Question: Doorknob VS Ceramic

[Ed/KB1HYS]

IF the cap value and the voltage rating are the same, why use a Doorknob type vs a Ceramic?

Doorknob caps seem to be mighty expensive compared to Ceramics.

500 pf  10kV doorknobs   10-35 dollars each depending on source.

Ceramics under a buck each. 

Yea, I'm collecting parts for a project that requires 4 500pf 10kv caps. And was just wondering.

[Mike/W8BAC]

Is it RF or DC Ed? The door knob type for RF, ceramic for DC.

Mike

[w8khk]

It really depends on where in the circuit you use the capacitor.  Doorknobs should be used in tank circuits where circulating RF currents are large, and for coupling from the plate of the tube to the tank or pi-network circuit.  This includes the function of loading capacitors in pi-networks.

Ceramics may be rated at the proper capacity and voltage, but are not designed to carry large amounts of RF current.  Ceramics in the above applications may fail with vigor.   Ceramics may be used for bypassing or coupling where the RF currents are not great, but DC voltages are high.

If you look at some of the old broadcast fixed capacitors, they have a capacitance rating, a voltage rating, and usually a current capacity rating at 1000 KC.

[Opcom]

for some reason the typical 0.01uF/6KV ceramic disc will be fine on a pair of 6146's or sweep tubes, but will go boom in a KW amplifier. That's why doorknobs. Could be the ceramic uses a poor dielectric never made for, or in insufficient dimensions for lots of RF current -but it's OK at 60Hz or 15KHz, etc. Even TV set doorknobs might not be enough. There are 5 (500pF 20KV) in parallel in the Tucker KW and they are cracked, perhaps from heat generated by RF current.

[N3DRB The Derb]

I always used disks at double and a half the voltage needed for DC bypassing and doorknobs for RF coupling at 3X the expected RF power. it costs no more to get way over rated parts. the dc blocking cap going into maul 1 is a Sprague rated for 18KV WVDC. it was 3 bucks and beefy. Why put a wimpy 5KV exact replacement back in the size of a dime?

Always overbuild. the price is almost always teh same or less.

The exception is when you start getting into vacuum variable class power. then good parts start getting expensive.

[WD8BIL]

Current,Ed. It's all about current!

[W3FJJ]

I could be wrong but I think Doorknobs are ceramic, they are not Mica which
would be the better type to use in an RF Tank Circuit. I tried using a single 500pf 20kv
doorknob  in a loading circuit for 200 watt transmitter and found it getting warm. I believe
a better way is to use 5 100pf  dorknobs in parallel to increase the current capability.
Doorknobs seem to work fine as coupling caps between the plate, and tank circuit, I don't
think there any circulating RF current like you have in the plate and loading circuit.

[Ed/KB1HYS]

Ok, so Doorknobs in High RF current spots like Plate-tank circuits, ceramics for low current like bypass or light coupling. Cool.

[WA1GFZ]

all dorknobs are NOT created equal. NPO is the best type in a tank circuit because the value is more stable. Bypass and coupling types can be any type that handles the current. Anything with XR7 dielectric is very unstable for value when it heats up.
rule of thumb if they heat up good chance they will fail in time. They blow up like a nice firecracker. 

[flintstone mop]

uh oh. The Derb uses vacuuuum variables?? I thought they were 'gay'.
Phred

[KF1Z]

Don't forget about the new multi-layer ceramic caps...

Very small size, handle large RF current, under $15. each.

Not sure if 2kv is high enough voltage rating though...

[Ed/KB1HYS]

2kV is not enough.  I need 4  500pf/10kV units for the QRO rig I am building.
There are some "deals" on the 'net. well $8 vs $35 (that place in Nebraska is NUTS!!)
Anyway much work to do before I'm ready for these anyway.

[WA1GFZ]

I've flashed across the electrodes of MLCC caps doing lightning tests. I would not run them over the voltage rating.

[John K5PRO]

Doorknobs are made with ceramic dielectric. But they have large electrode area on each side, hence the shape. A screw thread is soldered onto the metallization somehow. Ceramic disc capacitors, on the other hand, have small electrode areas with a wire connection. RF current through these causes heating, due to I^R loss. R being the series resistance (remember skin effect at RF too). Ceramic discs are rarely rated with RF current, only voltage. Doorknob caps are rated with both.

As someone pointed out, there are differences in doorknob capacitors. If the dielectric constant (K) of the ceramic is high, they can pack a lot of capacitance in a small unit. However, the tempco is going to be hell. They will drift in tuned circuits as they heat. Something else I found out last year, is that they can be used like varactors! High K ceramic capacitors, when biased with high voltage DC across them, vary in C depending on the voltage.

The cheap old plastic or phenolic doorknob cap that used to be seen in B&W television HV circuits (500 pF @ 20 kV) are NOT good RF capacitors. Hams have gotten away with them in parallel, but they often go up in a poof.

The best tempco ceramic doorknobs are < 100 pF, due to low K ceramic. These are sometimes NPO rated, and excellent for output networks in amplifiers. They can be paralleled to get higher C without drift.

As KF1Z pointed out, the multilayer ceramic caps are excellent, as they have the lowest L and the highest self resonant freq. But they tend to be limited in voltage handling. I used them in the loading capacitor position of a 8877 amplifier at work.

[WA1GFZ]

I melted to solder in a 1000pf 5 KV dork-knob in my linear when used as a loading padder. Trashed it on the first key up.

[Opcom]

A guy had a nice jennings 5-500pF 10KV ceramic vacuum variable at the gainesville fest for $40. I do not know how to tell however if it is good or bad. I've been warned to look in the glass ones into the gaps and see if there are spots on the cylinders, and if there are, it has been arced and is no good. Anyone comment on that?

[WA1LGQ]

         How would one test a doorknob cap, or other cap for that matter for current capability other than trial and error explosion test? Maybe some kind of tuned circuit/dummy load on the output of a transmitter with somewhat variable output power and an rf ampmeter? I've got a number of unknown doorknobs here and it would be nice to know more about them. ..........Larry

[Mike/W8BAC]

Opcom wrote

I've been warned to look in the glass ones into the gaps and see if there are spots on the cylinders, and if there are, it has been arced and is no good. Anyone comment on that?

I have 3 brand new Russian made 20-1000 Pf @ 10KV vacuum variables here. All three have lots of spots in them from arc over. The Russian caps still use glass envelopes. Reason given was quality control testing at the factory. I never saw a spot like that on a US made cap. I don't think a spot would be a reason to pass by a good deal but I would certainly look for any sign of smoke or debris.

Mike

[WA1GFZ]

You test a cap by putting it in the circuit and slowly increase the power to see if it heats up at each step. As a loading padder you will see the power output change as it heats up. This means the value is changing detuning the tank.

[w4bfs]

a vacuum cap is constructed much closer spaced than an equivalent air variable.  If the vacuum is lost or even worse, partially lost the cap will tend to ionize whatever gas is in there and can get localized heating. (the spots)  best way to test the vacuum is with a hi pot unit.  It will let you know ifn it gassy.  made a hi pot tester (to 4.5 kV) last winter    don't know how I ever got along without it ...  its a good use for a low current hv xfmr and small variac

[Opcom]

ahh thank you.. I have a 20KV "plastic capacitors" power supply I could use for hi-pot. Didn't think of testing them that way.  I guess a large resistor (series bank for voltage) maybe 20 megohms. would be safe. Good!