I'm looking through digikey's option on polypropylene capacitors.

None of them seem to specify dielectric absorption.

This is for a low voltage sample-hold circuit. Low leakage is good, but low dielectric absorption is more important. I cannot seem to figure out the difference between PP, Metalized PP, and Stacked PP capacitors.

I'm seeking advice for specific capacitors or a way to decipher the datasheet. For instance, does the voltage rating tend to matter? Does lower leakage/higher insulation resistance tend to mean a better dielectric absorption specification?

Thank you


Dielectric absorption in high K ceramic makes them poor for S&H caps.

Electrolytics all have this property but worse. Teflon is the closest to ideal cap with the lowest Dielectric Absorption. Followed by Polypropylene and NPO Ceramic.

enter image description here The error above is plotted for a 1V step signal into a S&H.

The above was reported by RAP in ED magazine but written in 1982. I have known this since the late 70's from Mil-Std Hdbk's on Capacitor material properties and had worked with long RC integrators and fast ADC S&H's.

If you need small tolerances use NPO ceramic. If you need better performance than above, negative feedback circuits exist to compensate for this property.

enter image description here

| improve this answer | |
  • \$\begingroup\$ The 'sample A/sample B' spread in mylar, polystyrene, NP0... is significant, in that graph. If one were to buy a PTFE (Teflon) capacitor today, it might be 'sample B'? Brandname/model info seems to be needed. \$\endgroup\$ – Whit3rd Feb 28 '17 at 3:13
  • \$\begingroup\$ Caps improved a lot since 1982 but D.A. may not have changed much. The old brands would be obsolete. Besides the spread is only <0.02% error after 100us \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Feb 28 '17 at 3:33
  • \$\begingroup\$ A recent TI App Note indicated "DA is specified as a percentage of the charging voltage. It can range from a low of 0.02% for Teflon, polystyrene, and polypropylene capacitors, up to a high of 10% or more for some electrolytics. For some time frames, the DA of polystyrene can be as low as 0.002%. Common high-K ceramics and polycarbonate capacitor types display typical DA on the order of 0.2%" in 6567.Dielectric-Absorption-in-Caps.pdf \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Feb 28 '17 at 3:39
  • \$\begingroup\$ It's not only 100 us; some sample capacitors will see low-slew inputs, so a bias voltage applied for many seconds is possible. A 'reset' phase between conversions becomes important (as in a dual-slope or Wilkinson converter). \$\endgroup\$ – Whit3rd Feb 28 '17 at 3:40
  • \$\begingroup\$ True but DA tests use a long soak time and this log scale makes it look worse, but is still a valid source of errors in high res. ADC's which ought to train young engineers to verify accuracy with a wide range of test patterns for repeatability errors with different input patterns. I recall using sawtooths and butterfly step signals to compare ADC & DAC for error differences and discovered many hidden flaws in BB ADC hybrids when they were one of the best in class at the time. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Feb 28 '17 at 3:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.