The AM Forum

I am in the process of restoring and upgrading a 500.
RF is OK, HV shorts have been dealt with
Peak modulation very poor with original driver transformer. Upgraded to a larger transformer rated single tube to PP grids.
Lots of audio now but having tube failures. Grid to cathode shorts. I have a feeling I need some loading and series resistance feeding the grids, maybe RF chokes?
Anybody had issues like that?
Keith
WA1HZK

hi Keith ... I assume you are referring to 811 failures ..... a few things come to mind .... the 811 has a Mu of 160 afaik which means a close filament to grid spacing ... ao a couple things come to mind ,,,, check filament voltage and parasitic activity .... this is a bit tricky in the 500 since the driver xfmr is on one chassis and the tubes on another .... what type of interconnect wiring did you use EFJ stock or something else?  snoop around with a ne2 neon bulb and see

Don't know what you're using in  the cable, but  I have some twisted shielded pair and I can send you a few feet if you want.   I just got done making cables for mine.

(https://photos.smugmug.com/EF-JOHNSON/Johnson-500/i-z8F9ZpX/0/825bb7e6/L/P1010005a-L.jpg) (https://mikeharrison.smugmug.com/EF-JOHNSON/Johnson-500/i-zqP4wRH/A)

I am working towards setting this one up with the power supply in the cellar and the RF deck in my office. For now I'm testing with the stock control cables but I had to remove the crappy HV cable set and use some China HV connectors and RG59 for the Plate and Screen power. The new cables will be 25 feet long. The Mod tubes are 572B's. I have -15 volts on the grids to tame the tubes down with the stock HV. Over the years I have removed the .005's that bypass everything in the decks ref. the modulator grids. I'm thinking some of those might need to come back, maybe .001's and feed the grids with some RF Chokes and .001 bypassing just in case RF is making it take off. Before I fire it up with my last set of tubes I will have a grid meter in line and be watching the grid to grid AC. Something is killing those tubes.

The mode switch was a source of shorts. Too much HV on old parts. Disconnected the un-modulated DC as it's not needed unless running CW. The audio has been completely gone through including negative feedback but I have not been able to get the thing stable enough to do some real testing on it with all these tube failures. This old leaky bucket will not defeat me. I'm interested to know what others have done for audio drivers short of using FETS. I want to stick with Glasfets.

My immediate reaction is that IF you are thinking of running "audio drive" 25 feet, that the best
way to handle that is with a transformer on the send side and a transformer on the receive side!

The question will be what impedance.

The instant thought is "600ohms".

Even IF the receiving side is a nominal 600ohm impedance, the send side
may be equal or lower.

IF you set it up for CT on both, you can kill any possible common mode crap along
the way. The receive side (the one that drives the 572Bs) will need a ratio that
provides the necessary voltage swing to fully drive the modulator tubes. Should be
easy enough to figure out.

Typically the grids of an 811A or 572B will need a few watts to drive into Class 2. Probably
AB2. So size the iron to fit.

This will of necessity require physically different transformers than you now are using as the
ratios will become different. But it ought to be substantially more stable.

                           _-_-bear

I was looking at an online manual for the Johnson 500 and from the schematic, it appears that Johnson put 2kv on the plates of those 811As
They have about -7volts on the grids to limit current, but that tube is designed to run a maximum of,1500 volts and plate dissipation is 65 watts for quality RCA tubes that have the fins in the anodes. These modern 811As are woefully inadequate to stand that abuse. Going to carbon anode 572Bs with the 150 watt plates and higher voltage rating makes a lot of sense, I would think.

Another thing that makes little sense to me is to drive a class B modulator with a transformer.
For class AB1, it makes perfect sense because the grids are always a high impedance load that requires no grid current and therefore- no power to drive. When the modulator tubes go positive, the input impedamce drops very fast and that shifting load must cause distortion in the output. Putting cathode followers or a MOSFET driver will result in much cleaner audio and still provide the power needed to drive that 4-400A rf amp.

