FIXED LOADING CAPACITOR FAILURE & REPAIR
FOR JOHNSON VALIANT, VIKING II, DX-100 & COLLINS 32-V
  
BACKGROUND

I have received numerous requests over a period of time for information regarding the fixed loading capacitors used in the "COARSE" loading circuit of the Valiant, Viking II, DX-100, and other transmitters. This article is a compendium of some correspondence that I have exchanged over the last couple years regarding this problem. I have referenced work by W8JI, KA3PTX, and N9PVF on successful repairs at the end of this article, with a photo and part numbers to order. Full credit is given to the source of the information. I am just trying to get all the true and relevant facts assembled in one place, and eliminate the misinformation on these subjects (and there is PLENTY of it out there on the internet).

All three of these rigs use the infamous "TREE" capacitor or a variation, which is a multiple element Mica capacitor providing the necessary step change in value to increment the capacitance in parallel with a variable "FINE" variable loading capacitor of approximately 250 to 350 pF. The object of this circuit is to provide a loading adjustment that is less touchy on the upper frequencies, while providing sufficient capacitance to correctly load on 80 and 160 meters.

Note that the Heathkit DX-100 B eliminated the fixed loading capacitors completely and went to a wide spaced 3 gang broadcast band tuning condenser (350 pF per section). They also sold an update kit for the DX-100 that made that modification to older DX-100s. I have considered this solution to the problems that come up in the E.F. Johnson rigs, since there is plenty of room to implement it. To avoid "touchy" adjustment on the higher bands, a Jackson reduction drive and pointer system is available as replacement parts from Ameritron that could be easily adapted to this application. Of course, the appearance of the rig would be changed. Another way to deal with it would be to use only one section of the 3 section variable capacitor and the existing "COARSE" switch sections to add in the other two sections, or additional padding (for 160) as needed.

Note that the 32V Collins rigs use a similar circuit, but due to the Pi-L configuration, much higher component ratings are needed than provided in this article. However, you may find that an appropriate network of parallel (to provide current handling) and series (to provide voltage handling) may be possible with the 32V. But you are on your own to figure out what exactly to use for parts. I have no rig to experiment with. If you compare the 32V3 with the earlier models, they did a series-parallel arrangement to improve the situation. Probably if you used a dipole similar to the broadband dipole shown in their manual, or the one shown on my webpage, this might be enough. But if you used an open wire fed flat top with a matchbox, it probably will not be enough.

WHY DO THEY FAIL?

As W8JI notes in his comments, they just get old, and they were not hermetically sealed. Eventually, moisture penetrated the square plastic enclosing them. Leakage resulted. Often even receivers will fail this way due to smaller "postage stamp" capacitors. The common failure in IF transformers is the same, since mostly the mica capacitors embedded inside the can get moisture induced problems. I NEVER throw away the IF cans; I disassemble them, remove the internal capacitors, and replace them with an external modern fixed mica capacitor, sometimes with small trimmer capacitors in parallel (if the slug is frozen in the adjustment coil).

In the Valiant, I place a 33K 2 Watt resistor across the antenna terminal. Many rigs used a 2.5 mH RF choke across the antenna terminal. This served to bleed off any DC leakage from the plate circuit, as well as any static voltages from the antenna. In cases where the input coil of a transmatch provided a DC ground, this was not needed, so the older rigs did not have one. The modulation voltage coupled through the plate blocking condenser can also cause some AC voltage across the loading capacitors, if the rig is connected directly to a dipole or antenna that has no DC return. Remember all these possible voltage sources are additive to the voltage caused by SWR in the antenna. Cumulatively, this can exceed the DC rating of the fixed loading capacitors and cause failure based on voltage.

Reactive loads from an antenna or misadjusted tuner can cause failure of the fixed loading capacitors due to exceeding the current rating. This is likely the most common cause of failure.

Old Dow Key relays can develop intermittent or dirty contacts. There is a way to adjust them and clean them, but I am not going into that here. They are notorious for developing a coating of oxide. I tend to avoid them despite their old buzzardly charm and use modern relays with gold plated contacts for better reliability. You do not have to give them a smack to get the receiver to come back to life.

