Capacitor Ordering Help

[lcdman]

Hello All,

Since I don't know all about selecting replacement capacitors as of yet I have some questions(other than trying to understand all thats talked about here). Was on Digi-Key's website checking out capacitor stats yesterday. And I thought I'd try picking their collective brains for knowledge for once. I asked one of their reps. thru the chat window about how to pick the best caps for my applications ie LCD monitors and computer motherboards. A question I posed to that rep. was about ESR vs. ripple ratings. Since most of the caps I looked at on their site didn't even have an ESR rating, was it better to select by ripple rating instead? The higher the better or lower ripple rating? That rep. yesterday told me that they weren't aware of how ESR rating related to aluminium electrolytic capacitors? They thought it was more akin to titanium capacitors? So my question to the veterans and the like on this site, would it be better to go by a higher ripple value or lower with vendors such as Digi-Key that don't list an ESR rating for a particular capacitor series,wv,dimensions?

Thanks As Always
Lcdman

[mariushm]

Re: Capacitor Ordering Help

Both are important.

( note: I'm relatively a beginner in this but I did successfully recap a monitor and went through university with some electronics classes. Still you may want to wait for a word from other more experienced guys, maybe they disagree with something I say)

You have to open up the monitor, look for any markings on the capacitors and then try to find the datasheet for that series of capacitors.

For example, in my case I found a capacitor that says on it CapXon, GL and 330uF , 50v. So I made a search for "CapXon GL datasheet" and found the specifications.

In general, what's important is the uF value, the voltage, ripple and ESR value.

uF value, you keep the same.
Voltage, you keep the same or you can go a bit higher but never lower, unless you're very experienced and can tell from the circuit that voltage is safe to use. Sometimes when you need a capacitor with higher ripple or lower esr, I believe you may find capacitors rated for higher voltage have better ripple or lower esr or both.
Ripple value should always be equal or higher than what the current capacitor is spec'ed for.
ESR value should be equal or lower. I have a mental note from a thread I've read on this forum (but I may be wrong about this or misunderstood what was happening there)... in some cases if the esr value is way way lower than what the original capacitor is spec'ed for, you may hear some other parts start to buzz, humm or vibrate.

Here's the steps I followed:

Pick the minimum dimensions for the capacitor you have and others slightly bigger ones that would still fit in the area (for example I picked 6x8mm, then I also picked 8x10 and 8x11.5 mm because they would fit in the space). Filter...
Pick the uF value because that pretty much remains the same. Filter...
Pick the voltage. Filter... (if no results or nothing on stock, go back and choose higher voltages and filter again)
Sort results by dimensions, open the results in several browser tabs starting with the exact dimensions of the old capacitor...
Check on each page the ripple value and the ESR, if you don't see the ESR look down on the page and try to find a link to the datasheet on the page - there should be a chart there in the datasheet where you find it along with the ripple.

Then pick whatever you can order in the small quantity you need and still is better than your current capacitor specifications. (and from what I learned here, Nichicon, United Chemi Con, Panasonic, Rubycon are good and safe brands)

PS. Here's an example... 330uF, 25v, found this one on Digikey: http://search.digikey.com/scripts/Dk...name=P12387-ND
Scroll down to "Datasheets", click on the link. You're on Panasonic's site... click on Catalog. You get a PDF, scroll down until you see on the left 25, the voltage of the selected capacitor, and there you get both dimensions possible for that voltage and you see this:

[lcdman]

Re: Capacitor Ordering Help

Hello & Thank You mariushm,

I'm aware of most of what you've posted as I'm not a total newbie (Respectfully). I currently have a small stash of capacitors. I too have repaired a monitor before. But I too have done a little research on and off this site. I don't size up every cap, every time. Maybe thats wrong? From what I can gather ripple being higher usually means a correlating lower ESR value. I have worked on over a dozen LCD monitors and come to the realization that a common set of dimensions starts to come into play. And thus mfgs,series,wv,uf.
Maybe this .pdf will back up my conclusion?

http://dkc3.digikey.com/PDF/US2011/P1691.pdf

And if not I'm I hope I will be kindly told I'm all wet!

