Electrolytic capacitors with -10 %, +75 % tolerance values

[ron_swanson]

Hello,

I'm looking to replace some electrolytic filter can caps in an old ( mid 1960's ) vintage guitar amp. Of course, the spec'd values are no longer available ( that I can find - 1250uF/50V ) in a single or multi-sectional 1" diameter can configuration so I'm left to abandoning ( disconnecting ) and leaving the old can in place. Then replacing the cap with a modern axial or radial cap of 1300uF/50V placed under the chassis.

One of the choices I have found has a tolerance of - 10 %, + 75 %. While I undertand what this means numerically, I can't seem to find any information to explain why I would or would not want to use something like this instead of something that is simply +-10 or 20%.

I'm curious about the best use for something marked -10 %, +75 %. +75% seems like a huge discrepancy on the high value end from the desired value and I'm guessing would likely require the circuit be designed with something like this in mind from the get go.

Would anyone here be able to explain the use here or point me to info in the Interweb?

Thanks in advance!

 

[JohnRoberts]

it depends on how/where used...

I the C forms a pole in the audio bandpass, the +/- 10 or 20% part is indicated. For reservoir and general purpose applications +75% is fine.

JR

 

[user 37518]

The smaller the tolerance, the better, but also more expensive. Back in the day, caps weren't as good as they are today. Like John said, in something like a PSU, a ripple filter cap, the more the capacitance the merrier, -10% tolerance is tight for an electrolytic and for a PSU filter cap, you really don't want the cap to go lower than what you designed it for, but having a cap that is +75% tolerance wont do any harm, in fact it can be even an advantage, you get more capacitance than what you paid for, that is unless the inrush current becomes so high that it can melt something upstream. For accurate stuff, the tighter the tolerance the better, so electrolytics are a no no when it comes to filters, eq's, etc...

Usually, electrolytics are +/- 20%, my best guess is that the manufacturer is doing something to provide a low negative tolerance in exchange for having a huge positive tolerance, as I said, with a PSU filter cap you really don't want it to go too low, but going high is not bad.

Also, remember that the tolerance rating is usually related to a normal (Gaussian) distribution, and the tolerance level quoted by the manufacturer is for 6 standard deviations (+/- 3 sigma), take for instance a 1000uf +/- 10% cap, it means that the manufacturer is telling you that there is a 99.7% chance that the cap labeled '1000uf" that you buy will fall in the 900 uf -1100uf range, there is a 95% chance that it will fall in the 934 - 1066 uF range, and there is a 68% chance that it will fall in the 967 - 1033 uF range. In my experience, the 68% range is what you will usually get with good quality caps, but that is just probability, you can definitely get a cap which is close to the the 1000uf value. In this case, it seems that the mean is negatively biased, that is, the mean is not centered at the middle but rather towards the high side, or maybe its not a normal distribution, who knows... If you could buy a few hundred of them and measure them, we could make a statistic analysis.

“There are three kinds of lies: Lies, Damned Lies, and Statistics” — attributed to Mark Twain

This is particularly important when manufacturing many units, designers design equipment with some tolerance in mind, thinking that as long as it stays within the spec'd tolerance, the equipment will still function properly, since tighter tolerance components mean more cost, they want to make sure they can manufacture something by spending the least possible and keeping their bosses happy. Usually, they will use (abuse) of something called Montecarlo Analysis to determine this.

 

[Newmarket]

In the majority of cases more capacitance in that range will not be a problem. A stated above, too much capacitance may give issues due to inrush current putting strain on the rest of the power supply.
But if that isn't an issue then any increase in capacitance is good esp given that the capacitance decreases with time and usage.
For "Belt and Braces" I'd suggest to measure the actual capacitance and have a think if it is actually more than +20%. Somewhat arbitrary but typical commercial electrolytic is +/-20% tolerance.

 

[ron_swanson]

Thanks all!