When to Choose ZL vs. MCZ

[PCBONEZ]

Re: When to Choose ZL vs. MCZ

As you are going up so far on the uF I'd forget about the empty positions all together.
That would just add even more uF.
Going to MBZ from KZG you are already lowering the ESR anyway.

I personally don't and wouldn't make those big of jumps in uF but I know others in here have and they say it worked. But on the other hand by human nature people don't pop into forums and brag on their failures very often. Hard to say how many tried and it didn't work out.
- Take your best guess on that one.

.

[gonzo0815]

Re: When to Choose ZL vs. MCZ

This would be a good board for solid polymer caps like Samxon X-con or Sanyo Oscon SEPC.

Even with the water block, this caps will getting hot. This boards design needs airflow.
Solid polymer caps are not that sensitive for heat and provide lowest ESR possible.

[Kaine]

Re: When to Choose ZL vs. MCZ

I'll be sure to let you know how it goes

[kc8adu]

Re: When to Choose ZL vs. MCZ

if that were my board i would verify the empty spots were vcore and mark them.
since i got a deal on a huge quantity of polymer caps i would use them on vcore and fill all positions there.mbz/mcz would be fine there too .filling the blanks has always helped any board i did it to.
the mbz you have are fine for vrm input.
just do a simple ohmeter test before stuffing caps into blank spots on a board.
i had an oddball once that had a blank in the vcore filters that was opposite polarity.
good thing i checkerd it as it was a vrm input with +12 on it!

Forgot to say the MBZs I have are slightly hugher capacitance than the KZGs i'll be replacing 1500uF 6.3v and 1000uF 16v KZGs with 2200uF and 1800uF MBZs. Stupid forum software edit time limit.

Speaking of which;



See the unpopulated spots? Any suggestions on what I should put in them? For reference it is an A8N-SLI Premium (socket 939). The heatpipes will be going pretty soon (having a waterblock made for the VRMs) and I figure since I'm going to have a few spare MBZs I might as well populate them.

[bgavin]

Re: When to Choose ZL vs. MCZ

I don't know where your reading.

My datasheet says PX only go to 3300uF in 6.3 and 10v and even then the ripple they handle is 1100 mA or less. The 6.3v MBZ 3300uF handles 2800 mA.

Rubycon PX datasheet.pdf, 30SEP2006

10mm can size up to 10v.
12.5mm at 16v
16mm at 25,35v

For Kaine, the 1000uF KZG in 10mm has the same ESR as the MBZ 1800uF. For my newbie point of view, this is too much of a jump. You can get MCZ 1000/10v/10mm with ESR of 12.5.

[gdement]

Re: When to Choose ZL vs. MCZ

MBZ are already better than good (not failed) KZG and will do fine there.

Assuming they're in equally good condition, then I don't see any difference between them on the datasheets. The only difference I can find is that one of the standard can sizes is slightly different: Rubycon uses 10x23 and UCC uses 10x25.

========
I agree with tracing out the caps, or at least the empty pads. It would be helpful to populate any extras if you know what circuit they belong to. Given the heat issues involved here, extra caps also might help dissipate the heat slightly better, allowing the next batch of caps to last a little longer before drying out.

You can't just guess though - the connections usually aren't all what you expected just from looking at it. It has to be checked with a multimeter. I've traced about 5-10 boards so far and I've never been able to guess all the pads correctly.


Re raised capacitance:
I've raised Vcore by maybe 50% or so on at least a few board models, and I don't think it ever caused a problem. However, on the first board I recapped, an ABit BX133-RAID, I tried replacing all of the 1000uF and 1500uF caps with 1500uF replacements. That didn't work out so well - the PSU couldn't turn it on anymore. It appeared to be overloaded, just twitching the fan and stopping right away. I had to put the correct sizes in the 1000uF spots to fix it. I'm guessing the input caps were to blame for the problem. I don't think it was the Vcore caps because I think ABit already used 1500uF at Vcore.

