The New WA1GFZ Super Valiant

[K1JJ]

I figgered you Johnson enthusiasts (and homebrewers) would get a kick outa Frank's modified Valiant.  It's like a Pro-mod race car.  As you can see he used four 6146s for the finals, four 36LW6 sweep tubes in p-p parallel as modulators, a homemade mod xfmr, full Teflon wire rewiring, all ceramic tube sockets, super-strapping tank components and bandswitches, etc.  It's not finished yet but is running and in the test mode putting out a well-modulated 250 watts of carrier.  He is planning on a solid state audio driver - no transformers.  All self-contained.  (The mod xfmr fits on the chassis when turned 90 degrees)

I will add more pics as he sends them.  I would love to own this Valiant in its modified form!  I really appreciate seeing a rig that has many weaknesses turned into a reliable, optimized workhorse as Frank has done.

Tom, K1JJ

---------------

Here are some of his comments that describe the engineering thoughts as he built it - over the last 6 months.   Those of you who know the Valiant well will recognize the tremendous amount of detailed work he put into it. Frank is a ham's ham when it comes to design and building...

----------------

Frank said:

  W1CKI turned me on to a course load switch wafer that I rebuilt with new contacts. Now I have two wafers in parallel to handle the current. I blew a couple fixed loading caps so built a module with door knob caps. Also replaced the fine loading cap with a larger 500pf Johnson, There is just enough room to install a change over relay next to the connector once I decide what to use.

I used a bunch of transformers that failed the hipot test. They were 2KV output DC transformers that had to pass 5KV. The parts that failed didn't have enough insulation between the primary and core. The primary would never see 5KV so it wasn't an issue. I gave Dean a couple, W1CKI has one in his 6 meter amplifier. I rewound one to make this modulation transformer. The core material was pretty high quality and good for audio. I studied a bunch of stuff on designing audio output transformers and figured this one would work. It will modulate 250 watt output carrier but I will be happy if it plays at 200 watts. I'm running 900 volts on 4 straight 6146s. I will swap over to 6146Bs or 6293s once I'm happy. I still need to design a driver. One option would be to copy the one in your rig. I'm driving the 36LW6 screens with an external output transformer backwards using a sanyo 50 watt module. The control grids are grounded. I have 10 ohm cathode resistors to balance the load and have a 5.1 volt 10 watt zener in the common cathode line for bias. I think I only need about 70 volts of drive because of the parallel sweep tubes. I figure the modulator will do close to 200 watts output. I was also considering a pair of 12B4s as cathode followers. Not ready to put on the air yet still need control relays and the HV is on a variac for safety reasons. You can share pictures if you want. Enclosed is a picture of the final. I rebuilt the whole rig with teflon wire and ceramic tube sockets. gfz

My last day of vacation I made great progress. I did the low drive on 20m modification and now have good drive on all bands. Then I was carefully looking at the loading cap area. As it was just determined I can fit much nicer Johnson 500pf variable cap in place of the stock cap. It has just enough clearance to get by the new dual course loading switch. The corner of the chassis left open I plan to build a LV supply. Maybe 28VDC for the relays. I blew up a couple padder caps the other day. I have some 470pf Russian door knobs that may be pressed into service.

I would be happy to do 200 watts carrier on AM. The HV supply is a bit over 900V, transformer rated for 750W CCS. We will see once I try 6146Bs or 6293 pulse tubes. It will definitely need a fan when it goes back in the case.
The A3898, 13H primary is way too big and the A3894 8H primary is too light. S21 and S22 have high resistance windings and will not like the current. CVM4 is too tall and would need to be sub mounted. Tim was going to trade me a transformer but the primary inductance was quite low so it would haul big current below 200hz. He thought 4 tubes would drive it but the efficiency would suck for low end.

