On Sun, 2 Sep 2012, Josh Dersch wrote:
Hi all --
Working on recapping an old (failed) 5V supply in that Microkit machine I
picked up last week; the main filter capacitor is rated at 16,000uF 15V,
axial. So far in my searches an exact (or even close) match in an Axial
form-factor seems to be unobtainium. I can find snap-in or screw-in types
but those are going to be clumsy to retro-fit given the space it has to
occupy. (plus they're pretty expensive.)
How far off the original capacitance value can I stray before it compromises
the original design?
What are the dimensions of the original capacitor (in millimeters)?
If it is being used at 5V then you could go down slightly on the voltage
rating, but I'm not sure I'd want to go up much higher because an even
larger plate surface will result in an even thinner oxide layer being
formed, which isn't optimal for electrolytics. The general rule of thumb
in an application such as this is to use a capacitor that is rated for 2x
the operating voltage.
As far as the capacitance goes, you can probably go a little higher
without too much worry (20% of 16,000uF is 3,200uF, so the original could
have been 19,200uF and still would have been within specs), but I'd keep
it under about 18,000-22,000 unless you know for certain that the existing
regulator(s) and/or reverse biased protection diode can handle it (you
don't want to burn out your regulator when you switch the supply off and
the capacitor still has a charge).
Axial style capacitors are available in that range, but you may not like
the price. They appear to be going for USD ~$10.00-15.00 each in qty of
about 50/pcs.
There are a number of capacitors in Nichicon's VX series that you could
consider. The VX series are general purpose 85c rated parts.
TVX1C153MDD 16V 15000uF 22x52mm
TVX1A153MDD 10V 15000uF 22x40mm
TVX1C223MDD 16V 22000uF 25.4x61mm
TVX1A223MDD 10V 22000uF 22x52mm
Another option worth considering is using a radial mount part. Just attach
it to the board with some silicone (which isn't a bad idea for large
capacitors anyway) and extend and insulate one of the leads with some
small teflon tubing. The Nichicon VY and VZ are 105c rated, and the VK and
VR are 85c rated.
UVY1C153MHD 16V 15000uF 18x40mm
UVY1A153MHD 10V 15000uF 16x35.5mm
UVY1C223MRD 16V 22000uF 22x40mm
UVY1A223MHD 10V 22000uF 18x40mm
UVZ1C153MRD 16V 15000uF 20x40mm
UVZ1A153MHD 10V 15000uF 18x35.5mm
UVZ1C223MRD 16V 22000uF 22x50mm
UVZ1A223MRD 10V 22000uF 20x40mm
UVK1C153MHD 16V 15000uF 18x40mm
UVK1A153MHD 10V 15000uF 16x35.5mm
UVK1C223MRD 16V 22000uF 22x40mm
UVK1A223MHD 10V 22000uF 18x40mm
UVR1C153MRD 16V 15000uF 20x40mm
UVR1C153MRD6 16V 15000uF 25x30mm
UVR1A153MHD 10V 15000uF 18x35.5mm
UVR1C223MRD 16V 22000uF 22x50mm
UVR1C223MRD6 16V 22000uF 25x40mm
UVR1A223MRD 10V 22000uF 20x40mm
UVR1A223MRD6 10V 22000uF 25x30mm
I might also consider using a Nichicon HE, PW, or PS series assuming I
could find a close value (these generally top out at 10,000uF or
15,000uF). These are rated for high ripple and 105c. I use tons of these
in all sorts of values in switching power supplies of all sizes and they
would probably work just fine in a linear supply too.
UPW1A153MHD 10V 15000uF 18x40mm
UPS1A153MHD 10V 15000uF 18x40mm
http://nichicon-us.com/english/products/pdfs/e-vx.pdf
http://nichicon-us.com/english/products/pdfs/e-vy.pdf
http://nichicon-us.com/english/products/pdfs/e-vz.pdf
http://nichicon-us.com/english/products/pdfs/e-vk.pdf
http://nichicon-us.com/english/products/pdfs/e-vr.pdf
http://nichicon-us.com/english/products/pdfs/e-he.pdf
http://nichicon-us.com/english/products/pdfs/e-pw.pdf
http://nichicon-us.com/english/products/pdfs/e-ps.pdf