A few questions about capacitors and Power Supplies

[lightcycle]

I intend to build a couple of Ray Wilson's wall wart PSU (http://www.musicfromouterspace.com/analogsynth/WALLWARTSUPPLY/WALLWARTSUPPLY.php), to power a Thomas Henry Mega Percussive Synth and a modular synth, but I have a few questions regarding capacitors in the power supply I was hoping to get clarified first.

The structure of the PSU design is AC wall wart - Rectifying diodes - Large electrolytic caps - Regulators
Around the 7812 and 7912 regulators there are a couple of protection diodes and 2 1uF tantalums, mounted close to the input and output pins of the regulators. From what I understand these caps are included to increase stability in the power supply. I'm wondering what the effect of replacing those capacitors with 1uF electrolytics, or (smaller value) poly caps would be? Or removing them completely?

The second question concerns the large electrlytic filter caps. Is there a way to estimate what amount of capacitance a given current consumption would require? (I expect the drumsynth would require around 100mA on each rail, the modular I'm not exactly sure yet but probably more)

Finally I have a question about some capacitors I got from a grabbag of electrolytics. I cannot find any markings indicating polarity, and was wondering if any assumptions can be made regarding the polarity of these caps. The text on them reads Richey 4700UF-40V 85C 52-9205 Taiwan. In the bottom there's a pin in the center, and three pins distributed around the edge.

[JKowalski]

I intend to build a couple of Ray Wilson's wall wart PSU (http://www.musicfromouterspace.com/analogsynth/WALLWARTSUPPLY/WALLWARTSUPPLY.php), to power a Thomas Henry Mega Percussive Synth and a modular synth, but I have a few questions regarding capacitors in the power supply I was hoping to get clarified first.

The structure of the PSU design is AC wall wart - Rectifying diodes - Large electrolytic caps - Regulators
Around the 7812 and 7912 regulators there are a couple of protection diodes and 2 1uF tantalums, mounted close to the input and output pins of the regulators. From what I understand these caps are included to increase stability in the power supply. I'm wondering what the effect of replacing those capacitors with 1uF electrolytics, or (smaller value) poly caps would be? Or removing them completely?

The second question concerns the large electrlytic filter caps. Is there a way to estimate what amount of capacitance a given current consumption would require? (I expect the drumsynth would require around 100mA on each rail, the modular I'm not exactly sure yet but probably more)

Finally I have a question about some capacitors I got from a grabbag of electrolytics. I cannot find any markings indicating polarity, and was wondering if any assumptions can be made regarding the polarity of these caps. The text on them reads Richey 4700UF-40V 85C 52-9205 Taiwan. In the bottom there's a pin in the center, and three pins distributed around the edge.

The protection diodes on the regulators and the low value capacitors are not absolutely necessary (it will likely work perfectly fine without them) but it is a good idea to include them anyways.

I would guess that the capacitors you have are positive on the single wire coming out the bottom, and the three prong ring around the can is the negative terminal. Sometimes the can chassis are grounded, so that would make the most sense. I like those capacitors, they look nice and mechanically stable, would be nice for amplifiers (where people usually solidify the capacitor mounting using globs of hot glue  )



As far as calculating the capacitance for the filter, try

(Capacitance) = (Current)/(2(Frequency)*(maximum ripple voltage allowable on output)).

Remember the capacitance is in Farads.

Usually you are fine just ball parking it, especially with a regulator to filter out any ripple left over.


Those caps should be fine for this application. You could get by with lower voltage ratings, since your output is only 12 and -12V... And those are 40V caps... It's always good to be considerably above the rating but 400% is quite a bit. If you are not concerned with space, then you shouldn't have a problem.

[lightcycle]

Thank you for the information JKowalski, much appreciated.
The formula for calculating capacitance led me to further reading on the subject and now I feel I have a much better grip on this subject.
Regarding the smaller capacitors around the regulators I think will include them also. I understand that tantalums are by many considered the best for this application, but after having tantalums mysteriously fail on me in much higher numbers than any other type of component in other builds I'm not very keen on using them. Would there be any drawbacks in using electrolytics or poly caps instead, or should I really use tantalums here?

[GibsonGM]

I always use the poly's, and never have any problems with them.  If you can't find 1uF, you can just use a couple of .68uF in parallel...

[JKowalski]

I always use the poly's, and never have any problems with them.  If you can't find 1uF, you can just use a couple of .68uF in parallel...

Or just one 0.68uF. The exact value doesn't really matter here either on this part. 0.1uF, 0.33uF, 1.0uf, I've seen them all used.

[amptramp]

Thank you for the information JKowalski, much appreciated.
The formula for calculating capacitance led me to further reading on the subject and now I feel I have a much better grip on this subject.
Regarding the smaller capacitors around the regulators I think will include them also. I understand that tantalums are by many considered the best for this application, but after having tantalums mysteriously fail on me in much higher numbers than any other type of component in other builds I'm not very keen on using them. Would there be any drawbacks in using electrolytics or poly caps instead, or should I really use tantalums here?

Dry-slug tantalums must never be used without series resistance.  Tantalum can be anodized, so it is used to make an electrolytic capacitor.  The tantalum pentoxide dielectric has a dielectric constant of 27, so these capacitors are quite small physically and the dielectric is almost pinhole-free, so the leakage current is small.  However, dry-slug tantalum capacitors should never be used as power filters because any local heating will turn the tantalum pentoxide into tantalum dioxide which is conductive.  If any part of the dielectric does turn into the dioxide, the local heating will cause the local area to turn into dioxide as well, so you can get a firestorm happening in the dielectric unless the current is limited so that the heat required to transform the pentoxide to the dioxide cannot be delivered to the capacitor.  The dielectric film cannot be made very thick, so tantalum electrolytic capacitors are usually limited to 75 volt sizes in most types with maybe a few 100 volt types.

[lightcycle]

Thank you all for your answers. I have now built the power supplies and they seem to work well.