The core I have has high frequency laminations and its size will not require a large number of turns to get good inductance. The present secondary is 2000VAC and measures over 100H, So one half that would be 25H then add the gap spacer.  I would like to wind the primaries side by side so they are balanced. I may make a bobbin out of G10. Another option is to wind all three windings side by side with the secondary in the middle.  I have an area of 4 1/2 by 6 inches for the mod iron where the old LV transformer and one rectifier tube sat. i'm thinking 1.3:1 or 1.4:1.
This will be the worlds most expensive Valiant when I am done considering man hours. gfz  

Happy New Year. A successful winter vacation. I'm making 200 to 250 watts on each band with straight 6146s. I picked up a second course loading switch to parallel the wafers. I'll start on the audio once I get the step start and antenna change over relay installed. HV is on a variac for testing.
My plan is to start with 4-6DQ5s and move up to 6LW6s if necessary. I have an A3898 for testing but plan to wind a modulation transformer. I'm shooting for at least a 16H primary inductance. I have a nice 500 watt core with good thin laminations from La Pointe days.
The final tuning in my Valiant does not give me peak power output at plate dip. It actually happens at slightly less C. I can't duplicate the same power at dip with heavier loading. It is acting like my Q is low but I am running 15 to 17. My bias is -86 volts increasing to about -117 with drive. My screen voltage is 150 to 160 volts. The plate is 900 or 950 VDC. I can get 200 watts at dip but I can get around 250 watts with a slight detune. I wonder if you noticed anything like that in your 4D32 rig? On 160 meters 1.8MHz my plate tune is about 1050pf and loading is almost 5000pf. It is lower at the high end of the band.
I think I will end up running at 200 watts output on AM. This is with straight 6146s so "B"s would be a bit more power. I would just like to understand why the power limitation.
I just received the wire for the modulation transformer. I can start on the audio section soon. gfz

I just fired up the HB iron 1.37:1 It appears to work fine at 250 watts output. I see power going up on the Bird when I talk. I'm using 4-36LW6s the B+ is 900 volts and hitting 400ma of modulator current on peaks.  gfz

gfz

[K1JJ]

More pics to come later...

[WBear2GCR]

Whoooo...

Smokin'!

Hope Frank runs the sweep tubes in AB2.

Can't wait to hear that.

[K1JJ]

It's been a year.  Frank has made more progress on his Super Valiant.   He will be adding the GFZ MOSFET audio driver when built.

Take a close look at the top and bottom.


(reposted with permission)  Frank said:

"Added step start, TR relay, better mode switch, filament transformer for the modulators, Mounted modulation transformer,  negative cycle loading with spark gap.
Still plenty of room for the audio driver and power supply. Still need to rectify the 36 volt CT of transformer grounded. I'll make plus and minus 24 volts for mic amp and relay power. I plan mic amp on a little board mounted to the shield for the mic connectors. gfz"


T

[Steve - K4HX]

Nice. I need to do something like this to my Globe Champ 300.

[K1JJ]

Here's an update on Frank's Super Valiant.   I especially like the first and third pics.  He stripped it and rewired with all Teflon wire, as well as wound his own mod xfmr with audio negative feedback around the modulators back to the SS driver input. He is running the famous GFZ solid state audio driver board.  I talked with him on the air and it's sounding pretty decent.   He needs to get the fine adjustments finished up and will probably have the best sounding plate modulated AM tube rig he ever had.  I'm pensively awaiting his acid-test scope picture of a 100 Hz triangle wave... :-)


T

[W2NBC]

FB Indeed!

Just goes to show that perseverance and careful design ALWAYS works in the end.
Yes, the three R’s in RRRF design..

Refine, refine, refine!!

Great work Frank

[K1JJ]

Hola Jeff,

Yep, refine, refine.   I took that to heart when you first suggested it a few years ago.   Beforehand, I found myself building a rig and then bringing it to a certain point less than perfection. Then tearing it down and stating over.  My latest rigs have been kept longest and refined until diminishing returns - and it paid off.  Frank is going thru that now.

