W8JI Johnson Auido mods opinions?

[Ed/KB1HYS]

Ok in all the reading I've done around here, the concensus is to clip the negative peaks at the mod tranny.

Well here's this fella actually recomending clipping the negatice peaks in the audio chain. Interesting!
he has some interesting opinions about he valiant which he seems to have done the testing to support.

I was wondering in cliiping the negative peaks in the early in the audio chain offers some advantage or does the 3 diode peak limiter offer an advantage that he overlooked.

Here's a link to his site:  http://www.w8ji.com/johnson_valiant_audio.htm

Opinions??

EDITED TO put the correct link in.  Sorry about that

[k4kyv]

Can you provide a direct link to that specific article?  All I could get was his main website page, and after several minutes of surfing through many pages of his technical writing could not find that specific article, or anything about AM modulation for that matter.  But here is my opinion, without having read the article.

Any clipping or other modification of the audio waveform, by definition, will generate distortion, and unless it is filtered out, spurious sideband products will appear.

One advantage of the low level clipping is that filters can more easily follow the clipper stages.  But you have to watch for phase shifts.  You practically need DC coupling between the clipping and the RF final to avoid this problem, and any type of passive filter is apt to generate its own phase shift.

Negative cycle loading, or ultramodulation, schemes that reduce the amplitude of the negative half of the audio cycle while leaving the positive half unattenuated, by definition generate even harmonic distortion, just like a single-ended class-A vacuum tube stage operating too far down on it characteristic curve.  The rectified audio contributes to the average DC voltage to the final stage, causing a rise in PA plate current under modulation.  If you think about it, this is exactly the same thing as running controlled carrier modulation combined with even harmonic distortion deliberately introduced to the audio.

The only kind of diode configuration around the high level modulation transformer that I would recommend would be a protective circuit to prevent the modulation transformer from losing its load and effectively working into an open circuit during negative overmodulation peaks, which could possibly cause failure of the transformer.  In its simplest form, this would be a diode in series with a resistor of the same resistance as the modulating impedance (plate voltage ÷ plate current), rated at maybe 5 or 10% of the audio power output from the modulator.  The cathode of the diode goes to the modulated +HV audio line to the RF final amplier, and the anode goes through the resistor to ground.  Or the resistor can be on the top side, with the anode side of the diode grounded.  Whenever the audio output voltage from the modulator exceeds the DC plate voltage in the negative direction, and cuts off the plate current due to driving the PA plate voltage instantaneously negative, the diode conducts and maintains the proper load on the modulation transformer through the resistor.

This can also form a very simple negative peak overmodulation indicator, by using an 866 or 866A mercury rectifier as the diode.  A type 83 tube could be used for lower power transmitters with lower plate voltage.  As soon as the 866 or 83  conducts, you see a flash of the familiar blue glow inside.  Mount the tube in a small box with a window that reveals the vicinity of the plate structure, paint the inside of the box with flat black, and the flash becomes very visible.  With the tube, it would be better to put the diode on the bottom side of the resistor, to avoid the necessity of a filament transformer rated at the DC plate voltage.

You can bias the diode to a positive voltage, using a small "keep-alive" DC power supply to maintain a little positive bias on the anode of the tube.  Then the flash will occur  whenever the modulated plate voltage drops below the bias voltage, before it actually goes negative.  For example, if you are  running 2000 volts on the final, a bias of 40 volts will  cause the flasher to indicate at 98% negative modulation.  200 volts of bias wall cause it to flash it 90% modulation.

The power rating of the resistor can be a small fraction of the operating power because it operates at such a low duty cycle.  A 50-watt resistor should be much more than adequate for a full KW plate modulated.  If you carefully limited your negative peaks, a 10w resistor should be fine.

The best way to maintain positive peaks above 100% while avoiding negative overmodulation, is to take advantage of the natural asymmetry of the human voice.  You need a good, wide range low distortion microphone and a speech amplifier with low phase shift distortion, that faithfully maintains the waveform coming out of the microphone.  With most voices, the ratio of positive to negative audio peak voltages will be between 50% and 100%.  It is simply a matter of maintaining proper polarity of the audio line so that the peaks with the greater amplitude modulate in the positive direction.

