input caps - time to ask the experts

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've looked high and low for info on input coupling caps (google, diyaudio, etc.). I'm doing Elliott's (ESP) 3A project (the cap in question is C1, just to make clear), and the best info I could find was right there on the linked construction guide.

Well, this is what I gathered: It should be "large" between 1uF and 10uF, polypropylene or polysterene, although polyester is okay. So I'm thinking, input doesn't need to be more than a few volts, right? I mean on the official page, it said the bootstrap cap (C7) has to be 35V, but the rest of the polarized ones can be whatever is on hand. That is a bit non-specific, no? I guess if I knew what I was doing I'd know what is sensable. So I browse around the construction guide, and contrary to his own recommendations, Mr. Thornblade used a 2uF 400V oil/paper cap. So I have been looking through cap catalogs and such, and have found some 400VDC (275VAC) metalized polypropylene caps for ~$13 each.

But, I have a hunch it doesn't need to be more than a few volts, so if I get a reasonable cap, say 50VDC, I can get a better poly-film/foil cap for much less $$$.

So, what I'm asking is: what is the optimum voltage for an input coupling cap? And I'm guessing a 10uF input cap would be best.

Hmm, now that I'm thinking about it, it doesn't seem to give the recomended voltages for the nonpolarized caps either. I'd guess they should be >=35V (rail voltage), but I might be wrong.
The cap size determines the low-frequency cutoff, so in that sense larger is better.

The only reasons I used the 400V oil/paper were

a) idiosyncrasy
b) DC blocker (150V) for my tube input stage

You can't go wrong with a PP film cap of 50-100V, although you are correct about the voltage rating not really mattering. Higher voltage caps just tend to be a little more linear (sometimes ^_^) Since film caps often start at 50V ratings, you can just use that rating or better for all your caps and have no worries.
poly cap vs eletrolytic

In an eletrolytic cap there is (because of the way it's made) some inductance. This can (at the right/wrong freq depending how you see it) cause large not wanted effects in a curcuit.

An other effect is that eletrolytic caps are slower, and have a DC leaking current running through them.

A poly cap is closer to the "perfect" cap, if only they came bigger !


If you are talking about the bypass caps, they should really be installed as close as possible to the board. mounting them any distance away increases loop size (hum!) and series inductance (filtering efficiency, stability). These are your high-frequency bypass, so keep them close by.

Sorry for overlooking the miller cap voltage ratings. I'll add cap voltage notations to my site this weekend.
Any idea why a 100V 10uF metalized PP cap is more expensive than a 200V or even 400V or the same capacitance? And is material (PP vs polyester) more important than type (metalized vs. film/foil)? 10uF film/foil PP caps are pretty hard to find. Also, is 10% cap tolerance for an input cap okay? all the 5% ones are too small. I guess I could look for oil/paper since they are 1%.
Altaic, I would guess you are comparing apples and oranges. There are many different construction methods and intended applications for various caps, regardless of the dielectric material. This is why selection can be so interesting--there is really no best cap for all applications.

Material is more important than metallized/foil, and tolerance does not matter at all in the input cap.

Hint: someone has been selling 12uF wound foil Aerovox PP caps on Ebay. Alf might sell 10% tolerance direct if you inquire.
Sweet! One last question, this time about the choice of a 5W resistor. Is it worth it to spend a few bucks on a thick film resistor rather than an inductive wirewound? In my experience (which is not much) inductance leads to oscillations.

Thanks so much for all the help. There is such an overwhelming selection of components.

The ESR parameter has do with the performance of the cap when cgarging/discharging it quickly. It has nothing to do with the inductance.

BTW the inductance is often not noticable under 50 khz when one sweep the impeedence vs freq. so unless you are building a monster amp with wery high specs, i would not worry about the input cap at first.

Another matter is the cap in the feedback system, I would try and find a low leakage NP electrolytic cap instead of the specified one in tha project 3 A

Capacitor voltage

Most circuits , tubed or transistor , have their inputs at almost zero volts . Most sources also have their outputs at close to zero volts dc. This would mean that you could use a low voltage capacitor and 50 volts seems to be Ok especially if you use large ones like 10 uF at the input. Personally if I use 1uF or lower I use 250 volt or 400 volt types . Larger than 1uF I use 63 or 100 volt types. The only exception is if the capacitor is connected to points with high dc voltages. In that case I use at least a 50 percent margin or more for the voltage rating.
For those who want to know :
With a 10uF capacitor and 10 Kilo ohm input impedance you will have a -3db point at 1.6 Hz. You can find the -3db point for other values by simple division of these figures.
Eg. For a 1uF capacitor and 22 K ohms input impedance

= 1.6Hz X (10uFx10Kohm)/ ( 1uFx22K ohms)
= 7.2 Hz.
Note : the text in the equation is only explanatory .
The only low-leakage electrolytic caps I could find all had 20% tolerance. Is that okay? It seems a bit high, seeing as you can get high precision e-caps with respectably tight tolerances. I guess a 100uF film cap would be out of the question...

I've got another question.. since I beefed up my input cap to 12uF (as opposed to 4.7uF), does that mean I have to lower R2 (goes from input, after the input cap, to ground)? And adjust C2, which goes in parallel with R2?
Input capacitor

I am assuming C2 is a low value cap. Keep the product of C2 and R2 same.
That is C2xR2=CnewxRnew.
You can determine the value of the input capacitor as explained in my previous post. You do not have to go any lower than -3db at 10 Hz or so. Depending on the powersupply and other factors a low roll off could cause problems if your source has plenty of very low frequency content.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.