SMD Caps: Large value ceramic vs. electrolytic

[Ripthorn]

I'm looking at a couple of designs that use things like 10 uF electrolytic output caps. However, I'm looking to go SMD and see that there are ceramic caps with those capacitances. Is there any reason to not use them for things like output or transistor bypass caps? I like the smaller packages of the ceramic chip caps, so if it's not a big deal, I would go with those.

[PRR]

Large ceramics distort too much for highest fidelity.

In guitar work we can call this "flavor".

[phasetrans]

I'm looking at a couple of designs that use things like 10 uF electrolytic output caps. However, I'm looking to go SMD and see that there are ceramic caps with those capacitances. Is there any reason to not use them for things like output or transistor bypass caps? I like the smaller packages of the ceramic chip caps, so if it's not a big deal, I would go with those.

MLCCs are a different animal than electrolytic Al caps. A few key points:

-Respect the mechanical design guidelines, and heat input guidelines, to keep these from cracking. Especially high capacitance MLCC, which have lots of internal layers

-It is worth paying more for the higher quality X7R (vs. X5R) ferroelectric dielectric, unless you are talking about pure bulk bypass capacitance

-All the ferroelectric dielectrics (commonly X5 and X7) have a de-rating of capacitance due to an applied DC bias. This is due to the behavior of ferroelectric materials, not a manufacturer issue. The de-rating can be substantial.

-Not all manufacturers report the de-rating vs. DC bias. It's not unreasonable to assume that you need 25% more nominal bulk capacitance than the design indicates, due to the DC bias. Off the top of my head, Kemet and TDK have capacitance vs. DC bias on the datasheet. I tend to use TDK for things that aren't deeply cost sensitive.

-If you are using SMD caps, then use the largest package size that makes design and cost sense, as the larger packages generally have less electric field per unit area than smaller packages, and therefore less DC bias effect.

-some people say C0G only dielectric in the audio path. I am also ok with X7R of moderate values, say 1u or less.

[vigilante397]

I use 10uF 0805 caps (X7R), and that is the biggest value I'll use in an 0805, everything bigger goes to a 1206. For really big stuff I still prefer tantalum over aluminum electrolytic, they only thing I use electrolytics for is 470uF.

My stance when buying MLCC caps is get C0G/NP0 when possible, and when not possible get X7R. Never had any complaints.

[Ripthorn]

I'm mostly concerned with values up to about 10 uf for decoupling caps and transistor bypass caps. For anything power I'll still go electrolytic, just wondering about those 1-10 uF values.

[Rob Strand]

I'm mostly concerned with values up to about 10 uf for decoupling caps and transistor bypass caps. For anything power I'll still go electrolytic, just wondering about those 1-10 uF values.

Regarding SMD caps and X7R's SMD caps, you might want to look at the end of this thread.

https://www.diystompboxes.com/smfforum/index.php?topic=124118.0

The point here is the cap is a non-linear function of voltage.  Non-linearity causes distortion, and any DC will change the value of the cap.    The more AC voltage across the cap the more effect the non-linearity has on the signal.

If you think about caps in transistor circuits and single-supply circuits, most of the time the caps will have a DC voltage across them.  An inter-stage coupling cap, or an output cap is going to have 4.5V across it.   Those small package ceramics start is misbehave, the capacitance is non-linear and the value is less than the label.

For anything but a distortion pedal you would want the signal to be undistorted by coupling caps.    For a distortion pedal you would need to evaluate if the effect is noticeable.

For emitter bypass resistors the voltage across the cap is lower and the voltage swing is typically low so you might get away with X7R's in those positions.

For filters, including simple high and low-pass filter, the fact the filter is filtering means the caps are usually subject to voltage swings *across the cap* and that by nature promotes non-linearity.

Really the only time it matters less is when the function of the cap is to do nothing and in that case you want to put a very large value in there.

As much as I hate the crappy long term reliability of electros they do fit the bill in many instances and have little detrimental effect on the signal.   For the smaller end of the scale, say 1uF to 2u2, it's not unreasonable to use films; provided they aren't the wopping old-school green-caps   .

