Is 750W enough?

[Yellowbeard]

FWIW, here's Jeremy's review unit from the JonnyGURU review:http://www.jonnyguru.com/modules/NDR...0/DSC_3300.jpg


And, be sure to get accurate information when deciding what to buy if you are going to examine the quality of the components inside: http://www.corsair.com/blog/capacitor/

[watercooled]

Thanks for clearing that up guys.

Sorry for confusing things, I'm admittedly not as up-to-date with new power supplies, and I'm still (probably overly) concerned about cap brands after the problems we've seen in the past. It was only a few years ago where it was nearly unheard of for reputable PSUs to use anything but Japanese caps throughput, with the occasional one using a respectable Chinese-branded primary. I'll make certain to read through that capacitor article though.

I thought (assumed) the JG unit was newer as it's a recent review vs November for the HardOCP one. I never thought to check the pictures for the S/N.

Like I said though it seems strange to spend a lot on a premium primary cap (which makes up a fair percentage of the entire BOM AFAIK) but use cheaper secondaries?

I don't mean to be anti-Corsair in any way, or to misinform, it's just I have come to expect solid build quality, support, etc from Corsair, and for all the products to be a safe buy. So it's just disconcerting when I start looking at recent reviews and see what appears to be a deviation from that. If that makes sense.

[jonnyGURU]

Like I said though it seems strange to spend a lot on a premium primary cap (which makes up a fair percentage of the entire BOM AFAIK) but use cheaper secondaries?

Then do read the capacitor blog I referenced and Yellowbeard linked to. It explains why a Japanese cap is used for the primary in almost every Corsair PSU. In a nutshell: While secondary caps are hardly pushed at all anymore, technology on the primary side hasn't changed all too much and can often be asked to run out of spec (temperature and ripple current).

[Yellowbeard]

Thanks for clearing that up guys.

Sorry for confusing things, I'm admittedly not as up-to-date with new power supplies,.

LOL....please keep ^^^THIS^^^ in mind when dispensing advice then.

[kalniel]

LOL....please keep ^^^THIS^^^ in mind when dispensing advice then.

Steady on I'm sure he will - he's one of our more conscientious posters and we all make mistakes, you should see some of the other posters we get!

[Yellowbeard]

I've seen quite a few of his posts over the time I've been here. Watercooled is indeed a good contributor for Hexus

[watercooled]

Then do read the capacitor blog I referenced and Yellowbeard linked to. It explains why a Japanese cap is used for the primary in almost every Corsair PSU. In a nutshell: While secondary caps are hardly pushed at all anymore, technology on the primary side hasn't changed all too much and can often be asked to run out of spec (temperature and ripple current).

Just read through the article, and it does clear things up, but I still have a remaining question or two; these cheaper caps are *specified* the same as equivalent Japanese capacitors, but are they actually tested to meet those specifications? MFR specs can't always be trusted.

About the secondaries being asked to cope with less ripple current in newer designs, how does that work for, for instance, the CS series which has >50mV p-p output on 12v? Yeah I know we're talking about ripple before the LC filter but for a given 'input' ripple and capacitor ESR we should get the same output ripple. In other words, at least the final capacitors are being subjected to similar ripple current as in another PSU using the same caps and similar output ripple (since output ripple is essentially a product or input ripple and ESR).

Again, apologies for any confusion, I was only going off my experience and knowledge of existing PSUs. I didn't (or didn't mean to) make out the power supplies were *bad*, just I had reservations about them for the reasons I explained and couldn't comfortably recommend them based on that.

[jonnyGURU]

Just read through the article, and it does clear things up, but I still have a remaining question or two; these cheaper caps are *specified* the same as equivalent Japanese capacitors, but are they actually tested to meet those specifications? MFR specs can't always be trusted.

You're absolutely correct. And that's a reason why we still use Japanese caps on the primary. If we know we're going to be at spec, or sometimes beyond spec, we don't want to rely solely on a paper specification. But since the secondary capacitors are so far from working any where near spec, they have a calculated life of almost 15 years in nearly every application. That gives one a lot of wiggle room for not being dead on the specs. But a lot of QC relies on the factory building the PSU rather than the actual cap manufacturer. At the factory, between the warehouse and assembly lines, there is a QC department that tests random samples from each batch and make sure they are within a predetermined tolerance. If a cap fails, the batch fails and they are not used.

About the secondaries being asked to cope with less ripple current in newer designs, how does that work for, for instance, the CS series which has >50mV p-p output on 12v? Yeah I know we're talking about ripple before the LC filter but for a given 'input' ripple and capacitor ESR we should get the same output ripple. In other words, at least the final capacitors are being subjected to similar ripple current as in another PSU using the same caps and similar output ripple (since output ripple is essentially a product or input ripple and ESR).

