Is Seasonic SS-660KM PSU ErP Ready

[momaka]

Re: Is Seasonic SS-660KM PSU ErP Ready

Anyway, these little FR guys look... well, plain. There are no other markings except the ones for the rating, temperature, series, and something that might look like a date code which gives little to no clue about the meaning behind it. For example, it says 681OL, or 78IIIGH.
I also checked them with the ESR meter. Capacitance is superb, while the impedance is lower than the one in the official specs. For example, a 50V47uF reading showed that its ESR is 0.09ohms, whereas the oficial specs say it is around 0.14. All the caps' readings are close to one another. The small ones have a flat top, while the larger ones have a T-shape indent.

Now, when I get some free time, I will take the PSU apart. My questions are:

1) Judging by the description, do you find these FR capacitors to be genuine (if necessary, I will post some pictures)?

Yup, they sound genuine to me. Panasonic FR caps do look quite boring indeed. But who cares, as long as they work as they should and last a long time.

2) Should I measure the old caps in-circuit? If they test OK, should I replace them with the new ones anyway (which brings us to question number 3)?

Well, if you have the replacement caps already, you might as well pull the "old" ones from the PSU and replace them.

But for the ones that you don't have replacements... sure, it wouldn't hurt to pull a few and measure them, if it isn't too much of an inconvenience (I'd personally only pull those caps that are near the hottest components and/or major/important filter caps).

And no, don't measure in circuit. In-circuit measurements are generally a bad idea, especially with those cheap component testers (which are bound to measure more than one component at the same time and then give bizarre results.

3) By comparing the spec tables, these FR capacitors have lower impedance than the original Rubycon YXGs. What are your thoughts, will the difference in the ESR cause this Seasonic PSU to misbehave?

Very unlikely.
Like I might have mentioned previously here (I don't remember anymore, it's been a while since I last posted on BCN), the ESR of a capacitor can vary quite a bit with temperature - sometimes as much as a power of 10. So most circuits are usually designed around the rather loose ESR tolerances of electrolytic caps.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

OK, regarding my question number 2, I meant the other caps in the PSU (larger ones), since I didn't measure those last time.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

Hey, guys, I got my pack of FR capacitors. Unfortunately, Farnell did not send the 50V22uF ones because they ran out them in the mean time. My PSU only needs one of those. Maybe I will check again to see if a back order is available. I would like to replace all the caps below 470uF as per advice and all at once, in order not to take the PSU apart multiple times.

Anyway, these little FR guys look... well, plain. There are no other markings except the ones for the rating, temperature, series, and something that might look like a date code which gives little to no clue about the meaning behind it. For example, it says 681OL, or 78IIIGH.
I also checked them with the ESR meter. Capacitance is superb, while the impedance is lower than the one in the official specs. For example, a 50V47uF reading showed that its ESR is 0.09ohms, whereas the oficial specs say it is around 0.14. All the caps' readings are close to one another. The small ones have a flat top, while the larger ones have a T-shape indent.

Now, when I get some free time, I will take the PSU apart. My questions are:

1) Judging by the description, do you find these FR capacitors to be genuine (if necessary, I will post some pictures)? I was amazed by the very professional packaging and the accompanying paperwork included.
2) Should I measure the old caps in-circuit? If they test OK, should I replace them with the new ones anyway (which brings us to question number 3)?
3) By comparing the spec tables, these FR capacitors have lower impedance than the original Rubycon YXGs. What are your thoughts, will the difference in the ESR cause this Seasonic PSU to misbehave?

Thanks in advance for your support and guidance.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

I bought an ESR meter with MK328 chipset similar to what was suggested by someone here in this topic. And yes, I calibrated it. I tried several used but good, new, and bad caps just to test it (I think I wrote about it here).

[Stefan Payne]

Re: Is Seasonic SS-660KM PSU ErP Ready

I have seen so little bad low-ESR 105°C caps from reputable manufacturers I'd count that on a single hand, so far. (Ultra-low ESR caps on motherboards are different matter.) It makes little sense.

