faulty power supply?

[999999999]

Re: faulty power supply?

After I looked everywhere around here in Sydney for a bunch of 1000uf 10V caps that were 8mm, I came up with absolutely no luck, only size was 10mm diameter.
I tried my last hope, the internet and found a store that sold 1000uf 10V caps in an 8mm barrel.
The caps that they are selling are Rubycon MBZ and MCZ series caps. Are Rubycon MBZ and MCZ suited for PSU's?? I know they are low ESR but that's about all I know about them.
I've only seen Rubycon caps used in a PSU once and that was a Compaq PSU. I'm only afraid that the Rubycon caps may not like heat and get damaged by heat.


Thanks.

Do not use a typical 1000uF, 10V, 8mm caps even if you find them. They'll either be higher ESR or too poorly constructed to last long term (part of why the originals probably failed, off-brand caps with unusually high mF per their voltage and physical size are among the most failure prone).

Decide what brand and model you want to use based on what your preferred vendor sells, and if a 10mm cap will fit, use one. If only 8 mm will fit, choose a 6V cap with the highest uF value (taller size) available. Since it is only a 10V part it can't be on the 12V or -12V rails so with any other voltage under 6V, a 6V (actually I mean 6.3V) cap should do fine.

Yes Rubycon MBZ and MCZ are suited for PSU, but are actually a bit of a luxury in a PSU as even good PSU brands often opt for lower cost caps instead (not junk, just not the extra cost for an ultra-low ESR 'lytic) which have stood the test of time.

If high heat is a factor, any lytic isn't going to do well long term. In that case seek to increase airflow, either increase exhaust rate out of the PSU or case intake on the case it'll be installed in.

[linuxguru]

Re: faulty power supply?

Rubycon MBZ and MCZ have excellent ripple-current handling capability, but don't necessarily handle high heat very well. I've seen a few that were slightly bulged in VRMs, usually when they're right next to the heatsink of a Willamette or Prescott.

For a PSU, I usually prefer non-aqueous (i.e. non-water-based electrolyte) electrolytics like Nichicon PL/PM/PW, NCC LXJ/LXF/LXZ, Panasonic FC/FK, etc., even though their ESR is slightly higher than the ultra-low ESR aqueous MBZ/MCZ. IMHO, they are likely to withstand the test of time better.

If you have no choice but to use aqueous electrolytics in a PSU (either due to space restrictions or availability), try to use them only on the output legs of the pi-filter(s) or in other locations which are not in close proximity to the secondary heat-sink. The combination of high ripple current and high temperature is a definite killer of aqueous electrolytics.

[999999999]

Re: faulty power supply?

1st, to establish that MBZ or MCZ were more intolerant to heat, we'd have to have the same board and environment with different caps installed to see if they fared better.

2nd, NIchicon PL/PM/PW at least are water based electrolyte, as are Panasonic FC. This is unquestionably true. Tear one open if you don't believe it but I ask why you would believe otherwise?

[linuxguru]

Re: faulty power supply?

NIchicon PL/PM/PW at least are water based electrolyte, as are Panasonic FC.

AFAIK, all of the Nichicon Px-series electrolytics are non-aqueous. Only the Hx-series are aqueous - and the only bulged Nichicons found in the field thus far are HM/HN/HD etc. Likewise, only the Panasonic FJ/FL/FM are aqueous.

All Rubycons are aqueous, but the Z-series (ZL/ZLH/ZT) have relatively less water, making them more suitable for long-life/high temperature applications, including PSUs.

[999999999]

Re: faulty power supply?

AFAIK, all of the Nichicon Px-series electrolytics are non-aqueous. Only the Hx-series are aqueous - and the only bulged Nichicons found in the field thus far are HM/HN/HD etc. Likewise, only the Panasonic FJ/FL/FM are aqueous.

All Rubycons are aqueous, but the Z-series (ZL/ZLH/ZT) have relatively less water, making them more suitable for long-life/high temperature applications, including PSUs.

It's very strange to me that you're trying to classify them differently than aqueous capacitors. What exactly are you suggesting, that they have a different liquid base to the electrolytic solution or that they are solid or dry? They are definitely not solid or dry. I'll have to root around more in my spare parts bins to see if I have any Panasonic FC but I just tore open a Nichicon PW and it looks like any other aqueous electrolytic capacitor inside.

If you're proposing the liquid isn't water, what shall I do to convince you it is? Perhaps a better question to ask is, where are you getting the information that these aren't aqueous? There's no mention of this in their respective datasheets or product family literature that I've seen.

[gonzo0815]

Re: faulty power supply?

It is based on the facts, that e.g. Panasonic FC are on the market long before any of water base low esr elektrolytics.
Your just confused by the meaning liquid=water.
Usual caps are filed with electrolyte based on different chemical solvents and solutions, which are liquid but does not contain any h2o.
On example of such a solvent is e.g. Gamabyturolacton to name one.

