High Voltage Ceramic Capacitor ESR

[ignilux]

Hi, all-

I'm in the process of troubleshooting a Tektronix 485 oscilloscope, and I currently have the transformer and inverter boards on my bench to isolate a fault (which doesn't appear to be present on these boards after testing). These can be found on pages 152 to 159 of the service manual. Anyway, they were a bit of a PITA to get out of the chassis, so while I have the boards out I've been testing any capacitors that regularly see high voltage. Most of these are ceramic, which AFAIK are touted for their superior ESR. So far, almost every single one has tested open circuit at ~100 Hz, ~1 kHz, and ~100 kHz. I should mention that I'm testing in-circuit, but the test signal applied by my meter is around 225 mVpp. Also, I could see in-circuit testing leading to artificially low values due to parallel impedances, but I don't see how it would give an artificially high value. I've used the same meter to find a dried up electrolytic filter cap in a 475, and have tested a number of resistors for calibration, and known-good and known-bad caps for confirmation that it functions as expected.

I find it hard to believe that every single one of these have failed, especially since the 'scope powers up and works mostly okay. I.e., the major functions all operate, but with some eccentricities that need resolving. So anyway, what gives? Is this operator error? Do high voltage ceramics have unusually high ESR?

Thanks!


Edit: I should also mention that I've found one film cap that measures open, in addition to one that measures sub 1 ohm on the inverter board. The plot thickens...

[Kleinstein]

The high voltage ceramic caps are usually low capacitance. At this capacitance range one normally uses the loss factor tan delta and not the ESR. They usually don't fail with an increased loss.

The usual ESR meters are made for much higher value electrolytic caps, so more like > 10 µF to get a correct reading.
In addition the meters are usually not good enough to measure the very low loss values of the small ceramic caps.

[ignilux]

Interesting. I'm usually more concerned with the reciprocal of dissipation factor, but it makes sense here. I'm also just now realizing that measuring a cap essentially places a high-pass filter in series with the test signal. The relatively low frequencies that I'm using may simply be attenuated too much. Sounds like a good excuse to invest in one of those shiny GenRad / Wayne Kerr bridges...

[TimFox]

A note about series and parallel models for physical capacitors.  If a very low resistance (from the rest of a circuit, in this case) is put in parallel with a capacitor (low compared with the reactance of the capacitor itself), the values in the series model will show a high series resistance.  The low parallel resistance and high series resistance give a low Q < 1.
See the usual sources for series/parallel conversion, but avoid the common approximations for high Q, where the parallel resistance is much higher than the capacitor’s reactance.

[Kleinstein]

The ceramic caps are usually very robust, so not really the parts that need to be checked. Chances are it gets expensive to get a capacitor bridge good enough to really measure the Q of the capacitors. It also depends a lot on the frequency the capacitors are used at. Losses may go up quite a bit to higher frequency.

In circuit testing is expected to give good results with such small capacitors.

[floobydust]

Ceramic capacitors (Class II dielectrics X7R, X5R, Y5V) have an aging mechanism and their value drops with age. I ran into this repairing vintage gear from the 1970's, after thinking it was only electrolytics and tantalum capacitors that fail with age.

"Under the influence of thermal vibration, the ions in the crystal lattice continue to move to positions of lower potential energy for a long time after the dielectric has cooled down below the curie temperature. This makes capacitance aging, whereby a capacitor's capacitance continually decreases. "

Y5V will decrease 15% over 10,000hrs: http://www.rfeinc.com/pdf/support/Ceramic%20Capacitor%20Aging.pdf
https://global.kyocera.com/prdct/electro/product/capacitor/ceramic/mlcc-handling/design/contents09.html

[David Hess]

Interesting. I'm usually more concerned with the reciprocal of dissipation factor, but it makes sense here. I'm also just now realizing that measuring a cap essentially places a high-pass filter in series with the test signal. The relatively low frequencies that I'm using may simply be attenuated too much. Sounds like a good excuse to invest in one of those shiny GenRad / Wayne Kerr bridges...

Even an impedance bridge is the wrong instrument in this case.  The common failure for high voltage capacitors is leakage which may only show up at the operating voltage.  This goes double in oscilloscope DC restorers and voltage multipliers.

Instead of testing the capacitors for leakage at high voltage, it is usually just better to replace them if leakage is suspected which does not sound like the case for you unless the blanking is not working correctly which indicates a fault in the DC restorer.