BTW, I listened to your audio archives of the various AMers on air. These guys sound tremendous!  This is something that should be listened to by beginners who are still figuring out their AM goals and what they need to do to get there.  Your recordings off air are very transparent and make it an easy choice to consider high fidelity AM.

http://www.dealamerica.com/audio/recordings.php

T

[WD8BIL]

Sounds HiHi Fine Business too!

[K6IC]

  "   ...   BTW, I listened to your audio archives of the various AMers on air. These guys sound tremendous!  This is something that should be listened to by beginners who are still figuring out their AM goals and what they need to do to get there.  Your recordings off air are very transparent and make it an easy choice to consider high fidelity AM.

http://www.dealamerica.com/audio/recordings.php   ...  "

Thanks Tom,  for posting the Link those recordings of stellar AMers.   Had never before heard most of those stations.

We usually only hear the Tallest of Ships out here on the Correct Coast.  And,   thanks,   Jeff,   for collecting and providing access to them,   as well.

73,  HNY,   Vic

[KK4YY]

I heard the Super Valiant on air a few days ago. Frank was discussing how he's driving the screen grids of the 'LW6's. He mentioned that he was using a "screen bypass" cap. I think this is ill-advised as the screen is not acting like a screen when the control grid is grounded and the screen grid is driven.

I'd rather use the terms G1 and G2 for the grids instead of control and screen when this is done. So, with G1 grounded and G2 driven, the tube is acting like a triode and there is no "screen" grid. G2 effectively becomes the control grid as it alone controls the varying flow of plate current.

The very low impedance of the source follower being here used could facilitate driving both G2 and G1 with correctly selected proportionate voltages by using a resistive divider between G2, G1, and ground, while driving G2. With both grids being driven in this manner, there's some evidence that the total drive voltage may be reduced and a more linear overall response could be achieved. This concept is discussed in this thread:
http://amfone.net/Amforum/index.php?topic=44390.0


Don

[K1JJ]

Hi Don,

Wow, that is some thread.   Interesting points on the screen...

http://amfone.net/Amforum/index.php?topic=44390.0

I ran thru it and noticed the Tron wrote a dissertation on the subject.  I lean towards what he said.  I have had a lot of experience driving the grid and screen  (G1/G2 tied together) in 813s, 4-1000As and a few others.   Personally I do not think this is as clean as running them as grid-driven  (G1) as tetrodes using well regulated screens and regulated grids. (AB1)   This is the way the engineers designed them to run (best linearity) and hams often change the config to make it a little easier to implement.


But all in all, there is one way to find out which is the best config. All tubes are different from one another and were designed to run a certain way. Experimentation is cool. All one needs to do is run IMD tone tests in all configurations - grid driven as tetrodes as intended -  or screen driven -  or triode connected G1/G2 and see what the 3rd order IMD looks like.

I just ran some IMD linearity tests today on my 4X1 plate modulated rig (using 4X1 tetrode-connected AB1 modulators with solid G1 and G2 voltage regulation). I had some sudden audio parasitic problems (splatter) that I solved today.  Once fixed, I obtained -35 to -40 DB 3rd order IMD results, which is outstanding for a plate modulated rig. I once had the modulators connected as triodes, but the results were not as clean as tetrode-connected.

I find the 813s work very well triode connected but never did a careful comparison against tetrode connected. (must use well regulated grid and screen voltages under full drive or all bets are off)

So bottom line is that I have been bugging Frank/GFZ to run the IMD tests and since you brought up this good point, I will see if he is willing to connect them up as a triode G1/G2  and as a regulated tetrode to see the IMD difference.   There WILL be a difference.  My feeling is the stock tetrode config will be the best since those sweep tubes were designed by smart engineers trying to get the best TV picture fidelity that they could... :-)

I'll let ya know how Frank makes out once he starts testing.

