Job lot PSU no.3: SolyTech SL-8500BTX

[Khron]

This one had a post-it on it saying "Dead". Visual inspection through the rear grille - fuse blown to smithereens This, disregarding the fact the silkscreen only specs the platform up to 400w (despite it being marketed as a 500w unit)

After removing the board from the casing, first step was removing the fuse (see picture no.1), and soldering in a pair of those metallic fuse holders. Popped in a 6.3A fuse, and hooked up the whole thing through a bulb-tester. No flash, no nothing; measured the 5Vsb, meter said 1.8v

Moment of revelation - with the passive PFC coil removed, there was only a 1.8u film cap in series, supplying AC to the rectifier bridge Soldered in a jumper. Result = 5Vsb = 5.063v, but bulb is lit quite strongly, and flashes at some frequency (~50Hz-ish or so). Voltage across the pair of main caps is around 250v DC, which is on the worryingly low side (should be around 320v or so).


Platform data:

PWM = 2008Z
Passive PFC
2x 80mm fans
Caps = YC
5Vsb = two-transistor (PN2222A / 2N60)
Switchers = 2x TOP3 13009's (half-bridge)

Quite a "vintage-design", still has silkscreening for -5v

Without having removed the primary silicon, nothing measures shorted (switchers or 5Vsb fet).

I should mention i've already replaced the 10u "critical"(?) cap on the 5Vsb primary, the two filter caps for it on the secondary, and a 47u (filter for the base drive transformer supply) and a 10u (PWM supply filter).

Any thoughts?

[Khron]

Re: Job lot PSU no.3: SolyTech SL-8500BTX

Same "flashy" behaviour even with all three primary transistors out...

Yep, primary silicon (on the extracted heatsink) checks out just fine.

Culprit: one shorted diode in the 6A 600v bridge (1,6mV drop, as opposed to ~550mV for the other three healthy ones)

Popped another PBU605 (6A 600v) bridge in, normal "single flash" upon plug-in

Primary silicon/heatsink back in, no lit bulb, 5Vsb = 5,063v

Plugged in a fan, shorted green to ground, et voila, it lives once again 5.17v, 3.335v, 12.325 (with only a fan loading the 12v).

Recap coming up Some Panasonic FK's on the 12v, and i'll see what i can scrounge up for the 5v & 3.3v rails :P:

[everell]

Re: Job lot PSU no.3: SolyTech SL-8500BTX

Your psu is similar to my Solytech SL-8460 BTX. To compare the two check out this thread:

https://www.badcaps.net/forum/showthread.php?t=11416

Toward the end of the thread I made a drawing of that 2008 chip.

[Pentium4]

Re: Job lot PSU no.3: SolyTech SL-8500BTX

Thing doesn't look bad except for the trashy YC caps. I'd recap it! Nicely sleeved cables too. Did you solder those Sanyo caps onto it?

[Khron]

Re: Job lot PSU no.3: SolyTech SL-8500BTX

Yep, slapped that pair of Sanyos on the 5Vsb, a couple of Panny FK's on the 12v, and a pair of Ruby MBZ's on the inputs of the 3.3v and 5v pi-filters, and another pair of Sanyos on the outputs. I know, i know, MBZ's might be a bit too low-ESR, but i didn't have anything else low-ESR at 2200u in 10mm cans (and i figured going down to 1500u wouldn't have been too good, for the pi filter inputs)...

Slipped in a Panny FR to replace the -12v cap, although i skipped a few of the tiniest caps (3x 0.1u near the 2008Z, the pair of 2.2u in the bases of the main switchers, and one or two more i couldn't see the value of, in the middle of the PCB)

[everell]

Re: Job lot PSU no.3: SolyTech SL-8500BTX

Possibly of interest, the 2008Z chip is an extension of the 2003/2005 chip series in the older L&C power supplies. I have a Raptor brand 400 watt psu which is a L&C OEM, model number LC-B400ATX, circuit board marked X2005B, ATX ver 1.2. Interesting thing is that the silkscreen printing on the component side indicates that the pwm chip can be either a 16 pin or a 20 pin chip. The Raptor has the 16 pin 2005 chip installed.

I did some experimenting on my LC8460BTX and removed the 2008Z chip and installed a 2003 chip. It worked. So it appears that the 2003 and 2005 chips are interchangeable, and the 2003/2005 can be installed where a 2008Z was.

The extra 4 pins on the 2008Z are for overcurrent protection for two 12 volt rails, the 5 volt rail, and the negative rails. So putting a 2003/2005 chip in its place removes the overcurrent protection for the power supply. Bet they charged lots of extra $ for that feature!

[Th3_uN1Qu3]

Re: Job lot PSU no.3: SolyTech SL-8500BTX

Just to chime in as me and the OP had a chat on YM during the recap - i measured a good YC 2200uF 16v cap and it came in at ~19mOhms, same for a 2200uF 10v. So Rubycon MBZ isn't too low ESR, especially since i did my measurements at room temperature, and ESR decreases with temperature rise.

Before you ask... Yes i do have an ESR meter which goes that low. As a matter of fact, it's got 0.1mohm resolution. I made it myself, i talked about it a while back. Still haven't gotten around to a redesign and a case, it's still finicky. It requires wifi coax for the probes. The 4 cables have to be run next to each other, only separated at the tips as otherwise it won't show 0 with the probes shorted due to electrical induction. But it works.