HEC Orion HP585D

[370forlife]

A freebee. Usually $30 on newegg, lets crack her open and see what she's made of...



Oh dear. Not that good. Those are the tiniest input caps I have ever seen! They are only 18mm diameter! No way they are 680uf. At least that is a decent actual 35 sized transformer. I can't remember the primary's part number, but it is in single forward rated at 600v, 10A continuous, 36A pulse (mode used here,) with a copper plate to help with heat. Fans are by Young Lin Tech, so not totally skimping here.



Chart shows no combined power, just the max for each rail. Hmm....



Hello, anybody home? And what the hell is that, is that supposed to be a bridge rectifier. I've only seen ones like these used in audio amps. No provision for heatsinking, and is rated at 4A, 600V. 480W is not near 585W, and this is rated at 75% efficiency.



Teapovision. All the secondaries consist of teapos, except for the small ones for the -12v and -5v, which are su'scon. Super group regulation. All three rails on one coil. Ugh.



Secondary heatsink components. One on the right is a 30A, 45V part on the 3.3v (rated at 36A). Middle is a 40A, 45V part for the 5v rail (rated at 36A). On the other side of that, not visible due to the transformer, is a lone 20A, 40V part for the 12v ("2" rails, even though all 12v come from same holes, with one at 19A, one at 20A, no combined rating).

[370forlife]

There, now a full input filter, 6A 600v bridge rectifier, and a full input filter with 2 panasonic .33uf X caps, 1 .1uf teapo X cap, 2 2200pf Y caps, and 2 coils. Also, those new Nichicon 680uf, 200v caps. Yay.



New secondary silicon. 3.3v now utilizes a 60A, 45V part. 5v uses a 70A, 30V part, and the 12v now has a 60A, 45V part. Much yays were made.

I elected not to recap it for 2 reasons. 1. teapos aren't that bad in psu's, and 2. They are all 10mm and 8.5mm sizes, which I don't have any 8.5mm caps near their ratings, and no 16v 10mm caps.

[Toasty]

A 12 minute makeover!!!!!!!!!!!!

HOLY CRAP!!!

[ratdude747]

that is one POS PSU!

free, but shitty as hell. glad you could fix it.

[370forlife]

Really not that bad of a unit. No 2 transistor 5vsb, medium grade caps from the start, wouldn't be so bad if it wasn't for two things. One being advertised as Intel 2.01, while the PCB is dated for 08/07/2008 and marked as Intel 1.2, and the other is the super group regulation, as well as the transient filter.

Single forward primary switcher is a more efficient design, but not really viable for higher wattage units, although the primary switcher in this one is capable of about 450W. Secondary heatsink is thicker than most delta units I have used.

[kc8adu]

it will still go BANG! if loaded to the label ratings.probably ok for most home pc use though.
the mfr counts on the fact that very few pc's will need that 585w!

[370forlife]

Originally Posted by kc8adu

it will still go BANG! if loaded to the label ratings.probably ok for most home pc use though.
the mfr counts on the fact that very few pc's will need that 585w!

Some tests of the HP585D have shown that the OCP trip point is just above 430W.

Not surprising, this is the same unit used in those thermaltake XP550 430W psu's without the transient filter, which can do 430W in spec.

[kc8adu]

the fcc logo and ul are forged.
no way it will pass either spec.well maybe now that you added a bunch of parts.
did you use x and y rated caps ?

[370forlife]

That is compucase's ul number. Fake UL logo is correct, same with the FCC logo.

Yup, the X caps are X2, Y caps are safety series X1 Y1.

[Newbie2]

Doesn't look like a bad power supply now that you've added better components to it.

Well done 370forlife!

[momaka]

Originally Posted by 370forlife

Some tests of the HP585D have shown that the OCP trip point is just above 430W.

Not surprising, this is the same unit used in those thermaltake XP550 430W psu's without the transient filter, which can do 430W in spec.