For some time the big manufacturers have found a way to inhibit the corrosion of the aluminum by water.
Then they were able to develop caps with much lower esr by adding it to the electrolyte.
So, i think if you compare the data sheets, the size /uF rating, the esr and endurance it is safe to conduct which cap series are based on a water based electrolyte formula and which are not.
This topic was broadly discussed here at BC.net forums.

[999999999]

Re: faulty power supply?

I'm not confusing liquid to mean it must be water based, I'm asking for some evidence instead of an assumption that it had to be nonaqueous based on the specs. How about the Nichicons, since they aren't particularly low ESR for their size?

Can you give a link to the broad discussion please or at least keywords to look for? Aqueous electrolyte didn't turn up any evidence, only people repeating what appeared possibly an unsupported rumour (which, even if repeated enough times doesn't really prove it) unless there is some real evidence.

[linuxguru]

Re: faulty power supply?

Check this link out: http://www.faradnet.com/deeley/chapt_10.htm

Note that this was the state of the art 20 years ago - a lot of improvements have occurred in the specific formulations used, although the principles remain the same. The idea is to use compounds that have very little liquid water, but can still ionize to give a hydroxide ion and are also weakly conductive. In some formulations, it is possible to reduce the amount of liquid water to almost zero, while still retaining electrolytic characteristics.

Various non-aqueous electrolyte formulations are given, usually based on a weak acids like Boric acid, Acetic acid, Lactic acid, Tartaric acid, Malic acid, etc. or hydroxy-alklyamines, like Monoethanolamine, H2N-CH2CH2OH or Diethanolamine HN(CH2CH2OH)2 or Triethanolamine N(CH2CH2OH), etc., or alkyl glycols like gylcerol, ethylene glycol, etc., along with ammonium salts of the weak organic acids., or combinations of the above.

[999999999]

Re: faulty power supply?

Linuxguru, while I do appreciate the informative link, this is not evidence that specific caps have a specific formulation, which was the crux of the issue. I don't claim there are no non-aqueous forumulations but rather, specific information about specific caps seems to be lacking.

[gonzo0815]

Re: faulty power supply?

Well, there are no known hard facts, and i think the only way to get this issue cleared would be to ask the manufacturers.

The discussions here about this is distributed over a lot of topics and sure it is all based on the understanding and knowledge of a few people here.
You may do a search with "water content" or "aqueous" to get the picture.

For me, this view, based on discussion and my own understanding and facts about caps is conclusive enough.
Despite that, i appreciate any hard facts and new considerations.
But sure, i think a dissection without any detailed analysis would not proof anything there.

PS: one thing i want to point out, usual electrolytic caps does not contain significant amounts of water.
The only purpose of water in the solution is to lower resistance and allow for smaller caps with very low esr and high current handling capability.

[PeteS in CA]

Re: faulty power supply?

1st, to establish that MBZ or MCZ were more intolerant to heat, we'd have to have the same board and environment with different caps installed to see if they fared better.

2nd, NIchicon PL/PM/PW at least are water based electrolyte, as are Panasonic FC. This is unquestionably true. Tear one open if you don't believe it but I ask why you would believe otherwise?

In reverse order ...

I've taken apart scores, maybe several hundred lytics. One of the tests I do is for the presence of water, and Nichicon's Px series, Panasonic's FC series and UCC's LXx series caps don't have water. The liquid is a solvent, probably gamma butyro lactone (I won't swear by the spelling); ethylene glycol is another commonly used solvent.

Another way you can test the heat tolerance of aqueous and non-aqueous caps is to put a bunch of them, connected electrically to nothing, in a temp chamber and heat them up. I've done this several times and aqueous caps start swelling at 25C or 35C lower in temp than do non-aqueous caps. Being connected to nothing, only the ambient temp caused the swelling, and I brought the temp in the chamber up slowly to get decent resolution and avoid temperature gradients within the chamber.

An interesting experiment might be to put an MCZ, an FM and some aqueous Lelons or FUs (or your favorite crap-cap brand) in an oven with an oven thermometer and turn up the heat slowly (start at 200F). My guess is that that crap-caps will vent before the good brands.

[999999999]

Re: faulty power supply?

In reverse order ...

I've taken apart scores, maybe several hundred lytics. One of the tests I do is for the presence of water, and Nichicon's Px series, Panasonic's FC series and UCC's LXx series caps don't have water. The liquid is a solvent, probably gamma butyro lactone (I won't swear by the spelling); ethylene glycol is another commonly used solvent.

Another way you can test the heat tolerance of aqueous and non-aqueous caps is to put a bunch of them, connected electrically to nothing, in a temp chamber and heat them up. I've done this several times and aqueous caps start swelling at 25C or 35C lower in temp than do non-aqueous caps. Being connected to nothing, only the ambient temp caused the swelling, and I brought the temp in the chamber up slowly to get decent resolution and avoid temperature gradients within the chamber.