Well, the ripple current is way within spec for the cap, and any other cap isn't going to filter it any better and nothing is going to filter all of the ripple. Different caps won't change this. More caps might, but the problem is that each cap adds resistance and then the unit wouldn't be 80 Plus Gold and more expensive. That's what happened to TX. In this day and age, a Bronze unit with better ripple suppression just isn't as big of a market as it used to be if you can do a Gold PSU with "decent" ripple suppression. And if you want better... that's why there are better PSUs.

[watercooled]

You're absolutely correct. And that's a reason why we still use Japanese caps on the primary. If we know we're going to be at spec, or sometimes beyond spec, we don't want to rely solely on a paper specification. But since the secondary capacitors are so far from working any where near spec, they have a calculated life of almost 15 years in nearly every application. That gives one a lot of wiggle room for not being dead on the specs. But a lot of QC relies on the factory building the PSU rather than the actual cap manufacturer. At the factory, between the warehouse and assembly lines, there is a QC department that tests random samples from each batch and make sure they are within a predetermined tolerance. If a cap fails, the batch fails and they are not used.

Gotcha.

Well, the ripple current is way within spec for the cap, and any other cap isn't going to filter it any better and nothing is going to filter all of the ripple. Different caps won't change this. More caps might, but the problem is that each cap adds resistance and then the unit wouldn't be 80 Plus Gold and more expensive. That's what happened to TX. In this day and age, a Bronze unit with better ripple suppression just isn't as big of a market as it used to be if you can do a Gold PSU with "decent" ripple suppression. And if you want better... that's why there are better PSUs.

I wasn't disputing the filtering performance of the caps, rather, wondering about suitability of a given cap given at least some of them appear to be subjected to similar amounts of ripple current. Or, why are cheaper caps now suitable in locations they previously weren't? Increased quality is one explanation of course.

It get it for PSUs like the RM series with low output ripple; either the filter is more effective, or they have less ripple to deal with, e.g. faster FETs, switching diodes among other things? However, for the likes of the CS series, at least the final smoothing cap appears to be dealing with a similar amount of ripple as any other PSU with the same ripple. But again, if improved quality is the reason, that makes sense.

Edit: Note: I'm more used to working with non-isolated DC-DC converters so I may be overlooking something in this type of SMPS.

[jonnyGURU]

Well.. not so much as you're seeing the ripple after the fact in the reviews. The ripple prior is still less than it was in the past. I admit that I don't have the means to measure ripple prior to filtering, but the engineers at the factory do.

It's funny that we used to judge a PSU by its weight. But designs have become so efficient that some "lighter" PSUs are actually better than some heavy ones. More and more, we have to rely on reviews to tell us if the product is worth a damn.

[watercooled]

Yeah that makes sense. I guess the 'reservoir' cap, before the inductor, is affected more by the starting ripple (which we don't see) than the final smoothing cap, assuming that's the sort of filter used.

[andrewhutt]

I think it is safe to say I know relatively nothing about psu after the last few posts but thanks for informing me.. steep learning curve building a pc from scratch!

[watercooled]

I think it is safe to say I know relatively nothing about psu after the last few posts but thanks for informing me.. steep learning curve building a pc from scratch!

Well I've learned a few things, and you've gotten some input from one of the most reputable PSU reviewers, and from the MFR of the PSUs.

LOL....please keep ^^^THIS^^^ in mind when dispensing advice then.

I do try to, but things change which I'm not always aware of. I'd generally rather recommend something I'm comfortable with than something I'm unsure about. I've no quarrel with Corsair, and I did continue double-checking after my post, hence how I stumbled on JG's review. TBH I'm probably far from the only one to perceive some of these things (e.g. cap selection) as negative given the inertia of the cap plague suspicion.

[Yellowbeard]

Well I've learned a few things, and you've gotten some input from one of the most reputable PSU reviewers, and from the MFR of the PSUs.

Very cool, I'm glad JG chimed in here. He's forgotten more about PSUs than I'll ever know.

I've no quarrel with Corsair, and I did continue double-checking after my post, hence how I stumbled on JG's review.

I did not take anything you said as a negative. It's all good by me.

[Enter-77]

I'm running an i7 3770k and 2 7970s of my 850w and it's having no problems. 750w Should be enough. But always best to buy up when it comes to psus

[andrewhutt]

I have spoken to OCuk and they also suggested using the 850W but the only problem is I am wanting to keep quite strictly to my budget and 850W units ate generally outside of it