Radeon X1650pro stops motherboard power up

[pfrcom]

I should have known better than to buy an "untested" video card, and find out the hard way that untested IS a euphemism for faulty

With this card in its AGP slot, a motherboard won't even power up

Is there any diagnosis an amateur like myself can do, or is it more fodder for the garbage ?

I've tried it an Asus P4P800-E and an A7V8X

In the P4P800-E:

If I leave the video card's Molex disconnected, the motherboard powers up, but its voice diagnostic says “Failed video test” or equivalent

With the Molex connected, the motherboard won't power up

In the A7V8X:

With the Molex disconnected, there was a crackling noise which I think came from the components circled in the upper left corner, before I cut power

Again, with the Molex connected, the motherboard won't power up

[momaka]

Re: Radeon X1650pro stops motherboard power up

When you say the motherboard won't power up, do the fans on the computer "twitch" or kick a little when you press the power button as if trying to start?

If so, then perhaps something is shorted on the video card, thus preventing the PSU from starting up (PSU detects short circuit). With a multimeter, check resistance between the 5v pin on the connector and ground as well as 12v pin and ground. Post back what you get.

The components you circled look like protection diodes. They may have been damaged by a bad PSU. This could be a real easy fix if they are the only bad parts.

Also, can you post a picture of the top of the video card as well?

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

The processor fan gives the very slightest twitch - not noticeable unless you're looking at it

Picture coming later

[PCBONEZ]

Re: Radeon X1650pro stops motherboard power up

You have a short on the card down stream from the molex.

[Th3_uN1Qu3]

Re: Radeon X1650pro stops motherboard power up

Check those MOSFETs on the back, if something's shorted then those must be it.

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

Here's a picture of its front - the three black caps are Panasonic FJ 1000uf 10v, and the smaller green is a Sanyo WG 470uf 16v

There IS continuity between both ground connectors on its Molex, and the 5v and 12v connectors

But no continuity between the 5v and 12v connectors

I've already tried the Mosfets in situ - no continuity on any of them between their two legs

All, with one probe on the base soldered to the PCB, give a brief buzz with the other probe to one or either of its legs

With one of the Mosfets, it's a brief buzz with either leg - all other Mosfets, it's a brief buzz with the left leg, and continuous with the right leg

It makes a difference whether the positive or negative probe is to the base, also depending on which Mosfet

Meter is a Fluke 77, bought in the late Eighties, with diode test and continuity on the same setting

[kc8adu]

Re: Radeon X1650pro stops motherboard power up

check the 2 mosfets above the inductor near the molex.
the single pin is the gate.source and drain are either 3 or 4 pins tied together.

[momaka]

Re: Radeon X1650pro stops motherboard power up

There IS continuity between both ground connectors on its Molex, and the 5v and 12v connectors
But no continuity between the 5v and 12v connectors

How many ohms?
Some meters will show continuity for as high as 100 ohms. To me, that's not continuity at all (depending on the circuit of course).
Try using the 2000 Ohm scale instead. Also, after each reading, reverse the probes and do another reading.

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

check the 2 mosfets above the inductor near the molex.
the single pin is the gate.source and drain are either 3 or 4 pins tied together.

Is the inductor the component marked 1R5 ?

Assuming so, I checked the two eight legged components just above it in the picture

They're marked 119N3S G0701, and one is mounted at 180 degrees to the other

I tested with one probe touching the yellow 12v connector in the Molex

Got a continuous beep with the other probe touching seven of the eight legs

With the negative probe in the Molex, got nothing when the positive touched the exception leg

With the positive probe in the Molex, got a short beep when the negative touched the exception leg

I think the exception leg is No. 4 - it's on the same side as the leg with the dot, but on the opposite corner

Using another video card with Molex as a reference, a similar test gives a short beep on most of the legs

How many ohms?
Some meters will show continuity for as high as 100 ohms. To me, that's not continuity at all (depending on the circuit of course).
Try using the 2000 Ohm scale instead. Also, after each reading, reverse the probes and do another reading.

My Fluke 77 is an autoranging meter, and I've used its audible continuity setting, which is also diode test - PDF manual attached, if any relevance

While it's beeping, the meter settles at .001

Makes no difference which probe goes to which Molex connector (excluding no beep between 5v & 12v connectors)

[momaka]

Re: Radeon X1650pro stops motherboard power up

Ok, let me get this straght because I'm getting a little confused here (please bear with me)...
There is continuity (meter reads 0.001 or close) between 5v pin and ground and there is also continuity between 12v pin and ground on the power connector, however there is no continuity whatsoever between the 5v and the 12v pins?

Is the inductor the component marked 1R5 ?

Yes.

I tested with one probe touching the yellow 12v connector in the Molex
Got a continuous beep with the other probe touching seven of the eight legs

What do you mean 7 of the 8 legs? Is that on each component or the 7 pins that are tied together on both MOSFETs and connected to the inductor 1R5?