Thanks To All As Always
Lcdman

[PCBONEZ]

Re: Capacitor Ordering Help

Hello All,

Since I don't know all about selecting replacement capacitors as of yet I have some questions(other than trying to understand all thats talked about here). Was on Digi-Key's website checking out capacitor stats yesterday. And I thought I'd try picking their collective brains for knowledge for once. I asked one of their reps. thru the chat window about how to pick the best caps for my applications ie LCD monitors and computer motherboards. A question I posed to that rep. was about ESR vs. ripple ratings. Since most of the caps I looked at on their site didn't even have an ESR rating, was it better to select by ripple rating instead? The higher the better or lower ripple rating? That rep. yesterday told me that they weren't aware of how ESR rating related to aluminium electrolytic capacitors? They thought it was more akin to titanium capacitors? So my question to the veterans and the like on this site, would it be better to go by a higher ripple value or lower with vendors such as Digi-Key that don't list an ESR rating for a particular capacitor series,wv,dimensions?

Thanks As Always
Lcdman

The people you get on the phone at Digikey are salesman first.
It's like getting advice from a salesman or the Geek-Squad at Bestbuy.
They went to some 'training' to get the basics but they have little real world experience and often they know just enough to cause trouble for people.

Ripple is simply a current limit. [Max it can handle.]
It's like the amp limit on a piece of wire. - A current limit.
Replacing something with a part that has a lower current limit is a bad idea.
.

[PCBONEZ]

Re: Capacitor Ordering Help

From what I can gather ripple being higher usually means a correlating lower ESR value.
Lcdman

That is common but it's not like a rule or anything.
-
Don't count on it being true when the caps you are comparing are not the exact same brand and series.
.

[lcdman]

Re: Capacitor Ordering Help

Hello PCBONEZ and All,

So I gathered a list from Digi-Key of possible candidates for purchase. These are based on approximate equivalent sizes and uf/wv of capacitors that I've encountered so far from various LCD monitor pcbs,mfgs.


220uf/35v (P12411-ND) Panasonic FM Series
100uf/35v (565-1979-ND) UCC LXZ Series
4.7uf/50v (493-1889-ND) Nichicon PW Series
47uf/25v (P14411-ND) Panasonic FR Series
1000uf/16v (P14398-ND) Panasonic FR Series


I'm looking for another value and need assistance with it as well:?

0.1uf/50v 5mmX6mm or very close

From what I've listed, in all you fellas sometimes vocal opinions am I on the right track so far? Please remember that I'm trying to do this right the first time and as economically as possible. Feel free to politely chime in!

Take Care
Lcdman

[smason]

Re: Capacitor Ordering Help

You could sweat the details and specs all day, but many of us, including the great PlainBill just go with Panasonic FR, FM and FC capacitors. In monitors, they're more than adequate, and I've done over 20 monitors, with no problems.

That said, there's nothing *wrong* with chasing specs and learning all you can about capacitor ratings.

[lcdman]

Re: Capacitor Ordering Help

Hello smason and All,

What about : I'm looking for another value and need assistance with it as well:?

0.1uf/50v 5mmX6mm or very close (105*C)

I didn't see any of those Panys or any other of the "good" cap series on Digi-Key? Is there another place online that carries FR,FM,FC in that value? And as far as my list some look like they'll pass?

Thanks As Always
Lcdman

[smason]

Re: Capacitor Ordering Help

Yeah, slim pickings on that value. A few SMT ones...
0.1 is not likely to be critical in terms of ESR and ripple, so whatever fits may have to do.

[PCBONEZ]

Re: Capacitor Ordering Help

You could sweat the details and specs all day, but many of us, including the great PlainBill just go with Panasonic FR, FM and FC capacitors. In monitors, they're more than adequate, and I've done over 20 monitors, with no problems.

That said, there's nothing *wrong* with chasing specs and learning all you can about capacitor ratings.

That plan works usually but there are times it will get you into trouble.
.
If you would like I can find the thread where I showed PlainBill [with the math] where replacing the crap cap in question with an FM would exceed the ripple rating of the FM.
I've done that twice in two months since I began reading the monitor threads occasionally.
FM and FR are simply not always good enough.
I found two instances where they weren't without even trying.
.
Further, there are LOTS [and I mean LOTS] of crap caps that have better ratings than FC.
Using FC as a default when you have no idea what the original specs are is FOOLISH.
.
"Chasing specs" as you say, keeps you out of trouble.
Having a pet cap you slap into everything without checking doesn't.
.

