About capacitors tolerance and readings

[euricogb]

Hi, folks!  First post here!

I'm trying to debug a pedal I'm building as a second project, so I'm really a noob about electronics and building pedals.

I bought a kit with several vintage Philips axial capacitors of 22uF and 4.7uF. Some of them have readings about 35uf (There is one showing 80uF!) of the first ones, and 6uF or 5.6, 5.7uF of the second ones. Way more then 20% tollerance!!

I wonder how could this affect the tone.

Also, some of them were really difficult to measure: I had to charge the capacitor (don't know why)  and then measure with my Mtester. Some shows V loss above 10%, is that an issue?

Does those numbers really mather tonewise?

FYI it's a Colorsound Power Boost.

Thanks a lot!

[Rob Strand]

I bought a kit with several vintage Philips axial capacitors of 22uF and 4.7uF. Some of them have readings about 35uf (There is one showing 80uF!) of the first ones, and 6uF or 5.6, 5.7uF of the second ones. Way more then 20% tollerance!!

Also, some of them were really difficult to measure: I had to charge the capacitor (don't know why)  and then measure with my Mtester. Some shows V loss above 10%, is that an issue?

Some electrolytics are more than 20% tolerance but even if they *are* 20% tolerance the way the cap is measured (and the test frequency) can affect the results.

What I suspect is happening in your case is your DMM uses an AC waveform with no DC bias.    This can result in erroneous readings with electrolytics.     Some DMMs have a DC bias and will measure electrolytics a bit more accurately.   

In other words you probably have a measuring problem more than a capacitor tolerance problem.

I wonder how could this affect the tone.

Does those numbers really mather tonewise?

In the circuit it will have a DC bias and the capacitance will be close to the manufacturers value.

Most positions where electrolytics are used the circuit is designed so the tolerance doesn't really impact the sound.  You really need to look at each position in a circuit to judge that.  A circuit simulator is the easiest way to get an answer.   In positions where it does matter, 20% might be good enough.  If you need better tolerance then you have to use a film cap.
---------------------
Here you go +50%, -10% on these,
https://www.amazon.com/dp/B00F2QKZP2?tag=duckduckgo-d-20&linkCode=osi&th=1&psc=1

[PRR]

Welcome!

> as a second project, so I'm really a noob
> I bought a kit with several vintage ...capacitors

Don't mess with "vintage" capacitors while a noobie. You have enough troubles ahead without getting into probably-bad capacitors.

My neighbor messes with vintage car parts, but he's had 60 years of experience. He can almost smell if a Hudson clutch is just dry or if it is worn beyond repair. You can't, yet.

[willienillie]

It may be more of a concern with high voltage circuits (tube amps), but in general NOS electrolytic caps are best avoided.  They have a limited life span in regular use, and my understanding is that they "dry up" even faster when unpowered for long periods.

[anotherjim]

Caps that old will probably have lost "forming" anyway. Simply, that's a how a polarized electrolytic capacitors get the polarity it has and reaches it's fullest capacity. That forming is lost over time without a polarizing voltage having been applied. In use, the forming is reinforced every time it's switched on. What I do is to put them in a breadboard circuit with a 1k resistor in series and apply your 9v with the correct polarity. Measure the voltage across the resistor - if the cap is formed and good, that voltage will have fallen to zero and the full supply voltage can be read across the capacitor. It may take some time to do this. If the resistor voltage never gets to zero then the cap is probably no good, although a few millivolts is ok so long as it's stable - electrolytic caps do leak a little current.

[mozz]

Caps reading high usually are leaky. Find a new cap and measure it to make sure your meter is working ok.

[euricogb]

Thank you guys!
´

What I suspect is happening in your case is your DMM uses an AC waveform with no DC bias.    This can result in erroneous readings with electrolytics.     Some DMMs have a DC bias and will measure electrolytics a bit more accurately.   

In other words you probably have a measuring problem more than a capacitor tolerance problem.

Maybe, is a cheap chinese MTester like this one:

Caps reading high usually are leaky. Find a new cap and measure it to make sure your meter is working ok.

As a mather of fact, a bought a few new chineses axial capacitors of 4.7uF that measure way better numbers then the vintage ones. That's why I assume those NOS might be bad.

Don't mess with "vintage" capacitors while a noobie. You have enough troubles ahead without getting into probably-bad capacitors.

Thank you for the tip, you are probably right.

Welcome!

Thank you!

A circuit simulator is the easiest way to get an answer.   In positions where it does matter, 20% might be good enough. 

Thank you! Culd you, please, recomend me one easy to use?



The schematics are these ones. Got from the internet and foruns including this one. Would you know, by their positions, which ones might be "tone changers"?
My gess is C5, C12 and C13.

[R.G.]

As electro caps age, their insulating layers get thinner, and the capacitance drifts up. The withstand voltage drifts down while the leakage drifts up. There are three lessons here.
1. Don't buy vintage electro caps. Electro caps can die just because they're old. It's a bad bet, no matter how "cool" or "original" you think your [whatever]will be when you put them into it.
2. Never ever use an electro cap for anything that needs the cap to have a specified value. High-, low-, and band-pass filters are a bad place to put an electro, period. Pick some other kind of cap for filter/rolloff/time constant duties. Film is good for this, or ceramic of the NPO variet.
3. Use electro caps only for positions where you just need at least X value of capacitance and it's OK if it drifts up.

[Rob Strand]

That's why I assume those NOS might be bad.

If they are NOS then that's likely to be the problem - see all comments already made.

FYI,

Maybe, is a cheap chinese MTester like this one:

The bias issue can affect good and bad meters.  It's just one of those things.
High-end LCR meters let you bias the cap but all that stuff complicates things.

Most meters are OK for parts identification provided you aren't trying to measure in-circuit.  Just don't expect the displayed number to always match the label - the parts are a bit off and the meter is a bit off.

Just how accurate they are and when they produce untrustworthy measurement is hard to tell from outside the box.
You have to learn from experience what things upset your meter.   I'd say in-circuit measurements are something you shouldn't expect the meter to do.   Beyond that it's usually technicalities which affect accuracy - probably stuff you shouldn't be overly concerned by.

-----------------
My only hanging issue is why biasing the cap "fixed" the measurement.   That probably wouldn't have occurred if the cap of entirely stuffed.

[MaxPower]

Free circuit simulators: Tina-ti is easy to learn and use. LT Spice is worth learning as it is more accurate/reliable.

However they can give you dodgy results sometimes so it helps to understand what each stage of the circuit being simmed is supposed to do.