The AM Forum

I'm almost finished with Fabio II and just awaiting parts.  In the meantime, I'm gathering parts and thinking about a new, smaller, plate modulated 200 watt transmitter, Rico Suave II. The main emphasis is on cleanliness.  It will be fed with the same Fabio II GFZ MOSFET audio driver and use about 8dB of negative feedback.

I have a broadcash quality 500 watt mod transformer that weighs 57 pounds.  The tube line-up will be either  two 813s modulated by a pair of triode-connected 813s at 1800V ... or   a single  4-400A modulated by a pair of 3-500Zs at 1800V.... or a single 3-500Z modulated by a pair of 3-500Zs at 1800V.

Has anyone ever run a 3-500Z in class C plate modulated service?  The Eimac spec sheet below shows it can be done.  Will I need to modulate the driver stage some?  That wud be a PIA. The 4-400A uses a screen choke instead.

The 3-500Z is exceptionally clean at 1500V when running at lower power.  As modulators they will be superb.  As an RF final the 4-400 final is easier to drive and modulate.  But the 3-500Z triode is appealing since I've never run one as a triode final.  It requires 125 ma !!  of grid drive vs: only about 15 ma of grid drive for the 4-400A.   Plus the 4-400A is more stable as a tetrode. The 3-500Z certainly needs a good layout and neutralization.

I want to run the rig at about 200 watts output looking for the cleanest sweet spots I can at reduced power for their capacity.

The 813s are a good choice too and will be loafing at 200 watts out, but not loafing as much as the 4-400A / 3-500Zs.

The 813's require less air, quieter.   I am also wondering if a mistake with the 3-500Zs modulator might crap out the 500 watt mod xfmr, even with only 1800V.

I'm leaning towards the 3-500Z modulated by a pair at 1800V - looking for opinions to persuade me one way or the other.

Tnx.

3-500Z tube datasheet:
http://www.g8wrb.org/data/Eimac/3-500Z.pdf

4-400A datasheet:
http://www.g8wrb.org/data/Eimac/4-400A.pdf

813 datasheet:
http://faculty.frostburg.edu/phys/latta/ee/wing813amp/813rcadatasheet.pdf

T
  

Here we go again

You haven't even gotten Fabio II fully working, and now start another xmtr.

Soon the threat for Fabio II will be on page 2, then what??

200 watter,, now you're getting down in my price range.

Nice to have a medium power rig for everyday use or when the funds are low and can't afford to turn on Fabio II.

I would go with the 3-500Z, now here are a few things I would consider if I were building a transmitter with that tube. First, since the 3-500Z can be used as a zero-bias tube, you can get away with using pure grid leak bias when in class C service with plate voltages under 3000 volts (this is mentioned right in the data sheet). Second, I remember seeing somewhere that in order to get that tube to modulate to and over 100% positive, you do have to modulate the driver stage slightly. How I would attempt to do that would be to use a dropping resistor off of the modulated B+ to the driver tube (if you use a seperate driver stage, if you use a ricebox or something else external for drive, then apply the audio to that). If the dropping resistor method is modulating the driver too much, then some of the voltage could derived from the modulated B+ and the rest come from the unmodulated B+, just like you would do with a plate modulated tetrode to apply the right amount of audio to the screen (there's a circuit on here somewhere that shows how it works). But if we are modulating the driver, then how about just forget about grid driving it and run the tube in class C GG, that takes care of the neutralization.

I guess some of the decision depends on your availability of tubes. You can't beat the 813. If you have a good supply I would use them. The 3-500 is now only made in China. Since it is used in so many linears good NOS ones are probably not that easy to find anymore. The 813's are more nostalgic if you are into that.  

Those are some interesting points.

I never thought of running the 3-500Z in GG for class C. As long as the drive is modulated,  can see how that wud work. Very stable too.  Also the zero bias not needing fixed bias.  And no screen supply and choke.

The modulators would also be run zero bias at lower plate voltage.

I am using an FT-102 as the driver. It is already modified for hi-fi AM. Maybe adding audio to that, in the correct phase would do the trick.

Very good food for thought.

I have five Eimac 3-500Z tubes that are in unused commercial linears, so can borrow.  

