Offline, outdoor LED driver datasheet claims unrealistically high efficiency?

[ocset]

Hello,
We dispute Efficiency claims of 89% in the datasheet of this 40W Phillips Xitanium outdoor offline LED driver….. 

Xi LP 75W 0.2-0.7A SN 230V S240 sXt

Topology:
However, when we opened it up, we found it was a L6562 controlled Boost PFC stage, followed by an L6562 controlled isolated Flyback stage. There is also a Viper controlled little Buck converter to provide bias power. The Boost PFC output capacitor is a 22uF , 450V electrolytic.

This surely could not be 89% efficient?
If it was that efficient, then that would mean the Boost PFC would be about 96% efficient, and the Flyback would be about 93% efficient. That surely could not be possible? In my experience, I find 90% tops  is about the  efficiency of low power offline flybacks, even with BCM operation

The L6562 controller operates in Boundary Conduction Mode, so turn on switching losses are about zero, but turn-off switching losses are high.

We won’t yet have time to test it for ourselves as we are way too busy at present.

Electrolytic capacitor
Also, why does the datasheet claim that its  has “ultimate robustness” and ” long lifetime” when it has a big internal electrolytic capacitor.?

High switching frequency at lighter loading
Also, its necessary for  offline LED drivers to maintain high power factor right down to 25W. Now,  if operating in BCM, then that means that as the load gets lighter, the switching frequency of a BCM controller will increase…..As such, switching losses will increase. If the switching frequency needs to go above 150kHz to maintain BCM operation, then that has severe repurcussions in terms of  conducted EMC.
The L6562 datasheet makes no mention of a maximum frequency limitation, or a minimum off time.
Maybe they are operating the L6562 as a constant  off time controller(?)   ..which is possible with addition of a few extra components….however, the L6562 doesn’t possess the necessary internal circuits to allow it to maintain high enough power factor when  operated in constant  off time.

Datasheet:
Xi LP 75W 0.2-0.7A SN 230V S240 sXt
Datasheet available here…..
http://www.lighting.philips.co.uk/oem-emea/support/technical-downloads

L6562 Datasheet:
http://www.st.com/en/power-management/l6562.html

Do you agree this LED driver couldn’t be that efficient? 
 

[tooki]

Why should anyone here waste time speculating? Test the damned thing, or look to see if anyone already has tested it...

[IanB]

Do you agree this LED driver couldn’t be that efficient? 
 

I don't know, but I don't think you used enough of these things:           

[BravoV]

Do you agree this LED driver couldn’t be that efficient? 
 

Nope, its more efficient than that, based on my "gut feeling" as you just did.

[capt bullshot]

Why not?
Every average laptop brick achieves this efficiency today, using this architecture (boost PFC and quasi resonant Flyback)

[Gyro]

Electrolytic capacitor
Also, why does the datasheet claim that its  has “ultimate robustness” and ” long lifetime” when it has a big internal electrolytic capacitor.?

A properly specified, good quality, large (low stressed) electrolytic capacitor isn't incompatible with robustness or long life. It probably soaks up a lot of the transients that make your products so unreliable too!

[Siwastaja]

Perfectly plausible numbers. Not every power supply with the same topology can do the same, but some do. I think that this particular topology, in this voltage configuration, has practical limit in about 90-92% range; over that, you need topology improvements.

If you want to dispute the claim, that's extremely straightforward! By actually testing it, you can simply prove/disprove the efficiency claim in much shorter time than trying to analyze the design and guesstimating how it "could" perform, compared to your own designs, limited by your design ability.

Also, no fundamental issue in using an electrolytic capacitor in a robust, long-life product. I would recommend designing in electrolytics for you as well; they are especially good for the "I don't know what I'm doing so I'm throwing in random parts" designers. What I mean, designing without elcaps requires more understanding, as just throwing in a random elcap often simply solves many stability, ringing and surge problems you seem to have.

[dmills]

Efficiency numbers are generally given at full load, as are the power factor values (Also, generally rated at a specified input voltage).
 
I see nothing incompatible here, any modern laptop power supply posts better numbers then that, and they are basically the same topology.

Nothing wrong with electrolytic caps, I have 30 year old T&M gear still on its first set. Understand them, read the datasheets and do the maths, life times are easy to calculate and are well qualified at this point. Of course if you cheap out or don't respect the ripple current and ambient temperate derating curves you get what you deserve.
Remember also that a 105 degree part is NOT automatically better then an 85 degree one, and that polymer types have different failure modes to conventional aluminium wet types.

Sorry, I think I believe the Phillips numbers, at least under full load.

Regards, Dan.

[kony]

Treez, are you really this desperate?

[tooki]

Treez, are you really this desperate?

Perpetually, it seems…