of them have to do with dropping resistor huge dissipation, others are the safety of the screen grid and the easiness of PTT.
Few of us who are using high level am have very late PTT and this is one of the problems we could solve with the way i propose.
My question is if with the proposed screen supply we can get the same percentage of audio as we get when we supply the screen voltage from the total plate high voltage supply by using a huge dissipation dropping resistor. In case by providing screen voltage as i propose in my schematic the screen grid will modulate as it should then this is the best way to supply voltage to the screen in relation with cost, easiness of PTT and safety of modulation parts including the tubes.

Stefano

 

[KD1SH]

Actually, I find that the burden of high plate dissipation required for screen modulation pretty much balances the burden of finding mod iron. Six of one...half dozen of the other. Ain't no free rides - nowhere.

Screen mod attracts those that have a hard time finding mod iron today.  It's a thing...

73DG

[AMLOVER]

You are absolutely right but it is strange that eimac suggests to obtain screen voltage with a drooping resistor from the unmodulated plate high voltage. I was wondering how does audio involves the screen feeding voltage in this case? That's why I thought that getting voltage for the screen from the same source (Hv plate supply) but from a lower voltage point it should work the same way but without needing a huge dropping resistor.

If your circuit supplies rectified, filtered, and regulated DC screen voltage, at the proper amplitude, a choke of the appropriate inductance would allow the screen to self-modulate, which is the only way in which I can see that your circuit would work.

This is obviously my case and I'll try it first. I'll get my screen voltage from the first of the plate Hv stacking transformers, I'll add after a powerful pot 1-2k in series to control the voltage in respect with screen current and after I'll add the screen reactor. I think that 5H would be enough as far I'll supply 600v/500mA which is a screen Rl=1200ohm
I have found in DigiKey the Hammond 193Q which is 10H/500ma and 193P which is 5H/500mA 1000V working.
Their prices are high with euro to dollar 1:1 but as you have said Ain't no free rides - nowhere.
Thank you for the very understandable explanation about the relation of plate and screen phase and the screen reactor role in the high level modulation.
I need 2 more sockets SK-510 in order to complete the construction and I don't see any available at ebay.

Thank you.

Stefano

[KD6VXI]

You can run a screen supply and then couple modulator current via a cap from the plate supply.

There are /is a thread(s) on the board about doing just this.

If you run a tube like a 4CX250 then you need a good amount of bleed R to take care of reverse screen current.

I bleed 15 or more watts in bleeder for each 250b.

--Shane
WP2ASS / ex KD6VXI