So in GG, how would I add in the grid leak bias?  Hopefully the grid is strapped directly to ground.  I read that  zeners in the cathode would work for class C too. I suppose that would have a self-modulating effect in the cathode no?    

Also, the input circuit could be like a conventional linear - An L/C parallel circuit peaked with a cathode that is already close to 50 ohms.


About triodes not hitting 100%...  I used to run a pair of 24G triodes modulated by a pair. Without modulating the driver I was able to modulate the heck out of them, way over 100%.    I have read triodes need driver modulation, but I wonder what guys running 833As, etc., have to say about this in real practice?

Also, I wonder if the 25 watts of cathode drive in GG, if left unmodulated, could be a natural negative peak limiter?

T

Tom,

I'm partial to the 813 myself, mostly because it's rated for natural cooling and that means quiet operation. It also doesn't require a very high plate voltage and that means smaller, less expensive components. At 200 watts output even a single 813 is within its CCS rating and is easy to drive too. I don't know if two 813s would make it any cleaner. In class C I didn't think it mattered, at least as far as IMD is concerned.

As for triode connected 813 modulators... I've never found any spec for them used that way. (It would be nice if you built it up and published some data though.) ;-) Still, you'll need alot of swing and a transformer to drive them that way. If you went to AB1, at that power level, I wonder if you could dispense with the driver transformer and have a no-iron all-tube audio chain. Maybe a 5687 could swing the 813 grids enough? Dunno, but that would be pretty cool and hopefully, clean too. Just picture it, an old school audio chain with a wee little 5687 driving a pair of giant 813s. Kinda gives ya the tingles, don't it? No??? Well, if you insist on MOSFETs...  :'(

So, what sounds like fun to me is a single 813 modulated by a pair. Silent operation and CCS ratings - a solid transmitter. If you go AB1, and the mod tranny is the only iron, even better! (One tranny should be enough for any man. There, I said it!) :o

Whatever you build, you'd better use high strength loctite on all the hardware or there will be a Rico Suave III in your future.  ;D

-Don

The 3-500Z will work FB. You will probably not be happy with flogging an FT-102 as a driver though.

All the modulating the driver jazz is for GG only. Grid drive it and make life simple.

I tried a 3-500z once, and it required a lot of drive. At the power level your running, probably 50-70 watts. I would run a SS driver or possibly a 6146 in GG class B ala my 833 rig (813 GG). At this power level, I hate to say it: as much as I love triodes, the 813 is the better choice at this power level. Bump up the plate voltage to 2000 v, and you'll run Globe chump power all day.

Tnx much for the info and comments, guys.

Mike, yeah, I think you are right... the 3-500Z RF final in GG would make life more difficult. BTW, you were running your 3-500Z grid driven class C (not GG) and it was hard to drive?

So, I'm now considering a single 4-400A modulated by a pair of 3-500Zs.      Or, a pair of 813s modulated by a pair of 3-500Zs.             Or,  a pair of 813s modulated by a pair of 813s.

I guess they wud all work, but can't make up my mind.  The silence of the 813s is a big factor though.

Don, I've run the 813s as triode-connected modulators in at least three rigs and they test out very well.  At one time I drove them with the WA3KLR SS driver and the results were flawless.  I remember Tron telling me it was the cleanest sounding plate modulated rig I ever had.  I was using the same 500w mod xfmr with heavy audio NFB as I intend to use now.

Dave, the old spurious apostrophe''''s habit sure dies hard.... ;D

T

The 3-500 grid driven would work OK with a pair of 6146s. I had one in my BC-610 for a while using a T-368 plate transformer for mod transformer. Pile of Wilcox chokes in series for reactor. 110 volt xfmr to Crest 4001, 1/2 6336 secondary to final. The pair of 807s drove it fine to a gallon input : )  If using the FT-102 with the 3 finals biased for linear operation things might be on the edge for the rig's iron. Dunno if the FT-102 driver would have the poop to drive final out enough if rebiased for class C.

My vote is to try a triode and experience perfect neutralization. You will not be sleepless over possible screen modulation non-linearity either.

813 hi-mu triode connected would work.

Yes it was grid driven at about 2.8 KV. Took a 100 watt class rig to drive it to full scrote.

I would try a triode, get out and live life a little; just not a 3-500z at that power level.

