The AM Forum

I was thinking of a new tube project that would have low power and clean audio. The attached schematic uses four tubes. The RF section consists of a 5763 driving a 7984 Compactron, which is series modulated by a 6L6. A 12AU7 is the audio driver.

I calculated 62 mA of plate current on the 7984 for 15 watts carrier output.

I've never built a series modulated rig before and I borrowed pieces of schematics here and there. Any glaring errors in the schematic?

Jon

** I did just think of something else. With series modulation, the power output is cut in half because of the class A modulator. If I want 15 watts output, I'll need to calculate for 30 watts output. That puts plate current at 122 mA and brings plate impedance down to about 1450 ohms.

I am not sure the 6G6 will handle the needed series current.

May I suggest a 6DQ5 for both the Final and the modulator if you want all tube rig? They are tough tubes and I have more than 8000 hours on one 6DQ5 final in a series modulated rig.

A better series modulated circuit IMHO is to use a HV power mosfet as the modulator since it has a lower series resistance.

Phil

I would use something like a 6DQ5, but it has the plate connection on the top of the tube. I want to have all the tubes exposed on the chassis for cosmetic reasons. The 7984 has the plate connection on the bottom pins.

I could parallel 6L6's or use a single KT-88 as the modulator.

Jon

hi Jon .... the limiting factor in the series modulator is its plate dissipation .... say with 900V available from the hv ps the modulator should drop 60% quiescently or so to be able to reach 100% modulation .... so this is about 540V and at 33 Ma this is 18W which is reasonablle for a hv 6L6 family tube ... the remainder of 360V tmes the 33Ma  gives 12W input to the rf amp and at 70% efficiency gives around 8 W rf out which is a good drive level for a rf amp ... making the Q point variable will let you shift output power and mod index if you wish

by the way I thot you had gone solid state for yer transmitters .... guess the soft glow of hollow state gotcha ...it has me  ;D

Yeah, I still have my 24-FET transmitter that I haven't finished testing yet. I've been thinking of a new project and I like exposed vacuum tubes without the danger of high voltage. I think I can build it to have a "wow, cool..." factor too.

I'm mostly concerned with the 'floating' 7984 above ground at high voltage. Anything look funky with the way I drew the schematic on that?

Jon

i apologize ....I didn't look at ur skizmatic before I spoke .... I looked at it and didn't see any show stoppers .... just a few things .... you could operate the 6l6 with screen and plate tied together or triode connected .... helps keep the screen out of trouble and is good at 400 V .... biggest thing is getting some isolation between grid and plate circuits on the rf final ... I imagine that neutralization and grid resistance willl be needed as well as a dividing plate between them .... maybe you'll get lucky and not need them ... if you could find a GE mastr pro low band tx you could see the 7984 applied AIRC

Here is a suggestion:

Page 2 is the Final.


Phil

What a Fun project!  Keep us posted on your progress. There are so many cool overlooked tubes like Compactrons,
Home-brew is great since you can use what you want, not just standard run-of-the-mill stuff.
Fun project!

Phil, that's... that's awesome!

I wanted to use tubes that were still in production for the audio section, and both the 12AX7 and KT88 are.

Thanks!

Jon

Sorry for going off subject...Phil, what do you use to draw schematics?

Ed

I looked at the Svetlana KT88 tube curves and they wanted a bit more control grid bias and a bit more screen grid current so those two values were modified and I added your "OP" switch at the cathode.

One thing about cathode modulated transmitters using tetrodes is they need solid screen voltages, low modulation stage series resistances, and plenty of RF drive.

CM transmitters have great audio.

Hope this helps.

Phil - AC0OB

I use LTSpice 'cause it's free and only takes about 20 minutes to learn. I have other CAD packages but I use LTSpice mostly.

I have most of the major components already and I just ordered a new Hammond 750 vct transformer. That'll be mounted on top of the chassis, so I wanted a transformer that was 'pretty.'

There's a hamfest coming up here the first week in January that I hope to pick up some other miscellaneous stuff for it.

I'm also going to use a Hammond chassis enclosure for it. I can't order the one I need until I get all the big parts together and figure out what size I need.

On a side note, I need a PW rig right now. I bought one of those LDG feedpoint autotuners rated at 100 watts. It's installed on my delta loop (my only antenna), and running 2000 watts AM is out of the question with it.

Thanks again for the schematics!

Jon

Hi Jon, I take you're PS will be a Full Wave Bridge design?

BTW, just the SA, Cathode Modulator, and Final require 195 mA by themselves.

The Power Output is calculated to be 45-53 Watts at an efficiency of ~ 55%.

I upped the SG current at the Final because GE preferred 8.5 mA and added a 5W power control pot at the cathode of the KT88's. Added bias protection zener.


Phil

Yes, I'll use a full wave bridge with the center tap at 350 volts.

I won't run the rig at that power though and will back it off a bit.

Jon

Phil

Can you give some information on a mosfet series modulator. Could it be done for a larger tube such as an 811A or a 572B? Or even a pair?

Thanks

Pat
N4LTA

Hi Pat,

Normally we would use 1 HV Mosfet for plate voltages up to 750V but in the case of say a 2000V plate voltage on a 572B we need a cascode arrangement which complicates things a bit but here is a suggested circuit.

The neat thing about using a Mosfet is it doesn't take much gate audio p-p voltage to modulate a final stage because of the Mosfet's low Rds.

So here is a suggested circuit:


Phil - AC0OB

Thanks Phil

I have some 4D32s that could work well at lower voltage.

Also thanks for your past help with voltage regulators and screen regulators. I am firing up one of your circuits today. A screen regulator for a couple of SRS 551 tubes.