With the negative probe in the Molex, got nothing when the positive touched the exception leg
With the positive probe in the Molex, got a short beep when the negative touched the exception leg

That reading sounds okay.
----
The "exception leg" is the Gate. MOSFETs are basically electronic switches, where Source is the input pin, Drain is the output pin, and Gate is the control pin which tells the MOSFET when to allow current to flow between Source and Drain.
The arrangement you see with the two MOSFETs on your video card is very common. One of them has its Source connected to either the 12v rail (or 5v rail), and Drain connected to the inductor 1R5. The other MOSFET has its Source connected to the inductor 1R5 (and in turn, to the Drain of the first MOSFET), and its Drain connected to ground. A controller then sends signals to the Gate pin of each MOSFET, telling it when to turn on. In this arrangement, they turn on and off out of phase (i.e. when one is on, the other is off, and vice versa). This is how lower voltages are derived from the 5v/12v rails.
----
With one probe on the 7 pins of the two MOSFETs that are connected together, you should not get a short circuit when you put the other probe on either 12v/5v pins or ground - i.e. no continuous beep for more than a few seconds or 0.001 readings.
If you do, state to which pin you got a short circuit. If both, those MOSFETs are likely bad.

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

Ok, let me get this straght because I'm getting a little confused here (please bear with me)...
There is continuity (meter reads 0.001 or close) between 5v pin and ground and there is also continuity between 12v pin and ground on the power connector, however there is no continuity whatsoever between the 5v and the 12v pins?

Thank you for bearing with me

Yes, as you've described is correct

No continuity beep with one probe on the 12v Molex connector, and the other probe on the 5v

But a beep with every other combination of the four Molex connectors

What do you mean 7 of the 8 legs? Is that on each component or the 7 pins that are tied together on both MOSFETs and connected to the inductor 1R5?

I mean one probe on the Molex connector, and the other probe, in turn, to each of the seven legs

With one probe on the 7 pins of the two MOSFETs that are connected together, you should not get a short circuit when you put the other probe on either 12v/5v pins or ground - i.e. no continuous beep for more than a few seconds or 0.001 readings.
If you do, state to which pin you got a short circuit. If both, those MOSFETs are likely bad.

With the positive probe on No. 1 leg (the one with the dot beside it) of the left Mosfet, I get a continuous beep to both the 12v Molex connector, and the two ground connectors

Same with the negative probe on No. 1 leg

And same from the No. 1 leg to all of the other legs, apart from the "exception leg" on each Mosfet which, thanks to your explanation, I now know is the Gate

[Th3_uN1Qu3]

Re: Radeon X1650pro stops motherboard power up

So the gate is not shorted, whereas the other two (drain and source) are shorted. That means the controller is okay, it's just the MOSFETs itself that fried. Replace them and you should be all fine.

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

So the gate is not shorted, whereas the other two (drain and source) are shorted. That means the controller is okay, it's just the MOSFETs itself that fried. Replace them and you should be all fine.

I'd love to try replacing them, but ...

... only supplier I can find, Mouser, wants $2-62 for the parts plus $39 for shipping

I think the original MOSFETs, marked 119N3S G0701, are Infineon part number bso119n03s, for which Mouser part number is 726-bso119n03s

A possible compatible, which I may be able to get, is sts11nf30l

I've attached datasheets for bso119n03s and sts11nf30l - could someone please advise if they're compatible

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

Bump - advice please

[PCBONEZ]

Re: Radeon X1650pro stops motherboard power up

Never piss in the wind. - I mean spit.
Well,,, either one...

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

Never piss in the wind.

I hope that's not what I'm doing, if I replace those MOSFETs

[Pyr0Beast]

Re: Radeon X1650pro stops motherboard power up

Bump - advice please

Remove those two mosfets and try to start up (just don't leave anything bridged with solder)

And check the tags on controller first. If it smokes, change it.

Also recap the damn thing. I do not trust those caps near the coil even one bit, probably cooked to death.

[pfrcom]

Re: Radeon X1650pro stops motherboard power up

Here's a pic showing the MOSFETs removed, and their solder pads

What I don't understand, for each set of pads, is continuity between any of the three pads for the Source legs and all four pads for the Drain legs, and vice versa

I expected, with their MOSFET removed, that Source pads would be isolated from Drain pads

Does this mean there is likely to be something else wrong, and replacement of the MOSFETs might be futile ?

Pyr0Beast, you're dealing with an amateur here - I do not understand "check the tags on controller first. If it smokes, change it."

[Pyr0Beast]

Re: Radeon X1650pro stops motherboard power up

Here's a pic showing the MOSFETs removed, and their solder pads

You've cleaned everything quite nicely.


I expected, with their MOSFET removed, that Source pads would be isolated from Drain pads
It must be different, despite your meter not showing it, otherwise something is shorted as well.

Stick the card in the slot and start it up. If there is nothing (computer still halts), unsolder the other two mosfets on the other end of the card, near the choke, as well. Check if one is connected paralel on the bottom.


I expected, with their MOSFET removed, that Source pads would be isolated from Drain pads
They are, but capacitance still couples them to some degree.

When you removed mosfets you disconnected a whole power line. If there was a short it would not reflect as such on the PSU now.
(SO8 stuff burns quite often)


Pyr0Beast, you're dealing with an amateur here - I do not understand "check the tags on controller first. If it smokes, change it."
Read what it says on the controller. Probably something like ISL6...
If the controller will go up in flames it was burned before.

[Pyr0Beast]

Re: Radeon X1650pro stops motherboard power up

I'd love to try replacing them, but ...

... only supplier I can find, Mouser, wants $2-62 for the parts plus $39 for shipping

Check farnell.

http://au.element14.com/

Something like this should be fine;
http://au.element14.com/vishay/si413...so8/dp/1779239

Did you completely destroy those SO8 mosfets ?

Shipping is around 5-10$.