Whether or not Bateman's test setup is the best or not, all of the caps tested were tested on that common denominator, and valuable comparisons could be made.
The NP0/C0G caps looked really good in all aspects. They were at the top of the heap in just about everything. They practically appear to be the perfect cap.
However, he did not test microphonics. Whether or not this is a real issue, I have not been able to determine. Because Class 1 caps do not use barium titinate, a piezoelectric material, they are orders of magnitude better than Class 2 ceramics that do, which should not be used for such critical applications such as a feedback compensation cap. The best info I could find about microphonics wrt Class 1 caps is that microphonics are "low". However, while all caps have some degree of microphonics, determining the relative value of Class 1 ceramics vs. caps other than Class 2 ceramics has not been easy to find. If the microphonics of Class 1 is less than, say, silver mica, they could be a perfect low cost, easily available cap for those critical applications such as feedback compensation, etc.
The NP0/C0G caps looked really good in all aspects. They were at the top of the heap in just about everything. They practically appear to be the perfect cap.
However, he did not test microphonics. Whether or not this is a real issue, I have not been able to determine. Because Class 1 caps do not use barium titinate, a piezoelectric material, they are orders of magnitude better than Class 2 ceramics that do, which should not be used for such critical applications such as a feedback compensation cap. The best info I could find about microphonics wrt Class 1 caps is that microphonics are "low". However, while all caps have some degree of microphonics, determining the relative value of Class 1 ceramics vs. caps other than Class 2 ceramics has not been easy to find. If the microphonics of Class 1 is less than, say, silver mica, they could be a perfect low cost, easily available cap for those critical applications such as feedback compensation, etc.
pooge said:
Believe me microphonics are there and are very real. There are in direct proportion to the amounts of Barium T used in the caps. I was involved with a corp in developing a demodulation IC for TV settop boxes. And the analog tuner ahead of our chip was giving us high BitError rates in proto testing. Along story short the tuner's LO PLL had an X7R cap in feed back of loop filter integrator. You could tap with your finger on the bench and count the errors.
Had the tuner vendor change to film cap there you go all fixed.
My belief is harmonic distortion and microphonics are related to piezo mechanics of Barium T parts. Test crystal's with same set up.
np0/c0g are very good but not quite perfect:
I have seen ~ -110 dB 2nd order difference in an amp with 2x series 100V np0 caps in a sallen-key 40 KHz low pass with 20 + 21 KHz 1:1 @ 8V pk sine drive
after replacing the np0 with polystyrene the 1 Khz diff dropped below my~ -160 dB measurement limt ( ~ -130 dB noise floor in a Lynx22 sound card + 30 dB gain @ 1 KHz)
I have seen ~ -110 dB 2nd order difference in an amp with 2x series 100V np0 caps in a sallen-key 40 KHz low pass with 20 + 21 KHz 1:1 @ 8V pk sine drive
after replacing the np0 with polystyrene the 1 Khz diff dropped below my~ -160 dB measurement limt ( ~ -130 dB noise floor in a Lynx22 sound card + 30 dB gain @ 1 KHz)
Hi jcx, Bateman's tests show that polyproplenes beat NP0s, and styrenes beat propylenes. Teflons are currently at the top of the heap for distortion, DA, etc.
Note that Bateman's test setup is not ideal for testing the better capacitors---though it does give some relative indication---as the measuring instrument he designed uses less than the best capacitors in all parts of the circuit.
Note that Bateman's test setup is not ideal for testing the better capacitors---though it does give some relative indication---as the measuring instrument he designed uses less than the best capacitors in all parts of the circuit.
Thanks for your help concerning FFT software.
JCX
---Cyril's circuitry isn't the best for the application - considerable simplification can be had with indirect nonlinearity measurements of IM differences---
Cyril is currently doing experiments with TDFD tests, a kind of IM measurements whose idea comes from Neville Thiele (1975) : almost unknown (at least to me till a few weeks ago), its sensitivity to non-lineraity can beat THD.
