Corsair CX550M – farewell group design

[mockingbird]

http://www.hardwareinsights.com/corsair-cx550m-farewell-group-design/

[powernod]

With this hold-up time being so short, not even a UPS (neither a cheap/off-line model, nor an older one with worn out relays) may help you under full load, as its own transfer time may be longer than this.

-You mean that the pc will shut-down before the UPS manages to activate?
-New "cinewave" & " online" UPSs would react properly? If yes, i think that you 've been a little harsh on this unit. (*After all, we shouldn't rely on previous decade's UPSs to protect our system right? This wouldn't be safe as well )
-Also, about protections: for such a low budget-unit, we shouldn't really expect to have all the protections working, should we? I was once told by Thermaltake that :

Single set of 12V design will result OCP protection will be very close to the OPP (Over power protection)
So the design is using the OPP protection!!

. So, even though the OPP isn't present, the OCP is working, right?

[Behemot]

On-line UPS by principle does not have this problem. Line-interactive depends; old ones where power goes through all-time magnetised core should help. Newer ones use different technique to save power so the transfer time may also depend on the speed of its relays.

Remember the ATX spec calls for 16 ms, I have arbitrarily set the absolute minimum to 10 ms, that is over 1/3 lower. So they piss on ATX spec, and they don't even deliver as little as 10 ms?! And the unit is really nowhere so cheap here in Europe. It's their decision to sell that for the shipping costs&few bucks in the US. But here the customers finance for their US prices with premium.

As for protections, I demand OPP from mainstream units. I would not say this one has that working.

All in all, the Cooler Master GM series has no such problems. It works OK, delivers long enough HUT, all the protections are working incl. OCP and OTP. And it costs the same. This time Corsair just did not cut it. That happens, nobody is perfect all the time. Good for competition

[LongRunner]

The results of the primary capacitor upgrades are quite interesting, and based on them, 1μF/W (or at least 0.8μF/W) would indeed seem worth paying for.

I must note, though, that most of the marketing claims for "audio grade" capacitors are at best contentious. Still, I would very much like to test a variety of capacitors (plastic film, C0G/NP0 ceramic, X5R and X7R ceramics, electrolytics of all kinds; general-purpose and low-ESR, aqueous and non-aqueous, "audio-grade" vs. "standard" types, hybrid polymer and even solid polymer although the latter are potentially undesirable due to high leakage currents) under realistic audio signal conditions (including both biased and unbiased operation) and compare their results (for noise, harmonic distortion, etc.), when I get the opportunity…

[Behemot]

That would certainly be interesting.

[wolverine9]

Hello Sir ,

Thanks for the awesome review !

I have a question , i hope you will answer

The hold up time of this PSU was measured at 100% load , now if the psu is run at 30 % load , the hold up time would rise a lot... >16ms , so ups would work alright at low loads

now if somebody pairs this PSU with a non-sinewave back UPS , the APFC will draw a lot of inrush current and run the PSU at close to full capacity
( as per apc http://www.apc.com/us/en/faqs/FA158939/ )


my question is "generally" what is the typical duration of inrush current in "ms" during which PSU is run at 100 % ? assume ups transfer time is 10ms

Basically what i am trying to ascertain is , if somebody buys this unit cx550m for only running at low loads , would it work alright with non-sinewave UPS or would the inrush current deplete all the saved up power by running psu at 100 % capacity before the UPS has the time to switch ?

So the low hold up time of 8ms of this PSU continue to remain a problem even if the PSU is run at low loads because of inrush current ?

This is more of a general question ..

Thanks

[Behemot]

Those are several indepedent issues. As I understand that, what the article is mostly about is that when the UPS switches to inverter, the connected eqiupment gets no power thus it discharges its internal capacity. After the inverter kicks in, the bulk capacity is charged. But a) this charging is with up to tens of A (from discharged level, with some remaining charge it is much less), see Aris'es measurement of inrush current at TPU. However, b) this happens no matter what (if any) PFC is there so I think APC somewhat did not manage that article.

Another information, correct this time, is that you should always calculate PSU input, that means consider it's own inefficiency to it's output and calculate the UPS with this complete value.

As for hold-up time, yes, generally with decreased load it increases, as I have actually even suggested in the conclusion:

Of course you can just buy a much higher-rated model, like the CX750M. But why do that? Why not just get a higher-end series with lower power for at a similar price point?

[wolverine9]

Thanks sir for replying