Screen Modulated 4-65A's

[KI4YAN]

I am looking at a no-mod-iron transmitter idea; one that I'm sure has been brought up before. Heck, I'm sure it's got a name, even, but it was a pretty intense thinking session.

Anyway, it's a pair of tetrodes, screen modulated, tied in parallel. Between the plates there is a 90* phase shift network, and the two tetrodes are grid driven 90* out of phase.

Tetrode 1 has the screen biased such that it's in saturation for all levels of positive modulation, but negative modulation works as normal.

Tetrode 2 has the screen biased such that all levels of positive modulation increase RF output, but with no or negative modulation, the tube is "cut off".

Now, when they're re-combined in the plate circuit, the phasing network lets it all work AND should keep the plate voltage pretty constant, so I can run the tetrodes at full CW scrote, not having to back the voltage down to account for the modulation peaks.

Anyone know what I'm rambling about? Is this done by any manufacturer? I think it's pretty similar to the Doherty amplifier; but I haven't seen one done with tetrodes with this kind of screen modulation. Heck, I am not all too familiar with screen modulation to start with!

I do have a pair of matching MOT's here, and will be providing the final plate voltage with them. I've wound another 40 turns on the primary, and with a 20uF cap on the primary side, they're not too bad, power-factor wise. Still hunting up some diodes for a bridge rectifier to get about 2200-2500V.

[w4bfs]

rf wise, seems to be a tetrode twist on the triode Doherty ... the 90 degree phase shifters at rf power are difficult and single frequency

Timtron has a non-Doherty notion for a screen mod rig ....maybe we can entice him to disclose .... also super modulation method

the mots usually do not have adequate insulation to lift the grounded end of the secondary to use a fw bridge ... are yours made differently ?    congrats on finding a matching pair

[KI4YAN]

I've got six matching ones; and using two in a center tapped configuration solves the winding issues.

I've also got one that has two HV taps, and is grounded to the core...Interesting as it is, it provides 2200VAC and 1150VAC on the two taps, unloaded.

If not, just run it half-wave and use a choke and a few big filter caps. It's less than ideal, but the only real issue is expense on the filter caps. 35uF 5000V PIO caps are not that expensive if you know where to look these days, and bigger values are out there too!

I did some research...CE built these for commercial stations; it's a Sainton amplifier. K9QI has a page on them, showing how he did it using a converted police base station.

To generate the quadrature clock; the drive side is easy, just go digital. Not sure about the plate side...I think an L-network between the two plates could do it, and variable caps could tune the 90* phase shift. QSY between bands might be a slow move, although I wouldn't think it'd be worse than a unit using plug-in coils.

[steve_qix]

I'm trying to remember - I think the screen modulated arrangement is Terman-Woodyard and Doherty is a linear - one or the other.

Another system you could try is outphasing modulation.  Back in 1973, I built an outphasing transmitter and used it on the air for about a year.  Mine used sweep tubes in the output.

With outphasing, you have 2 phase modulated rf amplifiers using 2 PI networks, and a common loading capacitor.  There is a 135 degree angle between the RF amplifiers at carrier.  So, for 100% positive modulation, the phase angle drops to 90 degrees.  For 100% negative, the phase angle increases to 180 degrees.  The phase shift modulation takes place in earlier stages and both sides are modulated equally.

There are variations of this system using a push-pull output.

[pa0ast]

Do You may be mean the amplifase priciple ?  http://www.rossrevenge.co.uk/tx/ampli.htm
Anton

[kb3ouk]

Ampliphase is RCA's name for it, outphasing modulation is the technical term for it.

Back to the original subject, here's a page on what I think you are describing (Continental Electronics made broadcast rigs that had a Doherty final that used tetrodes, that page makes mention of that) http://k9qi.com/page11.html and here is a schematic for such a transmitter (well, a simplified schematic) http://k9qi.com/image/317csch.jpg

shelby

[KI4YAN]

Yes, that schematic is what I am thinking of. There is anther based on a Motorola Micor transmitter using an 829B to screen modulate two 3CXsomethingorothers on that page as well, demonstrating the use of a Johnson counter to generate the quadrature signals needed to drive the finals. The downside is I need the LO to generate 4X the frequency I want to transmit on; as the counter will divide by four to produce the phases.

Generating the 0* phase and the 90* phase isn't too difficult, but what gets me is the plate network. It has to maintain 90* phase shift between the carrier and peak tubes.

Ampliphase is NOT the same system; as this does not vary the phase to generate the envelope. This is efficiency modulation; same as normal screen grid modulation, but the plate efficiency remains high at all times, not just on the positive modulation peaks.