It could make sense if you measure the temperature near those caps and/or try to measure ripple on those caps.
Its entirely possible that they might fail prematurely due to a design flaw.
Maybe a manufacturing error...

I'd still suspect the +12V FETs to cook the caps around it. And the +5VSB is pretty close to the +12V FETs.

On Super Flower Leadex the +5VSB is on the other side of the PCB as is the case with the newer FOCUS plattform...

[Behemot]

Re: Is Seasonic SS-660KM PSU ErP Ready

Customised values. Guesstimating by interpolation, it should be about 0.5 ohm for 47uF and 1.5 ohm for 10uF.

I have some doubts about the accuracy of that measurement though, don't you use some of those chinese garage meters? Is it calibrated? I have seen so little bad low-ESR 105°C caps from reputable manufacturers I'd count that on a single hand, so far. (Ultra-low ESR caps on motherboards are different matter.) It makes little sense.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

Well, I was not able to find all the reference values on Rubycon's website about its YXG series. All the Rubycon caps in my PSU are YXG series rated at 50V, yet the YXG data sheet only has ESR values for 22uF, no 10uF or 47uF. And according to the value for 22uF, I see that it should be around 0.70 ohms, way bellow the values I measured. Is there another source of these data sheets where I can see the 50V10uF and 50V47uF impedance values? How come no values are listed for these Rubycon caps in the official data sheet?

[momaka]

Re: Is Seasonic SS-660KM PSU ErP Ready

... but several of the Rubycon YXG 50V 47uF were measured at 3.4 ohms ESR by the ESR meter. I know that these are small caps, but that looks a bit high.
...
Strangely, I couldn't find all of the YXG cap ratings in the Rubycon's YXG sheet. Only 50V 22uF is listed, no 10uF nor 47uF. But for the 22uF it says that ESR is ~0.70, and my ESR meter measured the one from the PSU to be at around 1 ohm.

Just measure the case size on your caps. If it's 5x11 mm, take any 5x11 mm cap in the datasheet with the same voltage and the ESR should be the same or close to that value. Also remember the 20% tolerance and that temperature affects caps. So if you measured the cap while it was at any less than 25C room temperature, the ESR could be higher. Add the 20% tolerance on to, and you just may have a cap that is marginal, but still not bad.

In any case, I guess we will know once you have those caps replaced, now that you have ordered new ones. 3.x Ohms is a bit high, so maybe that was the problem, after all.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

Hello, guys.
It's been awhile, but there is development.

I finally got some time to deal with this PSU a bit. Right now, my PC is running on the SS-400ET PSU.

I removed the screws and then the bottom panel of the SS-660KM PSU case as per suggestion, and then I removed all the capacitors under 220uF. That particular one (KY) was fine, but several of the Rubycon YXG 50V 47uF were measured at 3.4 ohms ESR by the ESR meter. I know that these are small caps, but that looks a bit high.

I have ordered the Panasonic FR equivalents for all the ratings of the YXG and KY from Farnell and they should arrive in a week or two. Then I will replace all of them as soon as I grab some time and hope for the best. Strangely, I couldn't find all of the YXG cap ratings in the Rubycon's YXG sheet. Only 50V 22uF is listed, no 10uF nor 47uF. But for the 22uF it says that ESR is ~0.70, and my ESR meter measured the one from the PSU to be at around 1 ohm. The FR equivalent of that one is rated at twice as low the ESR (0.34).

[Wester547]

Re: Is Seasonic SS-660KM PSU ErP Ready

Now, I also tested several 6.3V 3300uF and 16V1500uF Rubycon MBZ capacitors. I looked into their datasheet and saw that their ESR for a given size and value should be around 12mΩ, or 0.012Ω. This ESR measured them at 0.5-0.6Ω, which was 5 times the normal ESR value. These are dated 2006 and have never been used, they are still attached to their paper strip. Their capacitance was also slightly higher than the rated one. I guess I can discard those as well.

