Higher anode voltage possible with screen grid modulation?

[PA1JO]

Hello colleagues, I am looking for your advice for the following. I built an AM transmitter with a 6146B tube. Combination of plate and screen grid modulation using a Hammond 1630SEA transformer (SE 30 watts).

It works well. With an audio tone of 700Hz and 100% modulation, I receive a 39 V signal (78V PP) in my 40-ohm antenna. Not bad I thought.

Now I want to double the output of this transmitter by connecting a second 6146 tube in parallel. The challenge now is the modulation. Hammond does not have modulation transformers of 50 or 60 watts, and I do not see any other manufacturers offering such a transformer.

Screen grid modulation therefore seems to me to be an alternative, but screen grid modulation produces much less antenna signal. So, what I gain with my second tube I lose in the modulation ☹.

Now I thought. Let me increase the anode voltage considerably. From 600V now to at least 1000V later. That could significantly improve the output, the antenna signal.

But,
The Datasheet of the 6146B tube indicates a maximum voltage of 750V (icas). In my current transmitter, the anode with plate modulation also experiences peaks of over 1000V. That doesn't seem to be a problem.

Does anybody have experience with this? Please advise, how do you view the 1000V anode voltage for this tube in combination with screen grid modulation. Let’s say a quiescent bias voltage of 125V on the screen grid (modulation 250V to 0 volts).

Thank you in advance for your input on this question 

[K8DI]

My first thought..  78vpp, 39vp, 27vrms is way less than I’d expect.  I’d expect 50 volts rms, 70 volts peak, 140 volts peak to peak of just carrier. At 100% modulation, v peak or peak to peak would be double that.

Are you sure about your measurements? Unless they are very far off, you need to find out why it’s working so poorly before you start adding tubes to it.

Regarding voltages and ratings, the datasheet value of 600 volts dc on the plate when modulated takes into account the modulation…so the added voltage during each audio cycle is perfectly normal and accounted for.

Ed

[KD1SH]

  You might want to check out Hammond's 1642sea, rated for 75 watts of audio and 300ma on the primary. Like the 1628sea and 1630sea, it's probably overbuilt and the numbers are conservative.
A good discussion about it here: http://amfone.net/Amforum/index.php?topic=47350.0

  Screen modulation can deliver superb audio quality—like broadcast quality—but keep in mind that the penalty isn't simply lower power output; it's lower efficiency, in the area of 35%, which means that for the same plate voltage you'll need more current. But, your plate voltage won't be doubling with 100% modulation as it does with plate modulation, so you'll probably be able to get away with pushing the plate voltage envelope a bit more.

  RCA, actually, specifies 750 volts ICAS for linear operation or class C FM or oscillator applications, but not for plate modulated AM, for which they specify 600 volts ICAS.

Now I want to double the output of this transmitter by connecting a second 6146 tube in parallel. The challenge now is the modulation. Hammond does not have modulation transformers of 50 or 60 watts, and I do not see any other manufacturers offering such a transformer.

Screen grid modulation therefore seems to me to be an alternative, but screen grid modulation produces much less antenna signal. So, what I gain with my second tube I lose in the modulation ☹.

[PA1JO]

Thank you Ed and William,

Ed, you are right about the power output but we are talking about a local station in a small Town. Operating 18 ours a day. We don't want to push the tubes to the limit.
But then, the town is expanding and to have quality sound in the suburps we need more power.

That is why I was thinking about a extra tube in the PA.

William, You got me thinking. I had seen the 1642SE but thought 5000 ohm inductance is way to much. I gather for two 6146B tubes I need maybe 1500 ohm. But I gave it e second thought. Maybe if I load the transformer with 4 Ohms on the 16 ohm tap the secundary will decline to 1250 ohm.

What do you guys think?       

[KD1SH]

  Well, I built a single 6146B transmitter—somewhere around 1800 ohms—and modulated it with great success using Hammond's 1628sea with a 5000 ohm impedance. There's an efficiency loss, obviously, but we're talking such low power that it's not a huge hit.