T

 

[KK4YY]

Hi Don,

Wow, that is some thread.   Interesting points on the screen...

http://amfone.net/Amforum/index.php?topic=44390.0

I ran thru it and noticed the Tron wrote a dissertation on the subject.  I lean towards what he said.  I have had a lot of experience driving the grid and screen  (G1/G2 tied together) in 813s, 4-1000As and a few others.   Personally I do not think this is as clean as running them as grid-driven  (G1) as tetrodes using well regulated screens and regulated grids. (AB1)   This is the way the engineers designed them to run (best linearity) and hams often change the config to make it a little easier to implement.


But all in all, there is one way to find out which is the best config. All tubes are different from one another and were designed to run a certain way. Experimentation is cool. All one needs to do is run IMD tone tests in all configurations - grid driven as tetrodes as intended -  or screen driven -  or triode connected G1/G2 and see what the 3rd order IMD looks like.

I just ran some IMD linearity tests today on my 4X1 plate modulated rig (using 4X1 tetrode-connected AB1 modulators with solid G1 and G2 voltage regulation). I had some sudden audio parasitic problems (splatter) that I solved today.  Once fixed, I obtained -35 to -40 DB 3rd order IMD results, which is outstanding for a plate modulated rig. I once had the modulators connected as triodes, but the results were not as clean as tetrode-connected.

I find the 813s work very well triode connected but never did a careful comparison against tetrode connected. (must use well regulated grid and screen voltages under full drive or all bets are off)

So bottom line is that I have been bugging Frank/GFZ to run the IMD tests and since you brought up this good point, I will see if he is willing to connect them up as a tetrode G1/G2  and as a regulated tetrode to see the IMD difference.   There WILL be a difference.  My feeling is the stock tetrode config will be the best since those sweep tubes were designed by smart engineers trying to get the best TV picture fidelity that they could... :-)

I'll let ya know how Frank makes out once he starts testing.

T

 

Tom,

I learned of this type of proportional grid drive concept from postings I read in diyAudio.com. I was trying to bring their findings to this group in the thread I referenced. They have members there that are very knowledgeable in tube theory and audio applications. I really can't give them enough credit for the work some of them have done. Honestly, their technical expertise is way over my head and I just try to pick up whatever tidbits I can. You'd have to read the threads on that message board to get a good idea of the work that's been done there. It will take some time reading through it all, it's been in development there for some years. If you look at some of the curve trace photos that have been posted there, you just might change your mind. For your consideration, I submit a photo of the curves generated by a 36LW6 using G2/G1 proportional drive. That photo is found here: https://www.diyaudio.com/forums/tubes-valves/323860-question-rla-lw6-triode-connection.html#post5461899

The key to making the proportional drive work is using a tube with a very low G2/G1 mu factor. Most sweep tubes come in with a G2/G1 mu of 4 whereas an 807 has a G2/G1 mu of 8 and the 813 has a G2/G1 mu of 8.5. The 807 and the 813 are not a good candidates for that reason. The 6LW6 has a G2/G1 mu of 3.7 which makes is a very good candidate for proportional drive.

I have great respect for the engineers that designed the old tubes we use. But every engineer had an engineering manager breathing down his neck to get the job done, and that manager had the sales manager breathing down his. Then there's the customer who wants it now and wants it cheap. The design that wins the day is always compromise. And, in the end, I doubt that anyone who designed a sweep tube for a TV set imagined we'd be using them or audio. But, here we are.

I'd be very interested to see IMD testing results for this. To date, I've only seen the displays of a curve tracer.


Don

Edit: FWIW, Timtron was the very first person I brought this idea to before I made my original posting. I couldn't sell him on it. Its not gotten much easier since then either. But I persevere.

[K1JJ]

Hi Don,

Frank mentioned that he looked at the curves to decide about driving the screens.  

"I believe the 6 Mother Humper 6 has the same guts as the 6LW6 with a newer base. Look at the curves with g1 at zero. Notice the clean parallel evenly spaced curve lines. Yes the tube was designed to be driven via the screen. I think they look cleaner than driving G1 and since I only need about 70 volts the plate can swing lower before there is distortion. Steve has played with driving both grids of 6DQ5s but they barely modulate a Valiant. I have 4 bigger tubes loafing along. Everybody I know using a pair of 6DQ5s barely hit 100%."

