Odd Thermaltake TR2 550 problem

[Rusty!]

Well it's is to me, at least.

The aforementioned PSU powers my PC like a champ, however, it can't stop powering it! Tell the PC to shutdown, and it only half does... fans still spin, power LED remains lit. The LED on the PSU turns orange to indicate that it "thinks" it's properly shut down, and the EON Powerdown it's plugged into kills the monitor and speakers, so it has dropped its power consumption significantly.

The PSU fires up (well, not exactly, but can't think how else to describe it) with no motherboard connected, multimeter shows both 5v and 12v rails are live (tested at a molex plug).

Popped it open and found the entire cold side (almost) is under a huge heatsink, which was annoying. The only visible cap outside it was swollen so replaced it, but it didn't change anything.

Any ideas what I should be looking at here? Have attached a picture of the offending thing, but I figure it's not much use due to the heatsinks.

Cheers!

[c_hegge]

Re: Odd Thermaltake TR2 550 problem

Have you tried a different PSU? Could the problem be with the motherboard?

[eccerr0r]

Re: Odd Thermaltake TR2 550 problem

It indeed is strange if it will come on when the ATX plug is not connected. Usually there's either a pull up resistor on the input to a transistor on the green wire (!PS_ON) - so if the ATX is disconnected but AC is plugged in, VSB will make that wire close to 5V. Is this the case or is it down near grounded now?

[shovenose]

Re: Odd Thermaltake TR2 550 problem

The PSU fires up with no motherboard connected,

I'm fairly certain that is not a motherboard problem.

[c_hegge]

Re: Odd Thermaltake TR2 550 problem

Ah, whoops. missed that bit.

[Rusty!]

Re: Odd Thermaltake TR2 550 problem

Yeah, with another PSU (and a few adaptors) everything behaves as normal.

It indeed is strange if it will come on when the ATX plug is not connected. Usually there's either a pull up resistor on the input to a transistor on the green wire (!PS_ON) - so if the ATX is disconnected but AC is plugged in, VSB will make that wire close to 5V. Is this the case or is it down near grounded now?

Green wire is showing 2.41v in this weird "off but still kinda running" state, measured between it and either of the adjacent ground pins.

[Rusty!]

Re: Odd Thermaltake TR2 550 problem

Checked PWR_OK signal, it's 0.10v when in "standby" and 4.92v when fully on. Looked up the ATX spec and that seems to be within the thresholds they should be, so at least this thing is doing something right.

[eccerr0r]

Re: Odd Thermaltake TR2 550 problem

"!PS_ON" - yes this is an _input_ but you should still see a voltage on it (make sure you're using a 10M or better impedance voltmeter, the 1M cheapos may affect the results)... PWR_OK is an output of the PSU and motherboards would not boot without this signal working properly.

What happens if you hook up the +5VSB wire (purple) to the green wire? (may want to use a resistor from 470 ohms to 1K ohm or so) but don't think it should hurt anything)? Will it shut off? If this is the case then somehow that internal pull-up resistor blew...

If it dows not turn off, then the power on circuit is really broked and will need investigation (circuit extract/schematic) to figure out...

[Rusty!]

Re: Odd Thermaltake TR2 550 problem

PS_ON is showing 2.41v with ATX disconnected.

I don't have a resistor handy, but tried purple to green with a handy bit of wire, nothing happened. Will see if I can rob a resistor out of something and try that next.

[PeteS in CA]

Re: Odd Thermaltake TR2 550 problem

2.41V is probably in the uncertain state range for TTL, if that's relevant. But is that 2.41V measured with a DMM? If so an oscilloscope might show that 2.41V to be an average voltage of a line that is actually hashing between high and low (in which case PS_ON is getting repeated "On" commands). I don't recall whether the "On" command is a low or a high, but it sounds like the pull-up or pull-down resistor on that input may be open (or a bad solder joint causing it to be open).

[Rusty!]

Re: Odd Thermaltake TR2 550 problem

Yeah just a DMM, no scope here.

I can't see any bad joints, although some of them look more like MIG weld blobs where multiple wires are being joined.

[eccerr0r]

Re: Odd Thermaltake TR2 550 problem

If it's in that funky state and you connect the green to the purple wire, then we know that the first stage is probably messed up and not just a pullup problem. Probably need to do a full circuit extract here, perhaps that input transistor is bad, no idea...

I don't think that first stage has any oscillator on it... For a working PSU the input should be held high due to the pullup and the motherboard will pull that wire green wire low (just like the "short" trick). One ATX PSU I saw this was fed into an NPN transistor whose collector is pulled low when the ATX connector isn't plugged in or the MB doesn't pull low. With this output low, it's fed into a PNP transistor that will pull its collector high, and this high signal is fed into one of the PWM's voltage inputs - voltage input is too "high" which causes the PWM to shutdown and result in no voltage output, hence "off".

...but this is just one implementation of PS_ON.

[Rusty!]

Re: Odd Thermaltake TR2 550 problem

So in short, sling it out?

[momaka]

Re: Odd Thermaltake TR2 550 problem

So in short, sling it out?

It's a nice PSU, so I'd say no. Just needs a bit more dedication to troubleshoot/repair it, though.

[kaboom]

Re: Odd Thermaltake TR2 550 problem

Hi there.

What do you get when you measure from purple to black/GND?

If 5VSB is low, and also used to derive pull up voltage for PS-ON, it's being low will trigger the opamp/comparator that ultimately controls the main PWM. This is because that opamp is usually powered from the standby AUX supply (10-18V).

Bad 5VSB caps?

[Rusty!]

Re: Odd Thermaltake TR2 550 problem

It's a nice PSU, so I'd say no. Just needs a bit more dedication to troubleshoot/repair it, though.

Unfortunately bin day came and I figured I was never going to manage it, and let it go. Looking at it, pretty much everything is either under a large heatsink of surface mounted on a small daughterboard, I think me with a multimeter was never going to manage it once we'd got the basics outta the way