Re: Replace capacitors

The message <SP6dnYobLoL_dTDaRVnyjwA@xxxxxx>
from "Dr.Hal0nf1r£$" <femail@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
contains these words:

Lee wrote:
"M Jones" <matt73@xxxxxxxxxxxxxxxxx> wrote in message
news:tLqdnbB4K4jDJjDanZ2dnUVZ8qydnZ2d@xxxxxxxxx
Hi all,

I have been given a faulty PC that freezes and shuts off frequently.
The owner says the fault has been happening more frequently to the
point where the machine does not boot any more. I had a look at the
motherboard because I experienced the fault before in another
machine as leaking capacitors. When I looked I could see indeed 10
capacitors are leaking / bulging.

Since I am quite adventurous and the machine is decent spec. I am
going to attempt to replace the capacitors and use the machine in my
home. I have checked and see they are all 6.3V 1500uF. I've gone
into my local maplin and asked for new ones only to be told they are
not available but 16V 2200uF will do. Since I have never done this
before I thought I would ask if these will be sufficient
replacements for the ones that have gone or should I look online for
the correct ones? Thanks.






I have electronic experience mate been taught by some very experienced
engineers and I know its ok the cap has more voltage then original
just not less but not recommended to put a higher micro farad rating
in its place. cheers.

That would depend upon the capacitor's intended purpose and the circuit in
which it was incorporated:

Certainly you are correct in that it is never a good idea to put a
component
with a lower-rated working voltage in place of another as a general rule;
but the extra capacitance mey or may not be a problem.

If the capacitor's intended purpose is as a ripple-smoothing component
or an
AC decoupling capacitor then under most circumstances the difference in
capacitance rating will make negligible if any difference to the circuit's
function, working at 16 or less volts. However if the capacitor is
part of a
frequency-dependant circuit such as a stabilised ocillator or
pulse-switching multivibrator then the exact value is paramount; although
el;ectrolytic capacitors are so leaky in terms of voltage and have such a
wide tolerance margin that other types of more-stable capacitor are usually
incorporated in such a circuit.; such as mylar-film, polycarbonate, or
drilitic etc.

Having said that; the difference in capacitance between a 1500 microfarad
and a 2200 microfarad capacitor is quite substantial; and therefore to
be on
the safe side I'd probably only attempt repair using a maximum of 1800
microfarads. You will find that the physical size of a given rating of
capacitor generally increases in proportion to its working voltage; however
you may just about be able to squeeze a 1600uF 20v capacitor into the
available space, but don't quote me on that. Don't fall into the trap of
size though; as it also holds true that the size of a given rating of
capacitor reduces proportionally along with its working voltage, so
don't be
tempted to fit a 1600uF 12v capacitor - You *may* get away with it if
you're
very lucky, as the rating tolerance of some components differs
slightly from
that marked, but personally I wouldn't risk it. If you were to buy 1600uF
20v capacitors and they didn't physically fit in the available space
then it
may be possible to insulate the leads, (Remember to observe the correct
polarity when fitting them.) and use the insulated leads as "stilts" with
the capacitor slightly above the other components.

As a general rule though; try to never use a lower voltage rated component
than the original as our friend suggests. In the case of capacitors of this
type, a little extra capacitance is usually ok; but using 2200uF in
place of
1600uF may be pushing it, or then again it may not. (?) generally I have
found IMO that spotty teenagers and £5-an-hour assistants behind a desk in
electronics shops talk out of the wrong hole and get a power-trip from
doing
their job: Just as a dog isn't a horse if it's born in a stable, so a
school-leaver doesn't become a qualified engineer or technician by selling
electronic components. Seek the advice of an electronics engineer if you
wish, or wait until you have the near identical components and you can't go
wrong.

Normally, such repair work at the component level (the component, in
this case, being electrolytic capacitors), isn't usually considered a
cost effective option. However, winXP's WPA process makes such a repair
infinitely more attractive than the usual MoBo replacement exercise,
especially when your planned upgrade is still a good 2 or 3 years away
and you'd rather not have to invest in not only a modern MoBo, but also
new memory and processor right now.