Pat
N4LTA

hi Pat and Phil .... hope the holidays have been merciful to youall .... I have a question about your cathode modulator ... the skizmatic shows a nonconnection from the bottom of rv1 and coc ....is that correct .... and what kind of diode is d1 ...thanks  73  John

RV1 is a 2.5 Meg pot. that sets the Mosfet bias in a feedback bias arrangement.
This and the 10 ohm metering resistor provide feedback to reduce distortion.

In LTSpice anytime a solid line intersects or is orthogonal to another line it is connected.

The only time a cross-over or a non-connection is shown is when there is a dotted or dashed-line or a jumper is shown.

D1 is a IN4151, a later version of the 1N4148.

ok Phil thanks for the info .... so the .7V or so drop of d1 will enable the drain of M1 to be at half potential V of M2's drain ... at least I assume that is the goal .... in this fashion will create the higher Voltage series pass element ....clever, if it works .... have you tested this with several parts of different production runs .... how about temperature changes

4D32 tubes are very rugged tubes.

Here is a genral Mosfet modulator schematic for a final with up to 750 volts on the plate. I have used this as a modulator for dual 6146's, 6DQ5, and dual 807's in a friend's Globe Chief.

Depending on your tube or tubes, the only changes you may have to make are the L1 current rating and the fuse's rating.



Phil - AC0OB

I have used the HV version with an 813 at 1800+ volts plate-cathode voltage. (Future ER article).

Yep D1 drops about 0.4V to compensate for the Vds drop of M1 due to its Vds at the expected drain current.

I am not producing this commercially as it is for the benefit of the ARS community who wants to experiment. My original article in ER 359 gives the circuit theory, etc. for a 40 Watt CM transmitter using a 6DQ5 Final in which I had converted my Heising Modulated transmitter (ER 353) to CM.

The only drift I have found is due to tube element heating after warm-up.


Phil - AC0OB

Hi Phil,

How stout would the PS supply need to be here for the HV version. Say 1800V at 350ma be enough? I have an 8000 laying around that would be perfect for this I think (after the 813 rig is done, of course...have to stay on track).

John

Back on Oct 16, 2016 in another forum, Nico PA0NVD(SK) posted a cathode modulation scheme using a MOSFET that he built and tested. His design references the PA tube grid to ground, not to the PA tube cathode. This has the effect of modulating the grid along with the cathode. I don't know if this method has had any further development, but it does look very promising.

https://antiqueradios.com/forums/viewtopic.php?f=12&t=309046

One needs good PS regulation for any transmitter design.

I would think something like this should suffice.

In a CM system the average plate-to-cathode voltage across the final is about 0.70 X PSV to 0.8 X PSV depending on the tube gain.

The grid sees the same variable DC as does the cathode, minus the grid-leak bias. In a CM system one also needs some way to supply a protective bias in case RF drive fails which is why I included a zener.


Phil - AC0OB

I am a Hammond distributor, and am thinking a Hammond 284 X - same that I am using in the Nico Amplifier. It is CCS rated at 1200 VCT at 230ma  and I should get 600 volts at 430 mA or so with a two diode FW circuit. It also has two 6.3 V at 4 A  filament windings and a 5 volt winding that I won't use. I'll probably use another small transformer for the screens with the same screen regulator that I have used in the past. I think this transformer is enough for two 4D32s. Probably used a 10H 500 mA choke in the 600 volt supply to help regulate the voltage.

BTW - If you  need Hammond transformer , chokes,  or chassis email me with your best price and I will try to beat it with the items dropped shipped from Hammond to you.

Pat
N4LTA

The plate current required is about 440 mA@600V for a dual 4D32 and with the screen current of 50 mA that's 490 mA total. You need 7.5A@6.3V for two 4D32 filaments in parallel.


Phil - AC0OB

I see the with tetrodes the screen is fed with a large dropping resistor as in Plate modulation. Can this be done with a choke and a seperate power supply?


Pat
N4LTA

As I stated earlier, "One thing about cathode modulated transmitters using tetrodes is they need solid screen voltages, low modulation stage series resistances, and plenty of RF drive."

RSG1 is not a large dropping resistor. It is a resistor to meet the screen voltage and the required screen current needed for changing plate current per the Raytheon tube specs. The electrolytics are there to smooth out the voltage variations, to try and keep the Vsg constant.

That is, we don't want the screen to self-modulate.


Phil - AC0OB




 

Then using the Hammond 372KX transformer  (600 volt CT at 575 mA and 6.3V CT at 9 amp) with a full wave bridge should work?

The old "economy style " connection by using the center tap as a full wave half voltage (300 volts) connection can work as a screen source eliminating dropping resistor heat has  worked well for me.

Thanks

Pat
N4LTA

I'm still waiting on some other parts to arrive, but here's what I've got so far.

I'll have another meter to measure grid current. Audio gain and power control knobs will be below the meters.

The grid tank capacitor will be mounted about in the middle of the chassis and the 1/4" shaft will run up to the front panel as well.

Not being picky, but I did note that the screen dropping resistor for the KT-88s is shown at 1 watt. It should be at 5 watts or so.

Jon

Looking like it's coming together well.

Yes good catch, the dissipation is 2.97 Watts so a 5 Watt 3.3k will be needed. Bad holdover from another design.


Phil

Oh yea that should give you at least 100 mA of headroom. I am trying to envision your economy power supply. Have a drawing or a link?

I had plans one time for a dual 4D32 but never punched the metal for it. Should be good for at least 175 watts out.


Phil - AC0OB

I have used this connection many times. It was once called and "economy" connection.

Pat
N4LTA

ok, my foggy brain now remembers that one. Should work well.

Here was the PS design for the transmitter that I never built. Sorry, it was to use a competitor's transformer.  :-[


Phil - AC0OB

I'm still waiting on the variable cap for the VXO to come in, and some knobs, but other than that the metal work is done.

