Re: 2465B power supply - Sprague 290uF/200V big blue caps


On Wed, 21 Oct 2015, Stefan Trethan [TekScopes] wrote:

This one point below I find interesting amid all the mud slinging.

For any given section of foil the the voltage differential, and consequently the required dielectric thickness, should it not be the same?
Surely one spot on the foil does not know how large an area it is connected to, and would form (or rather thin down to) the thickness of dielectric prescribed by the voltage applied regardless of overall foil size?

Please discuss.
I'm done with this one I think. Until capacitor manufacturers begin mentioning minimum working voltages in their datasheets, some (not you, but you know who) will likely never be satisfied and will seek any means to justify their purchases ("audio grade" capacitors, anyone?). I've already spent far too many hours on this subject as it is and I still have a sizable stack of boards awaiting replacement of their single wipe DIP sockets that I really should have been working on instead. (Anyone want to argue that single wipe DIP sockets are reliable? :)

For those who find this this thread later on, I would personally recommend sticking with low-ESR parts, as the original PSU designers did, and not general purpose parts such as the Nichicon CS, which as I pointed out elsewhere in the thread, have a rather novel way of presenting their estimated service life in their datasheet (super long life...until you apply all the correction factors).

The real challenge I see with this PSU though, is choosing low-ESR components which have a similar ripple current rating to the original parts (for example, a modern 100uF 25V 6.3mm low-ESR part may well have a lower ripple current rating than the original 10mm diameter part). This is made even more difficult by the fact that some of these capacitors are of a critical value (at least one of the 100uF 25V parts /must/ be 100uF for the PSU to function correctly).

As an aside, this model of thinning dielectric leads me to question if capacitance rises in storage - which I am not aware of. It should, in view of the reduced dielectric thickness. Easily tried I guess.
That's an interesting question too. I'm not sure if any increase during storage would be easily measure given the already wide tolerance of these sort of components (just warming one with a hair dryer will tend to show a measurable change in capacitance). A larger value capacitor (larger foil surface area) might work better for such experiments than smaller values.

One that does tend to be easy to measure is the increase in capacitance as the capacitor ages in service. Some of this would be due to the dielectric thinning but a lot of it is also be due to erosion/pitting of the surface of the foils (however I imagine someone will argue against this, too).

...and since we are already on a contentious subject...the above is also why "reforming" of heavily used capacitors tends to be futile at best. Sometimes you can get enough dielectric growth during the reforming process to decrease the leakage current, but if the foil is already eroded from a long service life, it will be temporary as the dielectric thins when the capacitor is used at working voltage and/or stored (unused equipment sitting on the shelf). Better results might be obtained from NOS or barely used capacitors (assuming that the capacitor doesn't have other age-related problems, such as loss of electrolyte or corrosion of the foil to lead connections).

Join to automatically receive all group messages.