[PCBONEZ]

Re: Capacitor Ordering Help

Hello smason and All,

What about : I'm looking for another value and need assistance with it as well:?

0.1uf/50v 5mmX6mm or very close (105*C)

I didn't see any of those Panys or any other of the "good" cap series on Digi-Key? Is there another place online that carries FR,FM,FC in that value? And as far as my list some look like they'll pass?

Thanks As Always
Lcdman

Look at Panasonic KG series. [Good brand, low ESR, and not in the far outdated 'list'.]
That series in the 6.8uF and less-uF 50v caps has very good ripple ratings and they are low ESR.
10uF [50v] and up has good ratings too but they start getting large [can size] for their uF so there are usually better options.
http://search.digikey.com/scripts/Dk...l&name=P925-ND
Not to mention you can get 50 of them for $5.
.
.
Can't make a -realistic- recommendation on the others because I don't know what you are replacing.
No one can.
.

[PlainBill]

Re: Capacitor Ordering Help

That plan works usually but there are times it will get you into trouble.
.
If you would like I can find the thread where I showed PlainBill [with the math] where replacing the crap cap in question with an FM would exceed the ripple rating of the FM.
I've done that twice in two months since I began reading the monitor threads occasionally.
FM and FR are simply not always good enough.
I found two instances where they weren't without even trying.
.
Further, there are LOTS [and I mean LOTS] of crap caps that have better ratings than FC.
Using FC as a default when you have no idea what the original specs are is FOOLISH.
.
"Chasing specs" as you say, keeps you out of trouble.
Having a pet cap you slap into everything without checking doesn't.
.

The only problem with that thread is the math is based on unwarranted assumptions. For example, you ASSUME that when the ESR of a cap is reduced, it's allowable ripple does not increase. This is contrary to both logic and an empirical examination of the specs.

What is even worse is you make two very simple mistakes. First, you ASSUME a given design is uses the components at their limits - highly unlikely; even the greenest design engineer knows enough to allow some leeway in the design. The last ASSUMPTION isn't so serious, but is very common - it assumes the spec represents a 'critical edge' - that a cap rated for 2380 mA ripple current will fail if the ripple current hits 2381 mA.

PlainBill

p.s. Hot enough for you?

[PCBONEZ]

Re: Capacitor Ordering Help

The only problem with that thread is the math is based on unwarranted assumptions. For example, you ASSUME that when the ESR of a cap is reduced, it's allowable ripple does not increase.

That is NOT an assumption.
- ESR and Ripple do not always track in locked-step when you change the series.
Examples can be shown, thus it's not an assumption, it's a fact.

Your theory is that ESR and Ripple -always- change in locked-step and that it applies even across different series.

When you change series there are different electrolytes, different foil alloys, different foil thicknesses, different etching methods, different methods and degrees of creating the oxide layer, and usually different CAN SIZES,,,
,, all of which affect both the ESR and the Ripple rating - to differing degrees.

Here is one example: 1500uF in 8x20mm
MBZ - 1870 mA , 0.019 Ohms
FJ - - - 1870 mA , 0.016 Ohms
If you went from FJ to MBZ there would be a ~19% reduction in ESR and ZERO increase in Ripple.

Please Note: That example even even holds the can size constant.
Constant can size OFTEN does not happen when you are switch from -whatever- to a limited series list such as FC/FM/FR.
Examples of ESR not being in locked-step with Ripple can be found all day long when the can size isn't held constant.

Regardless: EVEN when the can size is held constant there are examples that disprove your theory.
Holding on to a theory that has examples to the contrary is defective reasoning.

Thus, my 'assumption' is not an assumption, it has known examples and is a fact.
.. And your theory to the contrary is debunked.
.
.
~~~~

What is even worse is you make two very simple mistakes. First, you ASSUME a given design is uses the components at their limits - highly unlikely; even the greenest design engineer knows enough to allow some leeway in the design.

That is not a mistake. - That is referred to as GOOD ENGINEERING PRACTICE.
[That is an actual term used in Engineering fields.]
And, it is the accepted way to do things in the Engineering professions.
- When you need to replace a part and you don't have access to all the Engineering details you choose the worst case for each parameter and that is what you base the replacement part on.
- OR: You reproduce all the Engineering details. [Which here means draw out the full schematic and analyze the circuit.]
-
That's how it's done IAW Good Engineering Practice.
- And you aren't doing either one in the threads with OTHER peoples 'stuff'.
-
Were the device in question a car.
What I just told you is how a Mechanical Engineer or [hopefully] formally trained and experienced Mechanic would do it.
The way you are choosing replacement caps would come under "Shade Tree Mechanic".
.
And you are doing it like that with other people's 'stuff'.
Were you only doing it with your own 'stuff' - I wouldn't care.
.
.
~~~~

p.s. Hot enough for you?

