[AMRadio] HV Capacitor question

Jim Candela jcandela at prodigy.net
Mon Jul 17 07:33:34 EDT 2017

I have an unusual question. 
At my work, we use a electrostatic clamp to secure a silicon wafer while it gets an ion implant. The clamp controller supplies +/- 1.44 KV under the wafer, and uses a thin insulator to keep the HV current draw zero.
>From each HV output, there are three .01uf  3KV Z5U ceramic capacitors in parallel. These caps feed an op-amp oscillator that makes a ~ 20.5 Khz square wave. The frequency varies with wafer on or off the HV clamp. The presence of a wafer capacitively bridges the two HV electrodes under the wafer.
When things are screwed up, the oscillator frequency is not repeatable. There could be many reasons, but for the sake of discussion, I've found it is usually the .01uf 3KV Z5U capacitors.
If you look up similar capacitors, the Z5U designator for the ceramic dielectric is most telling. We might use one in ham radio for a plate bypass capacitor, but never in a high Q tuned circuit. The capacitance change with temperature is great, and they cannot stand much RF current. These capacitors are also piezoelectric, i.e. a sudden change in voltage across the capacitor will make an audible click...like a speaker!!
I just learned another characteristic that I've not found anywhere in print. If I take a DVM, and clip the leads across the capacitor (used bad ones), I will measure a voltage!! Maybe something like 20mv, and varying with temperature by just grasping the device with my fingers. Now the kicker, wave the device with a heat gun, and as the device warms up, the voltage increases upwards rapidly to 1 volt!! A good (used) capacitor, same type, will never go beyond 20mv, hot or cold.
Using the same DVM with a capacitance range, the measured value moves a lot (> +/- 50%), but which way depends on the test leads (red, black) are applied. I dismiss the measurement as invalid due to the voltage generated by the heated capacitor messes with the DVM measurement method. 

There seems to be something going on with the dielectric, or a mechanism from dielectric absorption.  

One clue is with resistors. Some resistors such as leaded metal film type crimp the resistive element to the leads inside the package. Each crimp joins two dissimilar metals making a thermocouple (TC). Heating a TC generates a voltage. With the resistor example, two TC's are in series, and back to back, like +- in series with -+ such that the voltages offset each other so long as each TC is the same temperature.  
It would take dozens of TC's in series +-, +-, +-, etc. to make a volt...
What is going on with my "bad" .01uf 3KV Ceramic Z5U capacitors??

More information about the AMRadio mailing list

This page last updated 19 Feb 2018.