Screen or Plate Modulation? Which way should I go?

[w8khk]

Hi All, I am requesting input.  I started out to build a screen modulated 4-1000 rig for 75 and 40.  I planned to put it in a 4 foot Bud rack, but that was not large enough for the power supply components.  So I switched to a larger HP computer rack, which uses 19 inch panels, but it is taller and 33 inches deep.  So I got to thinking, maybe I should be building a plate modulated rig instead.  Here are the components I could use:

4-1000A, final deck almost completed
Temco plate transformer, 225 pounds, primary is 120/240, secondary is 3650 volts each side of center tap, at about 2 amps.
20 amp GR 240vac variac for plate voltage
10 amp GR 120vac variac for filament voltage control on 4-1000
Oil capacitor, 100uf at 4kv
Stancor A-3899 600 watt multi-match modulation transformer
Modulation reactor, 50hy at 1.4amps, 10kv rating, 20kv hipot
2uf, 6kv oil capacitor for modulation
A NOS pair of 4-400As, sockets, filament transformer, and plate cap radiators
A used pair of 4-250A tubes
Relays, switches, and all necessary meters

All this stuff fits nicely in the bottom of the rack, so I am wondering if I should proceed with plate modulation, or go with screen modulation.  At first, I assumed it would be a lot less work to build the screen mod rig, but it seems to be very little more incremental work to go with plate modulation.  I would still have to modulate the screens to achieve linearity with the 4-1000 tetrode, but that could be done with a high-quality audio output transformer wired in as a choke.  

 I already have a decent speech amplifier, and a Behringer VX-2000 speech processor, so I just need to press on and get the transmitter assembled and tested.  The final is almost completed, just a bit of wiring, and the tank coil and plate choke to be added.  I am attaching some pics...  I put a couple 807s in the photos just for a size reference.  So I am open for suggestions, opinions, etc. What would you do, and why?

[W2XR]

How about resurrecting and modifying that RCA broadcast rig I see in the background of one of your photos? That might be the easiest path to a hi-fidelity AM rig.

Just my curiousity & 2 cents worth.

73,

Bruce

[w8khk]

Hello Bruce,

Thanks for the suggestion.  The BTA-1MX on the left was purchased from another ham over a year ago.  I have been doing safety mods to it, and it is almost ready to go on 160.  It had no back door (very bad for the cats) no safety interlocks, and no PTT function.  I have fixed these problems.  It had vac caps for tuning and loading, with the turns counters mounted on top of the cabinet.  These did not clear my low ceiling, so I am reworking this problem with some u-joints and right-angle gears.  Hope to have it on the air in the fall.  The BTA-1M on the right was rescued last May from WABV in Abbeville, SC.  I learned on AMfone that this rig was to be scrapped, and made the phone call.  The engineer stripped it of tubes, transformers, etc, and the rest was in very bad shape, lots of JS mods over the years.  I am collecting the parts to put this rig on 75, but in the meantime I am enjoying homebrewing, which I have not been able to do prior to retirement.

73,
Rick

[w8khk]

Hi Brian,

Thanks for the vote for Plate Modulation.  There are still quite a few factors I am struggling with, and probably many that I did not even think about.  For example, screen mod can provide extremely high quality audio with excellent linearity, just listen to Slab Bacon if you are not yet a believer.  I have looked at the overall power consumption, and adding the filament power of a modulator and speech amp offsets the low efficiency of the screen modulated final; there seems to be no winner here.  Either way I could run the legal limit without pushing the ratings.  Plate modulation requires transformers in the audio path, whereas screen modulation could be done with cathode followers or mosfets, eliminating the iron.  Would screen modulation be cleaner and more linear without the transformers?   Will I need to add spark gaps to the mod tranny to protect it, like they do in most BC modulators?  What if the Stancor craps out while testing?    If I use screen modulation, will it be tricky as to loading in order to have a clean signal?  With screen modulation, will I cool the 4-1000 sufficiently without undue jet whine in the shack?  Dissipation would be lower running plate modulated class C.  So many issues to sort out, that is why I was looking for input from the experienced folks on this forum.......

Maybe I just need to build it to work either way, switchable, and compare the two.   But I would rather get on the air sooner, and take the best route.......  so opinions and suggestions are welcome!

[N3DRB The Derb]

there's no way to tell a properly built and adjusted screen modulated transmitter from a plate modulated one. The big advantage is not needing 200 pounds of mod iron if you don't already have one, not to mention the tubes and audio power required.