Of course, to achieve higher-than-100% positive modulation peaks, regardless of method, requires adequate peak power output from the modulator.  The peak power audio output increases as the square of the modulation percentage.  To fully modulate 1000 watts DC input with a sine wave requires 500 watts of audio power.  To modulate @ 110% requires 605 watts of audio.  To modulate at 140% requires approximately 1000 watts of audio.  To modulate at 200% requires 2000 watts of sine-wave audio!

The human voice does not generate a sine wave, so the average power output of the modulator with voice modulation will be much less than what it would be with a sine wave, but the undistorted peak output power requirement is the same, and this is true regardless of whether or not the carrier modulation is limited to less than 100% in the negative direction.

If the asymmetry is produced in the low level stages, or as a result of the natural voice waveform, the average power output will be the same as that of an unaltered sine wave, but the peak output will still be proportional to the square of the maximum modulation percentage.  This is evident with the use of a modulation transformer in a plate modulated transmitter.  The transformer cannot transfer DC.  So let us say the positive peaks reach 150% positive, but the negative peaks remain just below 100%.  The only way the transformer can pass this waveform without resistors and diodes (as would be in the case of negative cycle loading or ultramodulation), is to maintain the same energy output over the entire positive half of the cycle as over the entire negative half.  Careful observation of the waveform of the audio on an oscilloscope will reveal that the "area under the curve", defined by the boundaries of the audio waveform and the base line, will be precisely the same during the positive half of the wave as during the negative half.  In other words, the positive half of the peak, which exceeds 100% modulation, will be of shorter duration than the negative half of the wave, which remains at or less than 100%.  In an extreme case, the positive half of the cycle would be a very short duration high amplitude pulse, while the negative half would be of much lower amplitude but would dwell on the negative side of the baseline for a much longer time.

[Bacon, WA3WDR]

I've applied clipping at low level through the use of an audio side chain.  The side chain had about the same characteristics as the modulator, and the result was that the clipping was remarkable effective, even though it was done at low level.  I discovered this by accident way back when.

'Asymmetrical Peak Limiting' article from AM Press/Exchange issue 79

Full AM Press/Exchange Issue 79

[Steve - WB3HUZ]

His approach is poorly engineered. Doing any sort of clipping and then following it with a lowpass filter with a poor impulse response (like the one in the Valiant) creates overshoot and reduces or removes any advantage gained by the clipper. It's nothing more than a distortion generator.

Below is the response of the lowpass filter in the Eico 730 modulator with the input of speech with 6 dB clipping. Major overshoot! The Valiant is similar.

[Ed/KB1HYS]

Hmm, well I guess I don't understand what's going on here. Seems to me that Bacon is ok with clipping negative peaks early in the chain, but Steve/HUZ points out that clipping causes distortion.
To be honest, Don's suggestion about the diode with resistor I actually understood,   I had thought that limiting the negative peaks in the audio chain input would also limit the negative going peaks that the modulation section was applying.   
If we allow the positive going side to run normally and just limit the low/negative side to some voltage, where does the distortion come from?

 

[Bacon, WA3WDR]

Steve's point is that the low-pass filter in the Valiant will cause ringing on the clipped waveform, and reducing the level to keep that ringing from overmodulating will pretty much defeat the purpose of the clipping, and he's right.  If you want low-level clipping to work well, then you need to have a fairly good audio chain following it.

Any change to a waveform caused by a clipper, etc, will produce harmonic and intermodulation distortion.  The sharper the clipping, the more high-order harmonic energy is generated, and the worse and wider the splatter is.  Overmodulation produces about the sharpest and worst clipping, which is why we try to do something in the electronics to make any clipping softer than that.

A loading resistor on a diode at the modulator output is about as soft as clipping can get.  The idea is that because of the extra load on the negative half of the modulating waveform, the modulator will mush out at just less than 100% negative mod, and also the whole negative half of the waveform will be attenuated somewhat, depending on negative feedback, plate resistance, etc.  This is about as soft as clipping gets, because the clipping is only on one side of the waveform, meaning it is less frequent, and tube mushout is generally soft rounding, and not hard razor clipping, and any alteration of the waveform where the diode switches will appear as a minor change in waveform slope very close to the zero crossing, which will make it much milder than a sudden waveform jolt at some level further from zero.  It's also pretty simple to do.  Yes it causes distortion, but inherently the distortion has little high frequency component, and there is no ringing to produce new and tricky little peaks, so splatter is generally not an issue.  It's crude, but effective.

[Ed/KB1HYS]

Ok that makes sense now.  I guess I should try to read technical articles at 12:45 in the morning. Thanks Guys!! 