Bear
Hey, that's a good idea. With that 25 feed could be 100 feet, no problem. I figure I'm not the only guy that used to lift his 500 supply and now pushing 70 can't drag it. OK, 6B4, Plate Z is 800 ohms so first transformer needs to be 800 / 600 then 600 / two grids. The Class B grids of the 572 B's are what? I'm guessing 12K CT?

I'm going to try my best to get running the way it is but failing that your idea is good. Put a transformer on the PS Chassis with a split secondary and that gives me a chance to balance the tubes bias like a real broadcast rig.

Does anyone know of a transformer prototype shop that can custom wind iron? Would be nice to have exactly what's required instead of the usual junk box stuff.
Keith

I used to drag my 125 lb Dahl around with no problem.  It's a 3A xformer.

Today I had to have my son help me with a 2.5A.

I ain't getting any younger. I think it's time to hit harbor freight and get another engine stand.  Felt it in my knees more than the back, so even with proper lifting procedures it didn't help.

--Shane
KD6VXI

uh huh .... go ahead and get the engine lift too.  :o

You do not need exact impedance matches. All you need to do is to get enough voltage swing at the
grids, backed with enough current to run it into AB2, instead of AB1.

So the first xfmr can be a 1:1 but probably a plate (unless it is PP?) to p-p "line". Or actually plate to whatever you
can find that is not  super high Z. The secondary wants to be CT.

Then the next transformer needs to be similar but not LOWER Z than the first one's output Z to
whatever voltage ratio you need to swing the grids.

Don't think you need to custom build anything.

First step is to figure out the P to P voltage swing of the 6B4 and the required voltage
swing to really fully drive the grids of the modulator tubes. Then just fit the iron between
those two voltage ratios. Your design "anchor" works out to be the Plate Z of the 6B4 at
the plate voltage it's run at.

                     _-_-bear

Then the second transformer wants to be a large step up - EDIT, actually just the ratio needed to
drive the grids properly...

Brent and I are going to test it today. I removed some previous mods and the setup I have now seems to modulate OK. I did a lot of work on this radio in the 70's and it's coming back. The thing with these radios is that it's a struggle to ever get to 100% modulation with the existing mod iron. I had in the past rigged more voltage to the modulators that helped but turned out to be a tube killer with the 811's. With the existing driver transformer I was only able to get around 70-80% but after changing the iron there I can just about hit 100%. Sounds OK here but I need on air testing. 2-meter & 75 transmitter on bench with 528 & cheap china mic test setup.

That's exactly what I meant.  I'm finding as I get older the chair to keyboard interface needs some slowing down haha.

--Shane
KD6VXI

Made progress. Brent W1IA said the audio was all muddy. That was traced to caps for RF bypassing on the grids of the 572B Modulators. Once they were removed, the transmitter came to life for about 5 minutes before the plate relay crapped out and all the fuses blew again. Hey, the audio worked. I have a nice big contactor to replace the crappy 10 amp stock one so progress. Once I replace everything in this 70 year old radio that can't hack 2200 volts I should be good to go. It sounded good and looked good on the scope. I'll document everything that was done and post it here when I am done replacing high voltage parts that go bang. Last project, a 25 foot set of cables.
Keith

Ya know sumtin'?

If I had one of dem dar transmutters, I think I would put in a pair
of 813s, or if I had them 805s or 838s in the modulator holes (change sockets,
drill and blast) and run 2000v on the plates all day long.

                     _-_-bear

Believe me, I had the 810's & 813's in my hands and thought about it but the 572B's are the more logical choice. I did add some protection as I had that issue of destroying these tubes by cathode to grid shorts.
So I did this:
Driver Transformer has 12 K across the secondary, 2 X 6K to ground on the 572B grids.
Also 1K resistors in series from the driver transformer to the grids with a 2.5 Mh RF Choke in series with the resistors to the grids.
Not sure what stopped killing the tubes but my last of three pairs that I had to work with are in there now are not shorting out.
Bottom line is the rig has to have a better driver transformer for the modulators. Sometimes with used not totally matched parts some effort is required to actually make it all work without starting a thunderstorm in your tube grids!
PS, 70 year old radio and I'm almost there... sometimes I short out too.