I guess this particular system is not really designed for relatively rapid frequency changes; both the plate tank and the interplate phasing network would have to be re-tuned each time, plus the local oscillator and any kind of LC/RC phase network. The Johnson counter wouldn't care if the LO changed; though...bonus for that particular method, I guess.

If this idea isn't going to be usable, I can always keep hunting for a big hunk of unobtanium to make a modulation transformer from. I have an old, possibly no good, 120W multi-match transformer, but I have no data on it other than winding resistance and somewhere I have a chart I made of the voltage ratios between windings. It is not very big. I also have a UTC S-19 30W multimatch modulation transformer, it also seems small for its ratings. I have wound and can wind custom transformers, or I could just have one wound. I have a few contacts in the transformer winding industry.

It would require a bigger, more complex power supply, plus modulator tubes, plus another chassis, but I can build a modulator to do the job. A quad of 807's can produce 225W of audio without undue stress, that's enough for 450W plate input to the final of a triode. I'd have to figure out how to get the screens modulated as well, for tetrode outputs.

[steve_qix]

If you decide to go plate modulation, and don't want to use a transformer, you can do pulse width modulation.  This is very practical - I have done it with tubes in at least 4 different transmitters over the years.

Something to think about

[KI4YAN]

If I go plate modulated, I'll very likely use a transformer. I was trying to avoid it due to availability; but it's really not that big of a deal. The screen-modulation path was mostly just for the neat factor and I thought I could cram it all into one partitioned rack chassis; but I did some arranging tonight and there isn't any good way I could pack it all in there without going 90% solid state.

So, I'm back to start.

[W3GMS]

If you decide to go plate modulation, and don't want to use a transformer, you can do pulse width modulation.  This is very practical - I have done it with tubes in at least 4 different transmitters over the years.

Something to think about

Definitely the way to go!

Joe, GMS

[ka1tdq]

I built a plate modulated 4-65A transmitter a few years ago and I've attached a picture.  The output power was only 15 watts with B+ at 500 volts, but it did a pretty good job.

Jon
KA1TDQ

[W3GMS]

I built a plate modulated 4-65A transmitter a few years ago and I've attached a picture.  The output power was only 15 watts with B+ at 500 volts, but it did a pretty good job.

Jon
KA1TDQ

Very nicely constructed Jon.  Even though the tube was loafing along, its still fun to put all the components together and make them play together!  Like all projects, I am sure you have learned something from each and every one that you have done. 

Joe, W3GMS

[KD6VXI]

This thread comes right on time.  I am hopefully picking up a Henry 4-65 based driver out of a 3000D this weekend.  I wanted to scream modulate it with a mosfet based plate modulator I found on here a while back.....   For smallish tubes.

Any ideas and suggestions welcome.  It's an ISM 13mhz job currently,  so 20 will be it's first home.   I also have a 80-10 tank including bandswitch for a 4cx250 based amp parted out for later,  if if this proves fruitful.

--Shane

[KI4YAN]

Picked up an original 1970's rack to put the whole thing in, so I'm slightly less space-bound than I was. It's being stripped of the semi-rusty DuKane blue-grey and I am going to paint it back Black Wrinkle.

I thought about silver hammertone, too, but I already have two cans of black wrinkle paint on the shelf; part of being into cars. I always have Black Wrinkle and flat black paint on hand; and a few bright flashy colors of acrylic enamel.

Doing some more research, a quad of 807's will do 240 watts at 750V, set up as AB2 triodes, with the screen tied to the grid with a 20k resistor. Drive goes to the screen, needs 250 or so volts of grid drive. Will draw a bit of grid current...250mA...

That same quad of 807's wants to see a plate-to-plate load of about 3.6K; while my pair of 4-65A's will (tenatively) produce a plate load of around 12.5K as I will probably drive them in parallel, at 2500V and 100mA each, 200mA total. We'll see.

I borrowed the hi-pot tester from work and started checking the MOT's, so far every single one of them will take 10kV from either end of the secondary winding, I don't see any reason why i can't bridge rectify the secondary safely. I have some Kapton tape that I'll be slipping under the winding pack to provide even more core insulation, just to be safer. I'll probably use the big one with the tapped HV winding, but I will test each one for primary current and power factor and I'll use the biggest, heaviest, best unit out of the stack for the plate supply. Whatever voltage it happens to churn out will be just fine.

[IN3IEX]

I like the 4/65a plate modulated with transformer and with high current sweep tubes pp class ab1.

 http://amfone.net/Amforum/index.php?topic=30021.0

4/65a. (QB3/200) the best for 100W TX.

Giorgio

This thread comes right on time.  I am hopefully picking up a Henry 4-65 based driver out of a 3000D this weekend.  I wanted to scream modulate it with a mosfet based plate modulator I found on here a while back.....   For smallish tubes.

