Mounting caps slightly off the board

[Professor Peach]

Re: Mounting caps slightly off the board

It does seem to make a difference, but not very much.

[shovenose]

Re: Mounting caps slightly off the board

ok. but not enough to make a difference or to constitute using an over-kil esr cap

[PCBONEZ]

Re: Mounting caps slightly off the board

ok. but not enough to make a difference or to constitute using an over-kil esr cap

So in your infinite wisdom if the original cap reads ESR @ 0.01 on the meter using a cap that reads ESR @ 0.02 is okay??

Particularity as considering with the resolution of these meters the original that reads 0.01 might actually be 0.005 and the replacement that reads 0.02 might actually be 0.024.

Some how I think 'upping' the ESR by nearly 5 times will noticeably affect the ripple in just a few circuits.

.

[shovenose]

Re: Mounting caps slightly off the board

do yo have to argue about EVERYTHING???
ok so i dont know that much about esr, but im assuming tht 0.21 and 0.20 is a small differene

[PCBONEZ]

Re: Mounting caps slightly off the board

It does seem to make a difference, but not very much.

Thank you.

[Krankshaft]

Re: Mounting caps slightly off the board

Increased ESR does occur but it seems like it's just out of the resolution of the BP meter in most cases.

If anyone has an LCR meter nows that time to whip it out .

[PCBONEZ]

Re: Mounting caps slightly off the board

do yo have to argue about EVERYTHING???
ok so i dont know that much about esr, but im assuming tht 0.21 and 0.20 is a small differene

I argue things that are wrong because I think having the correct information on BCN is important.

Since we are asking questions like this,,
Why do you insist on posting things with your brain in neutral?
and
Why do you ask questions that have already been answered many times over?

ok so i dont know that much about esr

Yes. You don't [yet].

I was a READER here for a VERY long time [couple of years] in addition to READING [studying] in other places and actually working with affected equipment BEFORE I started posting advice about caps.
-
Yeah,,, I'd asked stupid questions too, but I at least TRIED to find the answer -before- I'd asked a question.


No one is always right but if I don't know [or think I do] and/or I haven't researched something then I DON'T POST AT ALL.
- I LAUGH when someone spends all their time building their 'post count' by throwing out wild guess answers.
- Post count means nothing.
- Helping someone does.

There is a learning curve.
You and 'certain other' people seem to want to jump to the 'status' at the end of it without having done the learning.

All posting bad information does is make it harder to find the good information.
- It is no better than spam.

-
In the 5 years I've been here BCN MODs have NEVER bothered to clean out or correct bad and/or outdated information or even update sticky reference threads.
It's been adding up for years.
It's getting so there is more bad info on BCN than good.
-
Since I often refer people [or businesses] from other places to BCN for the right information, I care that there is good information here.
- I seem to be the only one that -cares- about that anymore.


Insofar as -frequent- visitors that have gone through the learning curve and are willing to go -out of their way- to help other people,,,,
,,, I think BCN is down to three.
Poping in once a month isn't frequent.
Tossing out two sentances or giving an answer without explain why isn't going out of your way.
.

[PCBONEZ]

Re: Mounting caps slightly off the board

Increased ESR does occur but it seems like it's just out of the resolution of the BP meter in most cases.

If anyone has an LCR meter nows that time to whip it out .

You don't even need an LCR meter. - Just google.
A number of caps manufacturers have docs that discuss the effect on ESR due to a cap's construction and the effect of the small resistance in leads is generally covered.
.

[MXM]

Re: Mounting caps slightly off the board

^ thanks for the reply on my Q.

peace.

[severach]

Re: Mounting caps slightly off the board

There's a reason why filter caps are always placed as close to the load as possible. Increasing the lead length increases the ERP and ESL. If it weren't so then the 12v caps in the power supply would be sufficient for the whole 12v circuit. Testing with the ESR meter is useless since the leads are 2 foot long.

That said I've seen circuits running fine with ridiculous amounts of ESR. +-0.01 ohms from a quarter inch of leads shouldn't make any difference.

[PCBONEZ]

Re: Mounting caps slightly off the board

No one said 1/4" and two people -just in this thread- have confirmed the existence of the phoneomena I described.

There's a reason why filter caps are always placed as close to the load as possible. Increasing the lead length increases the ERP and ESL. If it weren't so then the 12v caps in the power supply would be sufficient for the whole 12v circuit.

The 12v caps are where they are because that's where the VRM is located. [A VRM -is- an SMPS and it needs it's own filters.]
-
Yes, ERP and ESL due to the length of the TRACES on the board matters too, which is why caps for 5v and 3.3v are spead all over the board instead of being in a grouped right next to the PSU socket.
-
Given the effects of ERP and ESL from trace length are tiny compered to the effect of .01 ohm ESR on a meter you have basically given further evidence that .01 ESR on the meter matters.

