Neutralizing cap voltage rating

[Barrie]

I'm getting ready to start drilling holes and mounting components for a 10, 15 and 20 meter transmitter.

I'm especially motivated since last Sunday I heard three different stations (including KH6U) calling CQ AM on 15M with no takers.

I've got three reasonable choices for the neutralizing caps for this push-pull rig.  Two are smallish piston style caps, and one set is disc.  The only part number I have is for the disc, which is National NC-800.

My question is relative to the voltage ratings for neutralizing caps in general. 

Nowhere (that I can find) is the voltage rating actually given in any of my handbooks.  The closest I see is "twice the spacing of the tune cap".

The grid to plate capacitance of the tubes I will be using (VT-127As, which are similar to 100TLS) is roughly 3pf (according to my AAD meter.

When I set any of the neutralizing caps at 4pf and hipot them I start getting arcing at about 8500 volts DC (slightly higher with the disc style.

Well, RF is AC, so I'd likely get arcing at a lower voltage than the above.

I was hoping to run the tubes with 2500 volts on the plates.  The power supply is variac controlled, so I could run higher, or lower voltage.

The tune cap is a Millen, rated at 6000 volts.  The rotor will be bypassed to ground.

Any hints or clues as to where I'm at here?

73, Barrie, W7ALW

[KM1H]

National lists the recommended tubes for all their discs in their catalogs and HB ads.

The NC-800 is for low power 809, 35TG, HK-54, 5514 and similar

The NC-75 is for 811, 812, etc

The NC-150 is for HK-354, 250TH, etc

The NC-500 is for 251, 450T's, 750T's, etc

The "similar" and "etc" are direct from the specs so its up to the user to do some homework

Johnson and Bud should also have ratings. Ive used the piston Johnson on 810's and 250TH's but dont remember the model, it was eons ago!

Ive also made my own from coax.

Carl
KM1H

[k4kyv]

The neutralising cap has to withstand the full modulated DC voltage PLUS the full modulated rf voltage. If it arcs over, the full plate voltage will be applied  to the grid  circuit, possibly damaging some of the grid circuit components and maybe the tube itself.

Running 2500 volts on the plates means the peak DC voltage is 5000 volts at 100% modulation in the positive direction. The peak rf voltage is somewhat less than 5000 volts, but it wouldn't hurt to rate it for the full 5 kv for a safety margin.  That means the cap should be spaced to withstand at least 10 kv.

A measure of protection, as used in the BC-610, would be to place a mica cap rated at least 5 kv in series with each neutralising cap on the grid side.  The capacitance is not critical; anywhere from 100 pf to .001 mf should do, but use the same value for each cap. This won't offer much protection against arc-over, but if it does happen, it will prevent a sustained DC arc from the plate supply from toasting grid circuit components and the tubes.

With the disc type capacitor, the spacing increases as capacitance is reduced, so the voltage rating is higher towards minimum capacitance.

My HF-300 rig uses conventional rotary bread-slicer neut caps,  rated at 12 pf each, with about 1/2" spacing.  I don't have any series caps, but never have had any arcing problems.  I run 2000 volts on the final, but the positive peaks may reach 125% or thereabouts.

The 8005 rig runs 1350 volts on the final.  The neut  caps are the piston type with about 1/4" spacing.  I don't use series caps with those, either, and never have had an arc-over.  I think in both  cases, the final tank cap acts as an arc gap for safety purposes. The plate cap arcs over before the neut cap.  I do occasionally got arc-overs at the plate  cap in both rigs. The plate cap spacing in the HF-300 rig is 3/16" and the one in the 8005 rig is considerably less.

I recall a circuit from an early 50's article, either in the Bill Orr handbook or in QST. They used the series mica caps as described above, plus a 1-megohm resistor across each neutralising cap.  The resistor must be able to withstand the full 10 kv, so something like a string of 100k resistors would be needed.  Check the string with the hi-pot before installing in the rig. The 1-megohm resistor protects against arc-over, by keeping the neut cap completely discharged at DC potential. Otherwise, the cap may develop a charge at near the full DC plate potential, reducing the additional amount of rf voltage it can withstand. If the resistors are non-inductive, 1 megohm across the neut capacitors shouldn't otherwise affect operation of the circuit.

[WD5JKO]

I can tell you with certainty that you should NOT try this:

"You mentioned using glass. I'm not sure this is relevant, but I once had a Push Pull 8005 cross neutralized 500 watt RF amp. I converted this over to 805's, and the higher Cgp capacitance of the 805's resulted in the neutralization caps being too close together to standoff the voltage. I got an idea, and took the glass out of two round panel meters and inserted the glass between the round disc neutralizing capacitor plates. Then I neutralized the amp for a nice output null. OK, so far so good. Next I keyed the B+ to see how the amp ran with this modification. The result surprised me. Both pieces of glass shattered immediately, the RF amp flamed out, welded the HV relay contacts such that I couldn't turn it off, and then I had to wait for the garage located 15 amp circuit breaker to pop."

been there, done that..No Fun

Jim
WD5JKO

[Barrie]

Thanks Carl, Don and Jim:

I was hoping to use the small neutralizing caps in this rig as just one way to keep the whole thing small.