I think there are so many variables; Mod iron, plate impedance, audio driver used, voltage stability, amount of power needed to stay within limits of saturation, class of operation, etc....   Simply setting up the rig and running the IMD tests will cut right to the chase to find what works best. It's easy enough to do.  IE, after I build a rig it always takes as long to troubleshoot and optimize as it takes to build it.

I have no debate about the audio guys and their opinions.  I'll bet they spent a lot of time testing and proving things out too. OK on the optimum G1/G2 Mu ratio.  Yes, proportional drive makes sense since it is a form of optimizing for a given set of conditions. If done correctly, it HAS to work better than just slapping on fixed voltages in a standard config.  Testing and refining is the key.

T

6MH6 tube:

[KK4YY]

Hi Don,

Frank mentioned that he looked at the curves to decide about driving the screens.  

"I believe the 6 Mother Humper 6 has the same guts as the 6LW6 with a newer base. Look at the curves with g1 at zero. Notice the clean parallel evenly spaced curve lines. Yes the tube was designed to be driven via the screen. I think they look cleaner than driving G1 and since I only need about 70 volts the plate can swing lower before there is distortion. Steve has played with driving both grids of 6DQ5s but they barely modulate a Valiant. I have 4 bigger tubes loafing along. Everybody I know using a pair of 6DQ5s barely hit 100%."

I think there are so many variables; Mod iron, plate impedance, audio driver used, voltage stability, amount of power needed to stay within limits of saturation, class of operation, etc....   Simply setting up the rig and running the IMD tests will cut right to the chase to find what works best. It's easy enough to do.  IE, after I build a rig it always takes as long to troubleshoot and optimize as it takes to build it.

I have no debate about the audio guys and their opinions.  I'll bet they spent a lot of time testing and proving things out too. OK on the optimum G1/G2 Mu ratio.  Yes, proportional drive makes sense since it is a form of optimizing for a given set of conditions. If done correctly, it HAS to work better than just slapping on fixed voltages in a standard config.  Testing and refining is the key.

T

6MH6 tube:

Tom,

I'm not that familiar with the Valiant but it's my understanding that the modulation transformer ratio is too high to get much more than 100% modulation, if that, using the common power supply. 6DQ5's, or any other tube, would suffer the same problem. You just can't swing the voltage. I would think the cure for the Valiant would be to replace the mod transformer with one of a lower step-down ratio, or raise the modulator voltage, or perhaps some modulation auto-transformer connection wizardry.

I used a little pair of screen driven 12GC6 sweepies to fully modulate a DX-100. The 12GC6 is rated at 175mA average cathode current with a Pd of only 17.5 watts. A 6DQ5 can supply 315mA average current per tube. I would think you could get 160+ watts of audio from a pair of 6DQ5's. Then, by extrapolation, a pair of 6LW6's, having 500mA of average cathode current, should go to 250+ watts output. That said, you'll red-plate all of them with a sine wave at those levels. A human voice may average a duty cycle of perhaps 25% of that, keeping them within their (measly by comparison to the cathode) plate dissipation rating.

From an email I received from Timtron on Aug 21, 2018:
"Frank WA1GFZ hot rodded a Valiant with 4 6146s in the P.A. stage and 4 26LW6s in P.P.P . He wound his own modulation transformer. He found he could pull two of the modulator tubes with little or no change in the modulation capability."

So, I think that homemade mod transformer is really getting it done for Frank, and the second set of 'LW6's is an insurance policy.

Currently, I have screen driven 6DQ5's in my DX-100 and plan to try G2/G1 proportional drive. It's a matter of getting around to actually doing it. So far, I have built the dual rail PS for the 'fet driver but the rest of the driver is stalled on the drawing board. If anyone gets to try it before me, I'd be excited to hear the results.


Don