Based on this discussion I guess I will look at putting in a diode/resistor combination on the rig here.

[k4kyv]

The most effective way to achieve symmetrical clipping is to do it right at the modulator stage, by limiting the voltage on the modulator tubes and deliberately driving them into saturation.  Use a variac to adjust modulator plate voltage right at the point of 95+ per cent modulation.

The problem with this is that you have to follow the stage with some sort of low pass filter.  I tried this for a while back in the early 70's, using a pi-section filter between mod transformer and load.  It worked great, with minimal ringing compared to what is in a rig like a Valiant.  But the problem was, that right at the knee of the response curve of the filter, near the cutoff frequency, extremely high voltages occur across the load with high amplitude tones.  I finally figured this out after blowing at least a half dozen kilowatt size and larger modulation transformers.

So I would not recommend trying it.

[Steve - WB3HUZ]

I'd guess that the death of you mod transformers was due to the poor impulse response of the lowpass filter. Those overhshoot transients exceeded the breakdown voltage of the transfomer and fried it.

[k4kyv]

I figured out what was happening when I ran a signal generator at 100% modulation and swept the frequency.  Just as I reached 3000~ suddenly the 833A's turned white hot, the audio went away, and smoke rose from the 200 lb, 1000 watts-of-audio Thordarson that I had acquired new in the box.

I took it apart and found that one of the leads to the terminals had arced through several layers of insulation to the top layer of the winding.  I was able to disassemble the transformer and rewind the outer 3 or 4 layers and it was OK.  At the critical frequency, the voltage exceeded the capability of the insulation.

I stopped using the clipping and removed the splatter filter and have never blown another mod transformer.

I believe this is why KW-1's are so prone to blowing the mod xfmr, while it rarely happens with Johnson Desk KW's, even though both rigs run the same tubes at the same voltages, using the same off-the-shelf Chicago modulation transformer.
The KW-1 has that crappy splatter filter and the JDKW doesn't have one.

That's why I always recommend to anyone using a KW-1 that they remove the splatter filter from the circuit.

[The Slab Bacon]

One more comment, the original post was based on an article regarding an audio modification for the Johnson Valiant. The whole design of the Valiant was flawed from the very beginning. The problem is that it has two 6146s modulating three 6146s. The only way to get 100% modulation from the Valiant is too overdrive and saturate the entire audio chain and modulator. It was doomed from the very start. The only way to get clean audio from this transmitter is to reduce the audio gain from the preliminary stages and just live with weak under-modulation.

And... I still agree with Don, remove the crappy splatter filters.

Brian W5HRO / 6

Not necessarily so!! I have run a Valiant for many years now and had no trouble making 100% with good audio reports. Just simply DO NOT load it for the 360 mils of plate current that the manual calls for. I load mine for around 275 mills (which still gives about 100-105 watts out) and you lessen the saturation of the undersized mod tranny and the audio perks right up!! (not to mention makes a lot of other parts happy)

                                                      The Slab Bacon

[Ed/KB1HYS]

Nice Necro!

Anyway Slab is right here.  Loading the Valiantly more lightly than the "book" indicates gets you much better audio. 

Of course the valiant (mine) that was subject of the original post has stopped modulating at all...

It's done this once before, and "came back" all on its own after I let it set on the shelf for a couple weeks awaiting time to trouble shoot/repair. Go figure.
Hopefully this time the failure is not intermittent and can get FIXED right.

[The Slab Bacon]

The only problem with reducing the plate current on the RF amp is that you are moving away from the correct operating ratio of the modulation transformer design. I'm sure there are a lot of people that remember when I was operating my Valiant years ago. That transmitter produces a lot of distortion no matter what you do it. It was poorly designed. The Johnson Viking II, Ranger and 500 are far, far, better transmitters. The key is to change all of the resistor and capacitor values in the audio chain to the exact values show in the RCA "Resistance Coupled Amplifier" charts. That will clean up most of the problems and yield far better audio. I think most of the transmitter maker’s back then were shooting for more restricted communication "Space Shuttle" or “Military” type audio and they never used the correct values shown in the charts.

Brian,
        eliminating the clipper circuit and reworking the coupling component values goes without saying. (not to mention wrapping a little feedback around it as well) But lighter loading really increases the frequency response of the undersized mod tranny.