Having stated why such repair work is even worth attempting, it's worth
remembering that we're talking about _electrolytic_ capacitors, not
mylar film or paper or polyester types and the first thing to keep in
mind is the standard value tolerance range of -20/+80%. The second thing
to keep in mind is that they are used in circuits where 'more is
better', i.e a larger value will do no harm, and will generally improve
performance.

It's also worth bearing in mind that the value worked out in the design
phase has been arrived at by making due allowance of that -20% tolerance
figure and picking the next standard available value above the minimum
desired value.

Since there can be a very wide range of values that can be used in any
one such circuit the manufacturer can often reduce manufacturing costs
by 'overspecifying' to a more restricted value range wherein a much
larger quantity of each capacitor value can be ordered into stock at
more favourable bulk discount rates.

The usual voltage ratings are mostly 6.3 and 16 volts with the odd 10v
rated capacitor used here and there. At these voltage ratings, a 16 volt
capacitor will be roughly twice as bulky as its 6.3 volt counterpart, so
it's worth avoiding the use of higher voltage versions of replacements
unless you know those replacements will physically fit in the space of
the originals.

Although it's quite obvious (or damn well should be!), that the voltage
rating of the replacement should match or exceed that of the original,
you can sometimes specify a lower voltage than that of the original, but
only if the originals were overspecified to begin with.

On MoBos you can spot this state of affairs where the manufacturer has
used 10v caps as a substitute for the more optimally rated 6.3v parts
since there aren't any other voltage levels between the 5 and 12 volt
buses. 10 volt capacitors can't be specified on a 12v line so just
_have_ to be substituting for 6.3v parts on the 5v bus.

I've repaired quite a few MoBos over the years and have never had to
actually buy any capacitors since I've acquired quite a large collection
of scrapped MoBos from which to recover perfectly good quality
capacitors from.

I can be pretty confident that this source of supply is no more likely
to fail than those bought in 'brand new' from a component supplier. In
fact, I can be more confident that they'll be just as good, if not
better than, 'brand new' capacitors!

Having said that, even if I were to specially order in supplies, I'd be
inclined to rather restrict the capacitance range to a very small one,
especially in the 6.3v range. Most likely, restricted to just 1800uF and
2200uF values since this covers most situations (the 1800uF would, for
instance, nicely substitute for 1000, 1200, 1500 uF values).

If I'm replacing 1800uF caps, and I can't find enough of that value
which would physically fit, I'd choose 1500uF as a substitute since (all
other considerations aside[1]), a slightly lower value is infinitely
better than the existing high resistance versions I'm trying to replace.

[1] The other considerations being that -20/+80% tolerance figure and
the manufacturer's choice of a value in excess of actual design needs,
which could well be 2 or 3 times larger due to inventory considerations.

HTH

--
Regards, John.

Please remove the "ohggcyht" before replying.
The address has been munged to reject Spam-bots.

.

Relevant Pages

  • Re: FX-6A flashtube availability and data
    ... all but the one designated as the sparker are used only as ... One clue is that the max anode voltage is 1,000 volts. ... an internal trigger takes less than the full holdoff voltage. ... I recall that the flash capacitors are charged to 600 volts; ...
    (rec.crafts.metalworking)
  • Would a micro-turboboost be practical?
    ... used to provide a very localized increase in voltage to increase ... spatially and temporally local power boost. ... computation where replay is used to when a result ... even with appropriate placed capacitors there would ...
    (comp.arch)
  • Re: B+ rectifier question
    ... I replaced all the electrolytic capacitors. ... diodes mounted on a strip and I replaced the original 2 diodes for the B+. ... I have to run the preamp on a variac at 110v because the filament voltage ... Do you recommend for the B+ using resistance before the ...
    (rec.audio.tubes)
  • Re: 9.5 questions on capacitor, LED, resistor
    ... Does output voltage stay ... LEDs require a resistor, but how do you calculate the required ... There are three limits on all components: voltage, current and power. ... Almost all capacitors are rated for some sort of overvoltage; ...
    (sci.electronics.basics)
  • Re: Tantalum cap failures
    ... Normal opperating supply voltage in our application is 28V. ... >We have had problems with some capacitors becoming short circuit during ... >The bank of caps is there to sink/source short-term PWM current ... >Total steady-state ripple current is around 0.5A. ...
    (sci.electronics.design)