[starfury1]

Re: When to Choose ZL vs. MCZ

Humm well I haven't read of a VRM failing due to extra capacitors or capacitance.
and there seems to be to schools of thought as to weather or not.
some say do some say don't.

I guess manufactures probably use the lest they can get away with for standard running conditions.
if there are extra positions it does allow them to use different values in combination.

I guess the real enemy is the heat they generate and are subject to.

From what I understand if one starts failing it will add extra load to the others therefore a domino effect follows and the next weakest in the chain will start to go.

So I suppose there is some advantage to having extra caps on the PCB to lengthen the time of failure (more caps less load on each cap)

anyway thats the guesstimated theory and it could totally be wrong!.

Extra capacitance would have to change the operation of the VRM in some way(s)

I think some aspect of timing does alter and possible the "voltage droop" are affected
This is from memory from a discussion awhile back and possibly from a doc I read on the design of VRM circuits
But not totally sure on this

If it is, it doesn't seem to effect the operation of the MB to any detrimental affect.(that I know of or have read about here)

But I guess its a moderation type thing, don't go way overboard
(as Bonez said theres probably a point where its pointless)
me personally I don't thing a little extra capacitance goes astray

I guess its of most advantage to those pushing systems to the max but doesn't go astray for a longer life of the VRM circuit on the average box


As I understand VRM can be run off 12V or 5V I think now 5V is the norm but suppose that depends on the VRM controller chip used (older stuff may well be 12V)
(I think there might be some sort of standard to this now?)

Yeah agree you must check that thats what the empty positions are for with a multimeter (continuity, low ohms)

Well thats my uneducated thoughts on it..I don't design them so don't tend to think too hard on the design aspects of them.

On PSU my understanding is that low ESR (or ultra low) such as MBZ MCZ wouldn't last as long and although low ESR is required for output it doesn't have to be as low as what you need for VRM use

what you should look for is a longer life Cap of preferably non aqueous electrolyte such as panasonic FC series.

If memory serves the ruby ZL series is suited for PSU use
(if I remember correctly from when I had a rummage around there site)

Still I suppose using MBZ is a hell of a lot better then the crap thats probably be in them and would still last a reasonable amount of time I think

like I said I could be wrong, so please correct it if it is

HTH its of some help

Cheers

[Brian C]

Re: When to Choose ZL vs. MCZ

I've never seen MCZ in PSU output filter. I found MCZ usage on only computer mobo so far... High grade PSU like the 700W (made by Delta) for HP-UX C8000 workstation uses all ZL.

As I understand VRM can be run off 12V or 5V I think now 5V is the norm but suppose that depends on the VRM controller chip used (older stuff may well be 12V)
(I think there might be some sort of standard to this now?)

Recent mobo VRM circuit powered by +12V supply for better efficiency. Consider a +12V @ 15A max, the output power = 180W, it is equivalent to +5V @ 36A.

If the CPU power consumption is 65W and the VRM efficiency is 100%, the +5V powered VRM will draw 13A from +5V line. For +12V powered VRM, it draws only 5.42A from +12V line. Hence, power loss across cable and connector due to heat is minimized.

This is just my opinion...

[starfury1]

Re: When to Choose ZL vs. MCZ

yep with you on the theory of that Brain C ,
Higher voltage less current same amount of power which makes sense when you consider connector resistance and wire resistance
Less that comes into play the better.
As I said I don't really know these days I don't do it day in day out like some here

My understanding is that MBZ MCZ is probably not the best choice for PSU use and the very low ESR is not really required for PSU output.
There is an article on Recapping PSU, I think OW did at jonny G's were the use of quality caps resulted in worse slightly noise & ripple I think.
(Id have to find it and reread it)

There is another post round here were the use of quality caps caused the switch mode psu to drop it bundle and shut down so go figure
(DVD player psu)

As I stated somewhere else on the forums, I wonder how much some circuits depend on the actual characteristics of the caps used
I don't design SMPSU so realty don't have a clue, just a suspicion

But the bottom line is I suppose you want caps to last a PSU.
(like i said, not to say they wont work but not the best choice for the situation)

Any as I said "horses for courses" and there are going to be a few exceptions to the rule of what should be an ideal better cap may not perform as expected in the circuit.

cheers

[Brian C]

Re: When to Choose ZL vs. MCZ

My understanding is that MBZ MCZ is probably not the best choice for PSU use and the very low ESR is not really required for PSU output.