Thank God for AC.
[You may pick whichever God isn't busy at the moment.]
.

[smason]

Re: Capacitor Ordering Help

Here is one example: 1500uF in 8x20mm
MBZ - 1870 mA , 0.019 Ohms
FJ - - - 1870 mA , 0.016 Ohms
If you went from FJ to MBZ there would be a ~19% reduction in ESR and ZERO increase in Ripple.

Would there really be an issue with the difference between .019 and .016 ohms of ESR in any consumer electronics circuit?

It's a serious question, really I'm not disagreeing, but seems to me that's pretty a trivial . I've replaced caps in a circuit that's just starting to have issues, where the ESR is 2 to 3 OHMS just a tad more than 19% out.

[PCBONEZ]

Re: Capacitor Ordering Help

Not likely to but it can be if that 0.019 ohm cap is remotely near it's Ripple Current limit.

First realize that the Ripple Voltage as measured on a cap goes down when ESR is reduced because more Ripple Current is passing -through- the cap.

Lowering ESR in a filter does not reduce the Watts of Ripple coming from the source, it only redistributes the -flow- sending more through itself (more Ripple Current -through- the cap) and away from the load.
[That's why lowering ESR is good - *IF* the cap can handle the Ripple Current.]

Point is the energy in the Ripple doesn't go *poof* - it gets rerouted through the cap.

Given the same circuit [thus the same Watts of Ripple to deal with] , Lower ESR = More Ripple Current [Through the cap]

If you are going to send more Ripple Current through the cap then the cap has to be able to handle it.

Also note: Ripple Current is the primary source of a cap's internal heat. [Temperature]

~
Using the FJ and MBZ from before...

A cap with an ESR of 0.019 Ohms would hit a 1870 mAmp Ripple Current at 35.53 mV of Ripple [which equates to 66.44 mWatts of Ripple.]
A cap with an ESR of 0.016 Ohms would hit a 1870 mAmp Ripple Current at 29.92 mV of Ripple [which equates to 55.95 mWatts of Ripple.]
-
In other words the Lower ESR cap can only handle 84.2% as much Ripple -> without overheating.

So, lets say the 0.019 ESR cap was slapped into a circuit that has enough mW of Ripple [56.47 mW] for 85% of it's rated Ripple Current to be present.
[Which is entirely believable coming from a company that used crap caps in the first place, and 90% wouldn't surprise me either.]
If we swap in the .016 ESR cap it is going to have just enough Ripple Current to exceed the rating.
- Not much over spec in this example but that is with a mere .003 Ohms difference [and the same can size].
If the 0.019 ohm cap was at 90% of it's rating [which is possible] the 0.016 ohm cap would be at 1999 mA. [129 mA over.]
-
When it come to changing can sizes the difference can be a bunch more than .003 Ohms.
Going to a much smaller can can get you into trouble even if you don't change the ESR.
-
Same-same when jumping from a 'just barely low ESR' crap cap to an FM or FR - the difference can be a bunch more than .003 Ohms.

Beyond just the heat from the Ripple, excessive Ripple Current has a thinning effect on the thickness of the oxide layer.
As the oxide layer thins the DC Leakage Current goes up creating even more heat inside the cap.

Overheat = Short Life. [A 10C rise reduces the cap's life by 50%.]

Now, if the 0.016 Ohm cap could handle 15% more Ripple than it does....
... which is often the case when ESR is lowered but, as shown by those example caps, is NOT ALWAYS the case ...
... then everything would be fine.
.
.
The main point is:
Don't assume things by default.
Whenever possible check the old specs before choosing replacements.
.

[smason]

Re: Capacitor Ordering Help

Not likely to but it can be if that 0.019 ohm cap is remotely near it's Ripple Current limit.

Interesting. Thanks for elaborating.