I never have had any problems with mine. I use RCA 811A'S. I have all stock iron except for a Peter Dahl mod transformer. Not one of the new ones that are made by some schyster using Pete's name.

I am mystified by the practice of driving a class AB2- B audio amp with a transformer to begin with.
The impedance reflected back to the 6B4 drivers must vary very widely from the full negative swing where no grid current flows and the load provided is very high to a much lower load when the 811As are pulling up to 100mA of grid current on positive peaks rising to 170volts P-P
For one thing, this bouncing load will cause distortion and on the other hand, the grid voltage will go quite high on the negative swings.
Not sure if the latter would account for the shorted cathode to grid problem with the junky tubes available, but the distortion must be noticeable to a critical ear.
Another mind numbing thing is that Johnson would put tubes in their modulator that have a plate dissipation of only 65 watts and put over 2000volts on them. How much quiescent current is present? RCA speced 32mA at 1500volts on the plated, or about 50 watts per tube. You must drop that down to about 22mA to give the thing a chance at 2200volts.
A pair of (good)572Bs , or 810s or triode connected 813s could handle all of this, though the Chinese versions of the 572Bs seem to short a lot with those wildly swinging grid voltages...
Why not build either a pair of 807s as cathode followers or build Frank’s FET driver that can provide the grid current needed and maintain a smooth drive voltage profile that will produce beautiful clean audio. A look at any good broadcast transmitter schematic would be instructive. Johnson designed some nice stuff, but in this case, they went wrong, and trying to fix this with new transformers is spitting into the wind, IMO.

The plan for this one was to get it running again and with a set of 25 foot cords, let me remote the power supply. As they told Luke Skywalker, there is another, a sister... That one does not have a Power supply and I have plenty of KW class broadcast parts available. But that will be another story.

Good point.
When I was young, I got a job repairing two way Radios and often got an impulse to “improve” them. My boss told me, “ just get the frikkin thing going, get it out and go to the next one!!
Got a design engineers job after that.... No cure for it, so fed the monkey out there in the zoo for 34 years....😉

"Also 1K resistors in series from the driver transformer to the grids with a 2.5 Mh RF Choke in series with the resistors to the grids."

The 1K resistor will drop a lot of voltage as the grid current swings upward to 100ma. If it ever did hit 100ma, that resistor would drop 100 volts!
You might try reducing that value to just the 2.5ma choke. The choke by itself might have 100 ohms DC resistance +/-.

Driving a varying impedance load is always a problem, and that is why EF Johnson used a low plate resistance driver tube (6B4). A step down from there would be nice so long as enough grid swing is available. One way around this issue is to use a Power FET as a Source follower as in the Millet A2 design:

http://www.pmillett.com/A2_buffer.html

I use this on a Viking 2 to drive the screen grids on the 6DQ5 P-P plate modulators. This simple circuit can easily provide 200v P-P to each grid, and source 100ma peak current or more. Some heat sinking will be required.

You might notice other similar circuits here on AM Fone, but those use a source resistor to the (-) supply instead of an active current sink. The active current sink offers advantages over a passive return using a resistor.

Jim
Wd5JKO

love it !   that is just how it goes so much of the time .... it ok until someone decides to spank that zoo munkee  ;D

by the way I have a 500 rf section very clean available (no HV ps) unmodified if someone is looking  :)

Yes, the 6B4 may be the biggest reason it works at all. Nice big triode. It is like a big current reservoir- within limits...
Still, if one must use hi mu triodes in the modulator, cathode followers- 807s will work- is the best way to keep the linearity during that transition from high impedance to a sliding lower impedance because it can provide the voltage and the necessary current.
I have wondered why Johnson did not just use a pair of 4-125s in AB1 in this rig. Then the grid voltage never gets to zero and the interstage transformer sees a nearly constant load. I say 4-125 because they went with the Eimac 4-400, so they must have been enthralled with them. They put a pair in the Thunderbolt, I think. The 4-125s would give more audio than the 811A and run cooler also. Downside is thd need for a stiff screen supply and -95 bias voltage supply.
Of course, if you just want to get it going stock, this is all heresy...😉