  I was assuming amateur radio type service, but with your broadcast application, full CCS duty, I'd be cautious about pushing the envelope too far. RCA says 600 volts on the plate for ICAS, but only 480 volts in CCS. Over-build the power supply and fan cool everything. Still, even if you run at the conservative 480 volts, a pair of 6146B's should be good for around 80 watts to the antenna. I don't know what your country's legal stipulations might be—or even if we're talking about a legal station at all—but in a small town 80 watts doesn't sound too shabby.

Thank you Ed and William,

William, You got me thinking. I had seen the 1642SE but thought 5000 ohm inductance is way to much. I gather for two 6146B tubes I need maybe 1500 ohm. But I gave it e second thought. Maybe if I load the transformer with 4 Ohms on the 16 ohm tap the secundary will decline to 1250 ohm.

What do you guys think?       

[DMOD]

Thank you in advance for your input on this question  

You're going to kill the 6146's for a Vp of 1kV. I would advise against it.

What is your antenna- half wave dipole or vertical (how tall if vertical)?
What is your frequency of operation?

A few notes:
1)Your present matching network design may not be efficient-is it a Pi-net? If so calculated at 1MHz a Pi-net would show the plate impedance of 2400 ohms for a single 6146;
C1shunt (plate Tune) = 750pF
L1series = 38uH
C2shunt (Plate Loading) = 4000pF;
Plate modulation impedance would be about 4800 ohms,
2) According to my calculations for two 6146's in parallel, the plate impedance would be 1200 ohms for two 6146 in parallel at 600V 250mA, calculated at 1MHz a Pi-net would show
C1shunt (plate Tune) = 1400pF
L1series = 22uH
C2shunt (Plate Loading) = 6500pF
Plate Modulation impedance would be about 2400 ohms for two 6146 in parallel at 600V 250mA.

Phil-AC0OB

[DMOD]

Thank you in advance for your input on this question 

Screen grid modulation alone for two 6146's is a simple matter with a 12AX7 and a 6DE7 modulation system.

Phil-AC0OB

[PA1JO]

Thank you William & Phil, You are a great help  :-)

We are considering the 1642 now althouw it is a lot of Iron and a bit expensive.

And we forget about the idea of increasing the voltage.

Yes the transmitter is legal with a licence and we may go to max 100W PEP.

The frequency is 1395 kHz and the antenne is a 5 meter vertical with a huge capacitive hat.

The pi filter calculations we apply look verry much the same as yours Phil.

Last doubt.
We will try Screenmodulation too. But not with a 6DE7 as used in the Heathkit DX-60 because we do'nt like NOS tubes.

Does someone has a schematic wiht tubes that are still produces these days?

[DMOD]

Thank you William & Phil, You are a great help  :-)

We are considering the 1642 now althouw it is a lot of Iron and a bit expensive.

And we forget about the idea of increasing the voltage.

Yes the transmitter is legal with a licence and we may go to max 100W PEP.

The frequency is 1395 kHz and the antenne is a 5 meter vertical with a huge capacitive hat.

The pi filter calculations we apply look verry much the same as yours Phil.

Last doubt.
We will try Screenmodulation too. But not with a 6DE7 as used in the Heathkit DX-60 because we do'nt like NOS tubes.

Does someone has a schematic wiht tubes that are still produces these days?

I don't know what you mean by not wanting NOS tubes. What is a 1642?

Calculations show that with two 6146's in parallel, SGM, you should be able obtain at least a resting 50W carrier. According to current mythology, that equates to 200W PEP at about 100% modulation.

I think you asked for a SGM schematic, so here is what I use.

Phil AC0OB

[PA1JO]

Thank you Phil,

WE have bad experience with NOS tubes not functioning according to specs and not living very long.

The 1642SE trafo can be bought brand new e.g. Mouser.com sells them. 

Also the 6146B tubes can be bougt brand new. They are Chinees and  so far they do there job all right.

The power you mention is right. We experianced max 50 watts rms in the antenne with 475V anode power supply and even more than 75 watts with 625V power supply.

[DMOD]

Thank you Phil,

WE have bad experience with NOS tubes not functioning according to specs and not living very long.

The 1642SE trafo can be bought brand new e.g. Mouser.com sells them. 