An interesting experiment might be to put an MCZ, an FM and some aqueous Lelons or FUs (or your favorite crap-cap brand) in an oven with an oven thermometer and turn up the heat slowly (start at 200F). My guess is that that crap-caps will vent before the good brands.

Thank you for adding this information. If you found no water I concede I am wrong, but might I ask what tests were done to determine this as I find this test more significant, as we can't necessarily assume all electrolytes will have (significantly enough) lesser expansion than water can we?

[stevo1210]

Re: faulty power supply?

Do not use a typical 1000uF, 10V, 8mm caps even if you find them. They'll either be higher ESR or too poorly constructed to last long term (part of why the originals probably failed, off-brand caps with unusually high mF per their voltage and physical size are among the most failure prone).

Decide what brand and model you want to use based on what your preferred vendor sells, and if a 10mm cap will fit, use one. If only 8 mm will fit, choose a 6V cap with the highest uF value (taller size) available. Since it is only a 10V part it can't be on the 12V or -12V rails so with any other voltage under 6V, a 6V (actually I mean 6.3V) cap should do fine.

Well.... I found some Hitano EXR series 1000uf 10V caps in my toolbox and they are 10mm in diameter, but they fit into the 8mm spaces so that's one part done. The last set of caps in there were Saturn brand and were for the +5V SB and had vented.... I think that's what killed the motherboard that this PSU was powering....
There are two 2200uf 10V Su'scon caps next to each other in the back section of the PSU. Should they be replaced too?? or can I leave them alone?? I can get 2200uf 10V 10mm Samxon GC? caps from my local electronics store.

[Newbie2]

Re: faulty power supply?

There are two 2200uf 10V Su'scon caps next to each other in the back section of the PSU. Should they be replaced too?? or can I leave them alone?? I can get 2200uf 10V 10mm Samxon GC? caps from my local electronics store.

Yes, they should be replaced. If you don't, they could be ticking time bombs, waiting to fail. Fully recap the unit. You can ignore those tiny electrolytics (2.2uf, 4.7uf, etc.).

[yanz]

Re: faulty power supply?

yep. i got L&C with bloating su'scon here.. beware of su'scon caps.. psu is exactly like this..

Btw i use MBZ (1500uF 10v) for many L&C PSUs, the oldest one has been running fine for almost 2 years.. I think it's safe to use MBZ on PSU recapping assumming that you also replace the fan with good reliable brand (panaflo, nmb, adda) and at least with medium cfm type (not low cfm type).

[stevo1210]

Re: faulty power supply?

I have a small problem now. I replaced a Canicon 2200uf 16V cap with a UCC LXJ 2200uf 16V cap which is 12.5mm in diameter and its getting in the way of wires and seems to be covering a resistor next to it.
I can't find 2200uf 16V replacements in a 10mm size, the closest I could find was 1800uf 16V Rubycon MCZ caps. Will a Rubycon MCZ 1800uf 16V cap work?? I believe it has an ultra low ESR rating which is more needed in a PSU??

Thanks.

[stevo1210]

Re: faulty power supply?

5) Populate the missing RFI/EMI compliance components if you have the patience.

What do the RFI/EMI components look like?? and where do they go?? I've never taught about what it does and what it looks like.... I'm guessing it stops interference noise or filters out ripple current??

Thanks.

[gonzo0815]

Re: faulty power supply?

All SMPS are using a square wave, as it is far easier to produce compared with a real sinusoidal wave.
The problem with square waves is, that they are producing a huge amount of harmonic, e.g they consist off some parts of far higher frequency waves.

The EMI / RFI filters are needed to prevent leakage of this and the normal switching noise into the line, not disturbing other devices.

Usually this is accomplished by a few capacitors close to the input caps, followed by a common mode choke ending in a kind of phi filter (consisting of a differential mode coil and a few caps) at the entrance of the power plug.
This filters have to be on all tree wires, which needs some of the devices requiring a safety class allowance like X1 or x2 caps between neutral and phase and y1 or y2 between safety ground and phase.

All this parts are not cheap, but required by law and regulations in most countries.
Thus, the cheaper units may not have those places populated to cut cost significantly.
Most X caps are either large rectangular plastic boxes, or there are some smaller, round ceramic types which should be marked by a standard authority like UL, VDE or similar.

[davmax]

Re: faulty power supply?

Stevo. I can send you a Samxon 2200u 16V 10mm in an envelope if you send address by PM. Any larger quantities can be charged.

davmax

[stevo1210]

Re: faulty power supply?

Stevo. I can send you a Samxon 2200u 16V 10mm in an envelope if you send address by PM. Any larger quantities can be charged.

davmax

Thanks Davmax, but today after searching around my area I found a store here selling 2200uf 16V caps in a 10mm size .
Rubycon ZL last time I checked. I think they should be suitable.