I have distinct separate sections for RF, audio and power. I installed a shield to protect the audio section from the RF. I also floated the 12 volt transformer for the 7984 filament using a piece of teflon. It seemed like a good idea.

I added a separate switch to lift the ground connection on the full wave rectifier bridge so that the tubes could warm up before applying HV.

Jon

There was enough room to put another wall in between the grid tank and the pi output network.

I tell you what, for PW and a rig that requires very little iron, real estate is still at a premium even with a large chassis.

Jon

All the drilling and blasting is done. All that's left to do is the wiring.

I haven't bought the audio tubes yet.

Jon

Most of the left side is done.

I used a large grommet hole to run the solid wire through in the center to attach to the plate/parasitic suppressor.

The more I build this rig, the more I'm excited about it and how it would perform paired with a big linear. I could build another triple 3-500 linear to be cost effective, but an Alpha or Acom would look really pretty.

Jon

Be careful with regards to grommets.

I purchased some from harbor freight for feedthroughs on the posts of a 70A 12v power supply.

Plug it in, turn it in and it's under load.

Take resistance readings from each post to each other and to chassis.

I have resistance! And not much of it.

Using an ohm meter I find a hundred or so ohms.

Using my VNA I find it's conductive to the a couple kHz to 1.4 ghz.

I don't use those grommets for anything but shielded wire now.

--Shane
KD6VXI

Thanks for the tip!

Yeah, Harbor Freight anything is bad news. Look at my holes for the larger tubes on the underside of the chassis. That was a Harbor Freight step bit. At a certain point the bit lost cutting ability and I finished the hole with brute force and the metal folded in from heating up. Anyway, it’s through... right?

Large washers on the top of the chassis made it clean up nicely.

Jon

It will be interesting to see what the DC voltage division across the Final and Mod tubes are after you get it set up.


Phil

Don't know if any one mentioned it but you need some B+ for the osc.

Terry N3GTE

Yes, and actually I had to make a few resistor changes in the original provided schematic since my lower B+ is 375 vdc and not 350. I hand drew a complete schematic yesterday. The oscillator supply is from the 375 vdc center tap and uses a series dropping resistor.

Jon

Hi Jon,
I am reading your schematic over coffee this morning, and I have a couple questions.

First, why no audio bypass caps for the last audio stage cathodes (and maybe why such a small bypass on the first stage)?

Second, the word around here seems to be that the screen bypass on finals should be smaller, like 0.001uF, to avoid bypassing the higher audio frequencies along with the RF. I’m not sure what the reactance target is supposed to be or how to determine the AC impedance at the screen so that one could choose the bypass to be small at RF but big at AF compared to the screen’s reactance in a given circuit, though.

I’m building my knowledge here....not questioning your decisions. Please explain so I can learn.

Ed

I'm building my knowledge too. To be honest, I was relying on Phil's schematic for those components. I'm not sure what changing their values would have on the circuit.

Jon

Modification to the schematic:

I'm skipping the voltage regulation on the 7984 screen and just using the series dripping resistor.

Jon

I just finished.

Now I just need to wait for the audio tubes to arrive and test.

Jon

In post #10 I explained the reason:


And in post #29 I stated:



Phil - AC0OB

The tubes came in and it is making RF and audio.

On first try, I was getting between 15 and 20 watts carrier with low power settings (series pot at the KT-88 cathodes at maximum resistance and close to fully meshed loading capacitor). The problem is that carrier power drops off slowly when transmitting and audio isn't completely clean. Audio pickup is definitely sensitive since I could hear in the headphones low background noise in the room. This was even with the audio gain control near minimum.

I am getting a definite plate dip with the plate cap about half way meshed. I do need to remove some turns from the grid tank inductor since the variable cap is fully unmeshed. Grid current is barely registering on the meter. The only way I could tune the grid tank was by monitoring power output. It drops off if you move off the fully unmeshed position.

I'll give it a rest for the night. First try was pretty good though since nothing blew up, there's close-to-expected RF output and it does modulate.

Jon

I think I may have found the source of the slowly dropping power problem. The plate choke coupling capacitor is only rated at 1kv and is getting really hot. I knew it was too small but I didn't have any others on hand and I wanted to test the rig.

I have removed turns from the grid tank inductor 3 times so far with maximum power output still being with the cap plates fully unmeshed. I don't want to remove any more turns because I'm starting to think the inductor is now too small for 75 meters. During the build, I had calculated the inductor value for 52uH.

But, I'll fix the plate choke cap first and go from there.

Jon

You'll get there.  :)

Can you show us your overall schematic, minus the power supply?

You may need to scope the audio chain if you continue to have distortion.

Also, it's important to check the screen grids to make sure the voltages don't vary more than about 15%. If so, you may need to up the cap values.    

Phil

I'll definitely get there with this rig. Sometimes you build something and it just seems to click.

I'll post a schematic after I do my next iteration. One thing I'm changing is to remove the B+ on/off switch. Things changed mid way through construction and the switch now basically does nothing. I'm putting a green neon lamp in its place.

I'm taking my time with this now though. It has a heartbeat so now I can just putter with it. I sort of blew the budget building it... new tubes, transformer, chassis, parts, etc. The watts per dollar ratio is a little out of whack. I'll order the next round of parts in a couple weeks.

The things we do to talk into a microphone and look at a filament. It's like a caveman and fire thing. We're no different.

Jon

Atook lava alana

Well, it keeps us out of trouble, most of the time!   ;D


Phil

I found the problem for the low grid current indication. I had a 10 ohm resistor in parallel with the Chinese meter. I assumed that a current meter would have a very low resistance, but this one turns out to be 25 ohms. I cut the 10 ohm resistor and everything behaved as it should. I'm getting close to 3mA of grid current with a defined dip being when the plates are fully unmeshed. Still not proper, but I am getting enough grid drive to the final.


