The AM Forum

Yeah, I know.  I'm kind of a pest but I've spent years collecting parts to build a "dream rig" and now it gets down to the most efficient combinations.  I will use a single 813.  I planned on modulating with a pair of 811A's...but wait!

Just for sheets and giggles, I have a couple 4-65A bottles.  When used in a modulator they can have some serious output and will work with a wide range of plate voltages. It takes very little to drive them.  Of course, you have to consider screen voltage as opposed to 811A's.  On the plus side a single 6V6 into a UTC S-8 driver I originally planned on building would be simpler and make a lot more sense.  This just keeps evolving.  Coming up with the screen voltage is no problem.  S-22 mod iron with the 813 PA.  

The P-P 6V6 driver described in my other thread is far more complex than this beast and it would work with both modulators.  You already commented on the driver and I changed the diagram to reflect your input. This kills two birds with one stone.  Don't use 811A's in the modulator, period.  Sometimes the obvious isn't.  

Comments?

Build both.
An 813 likes 2000 volts at 200 ma, so its handy to make a modulator that runs 2000 volts on the plates.
You also want a modulator that makes as much audio power as carrier output power of the RF deck.

A good AM rig would be a pair of 813's as they can run forever at 400 watts carrier output.

Modulators that work at 2000 volts:
813
4-125, 250,400,
4x150a/4cx250b (in AB1)
100th

4-65 likes 1800 volts in AB2 per the tube data and only does 270 watts.
A single 813 at 1800 volts will do about 275 watts and the 4-65's will do for that.

One 813 at 1800 volts and a pair of 4-65A's at either 1250 or 1500 volts is the limit.  I could use a pair of 811A's for the modulator at the same voltages but driving them would require more power.

Run the modulator on the same supply as the 813 if you can.  Your supply would need to supply about 300ma.  What plate xfmr are you using??  Use the 4-65As as they can run at 1800 volts.  811As are limited to around 1500 volts.

Running the 813 at 1800 volts and the modulator at 1250 volts you'll never make 100% modulation.

Fred

I think you could modulate an 813 at 2000 volts with an 811a modulator running  1200 to 1500 volts with the right turns ratio in the mod iron.

I do not think its the best setup, but it would work.
I like having about the same voltage on both sides of the mod transformer.

These days, the mod transformer seems to be the big limitation, a really good high power mod transformer is hard to come by and you do not want to push them too much or screw up and blow one.

I have a stanley mod transformer I got years ago, from fair radio I think, 8000 to 2800 ohms at 300 ma and 1750 working volts, marked 200 to 15,000 Hz and tried it with 811's and the 2x4x150 rf deck a few days ago.
It worked but the low frequency stuff sounded very nasty.
So its useless really.
The old Thordarson 11m77's were quite good I think, the cvm 5 was very poor on the low end response, and after trying various setups, I kind of like screen modulation.
No phase shift, and no response changes, the screen modulated rigs seem to need very little eq of the audio to sound very good.

If you are considering running a pair of 4-65A tubes as modulators, or even a quad in push-pull parallel, if you run them class AB1, no driving power is required.  This mode would allow you to eliminate the driver transformer.  This would reduce overall phase shift, and improve frequency response, while allowing more stability if you decide to use negative feedback in your design.

If you wish to pursue this approach, you might want to take a look at an article entitled "500 watts of audio from AB1" by Mandoli, ex W6HAM and Atkins, W6VX.  The article was printed on page 13 of QST for March 1948.

While the implementation used a pair of 4-250 tubes as modulators, it could easily be scaled to a pair or quad of 4-65A tubes.  It only used four receiving tubes from microphone input to the modulator grids, and no transformers.   Tube lineup was a 6J7, 6J5, followed by a pair of 6SJ7s in a phase inverter similar to the design you posted.  Hope this gives you some good ideas.  GL with your projects.

Note.... I altered the thread subject to 4-65A, not 4-64A....

A TX here uses two 4-65's in AB1 as modulators at 1500VDC with a 14K p-p load.
The modulation transformer is rated 90W.
The TX is rated 90W carrier, 100% modulation.

AB1 was how all the old tube hifi amps used to be set up.
You need to regulate the screen voltage, not hard with VR tubes.
My 4x150 modulator deck runs AB1 and sounds cleaner then the class B 811's to me.
4-65's would look great glowing behind some glass, pulsating with audio.

Hello
Why not a pair of 572b as modulators?