As far as Valiants, Rangers, VK2s, and 500s go, everyone that works on them has their own "recipe" for increased performance. Its kinda like making good barbeque sauce, everyone has their own "secret formula" and thinks theirs is the best. Some work well, and some dont.

                                                   the Slab Bacon
 

[w4bfs]

beware of reduced power out by unloading the final ... screen currents can go high and a 6146 is fragile in that department ....better to load to full power and reduce screen voltage to desired conditions ...be sure to check on the scope ..73 ...John

[The Slab Bacon]

I've been running mine loaded that way for close to 10 years now without any problems. I keep the plate current somewhere between 275 - 300 mils. I havent seen any problems with the 6146s, and it makes nice audio peaks on the scope.
It still has the same 6146s in it from when I restored it back then. Running AM without using a scope is like trying to drive blindfolded.

                                                   The Slab Bacon

                                                         

[w4bfs]

interesting thread ... I am not familiar with Valliant schizmatic....I assume the screen voltage is derived from modulated B+ thru a dropping resistor (like Ranger, V2, etc) ...  if so trying to draw more screen current would drop screen voltage and becomes in a sense somewhat self limiting ... can get away with this in class C op ... not so in linear service unless you like distortion ... still best to fully load, especially at higher freq to avoid destructive effects from ion bombardment from secondary emission ...73...John

ps.  Interesting discussion for neg peak indicator ... I will try to incorporate one in my rig rebuilds..tnx Don.

[Steve - WB3HUZ]

Excellent Valiant Audio Mods from Keith, WA1HZK at the link below.

http://www.amwindow.org/tech/htm/valiant/valiantmods.htm

[KM1H]

Would it help or hurt to remove one 6146 in the Valiant and then reduce drive and tune as if it was a Viking II?

When I was a HS junior and built a new DX-100 I was regularly cooking finals. Didnt understand screen current back then and blamed it on the rig. Dumped it and bought a used Viking I and use another one these days.  Never could hurt a 4D32.

Carl
KM1H

[WQ9E]

I remember a long time ago one of the locals tried to run his Valiant with two tubes and couldn't get it to neutralize so you might have to change a component in the neutralizing circuit to get it to work with one less tube.

The Valiant is my sentimental favorite since it was my novice rig.  I have two of them and they are both stock except for SS replacements for the 866A tubes and an inrush current limiter.  They still get a fair amount of AM use on 160 since my 500 and Desk KW won't go down there.  I think the only repair I have made is to one that I had left the 866A tubes in for looks and the plate relay stuck on which took out one of the rectifiers on power-up.  I replaced the relay and rectifiers and it continues to work.

Most of the AM roundtables I have been in always involve at least one or more weak signals with interference so rarely can I open up the bandwidth enough on the receiver to listen to "hifi" audio if someone has it.  When I want good sound I load one of my vintage tapes onto the reel to reel and listen to artists who really sound good; I have yet to meet any hams I really want to hear singing in hifi anyway although if Joe Walsh stops by I will withdraw that comment   

Rodger WQ9E

[Tim WA1HnyLR]