I'm still getting the power drop-off problem though during key down. I was daring and let things bake until power dropped to zero. That way I just needed to follow the heat. I found it and burned the back of my finger. It's the plate choke, marked 102k in the picture. Its value is 1mH at 1 amp, but has a ferrite core. I'll replace it with a regular wire wound one.

Jon

I replaced the plate choke as well as the two in the oscillator circuit with pi wirewound ones. The RF section now completely behaves. I have unlimited grid drive available with the grid cap tuning about midway for resonance. I have it detuned a little to give me a little less than 4mA of grid current on the 7984 final. Power output is totally stable with the minimum output power being 20 watts carrier. I can get more power by reducing the variable resistance in the cathode of the KT-88's or reducing loading capacitance. Right now the loading cap is fully meshed and the cathode pot is set for maximum resistance.

The audio section is going to need some troubleshooting. When I tune the plate cap for minimum plate current, the audio output seems to be cutoff. When I detune a little off resonance, the power output drops a little but the rig then begins to modulate. However the audio is distorted and average power output drops with modulation (you'd expect it to deflect a little higher with voice peaks).

Also when I crank up the audio drive pot, I get a screeching sound in my audio. The same thing happens when I detune the plate cap too much.

I've tried backing off and driving more grid current but that doesn't seem to have any effect.

Jon

Working from home is a dangerous thing. You have time to ponder stuff. I'm going to switch gears and plate modulate this.

I'll use the existing power transformer as the new mod transformer (using Tim's conversion) and I'll get a 600vct transformer for power. I'll use a Ranger's audio section as a template (12AX7 -> 12AU7 -> KT88's).

I'll run 600 volts on the KT-88's and 300 volts on the final. I'll have a negative peak limiter so the whole thing will give me plenty of audio.

The dual KT-88's were sitting there and asking for it.

Jon

I have the power transformer converted to a mod transformer with a properly phased negative feedback winding. The transmitter is sufficiently disassembled to rebuild as plate modulated. The power control knob hole position works out well to put another tube socket for the 12AU7.

To determine the phasing of the negative feedback winding I put a 2k ohm resistor between the B+ and final lead to simulate the resistance of the RF stage. I then put 6.3 vac into the leads for the two modulator tubes along with oscope leads for channel 2. I put the channel 1 leads on the 6.3 volt winding of the transformer. I then just had to compare the two waveforms to see when they were out of phase.

I think that's all right... correct me if I'm wrong.

The new power transformer and interstage audio transformer will arrive today with all the other components coming in tomorrow.

Jon

Almost done. Just waiting on a few small components for the audio section.

I'm curious how a Hammond power transformer will work as a mod iron. It is unbalanced, so who knows. If it sounds like crap then I might need to add a heising circuit. There just isn't much room to squeeze in any more transformers.

Jon

Hi Jon,

Looking good!  That rig will sound great once you optimize it and get the bugs out.  Audio negative feedback around the modulator is a good thing.   That was the biggest audio improvement I did to my 4X1 rig.  Plate modulation will be a challenge but give it a try.

I like reading threads like this that document the whole project as you are building it. 4,000 hits now will probably turn into 40,000 in a few years as it captures the web search engines.

You probably have quite a few rigs built up by now. You've come a long way in the last 5-10 years as shown by your hand-drawn schematics and construction work.  I remember your first 3-500Z linear and other early projects.

Continue to post evolving performance and rig features as you improve this transmitter and get it running right.  It's all a matter of constant rig improvement and refinement once it's running. That can be the hardest part of all.   Keep up the good work, OM!   I'll be following the threads.

BTW, what is that little enameled copper wire coil on the loading cap?  Is that a HV safety choke using a ferrite core? If so, I've never seen that done with ferrite, just air, but it shud work FB.

Tom, K1JJ

Thanks Tom!

Yeah, I thought about replacing that ferrite choke on the loading cap with an air wound one but it seems to work fine. Since there's no current flowing through it, it just sits there. If the DC blocking plate cap were to fail, then yeah, it would take the brunt. But, it could be good since the wiring is stiffer with a ferrite choke. It would get hot and pussy (pusss-eee) really fast and things would start to fail catastrophically at that point. Probably the fuse would blow pretty quickly.

I actually don't have that many rigs built up now. I could kick myself for tearing apart most of the things I've built, like my triple 3-500 amp. I'd like to have that right now as a nice companion for this transmitter. It's ok though, I was just talking with my local parts supplier (Dennis, W7TFO) who always has a large stock of iron. My rule as of late has been to keep everything I build, no matter how small. The problem was so bad that I actually got on the AM Rally this year with a barefoot Icom 7300 (how embarrassing is that...)

AND, after months of waiting, my HOA has granted me permission to stick my 75 meter dipole up again! I was fine for years but all of a sudden they had spotted me. I tore it down and hid the non-resonant dipole on the 2nd floor eves of my house with a low powered feed point tuner. Then, lo-and-behold I got the permission to put the old one back up again. Yaye...!

I'm rambling and off topic.

I'll post the results of this rig fairly shortly. I'm feeling good about this one!

Jon

Finished.

I'm getting 12 watts carrier and modulation peaks show 40 watts. The audio is a little distorted but definitely sensitive. I can hear background noise in the room on my headphones. I'm sampling the RF output and sending the audio right to my headphones.

I'm flat topping somewhere. When I whisper in the microphone everything is perfect. As I speak louder it starts to distort.

I need to putter and find where the distortion is coming from, but this rig will definitely be a keeper. I'll draw a schematic in a little bit, but here's the skinny:

300 volts B+ throughout. KT-88's are connected in triode mode with a 100 ohm resistor connecting the screen from the plate of each tube.