A pair of 3-500Z in GG linear mode will run a 200W carrier all day at 1800V (even at 1500V). Downside is a fan but not a noisy one. You might even do it with one tube at 1800V or a skosh more. Drive it with your Class A FT-1000.

Yeah, a triode in the final would be cool.

Problem is I don't really have many except for an 811A, 24G  (too small)  and some 833As.  I wonder how a single 833A would do at lower voltage, like 2KV?   The air is quiet too.   Modulated by a pair of 3-500Zs wud be pretty sick - and clean.

Don, yes, I understand that a class C final is really a switch and IMD is not a factor. Whereas, the modulator has to be very linear for good IMD and THD.

I'm making up a rolling steel/ aluminium platform to hold everything. Gonna use vacuum variables and turns counters in addition to a nice roller inductor.  So anything is possible at this point. Only thang is I want to keep the power below 250 watts and/ or 2KV to maintain a conservative modulation transformer rating. (500 watts)

The last time I ran this mod xfmr really hard, one of the outside leads became disconnected from the winding inside. Thank goodness it was accessible at the first layer of cardboard and I repaired it.  This xfmr is a beauty, so want to take it easy.

Carl, yes, a 3-500Z linear wud do it. I have two 3-500Z linears, but it gets so boring using them they are put in the cellar quickly.  I's needs some plate modulation action here, caw mawn.


I've been on a building binge since November. Must take advantage of the urge now cuz it might easily go away again for who knows how long.  ;)

T

Only high mu triodes need modulated drive. However, if you modulate the drive, you could probably get away with a lower amount of drive. The high amount of drive power is really only needed for peaks, so if you modulate the driver in phase, then it is putting out peak power at the same instant that the final is demanding peak drive. No matter whether you run the 3-500Z in GG or grid driven, you're gonna have to either modulate the driver or just drive it hard in order for it to modulate properly. If you have some linears there that run 3-500Zs, then right there is the perfect test setup, put some zeners in the cathode circuit to get the tube to run in class C, then hook up your modulator to the RF deck. 

Don, OK on only the hi-mu triodes  needing driver modulation.   I remember my 24G low mu triodes needing huge drive to get positive peaks. But the efficiency and power out was amazing for their small size.



Here's the class C specs for an 833A.  At 1800 volts I imagine an easy 250 watts. The drive seems reasonable for a triode at only 30 watts.  The 833A is a nasty RF tube lead-wise, however. Maybe good on BC band, but can I get it stable up to 40M?  I'm not building a Tron SBE rig.


I use my FT-102 to drive Fabio II also, so might bias it into class C to be more efficient.



Typical Operation (Carrier conditions, to 30 MHz)
CCS**                          ICAS***

DC plate voltage 2500   3000 V
DC plate current 335     335 mA
DC grid voltage -300     -240 V
Peak RF grid voltage 460  410 V
DC grid current* 75        70 mA
Driving power 30            26 W
Output power 635           800 W


Still thinking.

T

A single pube is pretty nasty like you said (RF layout wise), they work good push pull with a pair. You could go the route of another ZM innovation and place it sideways, with the plate cap on top and the grid lead on the bottom. Once again, that's how I have the big rig configured here (if I ever get it back on the air). That way the grid circuit could be mounted under the chassis and fed through keeping the grid lead incredibly short. Plate circuit become much easier to design as well.

There's pictures on Jim Hawkins' website of a 44 Mhz FM transmitter that was in use at WSM that used a pair of 833As in the drivers, so if they could get them to work at low VHF then they should work on 40 meters.

OK, here's where I'm at with Rico Suave....

I started pulling out tubes, sockets and fil xfmrs and realized I really like the 833A. I have about six of them - four NOS with Taylor ceramic plates.

The challenge of getting a sideways mounted 833A to work well (W2ZM style)  is appealing. I chose to build an 833A modulated by a pair.  160-40M w/ vac caps and roller inductor. What's not to like?  I will work hard to isolate the grid and plate circuitry and neutralize it well.

The 3-500Z modulators just didn't do it for me.  Though great triodes, they remind me too much of linears and riceboxes.  An 833A modulated by a pair is just too cool and very old buzzardly looking.   Still want to use just 1800V to 2KV to get 250 watts output or so.  Yes, the rig will do much more with higher voltage, but will run cool with little air at low voltage.  