I vote screen.

[w4bfs]

Hi Rick ... I know how you feel ... to get 400 Wts carrier with a 4-1000 screem mod requires 1400 Wts input and 1000 Wts plate Dissipation ... it will certainly heat the shack, also note that tube life may be less because emission drop will be noticed quickly ... even so, if the fil temp is held nearly constant between tx/rx cycles then fil life is a constant at a particular fil power ... If you can cleanly modulate the screen as well as find a way to diminish the carrier level say 6 db or so between words, then could drop 750 watts off the static plate dissipation and the screen mod could be a clear winner in the efficiency dept ... this has been done using controlled carrier modulation (which seems to use a roughly 9 db or so reduction - Heath DX 60) but can be annoying with some receiver agc circuits ... maybe a lesser ratio would be better, make it adjustable ...just some blue sky.....73...John

[w4bfs]

hey Brian ... watts is watts unless some don't count because of entertainment value ..beefus

[w8khk]

Hi Derb, thanks for your input.  Yes, screen mod is just as clean as plate mod if set up right, I have never done it so half the fun is trying it out.  From all I have read on this forum, it is better to get it set up linear and avoid the detected carrier feedback.  If I use screen mod, then I can save the heavy metal for the 304TL restoration.  Can't screen mod that guy!   

Brian, I was planning on using a couple parallel 1625s in cathode follower for a combined screen voltage regulator and modulator.  Still thinking about how to mod the control grid, probably some HF audio xfmr...   I am aware the screen load is non-linear, but it should not take much to wiggle the control grid and screen, certainly not anything as big as a 4-400.  I planned to use the reactor to keep the plate current off the mod tranny secondary, improving the low frequency response.  This should not endanger the mod tranny.....

John, the dissipation to get the legal limit out of the 4-1000 is one of my biggest concerns.  I have only one 4-1000, no spares.  Cooling is a concern, I wish I read Slab's discussion on cooling before I bought the chimney, his idea is really cool (squirrel cage under chassis for tube base seals, and muffin fan up top, switched on with the PTT, no chimney).  For a summer rig, I am thinking that this much dissipation may not be the greenest way to go.  Controlled carrier is a kwel idea, if not done to extreme, as it makes for annoyances on listening.  I suppose I could integrate this into the screen regulator/modulator.  BTW, this is a summer rig;  the RCA BC and 304TL rigs will heat the shack in the winter.

[w4bfs]

way to go Rick ... I think you are on to something with the cathode follower/modulator/regulator/voltage shifter .. I wonder if something else can be dreamed up for it to do ... I am most interested in this as I am considering how to refill my Gates 250 gy ... a 4-1000 is a def possibility as it is one of my favorites ... to make this even harder, I want to use year around on 160-10 meter... I am impressed with Bacon's hard work and results ....oh by the way Alan Bond gets some 4x1's occasionally ...last one I got from him was a Triton and was nearly new ...

[w4bfs]

super thread !!  ... it amazes me just how many ways it is possible to do a 2 quadrant multiplier (aka am modulator)  ...if you get really interested i'll tell you about using Walsh functions ...I will continue to consider the screen mod route ... sounds like a fun thing ...yeeehaw

[w8khk]

John, funny you should mention Allen Bond, WB4GNT.  Had an eyeball with him just last week.  He provided the chimney for the 4x1, genuine Eimac, for a very good price.  Allen is almost a neighbor, lives only 3 miles from my QTH.  He comes to all the local hamfests, IIRC I met him first in Orlando around 2006.  He sure has lots of good stuff!

Yes, If I go the screen route, I will definitely come up with something new for the modulation and screen regulator.  Main thing is to use parts I have.  Only things I bought new are the 4x1 toob, chimney, and socket.  Speaking of sockets, I first bought a chinese aftermarket job offa eBay.  Bad Idea!  Good quality, but the design blocks most of the airflow past the base of the toob.  So I found a genuine Eimac socket, NOS on ebay, paid the same price for it, much better.  Allen did not have any 4x1 toobs when I was looking, so I took a chance on one on eBay.  Got it from a ham club in SC, club was selling it for an SK.  They packed it very well and it arrived alive!

[ka3zlr]

Class E...Efficient...Parts very readily available..Todays game...

Stevie's a Great Guy....

My opinion...

