Subbing electrolytic capacitors with mlcc’s

[Locrian99]

I was curious if anyone has any thoughts on using non polarized mlcc's to replace polarized electrolytics.   Are there times this wouldn't be ok, or less ideal?  I wouldn't be doing any larger than 10uf.   And the tolerance of what I have on hand is the same.   Main reason for doing this is I have a lot of 1u-10u mlcc's I am never using, and we'll I just like the look of them on the board better. 

[iainpunk]

if i could, i would always pick a non-polarised cap over a polarised one, for pedals, its basically always acceptable or better than using polarised. voltage ratings and the like often prevent this in other projects of mine tho.

cheers

[Rob Strand]

Are there times this wouldn't be ok, or less ideal? 

That's about it.   Identifying the times it is OK requires a lot of explanation.

Not all MLCCs are alike, they have different dielectrics even for the same value.   For commonly
available parts the MLCCs in the larger values, say 1uF and over, tend to be the least desirable
for passing audio.   They are OK for bypassing supplies.    You can use them for audio provided
they are use where cut-off frequency for any filter it belongs to is much lower than audio frequencies.
(There's many posts of this forum about the finer points.)

[Locrian99]

Are there times this wouldn't be ok, or less ideal? 

That's about it.   Identifying the times it is OK requires a lot of explanation.

Not all MLCCs are alike, they have different dielectrics even for the same value.   For commonly
available parts the MLCCs in the larger values, say 1uF and over, tend to be the least desirable
for passing audio.   They are OK for bypassing supplies.    You can use them for audio provided
they are use where cut-off frequency for any filter it belongs to is much lower than audio frequencies.
(There's many posts of this forum about the finer points.)

I have looked through quite a bit on this.   I've seen comments basically saying 1nf and above use film only, it doesn't matter for guitar, etc.   im sure with like most everything there's plenty of opinions and it likely just comes down to trying and see if I can hear a difference.   That being said you mentioned passing audio.   I would assume something like this. 
https://images.app.goo.gl/QLqp2r7FH6H1dJJq9

Replacing c2 would be fine as it doesn't have any audio signal passing through it, simply is part of the lfo operating the led.   

[Rob Strand]

Replacing c2 would be fine as it doesn't have any audio signal passing through it, simply is part of the lfo operating the led.

That's the type of thing.   However I'll add a caveat in that if you have a crappy dielectric and cap has a small package there could be some effect.   What's happening here is the voltage across the cap changes a lot in the circuit and some of those MLCC caps with larger values like 10uF "lose" capacitance when the voltage across the cap increases.

Go to the section "Voltage dependence of capacitance",
https://en.wikipedia.org/wiki/Ceramic_capacitor

You can see a crazy reduction in capacitance as the applied voltage goes up (to the point where the rated capacitance and voltage are almost useless.)  You might find other such graphs on the web.  The graphs are usually presented as the applied voltage relative to the rated voltage.   However, what really makes the drop off bad is when the physical size of the cap is small and the value is high - it doesn't matter if the rated voltage is high or low.

[Locrian99]

Graphs always help.  Lol.  Thanks

[pacealot]

That probably explains why I've had different results in the same LFO circuit with nominally the "same" value caps of different types (that also read very close or the same on the meter), where with one cap the LFO will oscillate throughout the full rate/speed range but with the other it will stall at one extreme or the other (or both). I hadn't really thought through the why of it before — so thanks yet again, Rob!