Some good ideas. When we tested it we got fine audio for 10 seconds then another short appeared. That was rough on the plate contactor so swapped that with a 30 amp one that will not fail and isolated the power supply. Brought the DC up slow with some 100 watt 6300 ohm resistors in the HV secondary. DC came right up tp 2200. Put it together with the resistors out of the circuit, again, 2200 volts. So the power supply is OK. It's the 70 year old High Voltage wiring in the RF Deck. I already disabled the feeds to the AM-CW-SSB switch, those were a lightning storm. I guess the last thing is to check & replace HV bypass caps that look suspicious and hard wire the modulated HV right to the plate choke. 9:00, tired, been doing this all day..... 2200 volts, same as an electric chair! I'll try again tomorrow. At least I know what deck is giving me trouble.
WAIHZK

I agree with Jim on this.

Seems to me that the 6B4 triode is relatively low Z plate. But we're only sending a few watts to the grids.
And yes the load impedance certainly changes. But once ur sending watts to the grids,
IF the tube has negative bias on it, then the load Z changes at the zero crossing - if it's a "zero bias" tube, the
current is being drawn throughout the cycle. IN the case where the difference between
drive is going from Class 1 to Class 2, my understanding is that the tube is entering into a non-linear
region to begin with. One gets some extra watts out, and does not hit "hard clipping", which
is a plus overall.

Cathode drive is fine. There are number of different schemes that do this.
But it adds an extra stage, because the follower needs to be able to swing whatever the
grid needs voltage wise, so has to be driven with a full swing voltage amp. I like cathode drive.
It gets rid of the driver transformer.

                     _-_-           

Well I bought a set of HV connectors on E-bay, supposed to work with 5KV. First one connector vaporized now the cable flashed over. The ad for the HV connectors said use with RG-59.
Just making a note to all.... Don't. That's in the trash now. My last possible option is silver teflon PL's. Searching for better HV connectors that use RG-8 or 8X cable. That stuff looks a lot better for volts. This single holer can swing to almost 5K volts.
One of those projects where everything seems to break. At let the short is not on the decks.
Keith
WA1HZK

Found some 10KV connectors. https://www.ebay.com/itm/4Pcs-10kV-Lantern-Type-Banana-Cross-High-Voltage-Connectors-Plugs-Sockets/313322898733?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649
I can't believe what they are getting for HV wire these days. So I guess I'll try 8X. I have always used that in the past and been OK.
Anyway don't try those E-Bay advertised HV connectors that look like BNC's, not good at all. Advertised for RG-59 but might work with RG-6.
Trouble is, most of that Cable TV Coax is just junk. I have never had cable break down like that with a 2.2KV modulated signal.
So I'll wait for the next E-Bay delivery and try again. Back to making ammunition for the coming Civil War.
Keith WA1HZK

Hang on.

The HV BNC connectors fried or the coax inside fried??

The ones you show are NG, the chassis mount needs to be the one with the male plug
not the female!  You do not want HOT B+ exposed off the end of a cable!!

Can you post a link or example of the HV BNCs you had?

Regular BNCs should handle 2kv all day. Fwiw. As will RG-58 or 59.

                   _-_-bear

The coax should be of the solid poly, not foam.

IF you need HV wire I have some super high voltage rated wire, the OD is similar
to RG-8, but it is flexible. You can have a few feet to meet your build needs for
the cost of sheeping. Email me. BUT, coax with a shield is safer. FYI.

Again a problem is that the male should not be exposed with HV in any way that it could
come in contact with one's body - ever.