JCX
---Cyril's circuitry isn't the best for the application - considerable simplification can be had with indirect nonlinearity measurements of IM differences---
Cyril is currently doing experiments with TDFD tests, a kind of IM measurements whose idea comes from Neville Thiele (1975) : almost unknown (at least to me till a few weeks ago), its sensitivity to non-lineraity can beat THD.
There seems to be some confusion between microphonics and piezo in this thread.
Even an ideal parallel plate capacitor with vacuum dielectric would be microphonic if the plate gap can be modulated. This only happens if there is a dc bias. Piezo voltage is generated by stress so it happens even with no dc bias or gap modulation.
Microphonics by gap modulation will depend on the capacitor construction, pcb thickness and pcb support
Even an ideal parallel plate capacitor with vacuum dielectric would be microphonic if the plate gap can be modulated. This only happens if there is a dc bias. Piezo voltage is generated by stress so it happens even with no dc bias or gap modulation.
Microphonics by gap modulation will depend on the capacitor construction, pcb thickness and pcb support
Yes, that is my understanding too. The slight RF lossiness of the bigger ceramic caps means that parasitic impedances and resonances in PCB traces are not merely retuned (as an NP0 would do) but damped away. As always, use the appropriate cap for the job and ignore myths and fashions about 'good' and 'bad' audio caps.
Andrew
Check this out...
https://www.element14.com/community...-the-letters-stand-for?displayFullThread=true
RCE5C1H103J1M1H03A | Murata 10nF Multilayer Ceramic Capacitor (MLCC) 50 V dc 5% COG dielectric RCE max op. temp. +125C | Murata
This is a low price high value COG. Just bought 10 ( About £3 ), 10.18, 10.09,10.00,10.10,9.91,10.03,10.03,10.07,9.74,10.17 nF. measured to +/-0.5%. A good tight spec considering the 5% quoted. These would serve well for many uses. One day I shall try them as a RIAA preamp. NE5532 or better as the active part.
This is a low price high value COG. Just bought 10 ( About £3 ), 10.18, 10.09,10.00,10.10,9.91,10.03,10.03,10.07,9.74,10.17 nF. measured to +/-0.5%. A good tight spec considering the 5% quoted. These would serve well for many uses. One day I shall try them as a RIAA preamp. NE5532 or better as the active part.
The losses in X7R caps when use for decoupling help in reducing the resonance peaks when different values are used on a supply rail for decoupling this is a requirement for good power delivery system integrity, less lossy caps cause more peaks and lead to an inferior supply. Using very low value COGs next to oscillators as the primary local decoupler with an X7r ( larger vale) near by can be beneficial, this was for frequency hopping radios...
The other advantage of X7R caps is the small sizes achievable again a benefit for high speed digital decoupling (and always remember the speed of a digital signal is determined by its rise time, so even with a relatively slow clock frequency you may still have a high speed digital signal to work with).
The other advantage of X7R caps is the small sizes achievable again a benefit for high speed digital decoupling (and always remember the speed of a digital signal is determined by its rise time, so even with a relatively slow clock frequency you may still have a high speed digital signal to work with).
Many years ago I had on loan a 800 MHz scope. To test a theory I measureed the power rails in a cheap Teac CD player. Nearly every point I measured showed considerable noise which looked not unlike music albeit very high frequency. I wasn't really trying to see that problem, just testing the scope. The point to note is this noise vanished at the decoupling cap. It is tempting to think as Crystal recomended that soldering that decoupling cap by hand as close to the chip as possible would be worth considering. I am really glad this came up in this thread, I would never have guess X7R better for this reason. My brother said XR7 inductance was lower than most which he thought the real issue ( size ? ). Perhaps. He was a TV expert. The main reason I ask questions at DIY is he isn't here now. Always have a flu jab, he should have.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.