I actually looked at an old AM Forum thread that discusses 4-65A tubes as modulators.  I think I'll stick with the single 813 / P-P 811A combination using the P-P 6V6 driver which puts out 12 watts.  That amp was discussed on another thread recently.  I double-checked my plate transformers and can easily do 2000 volts for the 813 with 1350 for the 811A's.  If the 2 KV was rated at more than  300 MA, I could use one supply for both since the 4-65As wouldn't give me enough output at the lower voltage.

If you have 300ma to play with, the 813 at 200 ma and a modulator of 4-65's would be ok because its icas and the average current of the modulator is low.

Use the 2KV xfmr for the PA and the modulators with the 4-65s.  The modulators on average don't put that much load on the plate xfmr.  What plate xfmr do you have??  Some are much better than others.

The 811As at 1350 volts you will have a much harder time trying to make 100% modulation.

Fred

It's a Thordarson T-16P03.  It has a dual secondary.  With 120 on the primary, I've tested it with rectifiers (solid state), 12Hy choke input filter, 12MFD, and a 100 K bleeder. I get 2250 VDC on one secondary and 1840 on the other, no load.  It's 300 MA. It's 4600 and 3750 VCT.  I have both swinging and smoothing chokes plus plenty of HV caps so I could do a two-stage filter to separate the RF and Mod if necessary.  Sure would make life simpler.  I'm not sure if both the 4-65A's are good.  I know one is NOS.

Of course, I'd have to come up with a driver and wouldn't need driver iron.  I've seen a circuit for a real simple inverter used with a pair of 813's.  I don't want to use 813's in a modulator though.  I'd have to buy a pair and a new filament transformer and can't afford that right now.  I wouldn't use tubes in the screen supply.

if you decide to use the 4-65 be sure to test the filaments for normal V & I (6.3V and 4A) ... most I have tested have lost some vacuum and did not light up properly

I've heard they can get gassy.  Pun intended.  ;D

The 16 series Thordarsons are a little heavier duty than the later 19 series.  I have some 19 series xfmrs and I'm not too impress with them, not much iron in the core.  In my HB 813, 811As xmtr I use an earlier Thordarson P-6411, it's rated at 280ma, but has much more iron in the core.  The P-6411 has a tapped primary.  It's rated at 1000 DCV and 1250 DCV depending on which primary terminals used and after a two stage filter (LCLC).  Primary is 115vac, with 120 volts it makes about 1150 DCV with SS rectifiers, using the lower voltage primary terminals.  I increase the DC voltage by adding another 350VDC (from another xfmr) to the CT of the P-6411.  This brings up the plate voltage to about 1500 volts.  I use it for both the PA and the 811As.  The 811As are powered off the first choke and the 813 through the second choke of the LCLC filter.  Even though the P-6411 is only rated at 280ma, the xfmr doesn't even work up a sweat running the xmtr for hours.

You should be able to power both the 813 and the 4-65s from the 16P03.

Fred

Fred, would you recommend using a two-stage filter with the swinger followed by the smoothing?

Sure, if you have them.  You don't need a lot of filtering for the push-pull modulators, so feed that off the swinging choke and the PA off the smoothing choke.  About 15mfd on the swinger and more on the smoothing.  If you plan to key on and off the HV supply with xmit and receive the chokes help to reduce the inrush current on the supply.

Taking the mod current off the first choke helps improve voltage regulation for the PA stage.  You need chokes with a high enough voltage rating.  Also use chokes that do not have a high DC resistance, under 100 ohms for each choke would be OK.  Most xmitting chokes should be under 100 ohms.

You didn't mention what chokes you have, some are much better that others.

Fred

Also, keep in mind you can come off the rectifiers and go to each choke, then to a cap, and have one choke do the modulator and one do the rf deck, all the current does not have to flow though one or both chokes.

The DC high voltage does not need much filtering, but should be stiff and have as much energy storage as you can do (for positive peaks).
Step start is a must for energy storage, even if you leave the supply on and key the cathodes.
4uf per deck and you likely do not need step start.

Yeah.  Use the swinger on the mod and smoothing on the RF. I have a LOT of caps!  What about the bleeder/s?  

Referring to that modulator using 813's, would there be much difference in the driver if I went with 4-65A's?  Since the voltage to the plates of both types of tubes would be the same, it seems to me that the only difference would be making sure the 4-65A's had the correct bias and screen volts. The 813 diagram shows -68 volts through a 220K resistor to the 813's.  Bias is no problem since I have plenty of parts to build a separate supply for the modulator. The other consideration would be the feedback. Again, it's covered parts-wise.

