Hello hope everyone has a nice week-end !
I don't think one would want to load a cartridge with an electrolytic, but what do you recommend as the type or quality of film capacitor is best at this position? I would think quality matters here since that load is part of the "signal generator" circuitry, along with the load resistor.
Currently I use polystyrene, is that just overkill? Would a high quality polyprop work just as fine or the quality doesn't matter at all ? Although I didn't check I think teflon films are still in production, Solen maybe?
I'm guessing most current phono preamps use polyprops as the polystyrenes are basically extinct and a maker wouldn't use NOS stock on a new product.
Thanks for any insights.
-Joris
I don't think one would want to load a cartridge with an electrolytic, but what do you recommend as the type or quality of film capacitor is best at this position? I would think quality matters here since that load is part of the "signal generator" circuitry, along with the load resistor.
Currently I use polystyrene, is that just overkill? Would a high quality polyprop work just as fine or the quality doesn't matter at all ? Although I didn't check I think teflon films are still in production, Solen maybe?
I'm guessing most current phono preamps use polyprops as the polystyrenes are basically extinct and a maker wouldn't use NOS stock on a new product.
Thanks for any insights.
-Joris
Hopefully NOT Silver plated, Teflon insulated wires.
They are a source of noise when moved.
They are a source of noise when moved.
For cartridge loading / termination, I recommend 1% tolerance ceramic capacitors using the class-1 (high stability, low loss) dielectric named "NP0/C0G" . These capacitors cost about USD 1.00 in quantity one (example)
I trust your advice about ceramics but then in his book "Valve Amplifiers", Morgan Jones states:
These have no place in the path of analogue audio! [...] Commonly, ceramic
capacitors are made up of barium or strontium titanate, both of which are
piezoelectric materials. This means that they generate a voltage when
mechanically stressed (these materials were the basis of ceramic cartridges
used by inferior ‘music centres’ for playing vinyl records).
Ceramic capacitors excel as high frequency bypasses in digital or heater
circuitry where their poor stability of value and low ‘d’ are irrelevant.
Can't figure out what kind of "mechanical stress" is he talking about, not sure a cap sitting soldered on a PCB is much stressed 😎
Perhaps this applies only for usage in high voltages present in valve amps, or the C0G/NP0 formulations circumvent these flaws? I know these are recommended for MCU oscillators.
These have no place in the path of analogue audio! [...] Commonly, ceramic
capacitors are made up of barium or strontium titanate, both of which are
piezoelectric materials. This means that they generate a voltage when
mechanically stressed (these materials were the basis of ceramic cartridges
used by inferior ‘music centres’ for playing vinyl records).
Ceramic capacitors excel as high frequency bypasses in digital or heater
circuitry where their poor stability of value and low ‘d’ are irrelevant.
Can't figure out what kind of "mechanical stress" is he talking about, not sure a cap sitting soldered on a PCB is much stressed 😎
Perhaps this applies only for usage in high voltages present in valve amps, or the C0G/NP0 formulations circumvent these flaws? I know these are recommended for MCU oscillators.
Last edited:
I'm using a combination of Kemet PPS (polyphenylene sulfide film), and silver mica in the most critical places, with excellent results, but even a good polypropylene or polystyrene are preferred over electrolytics for this.
I can achieve near -130dBc distortion levels with COG or PP capacitors on RIAA stages. I've never tried PPS but they should be comparable.
Just a quibble, C0G and NP0 are categories of thermal performance of dielectrics, not a name. The actual ceramic could be many different materials (other than type-II ferro-electric ones), so long as its dielectric constant is constant enough over a particular temperature range.
Agreed, however, in exploring the distortion of several manufacturers C0G/NPO 4700pF 1% 25V or above caps in an RIAA feedback stage I could not detect any significant difference in offerings from Murata, Kemet, Yageo or Kyocera, with Murata perhaps being the best, but with a very small sample size.
Several units with Murata 0805 caps have shown performance through an entire RIAA stage like this:
Ignore the mains harmonics, that's an artifact of the measurement system.
This also includes the distortion of the measurement system.