73 Jack.

[John K5PRO]

Unless you drop it, or it is already spent, that 4-1000 should play for a long time for your TX. Dude, you got the mod transformer already, plus reactor. Put a 1 or 2 uF coupling cap in there, take the DC off the Polypedance, and get that baby on the air via plate modulation. Why were most vintage 1 kW broadcast rigs plate modulated? Cause they were easy to build this way, and they sounded great. Sure, if you run square waves into them, you might get trouble. So don't.

I think more fun would be had playing with 4-400A modulators or going to a pair of 833s in class B, than investing in screen modulator, but I hear you, the call to innovate and play with something different is there too.

Put spark gaps across the Stancor, and watch out during testing with tones not to run into negative cutoff (100%) frequently. You can always find vintage BC iron too, if you look on 'ebay and other sources, like location stations getting rid of tube rigs.

[N8LGU]

    If you're gonna do it...and you have the key parts...do it right...Plate Modulation!

[K1JJ]

Hi Rick,

As we all know, there's  many, many ways to get a good sounding AM rig on the air. If each method is set up properly no one will really know what you are running. Be it, grid, screen, plate modulation; class E, class D, solid state or tube; low level ricebox driving a linear, etc.

It becomes a "mind picture" of what you are running to the world. What you say is what the picture becomes.

It's the same analogy as riding into a Harley Biker club with a scooter. All bikes will get you there, but some bikes have more "image" than others.


I believe half the fun in ham radio is hanging with a bunch of like-minded ham buddies and then "riding" together admiring and encouraging each others creations and projects.  If it was a matter of simply putting out a signal and sounding good, a simple ricebox and linear is all you need. The overall efficiency it not too much different than other methods when everything is added up.

However, since we are in this for the fun and hobby, we restore old tube stuff, build big tube rigs and love to talk about what we've done ... just like any other hobby where there are builders.

With this in mind, I say build the biggest and baddest rig you can. It will give you much joy and pleasure - and something to talk about every time you get on the air. Your buddies will be inspired to build something as competitive and your interest in the hobby will grow.

I've always thought the 833A and  the 4-1000A were the most revered tubes (available too) out there. A 4-1000A plate modulated by a pair of 833A's is outstanding. It's riding a Harley.  Since you already have most of the iron and parts, stretch yourself a little more and put one together.   4-400A's are fine, but they are very very common and do not command the respect of 833A modulators.   

It's funny how we are usually remembered from a past one-time QSO not by our personality, call or location, but by the rig we were running. How many times have you heard an OT say something like, "yeah, he's the fella running the pair of 750TL's modulated by 450TL's.. now what's his call???"

Add a unique touch to the physical layout of your rig and post some pictures for us later. There is no better way to express your ham personality than to customize and run your own homebrew rig.

Good luck.

Tom, K1JJ   (Bought a kayak and been out paddling with Yaz almost every nice day for the summer)
http://www.pxlpilot.com/don/photos/135_01.jpg


Here's my own 4X1  X 833A's.  I have a short memory for past work I've done, so I built it under Plexiglas to look at...    It's an early shot before the full case was built and final shack positioning.

http://home.comcast.net/~k1jj/wsb/html/view.cgi-photo.html--SiteID-3019502.html

[w4bfs]

I like it Tom ... if you build it, they will come? ... very persuasive ...73...John

[w8khk]

Jack, thanks for the class E suggestion.  Maybe when I get burned out doing it the old fashioned way, i will try my hand at some real efficient solid state stuff.

John, thanks for the suggestions in caring for the mod tranny.  I am really leaning toward using it.  Yes, I have been thinking the 4x1 will last longer plate modulated.

Brian, that 4-400 modded by 833s sounds really kewl.  Plenty of strap!

Tom, thanks for sharing the info about your plate modded 4x1.  It answered one question I had.  Many of the guys use perforated metal cages to shield their class-C finals.  Plexiglas lets the entire spectrum thru - RF through visible and beyond.  I can probably get away with building a rig without lots of shielding, although not quite as open as the one in Quito....   I have side panels and a back door for my rack.  I was thinking of using 4-400A modders instead of 833As because the common plate supply would be in the correct voltage range for both mod and final.  If I used 833As I would likely need a separate plate supply.  Also I was thinking of running 4-400s AB1, with a very simple driver stage following my Behringer VX-2000.  

The kayak looks like lots of fun!

One question you folks might answer....  What method do you use to protect the screen of the final in the event the plate voltage craps out?   I thought I might sense the current on the plate volt meter at the end of the meter multiplier resistor string, thus releasing a relay to cut the screen voltage if I lose plate voltage.  Any other ideas?