So the question is, what changes would I have to make with the dual 6SJ7 driver if I substitute the 4-65A's for the 813's with 2 KV on the plates?  No problem with the screen supply either. Got that covered also...separate supply.

Yes, that is another way it can be done.  My 6146/6550s is set up that way.  Two chokes for the 6146 and a separate choke for the 6550s.  My 813/811s is also set up to run that way when I use 1800 volts for the 813 and only 1500 volts for the 811s, otherwise when it runs on only 1500 volts for both it runs with only two chokes.  The 811s run off the first choke, the 813 through both chokes.

Fred

If both the 813s and the 4-65s are running AB1 you may not need any changes to the 6SJ7 circuit, just look at the difference in drive voltage to the grids of both tubes.  If the voltages are close, make no changes.  You will need to check the differences in bias and screen voltage.  You may want to make a separate adjustable bias supply and a separate screen supply regulated.  Not too crazy about the PS shown in the schematic.

The two 220K resistors are just the grid resistors and they should fine for the 4-65s.  The other two 220K resistors are the plate load resistors for the 6SJ7s.

Fred

Fred,

Surprise, surprise!  I forgot that I have a completely-matched set of Thordarson iron. EVERY PIECE IS NOS.  I have the T-16P03, T-19C36 5-20 Hy swinging at 300 MA, 80 Ohms and the 19C43 smoothing at 12 Hy, 300 MA also 80 ohms. Like I said from the beginning, I've been collecting parts for over 10 years. All this iron is MINT.  No problem with power supplies for bias and screen.  If the chokes are in series that's one thing since I can tap between them and use just one bleeder.  Seems like a logical way to go.

MK

Then build the supplies.
Make it so you can adjust the output voltage using relays on primary taps, a variac, and or dual voltage outputs.
Then build an RF control deck, bias and screen supplies, metering, adjustable grid leak and fixed bias on the front panel, variac screen supply, big wire wound screen dropping resistor, adjustable high current overload for the screen.

Then you just need a simple RF deck and a modulator.
And you can build all sorts of RF decks to try, and modulators, and even screen modulation.

OK FB, those chokes should be fine, (I have a number of them also) they are rated for 2KV DC working.  The RMS TEST voltage is 5KV but that computes down to the 2KV DC.

Both chokes in series, use the one bleeder off the smoothing choke.  The 100K is FB, same amount I use (two 50K 100 watters in series).  Just remember even with bleeders on the caps, NEVER work on the HV supply without shorting all the caps to ground.

Fred

I have to buff the slight corrosion off the plate iron core and re-paint.  Here is my array. No doubt I'm serious about this...and REALLY "anal retentive" when it comes to waiting until I can find all the right parts! LOL.  I'm not a novice when it comes to building gear.  My build project in Electronics in my Senior year in High School (Class of '64) was a CW rig as a final exam.  Aced the course.  I'm not really a designer, but construction is no sweat.  I've already bread-boarded a high voltage supply based around these components. Don't worry about my safety around HV.  I'm totally "grounded".  That's a pun.  I always keep one hand behind my back because putting it in my pocket would be too much of a temptation!  ;)

If I don't use these choice pieces, a Wouff Hong is my only future.

For bleeder resistors, you can get the dale type ones that mount to the chassis for a heat sink, they are small and cheap and you can run them in series and parallel to get plenty of wattage.

Here's my concern about bleeders.  If the chokes were in series, then one bleeder would suffice.  However if the output of the rectifier were to be split to two chokes in parallel with one feeding the RF and the other feeding the modulator, would you need two bleeders and if so, they would be in parallel  raising the question of what value and wattage would they have to be to maintain 2KV on each leg?

If you use the two chokes separately coming off the rectifiers you would need bleeders on both caps to be safe.  This would increase the bleeder load on the power supply.

For an AM xmtr the bleeders are there to drain off any charge in the caps.  Don't worry about the voltage because the bleeders have little affect on the plate voltage you end up with.  The 813 will be drawing upwards of 200 ma.  The bleeders may only be drawing 20 ma.  Same in true for the modulator to a lesser extent, since the idle current may only be 50-80 ma.  If you leave the HV on during receive then you need to watch how high the HV will rise when in stand-by.  This is when you need enough bleeder current to keep the HV from rising too much.  This is true for both filter circuits (bleeders on both legs).  If you key off the HV supply on receive then the bleeders are not doing much since the PA and modulator keep a large load on the HV supply.

