Capacitors in parallel why?

93 views
Skip to first unread message

Kerry Borgne

unread,
Feb 19, 2016, 5:51:08 PM2/19/16
to neonixie-l
 A question came up in one of my tech forums that I don't have a ready answer to. 

 The questions was in regards to switching supplies (used in LCD and Plasma Tv's) and the tech was asking why manufactures use three 1000uf35vdc caps in 

parallel instead of using one 3300uf@35vdc.  Other than perhaps a difference in cost I couldn't come up with an answer. There are other examples of the use of multiple caps in parallel instead of using one cap of equal value.  I've seen four or more.


 Since I don't know a brighter, and more experienced group of guys than here on this forum, I thought I'd ask if anyone had an idea?


Kerry

M W

unread,
Feb 19, 2016, 5:59:44 PM2/19/16
to neoni...@googlegroups.com
On switchers you want the lowest ESR (Effective Series Resistance) between the power source (the switcher) and ground, so you bypass it to ground with low-ESR capacitors. Lots of good reasons; Sometimes the usual frequency (say 200 kHz or so, whatever your switcher's running at, frequency wise) of your switcher is just right for those particular capacitors, so using 3 in parallel will give you really GOOD results. Another is that sometimes cost for those caps is REALLY good this week, if you buy 10,000 or something. Or space in the project (sometimes 3 smaller caps FIT better on the PC board than 1 humongous cap.) And, also, what happens if one of the caps goes bad? If there were 3 caps and now just 2 work, probably it'll still work, if there was just 1, it quits working. Engineering is fun, but at times sorta a complex mess :)

  Mark

From: Kerry Borgne <deac...@teleweb.net>
To: neonixie-l <neoni...@googlegroups.com>
Sent: Friday, February 19, 2016 2:51 PM
Subject: [neonixie-l] Capacitors in parallel why?

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/4d084614-a799-425d-b9d4-954eae37463f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


M W

unread,
Feb 19, 2016, 6:01:56 PM2/19/16
to neoni...@googlegroups.com
Oh, also, I should mention heat; If the ripple current into/out of the caps is dissipating say 1.5 Watts (sometimes more, it depends on the project) and you divide that by 3 caps, they won't get nearly as hot as a single cap would get. Cooler electronics last longer, heat can cause problems for Caps especially.

  Mark

gregebert

unread,
Feb 19, 2016, 6:38:25 PM2/19/16
to neonixie-l
There's actually a net reduction in generated heat with paralleled caps; the reason is that heat is proportional to RMS-current squared

For example, if you had 2 amps of ripple current in a single capacitor with 1 ohm of ESR, it would generate 4 watts (P=I^2 * R) of heat. Splitting into 2 identical caps would reduce the current in each cap to 1 amp, hence each cap would generated 1 watt of heat, for a net total of 2 watts.

Note that switching supplies generate short-duration current spikes, which can cause significantly higher RMS current versus a linear supply, which means you need lower ESR. Generally, the added-cost of low-ESR caps isn't an issue because smaller capacitance values (hence lower cost) are used in switching supplies.
-------------------
BEWARE of connecting caps in series, though, to achieve higher working-voltage. You have to take into account the leakage current and variations in capacitance; you must use balancing resistors. This increases wasted energy (heat). Reputable capacitor manufacturers have app notes on how to do this correctly.
Message has been deleted
Message has been deleted

Jim_Z

unread,
Feb 21, 2016, 12:34:29 AM2/21/16
to neonixie-l
Reply all
Reply to author
Forward
Message has been deleted
0 new messages