~~
As for .01 ohms making a difference or not:

10x20mm
MCZ - .011 [.01 on the meter]
MBZ - .013 [.01 on the meter]
KZE - .023 [.02 on the meter - and exactly .01 different from MBZ]
WX - .023 [.02 on the meter - and exactly .01 different from MBZ]
ZL - .023 [.02 on the meter - and exactly .01 different from MBZ]

10x23mm
MBZ - .012 [.01 on the meter]
WX - .022 [.02 on the meter - and exactly .01 different from MBZ]
ZL - .022 [.02 on the meter - and exactly .01 different from MBZ]

10x25mm
MCZ - .009 [.01 on the meter]
KZE - .022 [.02 on the meter]

If you TRULY believe that .01 Ohms on the ESR meter doesn't matter then, by your own argument, the following statements are completely true:
~ You never need MBZ or MCZ grades of caps in 10mm sizes because KZE, ZL, WX are good enough.
~ You can replace MBZ & MCZ with KZE, ZL, WX without any problem because the .01 ohms difference just doesn't matter.

Personally I don't believe those statements are true but a number of people here are arguing otherwise.

If .01 ohms difference doesn't ever matter - then it doesn't matter ever...
Can't have it both ways....
.

[Wizard]

Re: Mounting caps slightly off the board

Between source (even before the SMPS!) and load there are bunch of different connections no matter what, soldered or sprung contacts and different kinds of conductors. Also to complicate this matter, components and joints, even capacitors do have these structures. All are piles of different resistances and this adds up.

With this said. Depending on some electronics low voltage at extremely high current (peak current can be very high for nanoseconds at each peak current demands pulled by CPU) can cross below minimum voltage of the CPU easily especially if the resistance of the too long leads of said CPU vcore capacitors (even using polymer caps) is sufficient to cause this. Also length of wire (even caps)/too narrow power trace can act as inductors which is most especially on low quality makers) PC crash. Enough said.

In my case it was AWG 26 (iirc, that was what fits through the mainboard solder holes) solid copper wires and was only 1 inch long or less with cap leads crimped securely right at caps bungs on those and properly soldered. Caps were proper caps for vcore caps. Freezes/crashes. Mainboard is Dell Dimension 8300. After correcting this with proper diameter caps and directly on the mainboard (sitting on PCB). Results: all is well, 100% reliable, and using 8300 PC for years now, going on 2 or 3 years now.

Hit the Vcore with 100MHz or higher scope. You will see hash. I see this especially worse on generic TV mainboards and in some cases with SMD electrolytics even on good brand maker TVs but this is their minimum specs designed to a cost.

ESR meter WILL NOT spot this. Not enough current. Spotting this requires electronics knowledge and scope to see this in action. This is reason why there is design guidelines set out by IC makers. I have read them.

PS even a design error of only a 1/2 mm long trace of wrong width or bad layout can screw up a circuit using *certain* IC. I have seen these on datasheet PDF files for ICs.

Cheers, Wizard

[Wizard]

Re: Mounting caps slightly off the board

One more thing. ESR meter is merely to check of the ESR capacitor ONLY. And understanding the limitations of different tools and electronics theory of operation stuff is key to success. Misusing the tools without understanding these ends in problems.

This is what set apart good designer and bad designers also in repair trade that seperates repair hack from good repair tech.

Cheers, Wizard

[Per Hansson]

Re: Mounting caps slightly off the board

I went into this a little here based on the fact that Antec released a PSU with capacitor at the ends of it's cables.
It's a design in afterthought, i.e. instead of fixing the real problem they put capacitors on the cables like 40' away from the Pi Filter itself...

If that where a good design choice we would see capacitors in a computer grouped by the edges of the board, where it is cooler, so they would live longer.
But that is not possible, increased trace lengths causes ringing and all sorts of other problems like mentioned by Wizzard

http://www.jonnyguru.com/forums/show...4&postcount=55

Testing with the ESR meter is useless since the leads are 2 foot long.

Nope, you forget that you zero the meter with the leads crossed before starting to measure caps
Else you get bogus readings...

[PCBONEZ]

Re: Mounting caps slightly off the board

Too low of ESR can also cause ringing.

Normally the effect of ESR is much larger than the effect of ESL which is why ESL is rarely mentioned at all.
Making ESR too small relative to ESL does effectively the same thing as keeping the ESR the same and adding ESL with longer leads or traces.

The usual method of suppressing ringing is to add resistance which is -ta-da!- the same thing as adding ESR.

.