stupid power filter cap question...

[LucifersTrip]

When there's a power filter cap (across the rails) in a pedal, I see polarized, electrolytics, usually in the 22uF - 100uF range.

Are they electrolytics because of physical size limitations or because they have to be?

thanx

[Kesh]

i guess because they are cheap and there is little gain in using anything more expensive.

[amptramp]

Electrolytics are the easy way to get large amounts of capacitance in a small package.  If you have a film capacitor, you have to either metallize the film or use separate metal sheets for electrodes and wind them around in a circle to provide a capacitor.  With an electrolytic, there is no ultra-thin piece of plastic film - the aluminum electrode is anodized, forming an aluminum oxide film on the aluminum.  This allows a much thinner film to be created, providing more capacitance.  A thicker film of aluminum oxide gives a higher voltage rating but lower capacitance.  It is normal in a given family of capacitors for the volume of the capacitor to be proportional to the product of voltage rating and capacitance.  The electrolyte forms the other electrode and is made of a liquid, which is an easy way to get the electrode to contact the film at all points.

Film capacitors of 10 microfarads are large devices, sometimes being an inch in diameter and several inches long for a 100 volt device.  The corresponding electrolytic is maybe a quarter inch diameter and a half inch long, if that.

Generally, if you can use something other than an electrolytic, you should, because the liquid core has rather high resistance and the impedance of the electrolytic begins to rise in the audio frequency band.  Also, electrolytics are normally polarized and film capacitors are not.  Reverse voltage on an electrolytic will remove the anodize which is the dielectric and punch through at any defect sites.

[LucifersTrip]

Thanx much for the details...though, I'm well aware of the smaller size, smaller costs and basic flaws of electrolytics.

...and I always do sub non-electro's  for electro's any time I can in the signal path. But my stupid question was whether I could do the same when dealing with a power filter cap across the (+) and (-) rails. I don't know why, but I thought it might be different...

So, from your answer ("Generally, if you can use something other than an electrolytic, you should"), the answer to my question ([1] "Are they electrolytics because of physical size limitations or [2] because they have to be?") would be [1] "yes" (and also because of other flaws & cost) and [2] "no".

thanx again

[R.G.]

So, from your answer ("Generally, if you can use something other than an electrolytic, you should"), the answer to my question ("Are they electrolytics because of physical size limitations or because they have to be?") would be "yes" (and also because of other flaws & cost) and "no".

In general, the capacitance values and voltages that are useful for power filtering are too large for non-electros to be competitive. I was very happy with the recent availability of 1uF and 2.2uF film caps in packages that will fit on PCBs. You can get a 1uF/50V film cap that's about 5mm (0.2") by 7.5mm (~0.3") by 9.5mm(~0.375") for $0.48 in ones, about $0.25 in hundreds. A 47uF cap made with the same innards would scale up from the 1uF with 47 times the volume and probably price.

So a 47uF/50V film cap made with that technology would cost about $22.56 and would be a cube about 1" on a side or a similar volume. It's possible that it might cost much more, because the manufacturing yield would probably be smaller and the sales volume would be much smaller.

The problem is that aluminum oxide insulation can be made vanishingly thin by electrochemical means in an electro cap. Plastic film is difficult (read that as "practically impossible") to make only 25-50V "thick" because it gets so thin it's hard to handle it without puncturing it. It's amazing to me that we get 1uF caps in such a small package in film.

However, you can get plastic film in 200-500V thicknesses that are easy enough to work with. These are the standard for HVAC motor run caps. In this voltage, plastic film in only 2-4 times the volume of an electrolytic cap. These make a good competitor for electro caps in the power supplies of tube amps.

[LucifersTrip]

thanx again for more details...


the main reason I asked that question was because I wanted to use this in a project as a power filter cap across the rails and there was no indication of polarity

[R.G.]

At 40uF/15V, those look like axial non-polar electros.

I am suspicious because of the presence of the orange and green dots, though.  Are those on all of them?

[LucifersTrip]

At 40uF/15V, those look like axial non-polar electros.

I am suspicious because of the presence of the orange and green dots, though.  Are those on all of them?

Yes, the dots are on all.  I believe the caps were for military use and were used at the Philco plant in the 1960's.

The larger 50V ones are actually polarized and marked clearly. On those, the orange and green dots are on the same side. I always guessed they were some type of quality or sorting marking like on transistors.


btw, specifically, what type of caps are these. Is there a technical name for the plastic outer cover/sleeve?

[R.G.]

The label says "GE" I think. The form factor was common on quite-old electro caps. I've seen a lot of these in surplus electronics shops, very few new ones. The pencil-thin electro form factor is one that I associate with decades-old stuff, although I don't really know if that's current.

From just the pictures, I'd say that the ones with no polarity markings are bipolar, as I guessed.

I get suspicious about positive any time I see a red/orange/yellow anything near a terminal.

As I said, they may be quite old, and perhaps leaky. Maybe not, but it's worth being aware of and watching for.

[LucifersTrip]

The label says "GE" I think. The form factor was common on quite-old electro caps. I've seen a lot of these in surplus electronics shops, very few new ones. The pencil-thin electro form factor is one that I associate with decades-old stuff, although I don't really know if that's current.

yes,  these are General Electric capacitors that Philco had at their plant in the 60's (pre - 1968)

As I said, they may be quite old, and perhaps leaky. Maybe not, but it's worth being aware of and watching for.

they are extremely accurate...most measuring between 39.5uF and 40.5uF with a couple at 41uF

thanx again

[LucifersTrip]

I am suspicious because of the presence of the orange and green dots, though. 

ps: resistors from same Philco plant, pre-1968