Let me first categorically state that of the three Johnson Valiants and the two DX-100s and three Viking IIs that I have owned I *NEVER* have had a failure of these fixed loading capacitors. However, I have abandoned the use of open wire and gone exclusively to coax fed resonant antennas with low SWR. The Johnson Valiant, Viking II, and Heathkit DX-100 all state that they will load 50 to 600 ohm antennas. The 600 ohm is not specified as resistive, but trust me, the Valiant wants a non reactive load. I had to develop a broad band dipole for 80 meters because the Valiant plate tuning range is so restricted, it will not tune properly at the ends of the band. It is also unwise to operate these radios directly into an 600 ohm antenna such as a windom fed with a single wire or open wire. This is true based on observed component failure. It is also true based on TVI or harmonic radiation owing to the smaller values of total loading capacitance to match 600 ohms. This loading capacitance shunts harmonics to ground and reduces to TVI. Yes, I know with modern digital TV, they will not actually hear your voice and identify you.

HOW CAN I REDUCE POWER DURING TUNE UP TO PREVENT FAILURE?

I use an MFJ-212 (click for link) for all antenna tuner matching adjustments. This is a $100 device that will absolutely prevent damage to your transmitter from the radically out of range loads caused by adjustment of a transmatch. The only other solution is to dramatically reduce the output power during transmatch adjustment to maintain the reactive currents and voltages within the ratings of the fixed value loading capacitors. Note that this also can protect the shunt capacitors in the Valiant plate tuning circuit on 160, 80, and 40 meters from damage from similar abuse.

The most common reason for highly reactive loads is during tuneup of the Valiant into a Johnson Matchbox driving open wire line; this is often done at full power output from the Valiant. The Collins 32V had a tune position on it to address this problem. It simply inserted a large power resistor in series with the AC line driving the high voltage transformer; but it was not a truly bulletproof solution. The Heathkit Apache has a tune position on it which is a cheaper and more elegant design - and it works! It places a much lower wattage power resistor to ground (in parallel with the clamp tube) to reduce the available screen voltage during tune up.

W1CKI RECOMMENDS MODIFYING THE "MODE" SWITCH "SSB" POSITION TO A "TUNE" POSITION ON THE VALIANT.

Here is my take on three ways to implement this idea. I no longer have a Valiant to try this on, so this is experimental and may require some cut and try.
  1. The simplest is to use CW mode for tuning. NEVER TUNE ANY VINTAGE RIG USING AM MODE, even into a resonant dipole! DO NOT USE AM mode to adjust a tuner since the modulator will not have a proper load on it. Load the rig into a dummy load on the frequency of interest. Then attach the antenna via the matchbox. Before keying up again, reduce the drive control to minimum. Then slowly advance the drive control until you get an indication on the most sensitive scale of your two indicator (cross needle or otherwise) SWR measuring device. This should show no more than 125 mA plate current on your Valiant. HOW THIS WORKS: The clamp tube detects the inadequate drive and pulls the screen grid on the 6146s low, reducing the plate dissipation of the 6146s to a safe level. Using the drive control to "fudge" the drive to a low level partially disables the clamper and uses it for a "power adjust knob". Once you are done loading with the matchbox, go back to full power (grid drive and plate current at specified levels). Since you started the Valiant with 50 ohm tuneup, only light touch-up is necessary. DO NOT TOUCH the matchbox setting without going thru the whole process again from the start. DOWNSIDE: causes a lot of wear in the unobtainium drive pot. Upside: no mods, all tuneup and bias settings stock. If the pot fails, you always can use the solid state fix on the AMfone website shown for the DX-100. It will work for the Valiant or Viking II.
  2. W1CKI increases the screen dropping resistor in the SSB position to the 6146s by inserting resistance in series with the screen dropping resistor on the SSB position. Could work OK, but screen voltage can go to very high in some loading circumstances. It does limit the power by limiting the available screen current by starving it through a larger resistance. Use the contacts that switch the VR tubes or the screen dropping resistor/clamp tube. BE SURE: The modulator tubes do not have screen voltage on them in any position except AM. Be sure the clamp tube is not disconnected in AM or CW or the new Tune position. Downside: must take a measurement for reference on the RF finals for correct DC voltage on the bias setting before this modification or number 3 below. Log this in the manual. The SSB position now cannot be used to set the resting current to 100 mA in SSB mode for RF finals. Use the Valiant 2 spec for resting current, not the Valiant 1 spec, as described in my Valiant article. (Modulator still can be adjusted per the manual in AM mode for 55 mA).
  3. Modify the SSB mode switch so that a resistor to ground from the 6146 RF screens is added, but all functions of the switch are in CW mode, including clamp tube connection (not the VR tubes used in the former SSB position). This is similar to the Heathkit Apache TUNE OPERATE toggle switch design. Scaling the resistors for three 6146s instead of two will be needed. Keep the EFJ value for the screen resistor (as modified in the TimTron mods), but make the ratio of that screen resistor to the grounding resistor the same ratio. This reduces the screen voltage and because of the divider, the screen voltage and current will be low enough to prevent damage to the 6146s and/or plate tank components during tune up if the keyed up plate current is less then 125-150 ma. Same precautions as in 2. above. Modulator 6146s always disabled by the methods for CW or tune mode, removing regulated screen voltage, and shorting out the modulation transformer secondary, etc.
HOW DO I REPAIR FAILED FIXED LOADING CAPACITORS IN A JOHNSON VALIANT, VIKING II, OR DX-100?