Fred

That makes it easier, just double the resistance of each bleeder section, and go light on the bleeder, it will be mostly for safety.

I once got a fist full of 50 K 25 watt dales and have used them in most of my supplies, series parallel, been working for 40 years so far, up to 3KV.
You can get them new at Mouser and other places really cheap, I got some 50 watt jobs (10K) for my screen modulators, under $5.00 each I think.

Dividing the rectifier output into two separate legs sounds like the way to go considering the entire current would not have to go through the first choke in a two-stage filter.  One little detail just so I can be sure: two resistors of equal value in parallel doubles the wattage of a single resistor?  I know that the total resistance would be half that of one resistor.  To come up with 100K total bleeder resistance, I'd need 200K on each leg?  That is based on ending up with 2250 VDC with only the bleeder for load.  I found that 60K for a bleeder gives me just 2050 VDC.  Under load, that would be closer to 1850 VDC so the 100K would be a logical choice to get me to about 2KV on each leg.

You can do series and parallel, to get the resistance and wattage you want.
Since I had a bunch of 50K ohm resistors I used a bunch.
Here is one supply that used four 50K 25 watt jobs for a total of 50K 100 watts.

http://n2dts.smugmug.com/Ham-radio/i-vcDSc2b/A


And here is a nice simple RF deck that will do 300 to 500 watts carrier easy at 1500 to 1700 volts.

http://n2dts.smugmug.com/Ham-radio/i-k3PMKRC/A

I looked into some power supplies and bleeders in the '64 Handbook and according to that, putting resistors in series doubles the wattage.  Now I'm confused.  If that's the case 2 x 100K at 100 watts each in series would give me 200K at 200 watts?  Nice station!

Correct,  two 100K 100W in series is 200K 200W,  two 100K 100W in parallel is 50K 200W.

Don't over think the power supply as to exactly what voltage you end up with.  You should build the best supply using the iron you have on hand.  Use the amount of bleeder current that is correct for the way you plan to key the xmtr on and off.

If you plan to leave the HV supply on and put the xmtr in stand-by using some method of cathode bias then you need a little more bleeder current to keep the HV from rising too high.  HV too high may exceed the voltage rating of the filter caps.

OTH, if you plan to key the HV on and off with xmit and receive. then you don't need that much bleeder current because the PA and the modulators will keep a heavy load on the HV.  The bleeders with this method are almost not even needed.

Fred

The "Ah Ha" moment!  I'll be keying the HV on and off.  This rig will be used primarily for AM.

Fred, I have a couple of 75K Ohmites @ 100 watt with taps here.  I'm thinking on getting a pair of 100K at 100 watts fixed.  That would be good for 175K @ 200 watts on each leg if I used a 100K fixed and 75K tapped in series on each leg with the variable on the grounded side of the bleeder.  Together I'd have 85.5K at 400 watts bleeder wattage once each leg is in parallel.

Given the variable taps, I COULD use the one on the modulator supply to provide screen voltage to the 4-65A's since the current draw is minimal.  The swinging choke would be in that leg. I wouldn't use the tap in the supply for the 813.  I can provide a separate supply, regulated for its screen in case I want to use the amp in a linear mode. I should end up with something in the neighborhood of 1900 volts under load on both the modulator and RF.

Though those Dale bugs are cheap, 2 x 100K Ohmites @ 100 watts in series with what I have makes sense and the price would be about the same as starting over with all Dales.

As for the modulator driver, I can loaf the 6V6 job through the UTC S-8 into the 4-65A grids.  Yeah, it makes for an extra transformer as a driver but it is more versatile for me because one speech amp / driver will work for everything and all the iron will be UTC for both modulators.  Further, at the higher voltage, Eimac actually recommends using a driver transformer, and recall that phase inverter resistance / capacitive coupled circuit was for a pair of 813's operating at 200 watts.

With close to 2KV on the 4-65A's and  280 volts on the screens, theoretically they can modulate an RF amp with 480 watts input.  I'm not looking for over 400 watts RF on the 813 and the UTC S-22 mod iron is rated at 250 watts.  1900 volts on both should rock.

One more question.  Filter caps.  With one plate transformer, how do the capacitor banks effect the system?  Standard procedure is load up the cap on the modulator to even the greater voltage swings relative to the PA.  I have a good variety of caps and can mix and match.  My original design with two supplies has somewhere around 18 MFD for the modulator and around 12 for the PA.