[Knightt150]

Hi: This the first time I have seen a need to reply to a question, I have and run a plate modulated tnx and have and run 3 screen modulated radios. The screen modulated radios look perfect on the scope the plate rigs looks almost perfect. If I was going to show someone how a AM signal looks (text book) perfect. I would show the screen grid radios. They require a slight differance in tuning the final but the efficency is very high. The carrier idles at a lower power level but when modulation is applied the power gose up and so dose your ac meter on the back porch. Belive it or not there are commercial short wave and broadcast transmitters out there running every day, the price of commerical electricty has driven them there.

[w8khk]

The carrier idles at a lower power level but when modulation is applied the power gose up and so dose your ac meter on the back porch.

John, thank you for your interesting info on screen modulation.  I wonder if your efficiency is accomplished by controlled carrier rather than pure screen modulation?  In any case, one reason I would favor screen modulation rather than plate modulation is the ability to eliminate the transformer entirely, and its inherent phase shift and bandwidth limitations.  Perhaps you could share more techical info with us on your implementation - sounds very interesting!

Brian, your textbook examples do not provide for simultaneous control and screen grid modulation.  The literature and practical experience indicate that this is required to achieve the optimal linearity, and truly compete with plate modulation.  Perhaps we should look at Bacon's post on "Scream Modulation", as he is using this technique. 

In summary, I have still not decided which way to go.....  So perhaps I will do both.  Either method requires stiff regulation of the control grid bias and screen voltage; adding modulation to either or both should be trivial.  I am thinking of building the modulator with 4-400As with a very simple resistance-coupled driver stage to interface the VX-2000.  In addition, I could use a professional solid state stereo audio power amplifier to drive two audio output transformers in reverse, one for the screen grid and one for the control grid.  Add sufficient swamping resistance on each transformer to provide a constant load impedance to the SS amplifier, level controls on the amp would allow continuously variable adjustment of the modulation level on each electrode.   A relay could be used to kill the 4-400 filament supply and short the mod tranny secondary and reactor, providing pure DC to the 4x1 plate when testing the screen mod option. 

Via this configuration, I could sweep the transmitter using a dummy load, evaluating each option, and determine how much we need to wiggle the control and screen grids to achieve the same, or better linearity, than plate modulation.  I thought of using 833As or 304TLs for modders, but this would require lower plate voltage for the mod, and thus a separate HV supply.  The single supply has sufficient scrote for both, and a separate supply just takes more space and adds weight.

I think most of my time will be spent on PS, control circuitry, step start, TR sequencing, etc, so I will probably just build to run this as a SKITZO 4-1000.  The information gleaned should be well worth the additional effort.  PS- I am still open for ideas, suggestions, and advice from all the experts. 

[k4kyv]

Plate modulation is much easier to adjust for good modulation linearity.  With screen modulation or linear amplification, you have to adjust the loading and grid drive just right, and supply just the right level of audio.  With plate modulation, as long as you are in the ballpark with grid drive and bias, and have enough audio to fully modulate, it will modulate with good linearity regardless of plate loading, up to a certain maximum, and to change frequency all you need to do is peak the grid current, dip the final and maintain the desired plate loading, and the modulation will be close to perfect.

As far as total efficiency goes, ratio of  power drawn from the a.c. mains to rf input to the antenna, there isn't a lot of  difference between low-level modulation (linear amplification and grid modulation) and plate modulation.

Plate modulation is a lot less critical to set up, but has the down-side of more components (and $$$), and the extra weight and space requirements.

[John K5PRO]

As for screen grid protection for your 4-1000A, the relay in the plate multiplier is a fine circuit to do this. It is about the most protective method, as just looking at the primary plate power to the transformer, and then assuming that there is plate voltage, is risky. For example, if the plate connection is left unhooked, or the rectifiers crap out, or a fuse blows, you might have screen voltage still going, and toast your screen. The trick to putting a simple relay coil in the multiplier string is that you need a high z relay, also called a 'sensitive' relay. a 5000 - 10000 ohm coil resistance that needs only a mA or less to close the relay is the ticket. That way, when the plate voltage gets up to about 1/2 or 2/3 of it's normal value, the screen voltage then switches on too.

[w8khk]

The 3 circuits are right from the 1959 Bill Orr handbook. They don't show the additional circuitry that would be required. They are just simplified examples showing the 3 different types of screen grid modulation configurations.