Hi Ed, Yep Valliant mods.Henry Nyellar style. First of all you have to take a good look at what it is. A transmitter that was built to give the amateur operator a little Xtra over the usual 100 watt class outpoot transmitters. 160-10 even the good buddy band too!. big silver plated tank coil.---overkill. The fact is you get a box that has all it can do to make 100% modulation. You can do basic audio deHnyellaphications to it that would make a DX 100 sound like a million bucks but improve the Valliant to the point where it is not that nasty to listen to. Forget any limiting schemes . The modulator has all it can do to deal with modulating the carrier fully.6146s are crap modulators. In the interest of not getting too carried away.I would do the following: first stage grid resistor-10 meg-22 meg. Coupling cap to audio gain pot .01 -.05 tubular cap . Both.1mfd  decoupling caps paralleled with 10-40 MFD 450 volt electrolytics. Removal of 6AL5 clipper stage and associated filter.Grid resistor on 6C4 changed to 470K or so .Coupling cap from plate of second half of 12AX7 should again be .01-.05 tubular. The plate resistor of the 6C4 should be 47K NOT 4.7K . Being that the need of a decoupling cap is not needed here just simply remove the .1 mfd cap from the circut . This will raise the plate circut resistance to 52 k or so. The grid resistor of the audio driver tube should be upped in value to 470k and coupling cap .0-1-.05 mfd. The audio driver circut in the Valliant is bogus. The driver transformer although physically large in core mass for the power levels involved does not have enough inductance for reasonable low end frequency response. There are circut fixes for this. (1) Change 12AU7 to a 12BH7. This tube is twice that of the 12AU7 .The 12BH7 has the same pinout as the 12AU7 and therefore is a plug and play. The cathode resistor can be same value as that used by the 12AU7 but dissipation should be upped to 2 watts.Make sure that the cathode bypass cap is at least 50 Mfd.@25 volt rating.  Connect a 100K resistor from the paralleled plates of the 12BH7 to the plate of the 6C4.this is a loop of negative feedback around the 12BH7 that reduces the plate resistance to a much lower value in order to work with the lack of inductance of the driver transformer.(2) change 12AU7 tube to a 12B4 low Mu triode. The cathode resistor my have to be as high as 1500- 2000 ohms 10 watts. The cathode bypass cap 50-100 mfd @ 150 volts.  The 12B4 is a better way to go but the socket has to be rewired for this tube .In both cases the secondary MUST be loaded with a swamping resistor. 47-68K 1 watt . The 6146 modulators must go away. A good replacement would be a pair of BIG tv sweep tubes. A pair of 6DQ5 tubes work well. There are others with similar characteristics that may be used as well. 6DN6,6EX6,6CB5,and 6CD6. Set the idle current for 50-70 MA .The screen voltage must be lowered to 150 volts. Simply change one of the VR tubes in the screen regulator to a VR150 and short out the other tube or replace the VR 105s with a pair of VR75s. It has been discovered that  the VR tubes do not provide a low impedance audio path to ground . It is imperative to add  a10-40 MFD 450 volt cap from the modulator screens to ground. A 150 ohm 2 watt resitor should be added in seriesed with the VR tube .This is too prevent the relaxation oscillator effect that happens with a gas discharge device and a parallel capacitor.In therory the 6146 tubes should make enough audio to fully modulate the carrier. The limiting factor is the ratio of the modulation transformer. It is deliberately mis matched. The load impedance of tetrode tubes is determined in part by how much screen voltage is applied. The higher the screen voltage ,with relation to the plate voltage ,the plate to plate load impedance drops significantly to secure the potential outpoot power. The available plate to plate monkey swing voltage is reduced as the screen voltage is increased. One way of getting more available audio from the given impedance is to go to triode modulators.IE: a pair of 811s. This entails much work. If you are going to go the extra mile, bulldoze the whole modulator and start with the right components. the alternative is to use tetrodes that have much more emission capability than what is needed to make the required audio power. Reduce the screen voltage as much as practiceable and still have plenty of audio power. As the screen voltage is reduced the available plate to plate monkey swing voltage increases therefore more audio power can be developed across the rather high plate to plate load impedance provided by the mod transformer in the Vallint. There will come a point when the screen voltage is lowered to the point that with consequent reduction in bias voltage that the modulator tubes will have to be driven into grid current.IE: AB2 operation. I have not determined where that point is You could probably get away with 105 volts on the screen and still be in AB 1. The audio driver using the 12BH7 in therory should have plenty of snot to drive the 6DQ5 tubes in AB2. The driver transformer does have a step up ratio which may cause some waveform clipping or limiting when driven into hard AB2 operation. Ideally one would want to use a step down ratio transformer.Negative feedback can be placed around the modulator by a 1 Meg resistor from the appropriate modulator tube plate to the cathode of the 6C4 (2200 ohm cathode resistor) Be careful not to accidentally have positive feedback . Things may pop and blow up. Be Careful. There are two bypass caps that need to be changed. They are .01 disc ceramic caps that are at the bottom of the plate RF choke . Change them to .001 3Kv disc caps.  The PA stage in the Valliant does not quite operate in class C Efficiency and modulation liearity can be improved by doing the following. The screen dropping resistor in the Valliant is 12 K . This is too low a value and places the screen input at the edge of max dissipation of the 6146 PA tubes. Add a 5K 25 watt resistor in series with the 12 k to get a total value of 17 K .A total value of up to 20 k is ok.   . Add a resitor of 6.8K 2 watt in series with the arm of the PA bias pot. This will allow for partial grid leak bias of the PA stage. Change the 500Pf TV doorknob cap to a pair of 1000Pf (2000 Pf total)transmitting ceramic doorknob caps. This improves PA efficiency on 80 and 160 meters.Once again the mod transformer is an issue. It is on the edge of saturation. A 10Hy 300 MA reactor may be used externally with a 10 Mfd 1Kv oil condensher as the components of a so called modified Heising or shut fed choke coupled modulator. The circut can be plumbed in through the 9 pin accessory plug in rear of transmitter.The reactor takes the place of the dc current handling of the secondary of the modulation transformer. The end of the secondary of the mod transformer that connects to the B+ connects to the 10 MFD cap . Other end of cap goes to ground. The end of the secondary that connects to the PA stage remains with one end of the 10Hy inductor connecting to this junction. Of course you may modulate the Valliant with an external modulator. A big solid state amp and a Hammmond 75 watt 5000 ohm@300Ma dc audio transformer as the modulation transformer and say the hell with it. Good luck Ed and all the othersthat read this post that have a Valliant and want to get ther most out of whats there . De Tim WA1HnyLR