Any ideas and suggestions welcome.  It's an ISM 13mhz job currently,  so 20 will be it's first home.   I also have a 80-10 tank including bandswitch for a 4cx250 based amp parted out for later,  if if this proves fruitful.

--Shane

[KI4YAN]

Doing some reading on Pi-network output circuits while I wait on the paint stripper to do its thing, and I suddenly realize that I may not have a tank cap that can handle the B+ AND the modulation peaks!

If shunt-feeding the output tank, will the air variable capacitors I have hack it? They are marked E.F. Johnson, 0.125", 5KV, 30pf x2, 60pf x2, and 200pf x2. The 200pf cap is as big as a loaf of bread, the other are much more compact.

I have a sack of doorknob capacitors around that are leftovers from my N2 laser; 590pF at 30kV. I intend to use two of them in parallel for the blocking capacitor; doesn't seem like that'll be any issue.

I am looking toward the Pi-network output for being able to match into just about anything, I assume link coupled outputs can do the same, just not as "neatly".

[KA2DZT]

1/4" spacing is good for 9KV peak.  I use one in my 813 plate modulated rig.  Plate voltage is around 1500 volts.  The cap will handle higher plate voltage, maybe 1800 volts with no problems.

Your cap is 1/8" spacing, so probably 1200-1500 max plate voltage will be OK.  If it isn't OK, you'll be the first to know

The two blocking caps you're using should be fine, I use two very similar caps.

Fred

[KI4YAN]

Being shunt fed, though, the plate voltage isn't developed across the tank capacitor...only the RF voltage. I intend to run between 2000 and 2500V on the finals; obviously if my capacitors can't handle that I'll need to get bigger ones.

[KA2DZT]

Being shunt fed, though, the plate voltage isn't developed across the tank capacitor...only the RF voltage. I intend to run between 2000 and 2500V on the finals; obviously if my capacitors can't handle that I'll need to get bigger ones.

Check in one of the Handbooks, there are some charts on what the voltage rating should be for the tuning cap in a plate modulated Pi network tank circuit.  I generally figure the peak voltage rating for the cap should be about 4 X plate voltage.

Fred

[KI4YAN]

After going through the whole stack of MOTs, none of them are really going to be suitable if I run them as a full wave bridge, capacitor input filter. All will produce about 3000V DC unloaded, with 122V on the primary+extra turns. Even the one with a "Hi" and a "Lo" tap on the secondary winding would churn out 3100V and 3000V...

That's higher than I need or want. The 4-65A's can take it, sure, but not for plate modulation...I'd be 100% committed to CW or screen modulation. Be right at the max limits too, another thing I'd prefer to avoid.

HOWEVER, if I was to run them as a full-wave bridge, choke input filter; (better load regulation, to boot!) I would end up with between 1800VDC and 1900V DC. This, I can work with. I'll give up 60W of output power, though, if I go by the datasheets. These things seem to be real touchy on plate voltage; they ramp up output power quickly with a small change in plate voltage. That's not a terribly big deal, and I figure I can stack a few MOT secondary winding packs on one core to make a usable filter choke.

[W9ZSL]

Are you still looking for power supply parts?  I was going to build a 300 watt AM rig but scaled down.  Meanwhile I have a NOS Thordarson 115 V primary rated at 1800 VDC under load at 200 ma.  I even have the original wooden box.  I will eventually list it here.  I too am looking at putting together a rig using a 4-65A with 1250 VDC on the plate modulated with a pair of either 809s or 807's.  Mike - W9ZSL

[KI4YAN]

Thanks; but I have an ample stash of transformer laminations in a large variety of sizes. Transformer winding is not much fun; but I have done it many times and probably will do it many more.

Besides, it's good to get a refresher course in calculus every now and again!

[KI4YAN]

Digging through the tube bin I found an Amperex 6155...the industrial version of the 4-125a. I will use this as the final amp and move the 4-65A's into modulator duty.

I have enough lams to build two 500VA transformers, or one 1KVA transformer. I have a plastic bobbin for the 1KVA model, so that is an easy choice.

The power supply should deliver 2000VDC at 500mA continuous, to supply both the modulator and the final amplifier, as well as 350V 90mA regulated for the modulator screens, -120VDC for final amplifier protective bias, -85VDC at 30mA for modulator bias, and 350VDC for the speech amplifier and exciter.

2000VDC at 500mA happens to be 1KVA, so that's settled. The 350V supply will be wound on a separate transformer along with the bias voltage supplies, and there will be 6.3VAC, 6.0VAC, and 5.0VAC supplies for filaments, each on their own transformer. Regulation will be Tektronix style, regulating the -120VDC rail and stacking the -85VDC, +150VDC, and +350VDC supplies on it to achieve regulated, tracking power supplies for all the "low voltage" supplies.