Jon

It looks like there's clipping on the output of the 12AU7 driver stage. The oscope picture shows the output (pins 1 & 6) in blue and the input (pins 2 & 7) in yellow. This is with no negative feedback. I connected the 'cold' end of R27 (grid resistor) to ground rather than the negative feedback lead. I wanted to eliminate the mod transformer as part of the problem.

The waveform is a 'Helloooooooooo.' I don't have an audio generator.

The schematic I'm using for the audio section comes from the Ranger. The only things I did differently was to change the audio coupling cap (C52) to .022 and use KT-88's in triode mode for the modulators.

Also, secondary tap of the driver transformer is grounded and I have the positive bias instead in the cathode of the KT-88's.

How do I move the bias point for the tube to give me more headroom?

Jon

...I've tried both raising both the grid and cathode resistance (separate times). No change.

One try, I doubled the grid to 300k. I then went to 2.4k on the cathode and then 1.6k.

Like I say, quiet audio is undistorted. As I speak louder or raise the audio gain pot, it starts to distort.

There's plenty of audio power as I'm getting excellent PEP peaks on the wattmeter.

Jon

Is your driver section capable of sinking power to the triode connected kt88s?

I've only used the kt88 single ended....  Replacing a 6l6.  Few circuit changes and it came right up. 

Other than that I don't have a lot of experience in those radios.  But I know in tetrode mode power sink is small vs triode mode requiring real power.

In a linear triode circuit you need a low q input network to provide the flywheel effect on the driving signal.

Could the lack thereof be causing your problem?

Again, not familiar with your circuit or the ranger very much.  So I could be off in very far left field.

--Shane
KD6VXI

That makes sense. I didn't think of the 12AU7 being loaded down.

I could drive the KT-88's as pentodes and that would make driving them easier. I'd need to up the B+ to 600 volts and run the screens at 300 volts. The 7984 would be begging for mercy at that point.

Instead I'll reconfigure for a couple 6L6's and see what that does.

Thanks!

Jon

I wrote this yesterday but it must’ve gotten lost on the way...

One thing I see in your scope photo is the result of triodes being inverting amplifiers. The plate trace is inverted relative to the grid. This makes it hard to see what’s really happening with an asymmetrical waveform like your voice. Your scope should have an invert setting on each channel, flip one and then look at it. It’s hard to mentally invert an image and compare it, but when I try I see some rounding, not flat top clipping. Is it possible you’ve misidentified the source?

Also... as an anecdote, I spent a week of hair pulling and circuit mods on a tube amp I built once....and then finally, while changing the circuit values yet again, realized that the input stage’s grid resistor ground end was in fact floating...resulting in high current draw, low plate voltage, limited output, distortion...in the end there was nothing wrong with my original design, just the dummy who assembled it. Check for simple stuff before you give up on it....

Ed

I did notice that the output waveform was the inverse of the input, but spaced on hitting the invert button. It would've made it a lot easier to analyze.

One thing I tried but didn't mention was to temporarily put a modified Heising circuit inline with a 20H choke and cap. I wanted to rule out the modified mod transformer being the issue with unbalanced current,  but it didn't make a difference.

I didn't realize it until after I looked up the data sheet but the original 1614 mod tubes were a version of 6L6's. I'll give it a go and see what happens after I put the circuit back to original from all my troubleshooting butchering.

Jon

Actually, I think I just figured it out and can maybe use the KT-88's after all.

I took a measure of the DC voltage at the cathodes of the KT-88's and it's only 0.28 volts. The grids are biased at ground since the center tap of the interstage transformer is grounded. There's not much voltage differential between the grid and the cathode for audio to go.

I'm going to bias the cathodes positive with a power supply.  I'll use the existing 12 volt filament transformer for the 7984 filament. I'll parallel in a voltage doubler circuit and inject >24 volts DC onto the cathodes.

Jon

Hold it - am I missing something? What happened to the series modulation? I was just following the thread because I also also have tried it and never have been successful, and I figured if any group can make this happen... I mean you see all kinds of schematics and threads about it, but somebody show me a proven circuit. I have made it work with small tubes like 6AQ5s and 50C5's stacked but...   

Good point. Mine wasn't working too well as a series modulator. I think to make it work you need to be a guy like Mr. Carlson's Lab on YouTube. He's smart, but he can't pronounce the word, 'solder.'

I'm more of a guerilla with a soldering iron, so plate modulation works well.  The schematic I used was from 1957.

Jon

Ahhh....

That Ranger schematic has bias, and the center tap is not DC grounded --it provides the bias connection. The cap grounds it for AC.

Another way to do this without another PSU is to add cathode resistors to the KT88s -- cathode bias.. Data sheets suggest someplace around 40v bias and 75mA per tube.... a couple of 470 ohm resistors will get you kinda close enough to see if its going in the right direction...maybe a pair of 1k 3w units in parallel per tube to keep them from burning up.  That should be a fast test for you!

Ed

I have successfully cathode modulated three transmitters, a 6DQ5 version, a Globe Champ 90 with dual 807's, and a vacuum Tube Broadcast transmitter so it can be done, but solid stating the modulator is the most efficient method IMHO.

Phil - AC0OB

The mosfet idea is awesome. It must be operating linear mode so you would need to balance a few things out to get the parameters optimized and a heatsink! Well basically the same with a tube I guess. Can you dredge up a valve circuit modulating a couple of 807's?

I assume you mean dual 807's modulated by the Mosfet circuit. We did that with the Globe Chief 90.

If you are referring to Tubes Modulating tubes then the dual KT88's modulating dual 807's is your cup of tea.

The Mosfet solid state modulator is the most efficient.


Phil

 

 

Phil, can you comment on a couple things here?