The overall gain was c. 46 dB in the input FET/current feedback/transimpedance gain cell block + 9dB in the RIAA stage +9db in a post RIAA stage gain block.
The RIAA compliance was also quite good using the 1% Murata caps.
The "straight line" was using Murata COG caps. The not so straight ones were a competing RIAA design.
These results are not cherry picked but are typical of half a dozen or so units that I have tested,
Several units with Murata 0805 caps have shown performance through an entire RIAA stage like this:
Ignore the mains harmonics, that's an artifact of the measurement system.
This also includes the distortion of the measurement system.
The overall gain was c. 46 dB in the input FET/current feedback/transimpedance gain cell block + 9dB in the RIAA stage +9db in a post RIAA stage gain block.
The RIAA compliance was also quite good using the 1% Murata caps.
The "straight line" was using Murata COG caps. The not so straight ones were a competing RIAA design.
These results are not cherry picked but are typical of half a dozen or so units that I have tested,
Thank you for these measurements. It seems I need to revise my pre-conceptions about ceramic caps after modern improvements in technology - I would never have considered using them in a feedback loop but that last plot is impressive.
I'm currently browsing Digikey to order some and I will be able to compare them to the polystyrenes that I used on previous builds with the same design and pcb.
I'm currently browsing Digikey to order some and I will be able to compare them to the polystyrenes that I used on previous builds with the same design and pcb.
There are type I and type II ceramics. Type I have always been limited in capacitance-per-unit-volume, but linear, type II have always been very non-linear (but much higher capacitance as they are ferro-electric materials). C0G are the best of the type I and stable for RF filtering/matching networks (where change in capacitance with temperature would be a problem). See https://www.electronics-notes.com/a.../ceramic-dielectric-types-c0g-x7r-z5u-y5v.php
Thanks for clarification. Could the class-1 "NP0/C0G" be suitable in coupling duty or coupling capacitor bypass as well?
I would think so since as described above @wynpalmer2 used them in RIAA feedback stage, which is akin to coupling.
I sometimes use 470nF C0G caps for coupling stages, sometimes multiple units in parallel. That's often for convenience and pricing benefits, but sometimes there are sound reasons to choose the C0Gs.
It's generally unnecessary as cap non-linearity is likely not a problem for that usage as the impedance should be very low relative to the load in the frequency ranges of interest, which results in a very small voltage across the capacitor, and the capacitance/voltage non-linearity will be inconsequential.
However, in some instances where the DC bias is large the actual cap value for the non-C0G caps can fall substantially from the zero bias case which is undesirable. This happens in locations where the value of the coupling cap can be important, such as in cases where the coupling cap is also used as part of a DC servo loop, for example.
Different manufacturers and cap types (X7R vs X5, for example) are more or less sensitive to this. I tend to Use Murata X7Rs for non-critical coupling (and decoupling/supply filtering) applications.
It's generally unnecessary as cap non-linearity is likely not a problem for that usage as the impedance should be very low relative to the load in the frequency ranges of interest, which results in a very small voltage across the capacitor, and the capacitance/voltage non-linearity will be inconsequential.
However, in some instances where the DC bias is large the actual cap value for the non-C0G caps can fall substantially from the zero bias case which is undesirable. This happens in locations where the value of the coupling cap can be important, such as in cases where the coupling cap is also used as part of a DC servo loop, for example.
Different manufacturers and cap types (X7R vs X5, for example) are more or less sensitive to this. I tend to Use Murata X7Rs for non-critical coupling (and decoupling/supply filtering) applications.
Agreed. "very low" is quite subjective, and as you say, it depends on the cap. In reality, I try to avoid coupling caps when possible, and those I do use are always C0G or PP because of the reasons you state, however I try to standardize to 470nF C0G wherever possible.
For example,
Although I have been known to use 220n and 100n values also, and as I said, 2x 470nF in parallel.
For example,
Although I have been known to use 220n and 100n values also, and as I said, 2x 470nF in parallel.
Yes ,but regular ones are also quite good, I use non-magnetic ones in combination with non-magnetic resistors.