The snag is the phrase "most linear". Please describe what you mean in that context. High DA and high linearity are kind of mutually exclusive. Many will take exception to saying caps with high DA (a non-linear phenomenon) are more linear than caps with lower DA, other things being equal. No argument if you say "best sounding" and leave it at that. "Linearity" can be objectively measured/proven; "sounds best" cannot - it's subjective.
Polystyrene have low DA and DF. They are very good at high frequencies but are only available in very small values at high voltages and get very large for even small values so placing them can be an issue. Sometimes I use them to bypass larger polypropylene caps. It gives the best of both worlds to my ears. Others think it’s crazy so try it and see how you like it. Most polystyrene are film and foil at high voltages. They don’t like extreme heat either. They are a cheap thing to try and can be undone easily. Capacitors are a very personal thing and can cause many an argument but I recommend trying things and making your own choices, after all, it is your equipment!
What's the resolution of that measurement? 1%? If you want to talk distortion let's at least use a modern measurement such as an FFT.
PIO has a pretty significant voltage coefficient. That'll generate distortion. I seem to recall that the temperature coefficient is mediocre at best too. You can read the first couple of chapters in The Art of Electronics X Chapters for more.
Tom
Last edited:
As much as I enjoy reading Stephie's (sbench) prose, I must agree reading X-Y display off an old scope is not exactly conclusive.
Another phenomenon not mentioned so far (I think) is the aging of all paper based capacitors. I ALWAYS replace the bumblebee capacitors whenever I see them.
Another phenomenon not mentioned so far (I think) is the aging of all paper based capacitors. I ALWAYS replace the bumblebee capacitors whenever I see them.
These are the only schematic I could find. I’m not sure if it is even relevant as i can see it showing 8 ohm and 4 ohm connections which was only on the mk1 of this amp. Mine is the mk2 unison research s6 which has 6 ohm connections.
Sorry for the dumb questions but these look so far removed from what’s in there that I don’t really know how to pick the right ones for the job. So if I get a code 940 series cap that is 0.047uf and 250v that I can make fit around all the other stuff on the circuit board it will be fine? Is the 250v critical or can I go the next voltage up if I can’t find the exact match? Do they need to be radial leaded or can I get axial leaded if space allows? Also do they need to be polarised?That`s your opinion, i think they muffled and lacking treble. Polypropylens are way to go. CDE 940 series are good caps.
Again sorry for the questions but I need some of your knowledge.
Cheers
Yes you can go next voltage up. Those are small caps what you are going to replace. CDE 940 are Bigger. Wima mkp 10 polypropylens may fit better. Good caps too. 0.047uF CDE 940 are 34mm long.https://eu.mouser.com/ProductDetail/Cornell-Dubilier-CDE/940C20S47K-F?qs=oz45iF9aKufHwZtwXmdzGQ==
Last edited:
Thanks zenzen .The voltage on the cde 940 caps you linked is 2kv, is this a problem seeing the ones I’m replacing say 250v? Sorry again i would prefer to get these if they are superior and the voltage difference is a non-issue but am a little unsure if this is dangerous as the voltage difference seems a lot? I can’t find film and foil pp caps that are 0.047uf and 250v. I can find metallized but would prefer to get the most of it with a film and foil if possible.
Thanks
Thanks
I think Tom is referring you to Horowitz and Hill, not to Bench.
kind regards
Marek
MetalMatt@ Here is 10mm caps what i found.
https://eu.mouser.com/c/passive-com...m-capacitors|~Lead Spacing|~Voltage Rating DC
Forget 940 CDE they wont fit. I just read hole thread, sorry.
https://eu.mouser.com/c/passive-com...m-capacitors|~Lead Spacing|~Voltage Rating DC
Forget 940 CDE they wont fit. I just read hole thread, sorry.
Hi,
Think the best capacitor is no capacitor at all, since it's actually an electrical current storage device.
In the case of a 0.047uF capacitor at 1ma per uF it's not a lot of current storage, fortunately.
The capacitor serves two purposes. One to block DC current from the AC signal path and two to act as a passive filter.
To improve the sound of a coupling capacitor simply halve the original value and double the value of the associated resistor to ground, for the same EQ.
Alternatively look at the filter it makes, at 50 Hz a 0.047 uF capacitor looks like a 68k resistor, now think volume control.
If you multiply a capacitors value at a given frequency by ten and now think that the capacitor acts as the top half of the volume control and the associated resistor to ground the lower half of the volume control. If the top half is our 0.047 uF cap of 68k at 50 Hz then the resistor has to be 68 x 9 or 612K to make a stepped attenuator of -1dB at 50Hz and -3dB at 13 Hz.
The same formula for a 0.022 uF capacitor requires an associated 1M24 resistor to ground. Thus 0.033 uF would require 1M to ground.
Does halving the capacitor value work ? Yes it does of the gosh that's better, sonic upgrade variety.
As to the type of capacitor, Polypropylene takes a while to bed in where as Polyester is what it is from day one.
Think the best capacitor is no capacitor at all, since it's actually an electrical current storage device.
In the case of a 0.047uF capacitor at 1ma per uF it's not a lot of current storage, fortunately.
The capacitor serves two purposes. One to block DC current from the AC signal path and two to act as a passive filter.
To improve the sound of a coupling capacitor simply halve the original value and double the value of the associated resistor to ground, for the same EQ.
Alternatively look at the filter it makes, at 50 Hz a 0.047 uF capacitor looks like a 68k resistor, now think volume control.
If you multiply a capacitors value at a given frequency by ten and now think that the capacitor acts as the top half of the volume control and the associated resistor to ground the lower half of the volume control. If the top half is our 0.047 uF cap of 68k at 50 Hz then the resistor has to be 68 x 9 or 612K to make a stepped attenuator of -1dB at 50Hz and -3dB at 13 Hz.
The same formula for a 0.022 uF capacitor requires an associated 1M24 resistor to ground. Thus 0.033 uF would require 1M to ground.
Does halving the capacitor value work ? Yes it does of the gosh that's better, sonic upgrade variety.
As to the type of capacitor, Polypropylene takes a while to bed in where as Polyester is what it is from day one.
Groner and Wurcer wrote an article for Linear Audio ("Quadrature Bridge Measures Harmonic Distortion in Capacitors", LA #12, pp 165-177) in which they measure distortion in polypropylene, polyester, polystyrenic and C0G capacitors. C0G came out on top.
I have no PIO caps, but would be happy to measure them with Sam's method. The AP2722 analog signal generator can generate a quadrature signal.
Bruce Hofer of Audio Precision reached the same conclusion. You can find his data here:
Tom
Thanks Zenzen. I’m not going to lie, I’m a little overwhelmed by choice and figuring out what makes one better than the other. How do I know which one out of those is the better cap to pick? Is there something to look for? If they are all pp film caps @ 0.047uf with a 10mm lead spacing what other spec can I use to seperate them? Lower tolerance or even try to find out which ones have copper leads instead of steel? Or is there something else?
It probably seems easy for you but I feel a little lost in all the choices.
But I am enjoying trying to learn.
Cheers