Stupid question

Shhhh... replacing such parts is what's keeping me in business.

Height. Flat panel LCD monitors tend to be rather thin. Everyone wants low profile thin products these daze. 3300uF caps tend to be taller than 1000uF caps. I don't know exactly why, but it seems to do with the way the foil and insulators are wrapped inside the can. I haven't seen any short and fat PCB mount caps. While it's possible to use axial leaded caps instead of radial, the robotic insertion hardware greatly prefers radial leads.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com

You get better ESR results with smaller cap values.

Jamie

More to the point, the difference in ESR is not directly proportional to the different sizes of capacitor, so 3x 1000uF has ESR of not much more than 1/3 that of a single 3000uF.

Easy enough to test, or so I thought. First, I spend 10 minutes rescuing my Dick Smith ESR meter from the Gordian Knot of tangled cables behind my test equipment wall. I then find an old PCB (Okidata

320 printer board) with 3300uf 35v and 1000uf 35v Rubyeon electrolytics. Carefully measuring the ESR, I get; 1000 uF 35v = 0.05 ohms 3300 uf 35v = 0.04 ohms So much for the 3x theory. I'll see if I can find some 6.3v or 10v caps, which should (hopefully) give higher ESR readings that are not at the extreme of the meter range.

Well, maybe not. The 3300 uf is only about 1mm taller than the 1000 uF cap, and a little less than twice as wide.

I would have thought the 3300 uF would be somewhat larger.

I just hate it when my measurements don't agree with my guesswork.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com

the

1/3

I wonder if your guess work and measurements may be correct Jeff. Your test subjects were are old capacitors in a scrap board. Their conditions may be questionable. Wouldn't you have to establish a "base line" with new parts? And then even given that couldn't the variations between new parts still come into question? Lenny

"klem kedidelhopper"

** This six toed, retarded freakoid has no other kind of question.

True. I should use new parts. However, having two caps, with the correct value, from the same manufacturer, was easier than trying to find a new cap in my collection. I have plenty of 1000uF caps, but no

3300uF. Also, I forgot to test what 3ea 1000uF caps in parallel will read.

However, that's not the problem. When the ESR values get below about

0.10 ohms, the readings are not very stable, heavily influenced by the test leads, and what I consider not very reliable. They're good enough for detecting a defective capacitor, but not really useful for obtaining a definitive reading.

I'll try to drop into the local electronics stores in the next few daze with my tester.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com

The proof of the pie is in the eating.

Try the 2 alternative arrangements in a practical SMPSU and scope the ripple.

True. I should use new parts. However, having two caps, with the correct value, from the same manufacturer, was easier than trying to find a new cap in my collection. I have plenty of 1000uF caps, but no

3300uF. Also, I forgot to test what 3ea 1000uF caps in parallel will read.

However, that's not the problem. When the ESR values get below about

0.10 ohms, the readings are not very stable, heavily influenced by the test leads, and what I consider not very reliable. They're good enough for detecting a defective capacitor, but not really useful for obtaining a definitive reading.

I'll try to drop into the local electronics stores in the next few daze with my tester.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com

"David"

** The penny finally drops.....

Now, imagine that the ripple current is divided three ways in the case of using 3 caps and the reason for doing it is obvious.

.... Phil

I hadn't thought of that Phil but it makes sense. Lenny

Thanks. That's the answer. 3 caps are better than one for ESR.

See column for "Impedance 100KHz".

I just noticed the "Endurance" column, which I guess means lifetime. Values run from 2000 to 7000 hrs or 83 to 291 days. That would suggest that if I leave my machines on 24 hrs/day at 105C, the caps would blow up in 3-12 months. Good to know.

Back to measurements. I couldn't find a 3300uF cap, but did manage to find some brand new 1000uF 10V and 2200uF 10V NHG caps. uF ohms

1x 1000 0.16 2x 1000 0.09 3x 1000 0.04

1x 2200 0.10

2x 2200 0.05

Difficult to be certain because +/- one LSB will ruin any definitive conclusion, but it seems like 2 caps are better than one big one. Two

1000uF caps in parallel yielded 0.09 ohms, while a single 2200uF cap yielded 0.10. Still, it's difficult to proclaim a winner with only a 0.01 ohm difference.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com

"Jeff Liebermann"

** No - it means that with the caps operating at 105C ( due to both ambient temp and ripple current ) - the rated life to when the ESR goes out of spec is 83 to 291 days.

Derating tables show that the expected life extends to many years at more typical equipment temps.

.... Phil

I hadn't thought of that Phil but it makes sense. Lenny

##########################

There's also the increased surface area of 3 smaller caps that can dissipate more heat and prolong life by running them that little bit cooler.