I guess you have 75K 100W Ohmites with the sliding adjustment.  You can use one 100K 100W Ohmite on each leg of the filters.  Then connect each 100K to the ends of one 75K 100W Ohmite.  Put the slider in the center of the 75K and then ground the slider.  This will give you a total of 100K + 37.5K bleeder on each leg.  Just so you understand how power resistors work,  if you use 1/2 of a 100W power resistor (one with a sliding adjustment) and leave the other half unused you now only have a 50W resistor.  The power rating is based on the entire length of the resistor being used.  So, using the set-up I described above you have 100K 100W + 37.5K 50W on each leg.  This is OK, because the power dissipated is the current squared times the resistance (IxIxR).

If you have about 2KV with 137.5K bleeder, the bleeder current would be about 14.5ma. So, .0145 x .0145 x 100K = 21 watts dissipated in the 100K resistor.  Doing the same for the 37.5K, .0145 x .0145 x 37.5K = 7.88 watts.  You can see that the power dissipated in each resistor section is well under the power rating of the resistors.

Now if you have two the same bleeders (137.5K), one on each filter circuit, you'll have a total bleeder current of a about 29ma (2 x 14.5ma).  This should be just right for what you're doing.  Meaning, you're planning to key the HV supply on and off with xmit and receive.

In general, I don't like using the bleeder resistors for anything more than bleeder purposes.  I would built a separate regulated screen supply for the 4-65s.  But, on second thought, this will create a major problem.  You can't leave the screens on (with screen voltage) and remove the plate voltage.  This will damaged the screens in only a few seconds.  Pulling the 4-65 screen voltage from the bleeders is what you need to do.  This way when the HV goes off, so does the screen voltage.

Fred

I don't really care if the 75K taps are used or not.  I have them on hand.  Here's the basic diagram of the PS.  I'm considering economy.  If I have 175K in each leg at 200 watts, the combined would be 85.5K which just "sounds right."    The 175 K would be a 100K and 75K in series.  Each would be 100 watts.  Any screen voltage will will not be applied before plate voltage is.  The only screen voltage derived from one of the 75K taps would be for the modulator alone if practical.  I can do a separate supply for the modulator screens keyed by the plate supply. 

The other one (variable) in the 813 bleeder is "just there" and would simply be treated as a fixed resistor for the 813 plate side of the supply.  I have more than enough iron to build separate supplies and tons of high-voltage diodes to build any power supply including bias.  All the supplies will be solid-state except for the exciter and the speech amp / driver.

Yes,  that will work.  With the 175K, the bleeder current will be about 11ma on each leg.

With 175K you wouldn't even need 100W resistors.  Two 50 watters would still run way under their rated wattage.

Like I said, that feels right!  You can do all the math, but sometimes intuition and a "gut" feeling makes perfect sense.  Will keep you posted.  Thanks all!

Have you tried Duncanamps.com PSUD? It is a power supply simulator and gets pretty close. It's made for ease of use. Then you can tell what effect various bleeders will have, etc.

I'll keep that in mind Op.  I'll probably bread-board the whole supply before I build it anyway.

Fred, I think my best bet is just go with the KISS principle putting the chokes in series and using the one 100K, 250 watt bleeder.  It will at least give me a start. I have to do this before I become a SK. The good thing about all the iron is it's not used.

Yes, that's what I would do.  When I built my HB 813/811s xmtr I used one HV supply.  It has a Thordarson HV xfmr (P-6411) and two Stancor 500ma chokes in series with 100K 200W bleeder.  I pulled the HV for the 811s off the first choke and the 813 through both chokes.  Works FB.  The power supply is about the easiest part of the project.  Put together correctly, very little can go wrong.  I designed the entire xmtr myself, from the VFO to the antenna.

I spend more time building stuff than I ever spend on the air.  I have thousands of parts here and I'm trying to use them up before I go SK.  I have probably more than 1000 xfmrs and chokes (counting the small ones) here so I have to keep going.

For the 813 final I just used a series screen resistor off the modulated HV B+.  This is the easiest way to modulate the screen.  When you shut off the HV so goes the screen voltage.  You don't have to worry about having screen voltage without plate voltage which will damage the screen in only a few seconds.

Try to keep the project simple.  If you're building an AM rig then just focus on AM.  I saw you mentioned something about also using the 813 final for SSB as a linear.  This will greatly complicate what you're trying to build.  Plus a single 813 as a linear is probably not worth the effort in added output power vs a 100 watt rice box.

Fred