[scottdarling1221]

Hi guy im new here . Im trying to modify a old Johnson viking valiant 1 so it has nice loud punchy audio.What do you guys think of this modification.I dont want to go crazy with mods i just want to keep it simple and not butcher the radio. Thanks

http://www.maineam.org/documentation/project-capkits-johnson-valiant-manual.pdf

[Ed/KB1HYS]

Nice Necro! 

     Looking at the kit, and the instructions, it certainly won't hurt and a couple of the mods are probably a good idea regardless (like replacing the 18K/2W Chernobyl resistor). Some of the components are prone to failure, like the resistor above, and the "special" loading cap, replace those if you do nothing else. The other component changes they recomend are real close to what Tim HLR advised above (I wonder where they got the idea )   I've done most of what Tim recomends to my Valiant, I get very good audio reports generally.  I couldn't find out how much they want for the "Kit" but replacing the components (reistors and caps) listed in Tims post above costs less than $20 including shipping from Mouser.
       Personally I think the Valiants get a worse wrap than they deserve, they weren't designed for broadcast audio quality, just telephone communications quality (kinda like trying to drag race in a stock '57 chevy and complaining about how slow it is).

I wouldn'd bother about filament conversion either, if you've got hum it's probalby easier to just bypass the fils with some ceramic caps. 

[WBear2GCR]

I'd not put the plate switch on the rear... makes it bad for tuning up.

I'd sooner use double shielded coax and properly ground it than move the mic jack... add a small copper box around the rear of the connector for improved RFI rejection, if that's really a worry - use roof flashing or some cut up PCB material soldered together for that trick...

              _-_-bear

[Ed/KB1HYS]

I wouldn't move the plate switch either, but owners choice there. Better to shield the audio like Bear said. Still if things are working right I don't think you will pick up a lot of hum in there.

You can also pull the "SSB" input 239 and you have a neat hole for something. I used it for a fuse holder. I thought it was better than drilling yet another hole.

Other things I've done to make life easier. 
Removed the crystal holder that was used as a relay socket on the back and replaced it with two pin type banana plug sockets. No metal work necessary and very easy to interface to as well (no special connectors needed).

Removed the 9-pin accessory socket and bypass caps/coils and hardwired the circuits that it jumperd. As long as you don't plan on externally modulating it this is fine. If you want to keep that, check the thing over for carbon tracks and signs of excess heat inside the chassis and out.

Check the calibration on the front panel meter. Sometimes they are way off, either because the orignial builder didn't do the shunts right or just from age this is helpfull to do Before you try toubleshooting so you can get acurate operating info... don't ask how I know 

On the base of the plate choke is a small molded choke. This can Zorch. consider replacing it and the associated caps with beefier components.

Some people recomend replacing the 866's with something (anything) else. IF you understand the limitations of MV tubes and like them no big deal. IF you want the rig to be more forgiving go with 3B28's as a drop in replacement or solid state with diode strings.  NOTE that the filament line for these rectifiers was pretty poorly insulated in some earlier valiants. Check the filament line for shorts to chassis

Put in a fuze on the hot side of the mains (only), and change the bypass caps on the mains where they enter the chassis. Go from HOT to Neutral to GND the original goes from HOT to GND, and Neutral to GND this is a potential problem.

[scottdarling1221]

do  c99 and c100 have to be a polarized capacitor or can it be a unpolarized cap.I noticed a black line on either cap on one end.I bought a couple replacment caps from allied elec. Both illinois capacitor  630 wvdc metalized polyester .1 uf and a .47 uf.Both unpolarized.