The bonus is I can run the rig silent and convection cooled - with just a breeze of air sucking out with a muffin fan.  I've run 833A's before, usually to modulate a 4X1, but feel they are a little light for that job.   A single one modulated by a pair is a great match for my mod xfmr too.

So, it took some time, but I'm happy with the selection. Gots all the parts needed in the junque box.  


Mike, I intended to use the standard 813 X 813's neutralizing scheme, but do you have a single 833A circuit you like?   Also, do I need to take precautions so that the filament circuit (cathode) of the 833A does not see the plate circuitry?  Usually the fil and grid are under the chassis, so no consideration.


Tnx for the advice and help!

T

The standard scheme of neutralization would be just fine, nothing fancy. I use cross nuetralization in the big rig, but am running shove yank. I mount the fil xfmr underneath the chassis, and mount the fil bypass caps right at the socket to groundium. I have actually run the big rig on 40 once or twice, never had a problem.

Glad to hear the you joining the ranks of class C triode mode. 833a is overkill, but I see where you are coming from, loafing along with convection cooling. Perfect tube for the job.

833's?  ....almost as boring as 3-500Z's, 4-400, and 813's as bellybutton tubes ;D Are those Taylors Chinese? Sacrilege if they are :o

Try for something different, 6C21 (run as 450TL), 450TL, TB5/1750, 810(s), 861, HF or HK something or other, or similar not often seen for RF in decades.

I didn't notice any comments regarding the need to partially modulate the driver of a GG final.  What with the feed through of the driver power you can't just modulate the final.  Stu, AB2EZ had done this with his linear when he plate modulated it.  As is with all his stuff - it sounded really great.

Good luck finding good ones.

There was a time when I had six working  750TLs. Now that's an animal tube.  But one by one the fragile filaments broke due to lack of tungsten from WWII.  

833As are what I have here so will use them.  They still look strange and cool to me. Haven't run them in years  now.

I dunno...  Lots of 4-400As, 3-500Zs and 813s  on the air.    Nothing wrong with those tubes, just very common.  The 813 is a very cool tube.   But very few 833A rigs; mostly BC ones which are not that common.

Al, yes, there was some discussion about modulating the driver for a GG rig.   I re-read Stu's two excellent threads about using a toroid as a mod xfmr and it also covered that well.  I decided to go conventional grid driven with the 833A.  

T

Go full scrote and CB mentality, 3cx25000's driven by 5 alternators and a Cummins.  ;D

If you read what Stu wrote on that subject, you'll actually see that it's not the feedthrough power that causes problems in a plate modulated GG final. With his GS-35B RF deck, it was doing full negative modulation, but the positive peaks weren't hitting 100%. What caused this was the input impedance of the amplifier was going to infinity at 100% negative modulation, and was half of the normal impedance at 100% positive modulation. What was happening in Stu's case was his Ranger was not able to deliver full drive into the input of the amplifier at 100% positive modulation, so by modulating the Ranger in phase with the RF deck, it is delivering a higher amount of drive at the exact moment that the input impedance drops. I included Stu's explanation below.

I love my pair of 813's but I would vote for the 833's

nice size jug and I would guess at 200w output it would cool itself. 

Tom,

  Perhaps you could start with a scale model run at reduced power. Many tube combinations come to mind, and here are two:

801 Triode RF PA, Modulated by a pair of type 46 tetrodes. This is what was in the BC-223, and good for 20 watts AM.

6S4B triode RF PA, modulated by a pair. This little triode is readily available and very inexpensive. Should be good for at least 10 watts Dc input, or 7 watts AM. This would be a QRO PW rig, and followed by an 8877 would make over 100 watts or so RF carrier.

   The idea is to play with different topologies without dimming the lights, or having 140 db ear shattering noise from large contactors "klunking".

Jim
WD5JKO

Im not against the 3 common tubes I listed and in fact have them all in use here in a TX or an amp where they just keep running.

I just like to see something different once in awhile

Hi Jim,

Yep, there are a lot of PW possibilities to try.  My favorite is the 1930's 24G.  Wish I didn't tear it down years ago.