1) operating/key down bias comes from where? Grid drive? The bias voltage while operating seems to be based on the 75v zener. Is the capacitor in parallel to stabilize the bias or allow audio to go around the zener?

2) you show plate current as 13mA less than cathode, yet the screen is shown providing 16mA. Is that an error, or is the remaining 3mA coming from the grids, and if so, isn't that low for a pair of 807s?

3) how scalable is this circuit, plate voltage-wise? Does the FET need to withstand the full plate voltage, or some known fraction? If I stick an 11N90 below an 813 at 2100v B+, will it reward my folly with a load of shrapnel?

Ed

Phil, can you comment on a couple things here?


Operating control grid bias comes primarily from the exciter's RF drive via grid-leak bias.

The short explanation is this. The Mosfet has a very low RDS or Drain-to-Source resistance at high audio level peaks so the cathode of the finals are near ground at this time.

At the other end, when there are audio valleys, the RDS is high allowing the final cathodes to go to approximately 0.35XVp for dual 807's. The high positive cathode potential represents a very low grid bias voltage. In some situations, the combination grid leak bias + the cathode potential can exceed the control grids' negative voltage spec. The "PB" components influence the total control grid voltage at the "valleys" to insure that the control grid's negative voltage spec isn't exceeded.

The power supply furnishes KeyUp bias which was a replacement/holdover from the Globe Chief 90 circuit. As such, there is at least -100 volts on the key circuit at Keyup. The 6.8k 2W resistor R2 does limit current at Keydown so as not to tax the Bias supply. This bias could also be used to cutoff exciter stages with proper dropping resistors.


The figure of IP originally quoted was a holdover from the Globe Chief 90 design in which the owner wanted slightly less currents. In this updated design, Ip has been corrected and is Ip = IK - 24 mA. Screen grid current = 16 mA and control grid current = 8 mA. Since the metering is at the Mosfet Modulator's Source, actual Ip is less.  


This circuit is good for 6146, 807 and similar type finals where the plate voltage doesn't exceed 900 volts. See Reply #16 for a Mosfet circuit for higher voltage tubes. This is the one I will be using to modulate my new 4-400 transmitter. An updated circuit with extra voltage protection is also included below.

BTW, the level of voltage to which a particular cathode voltage circuit may rise is dependent on the tube gain, the number of tubes paralleled, and final loading and may vary between 0.25XVp and 0.4XVp, depending on the circuit


Phil - AC0OB

OK, that clarifies things -- mostly...  I wasn't thinking of controlling grid to cathode voltage limits, I was thinking about how bias was working.  But if voltage between elements is concerning, the specs say heater to cathode is limited to 135 volts...and the cathode will be 200+ volts and the heater is grounded. Do we need to use a separate filament supply with the center tap connected to the cathode here?

Ed

Yes, if the cathode to filament potential may be greater than the spec I always use a separate center tapped filament transformer. For example as shown here, there is really no cathode in the 4-400 but this is a method to connect the modulator's Drain current to the filament:

Phil - AC0OB

I think I found it. After trying everything with the shotgun approach, I gave in and pulled out the scope.

There's flat topping on the 2nd stage of the 12AX7 (mic preamp tube). I took this reading directly at pin 6. I took another reading after C52, the coupling 500pf capacitor (I used a .022uF) and it's always a perfect sine wave. The amplitude does decrease as I go way up in frequency due to C53 but the wave never distorts.

Jon

Jon,

   Just because EFJ did it, does not make it right.

Looking at that circuit, as I recall you wanted to test it with the NFB disconnected. A good choice.

I just looked at the R-C Amplifier section in the 1973 RCA Receiving tube manual for a 12AX7 with 300v B+.

One concern is that RCA recommends the following stage grid resistor (R27) be 2-5X in value of the plate load resistor (R23). In your case it is 1/3X (150K/470K)which puts a nasty load on that stage. This not only reduces the gain, but also limits the undistorted P-P V output.

Suggestions:

Change Rp (R23) from 470K to 220K

Change Rg (R27) from 150K to 470K

Change Rk (R22) from 4.7k to 2.8k (2.7k close enough)

These values are from RCA, Page 645.

Jim
Wd5JKO

Sounds good, thanks!

I’ll report back in a little bit...

Jon

EFJ's speech amp is very nonlinear and their attempting to drive a 150k load with a second stage using a 470k plate resistor was ridiculous.

If you want to keep the feedback winding and still drive the driver stage with plenty of gain this is my preferred speech amp circuit:

Phil AC0OB  

I had enough resistors on hand to make it happen. I made R27 440k instead, but that's what I had available. It all seemed to help but it's not the complete answer yet. I'm still getting clipping.

Something strange is happening too. When I inject the 1000Hz tone I can hear it coming from something in the transmitter, but I haven't been able to locate the component(s) that are generating it. The level changes as I change the audio gain control so it's something after the first stage. It happens during transmit and receive.

I'm not opposed to scrapping EF Johnson's plan and going with something else. The only part that's concrete right now is plate modulation. Everything before that is up for grabs. It wouldn't take too much effort to switch to Phil's schematic.

Jon

Hi Jon,
I wonder what you are using for a scope probe.  It looks like you have a regular piece of RG-58 leaving  the scope input jack.  Are you just putting a coupling cap on the end of a 50 Ohm coax?  If so, this, would be a 1 Meg load with 120pf in parallel.  This is not good, but it would be best to have a 10:1 probe and use the dc coupling rather than ac coupling you have selected so then the true plate voltage is known.