This is the one that flashed over.
https://www.ebay.com/itm/5Pcs-RP-BNC-Female-high-voltage-power-RF-connector-SHV-5000V-RG6/201428261714?_trkparms=ispr%3D1&hash=item2ee60f5b52:g:AWUAAOSwFnFV9g95&amdata=enc%3AAQAFAAACgBaobrjLl8XobRIiIML1V4Imu%252Fn%252BzU5L90Z278x5ickkrmmq8hWevtuKE7%252F%252BLfVXto%252FVPbPR9YcnnRjCTIyH4Zg7OMogYHLskhOvU91VenaUhUwQlrFUL5CcSKpmCXVzpjI64NKA%252BI5kJ%252Bg6bUavabK1PGdyza%252Bh%252F2Ub63TNJvGxxS3hJrXqAB06yqNYUPvvVYlhAZ6jN4C3IKj9V4%252Fnh0GeaLeFl0H0y36%252FRZWj5EBZdKu3hGVNRjJq3cxLqazxAe6CuTC1gsSvK0EwA0knd7XF5TWJJdmbdZF%252FMkUezMgzMs2Ow%252FeJ024iLUi0wgB7fY2pFGqNb7%252FSrAANkvLGy1VJMXDyHPxIA1icEluMuIKu3NKZF%252Bxot3DTpal7Zyxn7hCLUfFZwL27whi40bjE7H6eIrbOyWb4DmQdVKTGxnjRMtZssJuKhgCXKlvLLzfKnZoZy0m46opyI2nNLEtRwmnRbF61sfG5AJzFzWjo%252BP2cirQvT9N%252BqBYHLH7xUJk%252F3%252B7tjqAaSKLi8mVHbcu7w15Fz%252FW%252B8VL%252FHY4o%252BBTMApdgT659jaVPDTfBs3OIsiH%252Bsu8v2KbUbG0Ks4POJP%252BoU1ORh60%252BEVp0dKUQuZfvA34f19a%252B2Bk5lmkrViBXDwOZ62%252FZnQHyJeIJ7L1emEjxeLuloERhIlgdRkhneJ3fkDLKwbkrc%252BXpi21Z8Z3%252FBd1zsq5uR1oeg1NZe1tVfNTJhD1vb0hzGl8%252BdewgSgcdD8y1JO13hXoDO66%252FdV1JLOfKF%252FmsWzJ%252B7%252FkXUmmoZA0oKjLiT%252F2LRz5tEmrXyazJLorHK%252BewYPINo5gEBCpSZ%252BM6oMa2n2275g25bxOqmM9EcxA%253D%7Ccksum%3A2014282617141588ac9589bb4342a2672b658e0987a4%7Campid%3APL_CLK%7Cclp%3A2334524

Damn, that's a stupid link.
The next ones I try will be very similar to the orig. stuff but rated 10KV. I'll just hook it up the same way. Only need the plate and screed feeds.
The above connectors might have been OK but were definitely not RG-59 Crimps as advertised.  The crappy RG59 was foam, that flashed over. Lasted a couple days. that's it.
Lesson learned, don't trust Chinese ads.

Outdoor luminous sign sign wire is rated 15KV and it's not that costly. Shield from RG-8 can be slipped over it.


https://www.amazon.com/dp/B000NUX4PK/ref=twister_dp_update?_encoding=UTF8&th=1&psc=1

Ancor GTO-15 is single conductor, stranded AWG14 copper wire. The GTO-15 indicates it is highly insulated (for up to 15,000 volts). The "marine grade" indicates that it is tinned copper wire.

I've been using those connectors for close to 3 years now, ever since a (not installed by me) Millen HV connector had abraded insulation on it and threw 6.2Kv into the back of my hand.  I've used those SHV connectors to 7kv without issue.

I DON'T use standard RG59 with them, however.  I use RG-59 RATED for the voltage, and not all is at the levels I play with.  At 2.2kv, you should have been fine.

Incidentally that 6Kv "reach out and touch me" landed me in the hospital for 2 nights.  Doc said that it could take up to a year for symptoms to appear, so be careful.

Sure enough, 13 months later I was in the ICU having heart surgery.

I suspect a bad install.  It's not hard to have ONE wire of the braid get screwy when you're installing those connectors.  I know this from experience.  I now ohm all of them out before applying voltage to them.  I've installed enough now that I have it down pat, but I did have to trash 2 in the beginning.