I agree... from what I know, many prefer non-aqueous caps in PSU, although its ESR not as low as the aqueous counterpart.
https://www.badcaps.net/forum/showth...ht=non-aqueous

It seems there is a trade off between ESR and the Lifespan (reliability). Not sure about it.

The secondary output of PSU uses heavy low pass filter formed by inductor and capacitor, being single stage or 2-stage. The ripple current is not so high compared to VRM for CPU. Hence it does not need ultra low ESR caps. High reliability type is prefferable.

Hope someone can explain further ... ...

[Brian C]

Re: When to Choose ZL vs. MCZ

Well... the choice of capacitors depends on the circuit design. I had a bad experience with LDO using too high capacitance at the output (high capacitance -> lower ESR). It caused severe oscillation.... 2 out of 100 sets produced.

Output capacitor ESR affects the stability of LDO regulator.

http://focus.ti.com/lit/an/slva115/slva115.pdf

[davmax]

Re: When to Choose ZL vs. MCZ

I've never seen MCZ in PSU output filter. I found MCZ usage on only computer mobo so far... High grade PSU like the 700W (made by Delta) for HP-UX C8000 workstation uses all ZL.



Recent mobo VRM circuit powered by +12V supply for better efficiency. Consider a +12V @ 15A max, the output power = 180W, it is equivalent to +5V @ 36A.

If the CPU power consumption is 65W and the VRM efficiency is 100%, the +5V powered VRM will draw 13A from +5V line. For +12V powered VRM, it draws only 5.42A from +12V line. Hence, power loss across cable and connector due to heat is minimized.

This is just my opinion...

I can confirm that ZL and non aqueous are best in PSU and VRM input. aqueous Ultra Low ESR or polymer caps best in VRM output.

Also the above statements re. 12V and 5V are very relevant, lower current in leads and connectors for 12V is a positive step in the right direction.

[bgavin]

Re: When to Choose ZL vs. MCZ

Rubycon says the ZLH is a long-life upgrade to the ZL. In my spread sheet notes, I have them specified for SMPS Output... but I can't find the documentation that spells out this specific use.

Davmax, do you have a URL or documentation from the manufacturer that specifies which models are optimal for a given SMPS purpose (i.e. input, output)? If so, I want to add to my library.

[Kaine]

Re: When to Choose ZL vs. MCZ

Rubycon PX datasheet.pdf, 30SEP2006

10mm can size up to 10v.
12.5mm at 16v
16mm at 25,35v

For Kaine, the 1000uF KZG in 10mm has the same ESR as the MBZ 1800uF. For my newbie point of view, this is too much of a jump. You can get MCZ 1000/10v/10mm with ESR of 12.5.

I am aware there are more appropriate cap choice available, people seem to be ignoring that I'm using the MBZ caps because that is what I have here, on my desk, pulled from another motherboard. I'm not ordering these in, if I were to order any I'd call on Big Pope and match the current caps with Samxon GCs.

1500uF 6.3v 10mm 0.014Ω 2280mA
1000uF 16v 8mm 0.012Ω 2220mA

Which is what I will do if the MBZ cause problems. I understand entirely that I'm increasing capacitance by at least 50% but this is a bit of an experiment for me. I'm not losing anything if the board won't boot because the increased capacitance increases the switch on load too high. I'll just pull the MBZs out and order the Samxons.

[bgavin]

Re: When to Choose ZL vs. MCZ

I don't know if Big Pope can distribute in the US or not. I remember something about Topcat being the distributor, so you should contact him.