Over the years, I have accumulated the exact replacement capacitors, so I have no problem. I also repeat that I never have experienced a failure of fixed coarse loading capacitors when operating with a coax fed resonant dipole. If you have the original spec part, or its modern equivalent dipped mica capacitor, I recommend that solution. Use the specs in the EFJ manual and you should be OK.

Some people recommend "doorknob" capacitors to replace the fixed coarse loading capacitors. This is overkill and may have some downside to it, due to Equivalent Series Resistance (ESR). Doorknobs are often used on the "plate" side of the PI network or for the plate blocking capacitor, but the impedances around the doorknob are much higher in that environment. Here are W8JI's comments on that method:

MORE BACKGROUND INFORMATION - CLIPS INCLUDED BELOW SO THAT YOU DO NOT HAVE TO READ THE IRRELEVANT 6146b AND NEUTRALIZATION DISCUSSIONS.
http://forums.qrz.com/index.php?threads/johnson-valiant-neutralization.425967/

W8JI, from QRZ.com forum states:

You do not need a doorknob!!! The Valiant is only a 150-180 watt CW and 400 watt PEP AM transmitter! We don't even use doorknobs in 5kW amplifier or transmitter loading systems anymore. At about 5kW large mica capacitors or ceramic take over, but below that other parts are better.

The primary problems with door knobs are temperature coefficient at high capacitance levels and high ESR.

I'd look into why the mica's failed, rather than using a part that is actually less suitable for the job.

What capacitance values do you need, and what were you using before that failed? What was in there, small silver micas or something else? Was that all to make one capacitor of 330 pF??

73 Tom (W8JI)



ALSO FROM W8JI on the same thread:

OK, those old micas just go bad sometimes. They are good caps, but have sometimes passed their life.

The lowest TC you can get in a doorknob is typically around N1000 for a 300 pF, and that probably is OK. If that is C39 for sure, across C8B (the writing is too fuzzy for me) then things are easy. Take the highest load impedance and calculate voltage and current. Let's say 400 ohms worse case.

That's 400^2 over 400 watts PEP (for AM peaks). 400 volts * 1.414 = around 600V peak.

Current is by average power. It is RMS voltage over reactance, but the available current from the pi network would limit things. The problem is we have to look at pi network limits of what it could match, but roughly it looks like 20 meters might be the worse case band. I'd say you are safe with a 5 amp part even with some pretty weird loads.

You could use a ~330pF 1.5 kV chip cap or HV snubber micas. Chips handle 15-20 amps if you don't go over a few kV, are NP0 (no drift) and are the size of an M&M plain. They are tiny parts. Snubbers are physically larger, but are less voltage and much less current, but two 160 pF in parallel will handle that.

The closest I have in chips are 360 pF 2.5 kV chips. I could send a 270 pF chip paralleled with a 68 chip. That combo would handle a few kilowatts, as would a single 360. I think I might have some with wire leads or foil leads attached.