[Andon]

Apologies for bumping a somewhat older thread, but it seemed the most recent and most relevant to a related question I have, especially when it comes to 4u7 capacitors per the prior-posted charts.

I'm migrating a RAT-esque build over to SMD and I'm wondering if I should switch the 4u7 and other electrolytic capacitors (in the audio path) to tantalum packages for the sake of space and sourcing (looking to try SMT services for some test boards), and I wanted to know if that seemed like a kosher move?

Unless I'm mistaken, the general consensus on SMD seems to be that using tantalum in lieu of electrolytic is fine, so long as it's not in the power filtering section of the pedal - is this correct? I've read that polarized electrolytics actually serve a function when it comes to power filtering, but elsewhere it's less of an issue (save for the aforementioned issues of using mlcc in higher values). Anything I'm missing here?

[amptramp]

One bit of advice I can give you is go to this site:

http://everyspec.com/MIL-HDBK/MIL-HDBK-0200-0299/MIL-HDBK-217F_NOTICE-2_14590/

and download MIL-HDBK-217F Notice 2.titled Reliability Prediction of Electronic Equipment.  There are probably a million man-hours of reliability testing and calculation help in this one military handbook.  This lets you calculate the failure rate in failures per million hours for most of the devices in use for modern electronics.  It also lets you check military part numbers so you can figure out what a device is.

There are a number of types of tantalum capacitors.  Dry slug and chip tantalum caps should never be used directly across a power source without series resistance.  They work great for timers and coupling applications where the low leakage afforded by the pinhole-free tantalum pentoxide dielectric is s benefit but if used across power lines, any defect in the electrolyte will cause a short which will locally heat up the dielectric and turn it into conductive tantalum dioxide.  This proceeds through the capacitor like a firestorm and the cap catches fire and explodes.  Series resistance ensures you don't have enough energy to create the necessary heat.

Wet tantalum can be used across power lines but must never see more than 0.3 volts of reverse bias or silver will migrate and cause a short.

The rare tantalum-case tantalum foil capacitors are what I used in power line filters for spacecraft.  Since tantalum is heavy, you can't just throw a bunch of them in - a filter has to be designed, simulated and tested.  But they have no vices.

[iainpunk]

quick question i'd like to add:
i haven't worked with SMD ceramics myself, but a class mate complained about microphonic behavior, especially in moving products (think construction site machines), that pops and crackles were really messing with control systems in his build, which went away when he used THT foil caps.

i'm wondering what the impact of microphonic behavior is in effects pedals, with regards to feedback loops from the speakers to the ceramic caps, and the effects and problems (touching noise? oscillation?) or if that's not something to worry about in pedal designs.

cheers

[vigilante397]

Most of the problems typically associated with ceramic caps I've personally never found to be a problem with multilayer ceramic caps, which is what most SMD ceramics are. I've been using exclusively SMD for my builds for about 3 years now and have yet to have microphonic or oscillation problems caused by MLCC.

[amptramp]

There is a D-Lab video here:

of the refurbishment of a Fender Bassman tube amp which shows how to use a plastic wand to check for microphonic components and that the 8:00 point of the video, he runs into a microphonic ceramic disc cap.  If the owner was using it as a bedroom practice amp, it may not be all that noticeable.  If he is using it at a gig, the volume levels will generate feedback and distortion.  Ceramic caps are not that different from ceramic microphone cartridges.  The higher the dielectric constant, the worse the effect.

Multilayer ceramic is not usually as sensitive.

Beware of assuming that if an amp or effects unit works at low audio level, it will work at high level.

[vigilante397]

Yup, ceramic disc caps are basically condenser microphones. Multilayer are a significant improvement.

[iainpunk]

Yup, ceramic disc caps are basically condenser microphones. Multilayer are a significant improvement.

i have a bunch of really large disc caps, rated at 450v i might give building a condenser mic a chance.
cheers