Also the 6146B tubes can be bougt brand new. They are Chinees and  so far they do there job all right.

The power you mention is right. We experianced max 50 watts rms in the antenne with 475V anode power supply and even more than 75 watts with 625V power supply.

I have had good luck with all 6DE7 NOS tubes I have ever purchased on the open market. You won't find anyone making them today. I have never had to replace a 6DE7 in any equipment I have upgraded.

12AX7A and many power tubes can be purchased today that were recently made in China but given the choice and if the cost isn't prohibitive I'll take a NOS tube anytime.

Phil - AC0OB

[K8DI]

The 6146 is much more likely to be discontinued than any of the common guitar/hifi tubes, and maybe you should be building based on a pair of EL34 or 6550 tubes. Less commonly used by hams running at frequencies ten or twenty times higher than yours, but will work just fine at AM broadcast frequencies, and will likely stay in production much longer.

The Chinese tubes you’re getting are substandard.  I think you’re misapplying the term NOS due to language differences, though.  NOS does not mean there’s no current production, it means an item that is new but made long ago.  I think what you mean to say is you do not want to use parts that aren’t being made any more.   A tube can still be being made today, so you have new vs. NOS, choices for the same tube, which is the case for the 6146.  NOS 6146 tubes easily out perform and outlast newly made Chinese 6146 tubes because those older made tubes were simply made better.  You would do well to try some NOS 6146 military spec tubes and see for yourself. 

Ed

[KD1SH]

  I think he meant to say 1624, an 807 derivative as I remember. I would imagine that choice would certainly lock him into NOS territory, as I doubt that tube is in current production, even in China.
  I think, Phil, that you might be a bit optimistic on the output of a pair of screen modulated 6146's. You've very likely spent more time with these than I have, but between messing with my QIX modified DX-60B's, and pine-boarding some screen modulated rigs of my own, I've found that an efficiency of around 35% is pretty much the norm, assuming the output is loaded so as not to squash the modulation peaks. With my rigs, I've consistently found that loading the output more lightly in order to improve efficiency very quickly results in distortion of the mod peaks.
  Playing with the numbers for a 50 watt output, even assuming an extremely optimistic efficiency of 40%, I come up with a plate dissipation of 75 watts, 5 watts over the 70 watt dissipation that a pair of 6146B's should be safely capable of. I absolutely love screen modulation, but I've come to have a grudging respect for the efficiency boogeyman.

I don't know what you mean by not wanting NOS tubes. What is a 1642?

Calculations show that with two 6146's in parallel, SGM, you should be able obtain at least a resting 50W carrier. According to current mythology, that equates to 200W PEP at about 100% modulation.

I think you asked for a SGM schematic, so here is what I use.

Phil AC0OB

[w8khk]

I believe the "1642" to which the OP is referring might be the Hammond output transformer, or perhaps an amplifier circuit provided by Hammond for their transformer.  Several references point in that direction. 

I do not think the 1642 reference is a tube type, or a misquote of 1624 or 1625.  The RCA 1624 is an ancient beam power tube, similar in appearance to the 807, but it has a directly-heated cathode (filament) at 2.5 volts, 2 amperes.

The 807 derivative from the '40s was the seven-pin 1625, with the companion 1626 triode oscillator and the 1629 magic eye tube, also probably not related to the OP's reference.

[KD1SH]

  Yes, a more attentive reread makes me certain you're right, Rick.

I believe the "1642" to which the OP is referring might be the Hammond output transformer, or perhaps an amplifier circuit provided by Hammond for their transformer.  Several references point in that direction. 

I do not think the 1642 reference is a tube type, or a misquote of 1624 or 1625.  The RCA 1624 is an ancient beam power tube, similar in appearance to the 807, but it has a directly-heated cathode (filament) at 2.5 volts, 2 amperes.

The 807 derivative from the '40s was the seven-pin 1625, with the companion 1626 triode oscillator and the 1629 magic eye tube, also probably not related to the OP's reference.

[PA1JO]

Thank you Guys,

You convinced me a higher voltage on the plate is no good idea. I'm going to try some other things.

Thanks very much for your attention and advices.

Koos