I've been thinking about trying some experiments with the NC-800 caps in order to increase the voltage rating.

Some of those thoughts are:

Spray the plates with several coats of clear Krylon.

Coat the plates with 3M liquid tape.

Coat the plates with epoxy.

Glue a sheet of Lexan or Teflon to one or both plates.

Actually, all these possible solutions would also increase the dialectic constant somewhat, so I could achieve the same capacitance with slightly greater spacing.  Sort of a "two-for-one".

I have the capability to hipot, so I should be able to determine which of these methods works best.

Or, am I just wasting my time?

73, Barrie, W7ALW

[WD5JKO]

I have the capability to hipot, so I should be able to determine which of these methods works best.
Or, am I just wasting my time?
73, Barrie, W7ALW

   Barrie,

   A DC or 60Hz AC hipot is not the same as RF. Between those plates the electric field is sky high. I'm pretty sure most of the options your pondering will repeat my test, i.e. holy sh_t turn it off!!!

   The one option you mention that I'd consider is a Teflon sheet. I would suspend it in the middle between the plates. You don't want any trapped air in their like you will have if you attach the Teflon to the plate(s). In my situation I found some 1 pf fixed vacuum caps that I installed in parallel with the existing neutralizing capacitors.  That was a bit bulky, but the result worked like a charm.

Jim
WD5JKO

[Barrie]

Jim:

That sounds like a good and simple solution.

I had considered that since I have a couple of 2pf vacuum caps here.  However, with the 2pf cap in parallel with the disc neutralizing cap I end up with 3.15 pf minimum capacitance.

That's almost exactly the same capacitance of the tube, so I don't have any "wiggle room".

Unless it's most likely that I'll be using more neutralizing capacitance than the tube has grid to plate capacitance.

Is there some rule-of-thumb regarding on which side of the tube capacitance the neutralizing cap will end up?

73, Barrie, W7ALW

[WD5JKO]

Is there some rule-of-thumb regarding on which side of the tube capacitance the neutralizing cap will end up?
73, Barrie, W7ALW

http://www.vias.org/basicradio/basic_radio_28_04.html

"Two condensers (CnCn) are needed to neutralize a push-pull circuit, as in Fig. 27 I. The capacitances of the two condensers are nearly equal to each other and to the tube's grid-plate value."


   I'd try it first. You could possibly alter the adjustable NC caps to achieve a wider gap? Alternatively, maybe use different tubes, or 4 in push pull parallel?

Jim
WD5JKO

[KM1H]

Try neutralizing it cold with a small variable, measure and replace with a small piece of RG-17 or home made coax.

Bet it works.

Carl
KM1H

[Barrie]

I want to thank everyone for their comments!

I'm not sure if I've fully solved all of my problems regarding the neutralizing caps, but there could be some usefull information for someone here.

The original hipot test on the NC-600 cap (flat plates), set at 3 pf, resulted in arcing at 8800 volts DC.

I mixed up a batch of plain-old Ace Hardware epoxy cement and coated both plates with as thick a coat as possible.

After drying, I again set the cap at 3 pf.  Before the coat of epoxy, the spacing was 5 mm.  Attempting to disregard the epoxy, the after spacing looked like 6 mm, which indicates that the epoxy may have increased the dialectic constant.

The arcing voltage after coating was 14,500 volts DC.  Watching the arc, it appeared that the cap was arcing from the top of the top plate to the bottom of the bottom plate.  So, I think that if I coated both the top and the bottom of both plates the standoff voltage would again increase.

I also tried something similar with a Johnson #159-250 piston cap.  Set at 3 pf prior to coating it arced at 8500 volts DC.

I ordered a bottle of "SUPER Corona Dope" from Newark Electronics.  The specs say 2100 volts per mil air dryed, or 4200 volts per mil heat dried at 275 degrees for three hours.

I put a good coating on one of the piston style caps and did the heat dry thing.  The XYL was not happy about me using her oven that way.  I explained that she must not stand in the way of science.

I then hipotted the thing.  Instead of arcing at 8500 volts, I was able to advance the voltage to 15,000 volts (the max my hipot will do) with no sign of arcing.

With the leakage meter set at 2 microamps, full-scale, the needle did not move off zero!

So, there you have it.  I hope that with this rather long-winded description in the archives it will save some other dummy some head scratching and trouble.

73, Barrie, W7ALW