X-Power ATX-600TD

[JEWilson]

This is a new PSU and is out of regulation as below.
I used a go/no-go PSU tester to establish the following;

Rail O/P as found on PSU label
+12.0V1 : 16.0 A max.
+12.0V2 : 16.0 A max.
+12.0V3 : 16.0 A max.
+ 5.0V : 24.0 A max.
+ 3.3V : 24.0 A max.
- 5.0V : Unused
-12.0V : 0.5 A max.
+ 5.0V SB : 3.0 A max.
All Combined - 600 W max.

I get readings of
+12.0V1 : 11.5 V ok
+12.0V2 : 11.5 V ok combined in ATX 8 pin with +12.0V3
+ 5.0V : 4.9 V ok
+ 3.3V : 3.4 V ok
- 5.0V : Unknown
-12.0V : 11.6 ok
+ 5.0V SB : 5.1 V ok
PG : reads occassionally 000 ms, reads also 160ms

When I apply a load (such as 4 HDDs) I get readings of;
+12.0V1 : 10.2 V out of regulation
+12.0V2 : 10.2 V out of regulation, combined in ATX 8 pin with +12.0V3
+ 5.0V : 4.3 V out of regulation
+ 3.3V : 3.4 V ok
- 5.0V : Unknown
-12.0V : 11.4 ok
+ 5.0V SB : 5.1 V ok
PG : reads 00 ms

So, both of the rails have a drop of p.d. which makes the PSU go out of
regulation.

I opened the PSU to establish if any problems;
Found a +12.0V rail yellow wire loose, reworked and is ok now.
Checked all O/P filter caps for signs of distress - appear ok
Checked ESR of O/P filter caps and +5.0VSB - all are ok.
Checked if O/P filter caps were s/c - they are not.

Funny thing is, I traced the filter layouts and cannot find
any load resistors for the +12V supply. There are only
two wire links - this is odd... or is it?

It may be, this is by design, dunno.

Any suggestions as to the load issue with the +12V rail?

[JEWilson]

May have found one problem

The load resistor for the +5.0V rail is 20.1R 5%
This is how it is colour coded. Red, Brown, Black, Gold

When measured, get a reading of 196R
So...

Either the resistor is incorrectly marked and was installed
as such during production or, the resistor has gone bad.

Seems to me it is no coincidence, the measured reading
is close to 201R!

Do load resistors go bad this way?

Alternatively, it may just be poor production quality checks.
Even so, all the labels on the PSU indicate it passed QC.

Clearly the above normal resistor measurement would account
for the p.d. drop under light load conditions.

Comments?

[linuxguru]

196R is about right for a bleeder resistor on +12v - 20R is probably too low. I don't think it's a fault.

[JEWilson]

Y - my mistake

One of these days, I'll recall how to read colour codes correctly.

[JEWilson]

Fixed.

O/P Filter caps were ok - the supply was brand new.
This is rebranded HEC - don't know much about these quality wise.

Replaced the small value caps in the primary - these had a very high ESR
reading - replaced with Pana NHG.

Also, found a loose +12V wire on the second +12V rail O/P, reconnected
to an available via on the PSU pcb. Poor quality termination issue.

Powered up and appears ok.
Connected 4HDDs and stayed within regulation - fine for +5 and +12V
Will try in a mobo next to get some readings - for +3.3V.

For future reference (for those who have the same problem)
Replace
1 of 3.3 uF 50 V Teapo SEK 105 deg C - H 11.0 mm x Dia 5 mm C6
1 of 1.0 uF 50 V Teapo SH 105 deg C - H 11.0 mm x Dia 5 mm C19
1 of 4.7 uF 50 V Teapo SEK 105 deg C - H 11.0 mm x Dia 5 mm C22
1 of 100.0 uF 25 V Suscon SK 105 deg C - H 11.0 mm x Dia 5 mm C11

Cheap fix. Got the PSU for £5 plus postage from eBay
These sell in Maplins (maplin.co.uk) for upward of £75! Unbelievable!

Thanks

[linuxguru]

HEC (Hiroichi?) is middle-of-the-road, usually worth recapping. Generally, small value electrolytics rarely go bad because they aren't in locations with high ripple current, and their ESR isn't critical - anything up to a few ohms is par for the course. My guess is that replacing the 100uf/25v Suscon would have been sufficient.

[JEWilson]

Y - agreed

I believe this was the cap for the TNY267PN, the switcher
This would account for the poor regulation symptoms in this particular
case.

[linuxguru]

Many designs omit the bleeder resistor for +12v because the fan is on +12v and draws anywhere from 100..200 mA. This also avoids wasting power in the bleeder and improves efficiency by a tiny fraction.