This is indeed the same power supply as the Thermaltake XP550-NP 430W, minus the input filtering. According to xbit-labs and hardwaresecrets, the limiting point for these power supplies is about 350 watts. Anything after that will overheat the primary switcher (from xbit-labs tests that is). There is OCP on these units, though, which puts them somewhat above regular gutless units.
Here's the two reviews if anyone is interested
http://www.hardwaresecrets.com/article/332
http://www.xbitlabs.com/articles/coo...oundup_18.html

As far as regulation, 3.3v rail is derived from the same transformer pins as the 5v rail. The reason you see only one coil is that the 3.3v rail is regulated with a voltage regulator rather than a saturation coil.
----
Now I have a curiousity about this power supply, though - how many ICs are there on the PCB, 2 or 3 (ID numbers would also help)? I'm wondering because I have the Thermaltake XP550-NP 430W version of this power supply, and I want to know how the 5vsb rails generated and regulated - wheather it's a 2-transistor desing or something else. Mine's still under warranty, so I don't want to open it just yet.

By the way, 370forlife, you did some great work on this power supply!
Should be pretty reliable now. I know that mine has stable voltage regulation (12v is a little high at 12.36v, but all other rails are nearly perfect).

[370forlife]

Well, not counting the three little 4 pin IC's near the 5vsb transformer, there is 1 8pin welltrend wt751002 930 j48a1g

The three little 4 pin IC's near the transformers, if they are even IC's, are all cosmo 1010 817 J18 C's.

I was thinking of getting some polymers for the secondary caps, as the highest uf 10mm 16v electrolytic cap I could find are all 1500. I was thinking of getting some nichicon 16v, 10mm, 820uf caps, but I don't think the ultra low esr of them would help any.

[Toasty]

(4 pin) They're the opto's.

http://www.cosmo-ic.com/object/products/K1010.pdf

The Weltrend is a supervisor chip. (The iMac supplies use the 7512)

http://pdf1.alldatasheet.com/datashe...S080WT-12.html

Toast

[momaka]

Thanks for the info.
I noticed there is one more IC behind the input caps. Probably a UC3842/UC3843 or some other similar variant. Unless this IC is something else, then it's very likely that this power supply uses a 2-transistor 5vsb design (The Welltrend wt751002 is similar to the TPS5310, and nither of those offer 5vsb monitoring/protection). The 3 4-pin ICs in the transformer area are optocouplers, like Toasty said. If there was 5vsb protection, there should have been one more optocoupler and maybe another IC somewhere (like the Bestec ATX-250 12Z).
With that said, it's probably a good idea to keep an eye out on this power supply (and definitely change the 50v, 10/22/47uF 5vsb cap if you find it).

[370forlife]

Opps, missed that IC behind the caps. 8pin ST UC3843B.

Don't know which cap is the 5vsb cap. There are many 50v 10uf and 4.7uf, and one teapo 6.3v, 2200uf cap next to the 5vsb transformer.

[Toasty]

That Teapo is a common failure. Every one I've seen in that position across various PSU makes and models is bulged. Some less than 2 years old.

Re: Polys in output - You'd need to double them up to get the oomph! provided by the higher capacitance of the regular lytics. Yes, ripple would be super low, but as soon as something drew current for an instant, the voltage(s) would sag to far. It's not something you may readily see with a meter, but on a scope you'd see the blip.

Think of it as trying to start the car with lawn tractor battery vs. a truck battery. Both may do the job, but the lights won't dim as much with the truck battery.

Toast

[qw3r7yju4n]

so any how to's? - would love to spend little money and upgrade my PSU - I have a little DC/AC knowledge. Witha quick walkthrough I would be a PRO

[qw3r7yju4n]

and how are you allowed to sell a fake logo? was this dated before there were laws?

[everell]

I have recently acquired a Thermaltake TR2-430W. PC board looks just like yours - except it already had the line filtering parts. GBU806 bridge diode. I think Bestec is better built.