Those are probably counterfeits if you sourced them from eBay (egecku?). That, or they may need to be reformed. But as Behemot said, the longer leads mean higher ESR.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

Is it even calibrated? The leads usually have up to 0.5 ohm depending on their quality (but even the best ones are 0.1-0.2 which is still zones times higher than of many low-ESR caps).

Yes, it has a built in calibration test and I did that prior to any measurement. Thn, I compared the capacitance reading between the ESR meter and the AN8008 multimeter of one and the same capacitor and the values were almost identical (+/- very little).

[Behemot]

Re: Is Seasonic SS-660KM PSU ErP Ready

Is it even calibrated? The leads usually have up to 0.5 ohm depending on their quality (but even the best ones are 0.1-0.2 which is still zones times higher than of many low-ESR caps).

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

Thanks, momaka. I have already found the datasheet PDF. You are correct, the bad Nichicons were indeed HM series from 2004.
Now, I also tested several 6.3V 3300uF and 16V1500uF Rubycon MBZ capacitors. I looked into their datasheet and saw that their ESR for a given size and value should be around 12mΩ, or 0.012Ω. This ESR measured them at 0.5-0.6Ω, which was 5 times the normal ESR value. These are dated 2006 and have never been used, they are still attached to their paper strip. Their capacitance was also slightly higher than the rated one. I guess I can discard those as well.
I did, however, find many NCC 6.3V 1000uF KZG caps pulled out from some old motherboards to be totally within their specs when it comes to their ESR and capacitance.

[momaka]

Re: Is Seasonic SS-660KM PSU ErP Ready

Now, I may try and remove the small capacitors from the PSU to test them. They are all 50V, only capacitance is 10, 22 and 47uF respectively. I have no idea what is the normal ESR value of these caps. Given that their capacitance is not big, they will probably have higher ESR. This ESR meter displays the following values: capacitance, Vloss and ESR. Do you have some ideas about the abovementioned normal readings of these capacitors, what should I expect? They are all Rubycon YXG series.

Get the Rubycon YXG data sheet and see what the impedance (ESR) is supposed to be for each capacitor. Use the capacitor physical size (diameter and height) when looking up your capacitors - this is because within a capacitor series, all of the caps with the same size will have the same ESR/impedance and ripple current. As for the capacitance - just make sure it's within 20% of the stated value. For ESR, within 50% on the high side is probably still reasonable, given the accuracy of your meter and the temperature of your testing environment.

I got an ESR meter at hand, a MK-328 model. I was playing with some caps pulled out from dead motherboards. I found two Nichicons rated at 6.3V3300uF. The multimeter measured their capacitance at over 4000uF, and so did this ESR meter, but it showed that the ESR was around 0.5 ohms. The capacitance looks very off, but I have no reference as to what should its normal ESR be.

Those sound like bad Nichicon HM, HN, or HZ capacitors from the 2001-2005 era (H01xx through H05xx date codes on each capacitor). If yes, they are BAD. Abnormally high capacitance indicates high leakage current within the cap. This fools the meter into showing that the ESR is low when it actually probably isn't. If you get a cap like that, it's junk/gone.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

No, I didn't try replacing those caps. I got an ESR meter at hand, a MK-328 model. I was playing with some caps pulled out from dead motherboards. I found two Nichicons rated at 6.3V3300uF. The multimeter measured their capacitance at over 4000uF, and so did this ESR meter, but it showed that the ESR was around 0.5 ohms. The capacitance looks very off, but I have no reference as to what should its normal ESR be. I would like to acquire some experience using and reading these caps before I move the actual disassembly process. So far, the PSU works.