I've seen nothing but excellent press (here) on the Samxon caps, and would be willing to try them also. I can't seem to find a bulk source for Rubycon ZL(x) series. I buy MCZ in bulk from Wai Fong in Hong Kong, but they do not carry the others I need for PSU.

[PCBONEZ]

Re: When to Choose ZL vs. MCZ

According to the ATX Power Supply design specs the PWR_OK signal must be sent in less than 500ms from the time the PSU is turned on.
If it's too slow the PSU won't start.
-
The PWR_OK means that not only are the voltages all within spec; but the PSU's capacitors are charged to a level that they can sustain in-spec power on the output for a minimum of 17ms with a total loss of input power.

Big caps take longer to charge.
If you have too much capacitance in there the PWR_OK will be late and the CPU isn't going to start.

I have tested the no PWR_OK signal condition by clipping the PWR_OK wire to see what happens. What happens is the PSU starts and runs normally (and does not shut off) but the motherboard apears dead. Nothing happens. Not even POST.

How much excess capacitance it would take to do that I don't know (probably a LOT) but even then every PSU is different and has other components that affect all this.
-
The 'stock' caps in some PSUs may already be near the max limit of capacitance to make it work. - Other PSU's may have room for 4 or 5 times as much.

I don't like guessing and doing things over 3 or 4 times.
I just like to fix it and be done, so whenever possible I stick as close to the 'stock' capacitance as I can.

.

[bgavin]

Re: When to Choose ZL vs. MCZ

I certainly agree with staying with stock specs.

Spec deviation isn't my problem, nor is one particular PSU. These boards are DOA with several different PSU.

For PSU use, I gather from all the above that the ZL(x) are high-end caps for this use. MCZ is low-ESR which appears to have a corresponding lower life span, which isn't really suitable for PSU use.

Trying to stay on topic, but:
1) Is there a bulk source for Rubycon ZL that anybody knows?
2) Is there a tutorial for basic doorbell bugshooting of the PowerOn circuit?

[Brian C]

Re: When to Choose ZL vs. MCZ

The POWER OK signal is connected to SMPS controller IC or the related protection circuit. Some PSU use comparator IC (LM339, LM393) to generate POWER OK signal, some use supply supervisory IC like the TPS3510.

Seems like this is a bit Off Topic...

[bgavin]

Re: When to Choose ZL vs. MCZ

I don't care about the PSU. I'm interested in the motherboard and the path through the ZL caps use by the power on circuit.

Somebody mentioned the ZL caps are close to the AT12V connector and an associated toroid on the board. I'm curious if there is a tutorial, etc, that shows this basic circuit, and how to make basic checks on a board that won't power up.

I have recapped a few of these dead Intel boards without success. It sucks replacing all those caps with high quality stock, only to have the board remain dead.

[starfury1]

Re: When to Choose ZL vs. MCZ

well there is the testing procedure Akor created for VRM

https://www.badcaps.net/forum/showthread.php?t=600

there is docs on how the VRM works too

thing is why a MB wont power up wont always be related to a power issue.
dead chipset etc

Its possible there is a common related issue to the MB model but that I can't answer.

I would treat a dead MB as a somewhat high risk in regard to recapping, although possible its caps only, seems there is a fair chance its not the issue depending on the MB (although it could have created the issue).


but again those that do on a day by day basis would be better to answer that question


To your specific question, I do remember reading that too but can't find it
I think it was in relation to the VRM input caps being near the PWR connector?
but could be wrong!!!

I don't know of any info on that particular circuit situation.
I suppose basic trouble shooting applies ...is there a short? is there an open circuit? etc and yeah what components make up the path for further analysis.

since you say you got a few of them there maybe a common failure mode
(if they are the same model or they use the same basic circuit design in that area)

So I guess its really up to someone who know the Intel boards well with lots of trouble shooting experience on them to answer that one.

Sorry no real help to you one that one

Cheers