  If your scope channel is not inverted, I have two interpretations of what is going on. (referencing your Ranger schematic) One is that the plate is going to cut-off on the 12AX7 but I doubt that.  You have 50 Volts peak-to-peak swing at the plate and the 12AU7 grid should not need more than 18 V peak to peak in the normal application.  I  think your test is way overdriving the 12AU7 and you have grid current flow on the 12AU7 providing a clamp on positive plate swing of the 12AX7.  Pull out the 12AU7 and see if the clipping goes way (and more swing).

Anyhow not having followed the whole thread at all,  I don’t know what is your problem, but I think you are getting off the track with the overdrive in the present test.

My 2 cents worth,
GL OM.

I am using a proper scope probe but I did switch it over to DC coupling, as you suggested. Plate voltage on the 12AX7 2nd stage settles out at about 213 volts.

Maybe I am overdriving the mic jack more than normal use. I'm attaching a YouTube video link to show what I'm talking about. This is after Jim's modifications and the clipping has gone down quite a bit.

When I pull the 12AU7, the remaining little bit of clipping goes away. The audible tone goes away as well, so something doesn't seem right with the next stage.

https://www.youtube.com/watch?v=cd5zrPRniv8 (https://www.youtube.com/watch?v=cd5zrPRniv8)

Jon

What are the DC plate and cathode voltages on each 12AX7A triode and on the 12AU7A driver stage?

Put about 50 mV p-p into the mic connector and measure the p-p plate voltage on the first stage, and then while scoping the second channel, turn up the mic gain and determine at what p-p voltage it gets to just before flat topping.

Phil

I was taking some initial voltage readings when I powered the radio on again and something started to groan. I'm not sure what it was so I powered it off.

I sat for a few minutes and stared at this thing.

I'm going to change gears and make it a hybrid. I'm going to install a solid state audio driver in the chassis to feed the driver transformer. There's 12 volts inside that I can use to power it. The KT-88's look cooler so I'll put those back in. I need to get creative on those two small tube sockets, but I'll put something there.

Onward to Phase 3

Jon

Hi Jon,


Especially when using plate modulation...

The tones you are hearing in the rig are often caused by the plate coupling cap, audio transformer talk back or other spots that torture parts. But to be sure, get an insulated piece of plastic pipe, maybe 3' long and 1" diameter. Put it to your ear and listen to the individual transmitter parts while the tones are causing the noise. You will find the source quickly.

Another technique is to use a long insulated fiberglass/ plastic rod to touch the components. You will often hear a pitch change or amplitude change when the noisy part is touched.    (by changing resonance mass or inducing the megaphone effect)

Sometimes it can be a very small sustained arc. Put a big mirror behind the rig and turn out the lights. Sometimes you will see it in the dark.


Be careful, of course.

T

Geez Jon you change circuits more often than I change socks.  ;D

If you do use those KT88's as the P-P modulators you'll have to reset the grid bias and change the screen voltages.


Phil

Jon, you embarked upon a series modulated PW rig, nice, neat self-contained package.  I understand why you dropped the series modulation and went to Class B push-pull plate modulation, probably a wise move.  You just have a few things to work out to get it running, probably much better than the Ranger audio circuits you are modeling with your design.  You are getting closer, probably almost there, so it would be a shame to require outboard audio to complete the project if it is not absolutely necessary.  

I have been a firm believer of testing a stage-at-a-time when puzzled by unexplained phenomena when first testing an new rig in its entirety.  You haven't mentioned whether you are trying to make the modulator and final work while you debug the audio circuits.....  If you are, it might be easier to pull the HV rectifier, and fine-tune the audio driver stages, then add further stages to the mix.  You posted the Ranger schematic as your reference, and I don't know whether that is representative of your power circuits for your modulator and RF final.  But assuming the HV rectifier supplies both plate and screen voltage to the final and modulators, there should be no risk in disabling the HV for the audio tests.

If the 12AU7 stage is being over-driven, or if perhaps the bias on that stage is not sufficient, then your squeal with increased audio gain may be coming from the driver transformer with non-linear input.  I am not sure whether you plan to run the modulator tubes AB1 or AB2, but assuming AB1, the tests for the audio stages should produce the same results whether the modulators are active or not.  If they are pulling grid current, then measuring with and without the modulator tubes in place, (no plate or screen voltage required) might shed some light on the issues.

As previously suggested, test the 12AX7 stages and thoroughly resolve any probems there before adding the 12AU7.  Then add the 12AU7, and if you see flat-topping at the input, the cause may be either over-driving the stage, or insufficient bias on the 12AU7, causing grid current to flow even though the peak-to-peak signal level on the grid is within the proper range.  Thus it is important to have an idea of the expected levels at each stage of the audio path.    If you find the levels are correct, and pull grid current on the 12AU7, try increasing the 12AU7 cathode bias resistor value, thus moving the operating point and allowing greater input swing before the grid goes positive with respect to the cathode.  After solving this, you should have a clean signal at the input to the 12AU7.  You might need to elevate the plate voltage for the 12AU7, using the HV supply with a larger resistor, and capacitively couple the plate to the driver transformer primary, thus taking the saturating DC current off the driver transformer primary.  Your squeal may be resolved by these steps, as they may be caused by over-driving the driver 12AU7, and the noise emanates from the driver transformer.  I think you mentioned unexplained noises and growls, which should be rather easy to identify the source.  

Now that the driver stages are working properly, move on to the modulator tube stage, still with no screen or plate voltage, and inspect the modulator grid voltage swing, with AND without the modulator tubes.  With no actual circuit to review for your build, I do not know if you are using fixed or cathode bias for the modulators.  If cathode bias, then you will need plate and screen voltage for the prior test, lest you have no grid bias and you will surely draw grid current on positive grid excursions, sans any bias, rendering this test irrelevant.  By the way, KT-88s seem overkill for this application, and 6L6s, or 1614s should be sufficient.

