What makes a capacitor pulse rated?

[ppppenguin]

I've been sorting out a box of odd components and found a number of MKP1841 capacitors: https://www.vishay.com/docs/26019/mkp1841.pdf

Does anyone know what makes them rated for pulse duty? Conversely why many ordinary film caps won't surive that kind of duty?

[Nick]

My understanding is the dielectric film is more robust.

Somewhere around I saw an electron microscope photo of the film from ordinary capacitors that had been abused with high current discharges - the film had loads of tiny holes...

[pwdrive]

Don't know the ins and outs but I suspect better dielectric stress shock, the old Radiospares ones looked like tubular ceramics but with one of the connections located approx. midway of the length, the Radiospares ones I remember when doing TV's were red (ish) in colour.

Lawrence.

[ppppenguin]

I remember those pulse ceramics. Still have some in a drawer.

My suspicion on the film caps is that they are just bigger and heftier when rated for pulse duty. Thicker film, thicker foil. I also suspect that the DC rating is very conservative.

[Mike Watterson]

As a rule of thumb decide what the peak AC voltage is and pick a DC rating a little higher.
So a capacitor on the anode of a valve with a transformer load might need HT x 2 plus a safety margin.

I did work in QA in a capacitor factory for while. The capacitors were tested with static DC and sometimes -55 to +125 °C. For some applications the DC voltage was twice rated and for an hour.

Pulse rating is complicated and depends on dielectric material, metallised or metal foil, aluminium or silver, frequency /power/duration, if the material/construction is self healing, goes open circuit or short.
See also capacitors rated X (across mains), Y (Mains to "Earth" or chassis) and some usually only ceramic types are X Y.

Also ceramic have both widely different dielectric types (temp co, piezo effects, K (how much times vacuum capacitance is multiplied, voltage rating per mm, loss at higher frequencies). Also ceramic can be multilayer which is much lower inductance than a rolled foil type). The thickness of the metallisation and area of the dielectric affects power handling.
I made a tuned 1 MHz 20 W SMPSU once and was surprised to see the ceramic capacitor suddenly glow and explode into sand. A high power class D amp needs carefully chosen caps on the first stage of the filter.

Capacitors have more range of values, materials, voltage and power handling maybe than anything? Printed coils do exist. But printed capacitors are common and at high frequencies the connection is to a point of a cheese triangle shape.

Film caps can be metallised plastic or using separate aluminium foil. Common materials are: Polystyrene, PTFE (Teflon), Polycarbonate and the most common is Polyester = Mylar (Magnetic tape and photo film) = PET (as in bottles).

Paper is rubbish unless in hermetically sealed cans. "Oil", "Wax" is all usually just variations of paraffins. Candle is a Paraffin, as is diesel and kerosene. They all gradually absorb water!

Electrolytic types are using a chemically created thin insulating film on the tantalum or aluminium, which is why reversal is catastrophic. It's why they go can high ESR with age or have high leakage and why controlled reforming often works. Like a lead acid car battery, the polarity can be reversed on some Aluminium types if you are careful and know how. Not recommended.

[Mike Watterson]

the old Radiospares ones looked like tubular ceramics but with one of the connections located approx. midway of the length, the Radiospares ones I remember when doing TV's were red (ish) in colour.

Those were ceramic and the place I worked sent any working but slightly outside of capacitance spec to Radiospares as they never ever made any returns. If they couldn't make production quota.
I discovered this when I was asked to help in production. QA allegedly didn't know!

[Kalee20]

My understanding is that the pulse rating is quite a lot down to the metallising.

Pulse currents flow from the leadout wire into the metallising, thence exist as displacement currents (I hate to say 'through the dielectric' but hopefully the meaning is clear!). So the bits of metallising remote from the leadout carry small currents. As you approach the leadout the current grows, and by the time you get to the junction between metallising and leadout, the current is at its maximum.

If the metallising can't carry this current, it degrades, stops conducting. In a stacked-foil type of capacitor (Siemens) the capacitance would reduce, because each layer is isolated from the rest. In a rolled-foil type, it wouldn't, because the layers are all linked, until the last death throes, but ESR would increase.

So high pulse rated capacitors need either foil electrodes rather than metallising, or lots more metallising.

Mike W's comments 'from the inside' are very informative - thanks!

[Mike Watterson]

My understanding is that the pulse rating is quite a lot down to the metallising.

I would suspect this, but I've not seen it written down, nor was it discussed in the capacitor factory, or if it was I've forgotten.