Thanks for the Orr reference, Brian.  I am familiar with that, as I have several of the Orr handbooks, including 1959 (fifteenth edition).  I recall reading about simultaneous control and screen grid modulation, and I was surprised that I could not find reference to this in the Orr handbook. 

A bit further research in the ARRL HB, 1965 page 265, with reference to screen modulation, I quote:

"With proper adjustment the linearity is good up to about 90 percent modulation.  When the screen is driven negative for 100 percent modulation there is a kink in the modulation characteristic at the zero-voltage point.  This introduces a small amount of envelope distortion.  The kink can be removed and the over-all linearity improved by applying a small amount of modulating voltage to the control grid simultaneously with screen modulation."

This is what Slab Bacon and others do, and it is impossible, at the receiving end, to tell that they are using screen modulation. 

With all the study and discussion, I am leaning toward implementing plate modulation, at least for the first pass.  With everything in a rack, it is easy to swap out modules and make changes; nothing is in concrete!

[w8khk]

Plate modulation is much easier to adjust for good modulation linearity.  With screen modulation or linear amplification, you have to adjust the loading and grid drive just right, and supply just the right level of audio. 

Don, thanks for offering your point of view.  This is exactly the type of information I was looking for.  If I had to select one feature of plate modulation to make my decision, it would be ease of operation in order to produce a quality signal.  Operating has to be fun!  All else considered, initial expense and operating costs are a wash in my situation.  And whether I use the mod tranny and reactor or not, the plate transformer weight has pretty much locked me into the basement shack.  I will initially implement the 4-400A modulator, but leave options open for experimenting with screen modulation in the future.

BTW, I have gained a lot of insight and understanding reading your posts that I have not found in many of the handbooks.  Thanks!

[w8khk]

As for screen grid protection for your 4-1000A, the relay in the plate multiplier is a fine circuit to do this. It is about the most protective method.........

John, thanks for confirming my suspicion.  I was going to use either a sensitive relay, or use a transistor or darlington pair biased by the plate multiplier to drive a 24 volt relay. Lots of control logic to work out with step-start, safety interlocks, and TR sequencing.  I am planning to leave the HV on during standby periods, and cutoff the modulators and final, leaving only the bleeders loading the PS.  100uf oil caps are a bear to charge initially, and I do not think I want to cycle them with the PTT logic. 

BTW, I still have to restore my dad's 304TL rig modded by push-pull parallel TZ-40s, and my grandfather's 450TH rig, modulated by 810s.  That would be quick to put back in original condition, but I feel that it requires overload protection, step-start, safety interlocks, etc. before putting all the old components and myself at risk.  The old rigs used plug-in coils, but at my age I would rather just get them running as individual single-banders and avoid messing with plate coils on a daily basis.

[w8khk]

John, my present goal is to eliminate as many transformers in the audio chain as possible.  I was looking at an article in QST for March 1948 page 13.  Here they run a 500 watt modulator with a pair of 4-250As driven by RC coupled 6SJ7s.  Using tetrodes in class AB1, no driving power is required.  The author said he left room in the design for inverse feedback, but it was clean enough that feedback was never needed. 

I expect to implement something similar, driven by the 600 ohm balanced output of the VX-2000.  We are down to one transformer, the modulation tranny.  With screen mod, I would do it with zero transformers, but that is a few months out.

Speaking of doing something with the screens, In the 60s I ran a pair of parallel 4-400s on 75 thru 10 using the G2DAF design.   No screen supply.  Just take the RF input to the linear and pass it thru a voltage doubler rectifier filter, and slap it on the screens.  Used it for years, still have it but I want to get on AM, not slop bucket. 

The reason I bring this up has to do with my thoughts on cooling the tube.  I used a pair of Johnson ceramic sockets for the 4-400s, and a single four-inch muffin fan to cool both.  The only air they got was thru the socket and thru the tube base.  I did not have any holes in the chassis around the tubes.  Being on a budget while in the Air Force, I used two Coleman kerosene lantern chimneys for the tubes.  The original 4-400s are still working like new, even after long key-down tune-ups and testing.  That is why I have the spare tubes.  But when testing the 4-1000A under just filament power, it seemed very hot, even with a large fan and Eimac air system socket.  So I got to thinking, with no spares, running it with variable efficiency modulation at high dissipation, might tax the tube, or require lots of air.  One of my ham friends once told me how you can tell if you have TOO MUCH air flow thru the tube.  It blows the tube out of the socket!  You can never have too much air, but you can have too much blower noise.  Dissipation - just another factor pointing to high-level plate modulation. 

73,
Rick