In summary, taking the troubleshooting and fine-tuning stage by stage should enable you to identify issues and resolve them by focusing on a smaller section of the entire project at one time, then moving ahead with each additional stage with confidence.  I hope you do not have to shift gears and 86 all your work invested in the audio stages.  I think you will find you do not need the excess baggage of a separate audio amplifier! 

This could become a very nice larger-sized PW transmitter if you follow through as originally planned.  All this may seem long and involved, but, taken one step at a time, it should greatly simplify your tasks at hand.  GL OM!

I am changing the plan a lot but this is getting fun now! On one of my first rigs, I used both push-pull KT-88's and 6L6's interchangeably by connecting the grid to the screen. It's a lot simpler and it sounds great! No grid bias voltage, no screen voltage.

I can appreciate the step-by-step method to get this working. I would've eventually gotten there but it was getting too tedious. This new plan is adding more excitement. And it will still be a self-contained transmitter without any external parts.

I'll keep the 12AX7 tube as the mic preamp and then drive this:

https://www.digikey.com/products/en?keywords=amp2x15 (https://www.digikey.com/products/en?keywords=amp2x15)

I'll use a backwards driven audio output transformer to then drive the KT-88 control/screen grids.

I also found a use for the previous 12AU7 tube socket. It'll be a glory hole with a countersunk 1/4" jack to break the KT-88 cathode ground connection for current measurement.  It also makes a convenient chimney to let the smoke out if that ever were to happen.

Jon

The input impedance of the Class D amp is rather low so one would need a speech amp similar to this since the output impedance is low enough to drive a 68k load.


Phil

Phil, not a cathode follower output??

Ed

The output impedance of the second stage is about 40k.

Here is the problem, if your mike only produces about 50mVp-p then the p-p voltage at the plate RP1 is only 3.35V p-p.

A cathode follower stage has a voltage loss of about 30% so at its cathode is only about 2.25Vp-p, not enough to drive most modulator or driver tube grids.

With 50mV input and the mic gain all the way up, you can get a maximum of 115Vp-p at RP2 and that's more enough to drive the Modulator or driver tube grids.


Phil

It's working and the audio is crystal clear. It's lacking a little in bass so I'll putter with that to see what I can do. Attached is the schematic I used to drive the solid state audio driver mounted on the side of the chassis. It's the Ranger's 12AX7 first stage mic preamp driving a cathode follower.

I'm using the 4 ohm tap on the driver transformer. I might try the 8 ohm.

Maximum power output is 14 watts carrier which is a nice level to drive a linear.

Jon

I played around with RC combinations to the input of the audio amp to bring the lows back in. It worked but there's a slight bit of distortion in the audio. I'm thinking that the root of the problem might be the unbalanced DC on the 'modulation' transformer. I'm going to experiment with putting in a modified Heising circuit. This rig will probably end up with enough iron and tubes on it that I could fetch $2 grand from the audiophile guys.

I'm also going to remove the KT-88 current tap point and put a line level RCA jack there instead. Measuring modulator current really isn't necessary on a rig this small and having an option for line level input makes sense.

BUT, I literally ran out of solder this morning. I'll play some more this coming weekend when my shipments of stuff come in.

Jon

And the finished product.

I added the modified Heising circuit and it helped the distortion a lot. I did need to cut back on my bass boost RC combination after I did because at that point I had too much bass.

After making so many circuit changes, pin 6 of the 12AX7 snapped off when I was making a change to the cathode follower. I had to solder onto the stub with a short piece of stranded wire and make a junction point for it on one of the terminal posts.

Anyway, this was a fun project and this will be my main 75m AM transmitter. I really would like to pair it with an Alpha.

Jon

Congrats on getting it built, Jon!  Mucho work and you learned a lot.


But...  "Finished Product?"   "Pair it with an Alpha?"    Not so fast.    ;D    Now the honeymoon is over and the hard work starts to see how clean it is and to optimize it.  Next: Some formal, systemized data tests.


1) Run a 50 Hz triangle audio wave thru and see how it looks, input compared to output.

2) Also, run a 50 Hz square wave thru to see how it handles it.

3) Then run a 2-tone  test to see how clean it looks up and down the band on your SDR spectrum display.

4) Do a sinewave audio sweep from 30 Hz to ~10 KHz and see how true the sinewave stays and how flat it is across the audio spectrum.  This will let you EQ it with confidence knowing that there are no band-aids covering up problems.

You may find you need some work before amplifying it thru a big amp. The 2-tone will tell a lot.    Once you get the results, there are a list of things you can look at to improve things. I can help you zero in.  (negative feedback, heavier loading, voltage stability, capacitive shunting, general IMD improvements, etc.)

Post the screen shots and data results of the tests above and put your BAs on the table, mate!

T

True, we are never really 'finished' with any rig. There's always a tweak.

It sounds good to my ears, with the sampled audio coming directly from the antenna jack and into my headphones. But my ears aren't calibrated test equipment either.

I'm going to compare the mic input audio (goes through the 12AX7) and the line level input (bypasses 12AX7) and see what if any difference there is. I just had to order a connector to make that happen (long story).

In the meantime, I'll keep browsing Alpha's website. They're not shipping right now, by the way. Corona scared them off.

Jon

that "12AX7" you're holding looks mighty nice...can you say a little more on how you're using it ? bridged mode ? what is the driver xfmr ? the modulation xfmr ?

and from what I've read it also has a mute...you're using it ?

Peter

I have a description at the end of the page:

on.radio/hamradio  (http://on.radio/hamradio)

I've attached the data sheet.

Jon

Polished aluminum, who knew?

I've got some custom engraved brass labels coming too.

I had it on the air a few days ago and talked with a couple guys in Tucson.

Jon

The brass labels finally came in. It's a nice finishing touch rather than using P-Touch stickers.

Jon