Mike/ZE:

Keith/ WA1HZK has a pair of 833As by a pair. He says they need BIG drive to get positive peaks. In fact he uses a 4-400A GG IPA driven by a DX-100 to get the necessary drive to his pair.

How much power and what grid current do you use for your pair?   I was thinking of using a single 813 in GG as an IPA for the single 833A..   It's slick that he uses no input or output tuning for his 4-400A. Just feeds the cathode and the output feeds the grids of his 833A where the tuning begins.



http://www.criticalradio.com/833%20Rig%20Project/Web%20Pages/Schematic%20Diagrams%20Page.htm

Look at "Article in Microsoft Word"

T

That huge amount of drive power is only needed on the positive peaks, I suggest doing what Gates did in the BC-1T. They used a pair of 807s to drive the grids of the 833As. The 807s were modulated 20% by a tertiary winding on the modulation transformer. Since the driver and final are modulated in phase, there is a boost of power from the driver at the same instant that the final is putting out peak power. Since you are only planning on a single 833A, it should be even easier to drive like that.

Yep, I read that in Stu's thread.  Wish I had one, but there is no tertiary winding on this mod transformer.   The alternative is to modulate the FT-102 with some audio.  Though, I don't mind adding a simple GG class C 813 as an IPA for more drive.

It seems there are always trade-offs in every design. This is the price we pay for no screen / modulating choke.

T

807s seem like big boo boo.  BC1-T used them (Value Engineering).  read this:
http://www.oldradio.com/archives/hardware/Gates/1T.htm

BC1F used 813.  RA-1000 used even more scrote. 

I'm not saying it's necessary to use the same tubes, I'm saying to modulate the driver like they did, that way the driver is putting out peak power when the final amplifier demands it. You wouldn't have to run an 813 for the driver, that's wasteful since you only require the huge drive power that it is putting out on the positive peaks. Or if you want to get really good peaks out of it, run the 813 as a modulated driver. It seems to me that by doing that it would improve the linearity, since the grid drive will be tracking along with the modulation (higher positive peaks=more drive, higher negative peaks=less drive).

Tom,

I use the same IPA in the big rig. I use a GG 813 IPA with a 2K dropping resistor, that drops the B puss down to about 1600v on the 813. Its simple, and 1 tuning control for the IPA. It's link coupled off the tank coil to the grid circuit of the 833a's.
Kieth is right to an extent. We have had this argument in the past. You need a lot of drive, but with the 813 GG, you get plently if you run a single tube. I ran a kenwood TS850 as the exciter with about 10-15 watts key down, well within the capability of the ricebox. The 813 at that voltage easily drove the pair of 833a's to 150 mils. You are going to need 75 mils of grid current. Use a combination of fixed bias of -80 volts and a 5k grid leak. The grid leak with drive will cut the tube off well past 4x times cutoff (needed for class C PM), and the -80 will let the tube conduct a small amount if drive is lost to provide the mod xfmr some sort of load before safety circuits cut in.
I can't wait to hear it actually. You might have the got the ol' juices flowing for me again.

Carl,

I have a HB single HF-300 modulated by a pair of 805's. Strange enough for you? ;)

Nope, the "value engineering" was the 6BG6G's used in early models and they had to go back to 807's which were fine. The sweep tubes didnt hold up well at 24/7 balls to the wall.


Not strange at all, I like it. My very first attempt at an amp in 57 was PP HF-250's and 100TH audio. The HF's were from a scrapped diathermy machine and soft. The next try was PP 250TH's and 810's for audio, pretty conventional back then and used mostly WW2 surplus from AFMARS at Mitchel Field

These days I have a pair of HK-354's modulated by 8000's which was used often a few years ago before I started playing with linears.
The other amp is/was 250TH's with 304TL's which is apart for 6C21's and Im still contemplationg the audio. Too busy doing customer work these days.

Carl

Dern son, you must not have younguns running around the estate!!

Ever thought about building these suckers on commission?

Yep, trying various techniques to improve linearity is on the list.  While monitoring for IMD and THD, adjusting the drive and/or modulation of the driver stage can make it cleaner.  From experience I've found that simple drive levels, screen voltage, loading and other tuning can change 2nd harmonic THD by as much as -15dB on a plate modulated tetrode, using a spectrum analyzer.  You should see the THD blips creep up and down. These are side splatter levels, really.