Now, I may try and remove the small capacitors from the PSU to test them. They are all 50V, only capacitance is 10, 22 and 47uF respectively. I have no idea what is the normal ESR value of these caps. Given that their capacitance is not big, they will probably have higher ESR. This ESR meter displays the following values: capacitance, Vloss and ESR. Do you have some ideas about the abovementioned normal readings of these capacitors, what should I expect? They are all Rubycon YXG series. I should probably gather all the information prior to disassembling the PSU as I will not be able to use the computer during that time for obvious reasons.

[Stefan Payne]

Re: Is Seasonic SS-660KM PSU ErP Ready

Spending your time trying to solve this issue is probably worth it, itsn't it?

Have you tried replacing the (small) Caps in +5VSB??
Or did you just not bother because of "Japanese!!!111" caps??

[Behemot]

Re: Is Seasonic SS-660KM PSU ErP Ready

All versions based upon the original or modified X series suffer from this (Gold and Platinum KM-KM3 variants). Maybe the Primes do not as they likely have totally different platform which addressed this, but who knows. They have so long warranty it does not matter as long as you keep a proof of purchase though.

[UserXP]

Re: Is Seasonic SS-660KM PSU ErP Ready

Thank you all for your replies, guys. I am still scaling my choices. I am at a loss. The PSU has been working fine for weeks now, it hasn't refused to start since the last time.

I also don't think that such quality caps as in this unit could have failed, so I still haven't bought the replacement caps. The fan should, in my opinion, be on as the PC case is in a closed space and near a heating radiator pipeline, which is why I got interested in modding the PSU to run the fan always at at least low speed (and thought that that extra heat might have somehow damaged a component inside the PSU). I mean, the surroundings are not warm or hot, but a bit more air circulation couldn't hurt.

@Behemot (and possibly others)

I do admit that this topic has gone a bit off topic. Apart from not having enough time to try any of the suggestions so far, I still am a bit reluctant to accept that this PSU is going to dye. It is out of warranty and I can not RMA it, only service it (or replace it with a new one). You've mentioned that X series PSUs suffer from the same issue. Is there a modular SeaSonic around 600W representative that does not exibit similar issues after some time of use?
I really do like how SeaSonic builds their PSUs, and all the protections it offers, all of which did step in once or twice so far and protected the computer from surges.

[Stefan Payne]

Re: Is Seasonic SS-660KM PSU ErP Ready

I have not taken one of these PSUs apart, so I don't know what type of fan control method it uses.

I have.
They use a dedicated IC that is made for Semi Fanless Operation. The Cooler Master Master Watt 550W has the same chip...

Look at Techpowerup for a similar there you can see one of those, don't know who made it. I think it was Anpec...

[momaka]

Re: Is Seasonic SS-660KM PSU ErP Ready

I still have one question, though. Does anybody know how this PSU's fan controller kicks in? Does it engage the fan when the PSU reaches certain load, or does it kick in when it detects the inner temperature? If latter is the case, maybe we could hook up something to that sensor, e.g. to attach its probe to one of the radiators. That way, the warmth of the radiator would keep the circuit on, and I would be spared from additional soldering and disassembling. Is something like this feasible?

I have not taken one of these PSUs apart, so I don't know what type of fan control method it uses. Older and/or more simple PSUs typically just have a thermistor on the secondary heatsink and regulate the voltage to the fan based on the temperature of the secondary heatsink. But with this being a semi-fanless unit, it probably uses an IC to control the fan speed and when it turns On. As to whether it's load or temperature dependent - I don't know. You would have to trace the circuit of the fan controller and maybe even draw a diagram for us (if there isn't one on the internet anywhere - which I doubt there is).

That's why I commonly suggest the 2-diode method - it's very simple and anyone can build it. The only controller modding that is easier than that is when you have the simple temperature-regulated fan controller with a thermistor, like I mentioned above - those are easy to tweak by just installing a resistor in parallel or series with the thermistor (depending if it's an NTC or PTC thermistor).