Also, I've received connectors that have an insulation bushing/washer that goes on the center pin and also received ones that didn't.  You install the center pin on the coax and then a flat teflon washer goes over it.  Then insert the pin into the connector and slide the ferrule up.  flare the braid back and crimp.

Sorry you had problems with your connector, but you're a lot better off safe than sorry.  I'm now taking meds to try to keep me from having to go back for another surgery.  Surgery sucks, but dying was the other option.

--Shane
KD6VXI

Your story made me change my mind about using quick connectors. Glad you're still kicking and I wish you the best.This is what I did with rg213 to bring 5kv to my rf deck.

I used CGB before.  They are available at depot, etc.  The tighter you screw them down the more they compress the rubber bushing in them.  It's what we have to use in the electric world.  They work pretty good.

I pulled a ton of RG8 sized neon sign cord out of a Target store I did the electric on for a remodel.  It's just like RG8 but the dielectric is something more akin to rubber.  It's great for tying chassis ground together and brining b+ to the RF deck.  Unfortunately the source wasn't infinite and I'm low.....  And haven't found a new source:  Hence my finding that rg59 and SHV connector.

Glad you made it safe.  Sucks the SHV didn't work out though.

--Shane
KD6VXI

The 10KV parts are too big. Save them for homebrew chassis. I bought another set of the HV 5KV connectors and the proper RG-6 Crimper. I'll try this again. If the High voltage stays in the RG6 where it belongs, I can start testing again. If it looks clean, I'll set things up for 25 feet away and see what happens. I took a month to reorder the connectors from China so that threw a wrench into the project.
Keith
WA1HZK

"I have removed some mods, and now it seems to modulate ok"

This sentence should be chiseled into stone somewhere !

God save us all from ill advised "mods".

else where I read... " I have no audio coupling cap and almost no modulation"
do tell...

Mods
dirty & old parts
misadjusted transmitters

will account for almost all modulator failures

although the goofy cables in the viking 500 design  probably figure in a few disasters.

It works. I ran it this afternoon and set up the audio with Brent, W1IA. Reports were good. Brent says I sound like me. Life is good.
The only issue is lack of good positive peaks on the signal. After dicking with the feedback and switching the phase of the audio the opinion was that I probably have weak tubes in the modulator.
Having ruined so many this winter I was down to my last used pair. Hey, if that's all I need, no sweat. Now to finish the 25 foot cables and if the radio doesn't puke about that, I'll finally get this installed the way I want it and write this up with some pictures.
Glass FETS, gotta love them.
KC

For coax used for HV, the polyfoam I am told is NG.
What is good is the solid poly.

I'd never use crimp connectors myself.
Never have.

Hope you get the positive peaks working.
I'd swap the modulator tubes first, as you were going negative just fine.

Also double check the bias and resting and peak current in the modulator!
See what the manual calls for with full snot modulation as far as current.
IF ur making that, there might be a bad winding in the mod iron. But that may
show up in a resistance check.

I was seeing momentary very high peaks in the positive, intermittently
and only with certain vocal sounds. Hard to say what that is.

In addition, if you are coming out of a line level signal path, and depending on
what that is, IF ur using two of three pins on an XLR out, swap to the "other"
pin phase and see if the problem goes away or the positive peaks change.

                _-_-bear

Remember the 1K grid resistors were added to solve parasitic oscillations. With Class B grid drive, the driving impedance needs to be as low as possible. Adding 1K to each side is huge. This will limit upward modulation. Back earlier in this same thread, I suggested removing the resistors and keep the 2.5mh chokes that were added at the same time as the 1K grid resistors.
If the parasitic was at RF, a neon bulb could be temporarily attached to each plate with one lead, and the other lead sticking up in the air. If the bulbs ever light up, then you have a parasitic. The glow will be orange at HF and purple at VHF.

Jim
Wd5JKO

Good point.
I'll wind some suppressors for the plates and install them. That won't hurt anything and jump over the resistors and see what that does to the positive peaks.
A new set of 572b's is on order.