There's an old thread on this BB where Dino / WA1KNX adjusted the amount of screen modulation on his plate modulated DX-100. (6146s)  He was able to find the sweet spot and the rig did sound choice on the air.  I plan to do the same thing here with Fabio II and Rico Suave.

After all, whose to say that when sticking a 20H screen choke into the circuit is optimum? Chances are it could be better with some tailored R/C network work around the choke or screen dropping resistor.

Same with triodes - picking a "20%" driver modulation level. Maybe 13% or 26% is optimum. We don't know until we do some tests and tinker around with our particular parameters and situation.  I would think all users of plate modulated rigs would be curious to know if theirs' is optimized.

Mike/ZE:  I might do a GG 813 driver and then look to add some modulation to it. Either more drive or driver modulation will do the trick. It will be interesting to see which technique works best for THD.

T

Tom,
If your looking for a rig to give 200W out, how about a single 813 modulated by a pair of 811's.  If you want to go up a step, your old pair of 813's modulated by a pair triode connected would be great also. 

Remember much above 100% causes distortion to many that are not using sync detectors.  I  have always preferred a well processed 100% and then leave things alone. 

As has been mentioned, lots of choices and all have their set of merits!   

Joe, GMS   

They got Joe too ''''''

The 20% number seems to be mentioned often, that was how much driver modulation Stu was using in his GG amp, that's also how much Gates used in the BC-1T. From what I've saw of Dino's circuit, he was not using any screen choke at all, what he was doing was using a dropping resistor before and after the mod iron and then combining the voltage to get the optimum amount of screen modulation. I believe he was using 60% unmodulated voltage and 40% modulated voltage.

Dino's experiment would be something I'd like to try sometime if I ever used a screen dropping resistor.  Using a single 813 at about 2200V would mean about 70 watts of heat in the screen resistor. That wud be my heat limit before switching to a choke.


T

Ya know, second thought, Mike / ZE,  I'll bet you could try modulating your 813 GG IPA driver.  You already have some HV DC to the plate now.  Then find the in-phase tap off the modulation transformer and sum it in too.   A big 100 watt power resistor might do it. Maybe need less drive. Look for the best THD and linearity.

Would the connection between the mod xfmr and the 813 need to be thru a 2uf HV cap?

I may set this up myself once settled on a plan.

T

Thomas,
What happened to all those 4cx 350's with their very low IMD's?  ( I'm a comma man and proud of it!)

In the class C, Rf chain, one as a driver feeding a pair in PP.
In the audio class AB chain, one as a driver feeding another pair.

Plate Voltages on the drivers about half the finals or what ever yields the best transfer characteristics.

If you've got a ton of 'em, then double up everything, PP parallel, etc.

Yours,
-Ricardo (not really related to Rico this go-around, but oh so envious...)

Its something to think about Tom. Right now the rig is literally a shell of its former self. I still haven't assembled it back together from the move 6 years ago. The HF300 rig is on the air, but the antenna came down this winter and I haven't messed with it at all. Very little time these days, My main focus on hobbies is fly fishing. Have debated over the years whether to do an INR flush or just hang on to it all. In the end I always tell myself that I'm not hard up for the money and I will return to it someday.

For the big rig, the priorities are as follows:

get it on the air.
build low power exciter driven by QS1R/QS1E combo
build/install GFZ driver
then.....modulate the IPA

Mike,

Yeah, I hear ya - it comes and goes in cycles here too.  

Hopefully you'll get that 833A rig back up and running.

Cool on the GFZ driver and fly fishing.


The dream of designing and building is a lot of fun.  Actually, after the rig is done, operating can be anti-climatic sometimes.

What we need is for some of the old crowd to get back on and mix it up again. Still, there's a lot of new people on too. I plan to do some coast to coast in the future and also see what's happening locally on 75M.  Hope you hear the action and jump back in.

Spent a few hours up at 140' today getting up the rest of the WNW 2el 75M wire Yagi. Ready to hook up the coax.  Hot, sweaty and windy. It's getting scarier and harder to work up there these days. If we don't do it often, the mind plays tricks and can get us thinking like a wimp sometimes...  :-)


Later -

 
T