The transmitting chips are a superior replacement for a ceramic doorknob. The voltage is less, but they are far more stable and handle way more current, plus have about 20X the Q. Plus you can hide them. :)



HERE IS MORE INFORMATION ON THAT METHOD, ALONG WITH A PHOTO OF A SUCCESSFUL REPAIR: CREDITS: Ed, KA3PTX AND Dave, N9PVF

MLCC caps used in the Valiant


PART NUMBERS FROM MOUSER:
Picture of what I did on my valiant. Got the parts from Mouser Electronics. Also, a couple links to caps. Might not be the exact part numbers I used, been awhile. But should work. GL.

http://www.mouser.com/ProductDetail/Johanson-Dielectrics/302R29N331KV4E?qs=sGAEpiMZZMvsSlwiRhF8qjXaA2deiFeEVLrxG5rhQeo%3d
302R29N331KV4E Johanson Dielectrics | Mouser www.mouser.com

http://www.mouser.com/ProductDetail/AVX/1206GA750JAT1A?qs=%2fha2pyFaduhcF9yxpAW%252bRpuevkBVDWQevvQEat%252bFGFL8D4gWjoVLmBxscsiBub4u
1206GA750JAT1A AVX | Mouser www.mouser.com

I don't have exact part numbers to order, but Mouser has some Multilayer Chip capacitors MLCC I think will work good in valiant loading circuit.. You want "COG" type, or "NPO".

I just did a quick search 3000v MLCC and here's and example of one they sell, 3000v and COG you will have to solder leads on them, they are very small.. and wont hurt to parallel them for extra current capability but maybe fine as is..

Here's the link: http://www.mouser.com/ds/2/40/aphvc-346708.pdf

REPLACING A BAD COARSE LOADING SWITCH

If you have the rotten luck to have a switch that is arced or burned, you are not completely up an unsanitary river with no means of propulsion. REMEMBER: NEVER MOVE THE COARSE LOADING SWITCH WHILE CARRIER IS APPLIED or you will cause arcing, and may damage the switch and/or the fixed coarse loading capacitors! DO NOT TUNE UP IN AM MODE! If you tune up in CW mode, let up on the key, change the coarse loading switch, then close the key. Adjust the FINE LOAD capacitor with the key down, while keeping the plate tuning resonant. If you are out of range, let up on the key, change the coarse loading in the correct direction, then close the key and continue, until you have everything to spec. I touch up the drive one last time, since plate loading can sometimes influence grid drive. Then one last tweak to the plate and loading. Do not keep the key down for long periods of time, give your finals a rest once in a while. Log the settings when you are done, so you do not have to take a long time the next time you go to that frequency.

If you are lucky enough to scrounge one from a "hangar queen" rig, they are all the same, Johnson and Heathkit. Even the "TREE" capacitors can be scrounged.

You can substitute another newer type of switch. Ameritron sells progressive shorting switches as used in the band switches of their tube amps. You can buy one of these, but you will have to play with the coarse loading capacitor values. The way the original switch works is like a sort of Binary Coded Decimal switch going 1 - 2 - 4 - 8. This configuration saved money by reducing the number of capacitors needed. The new switch will work differently, so you will have to provide a capacitor at each position. Each position of the switch will add one capacitor of the value of the variable loading capacitor, so they are all the same value value, instead of being in a ratio of 1 - 2 - 4 - 8.

You can remove the whole system and go to one like the DX-100B with a wide spaced 3 gang broadcast band tuning capacitor, 350 pF or so per section, as mentioned earlier in the article.

REPLACING A FAILED BANDSWITCH

Note: If you have a failed band switch, you can use Ameritron parts for fixing those too. But you will have to study the circuit, and you will need three wafers: one progressive shorting for the coil, one progressive shorting for the variable capacitors, and one individual switching for the fixed padding capacitors for the 160, 80, and 40 meter positions of the plate tuning circuit. Once again, avoid tuning up in AM mode or at full power; off resonance operation could damage the switch or plate padder capacitors while adjusting a matchbox. Use an MFJ-212 noise bridge to avoid this whole scenario.

Do not forget to tighten and solder the screws holding the large plate band coil; see the main Valiant article for photos and details. DO NOT OVER TIGHTEN THE SCREWS OR YOU WILL BREAK THE CERAMIC STANDOFFS! Intermittent connections to the coil, plate blocking capacitor, or plate RF choke can make bad things happen. See my Johnson Valiant Modifications article.


Click thumbnail to enlarge

73,
Janis
AB2RA
Back to AM Transmitters
Back to Projects
Back to the home page
 
Text size:  +     –