ATX PSU Tester Analyzer modification / upgrade

[dicky96]

Hi guys
I have one of these ATX PSU Analyzers/Analysers Tester

https://www.aliexpress.com/item/3280...73275478%22%7D

The analyzer is very good for testing ATX power supplies - if they work! But it's pretty useless for testing them if they faulty. Which kinda defeats the object.

The problem is caused by the fact the Analyser is powered from the 12V on the 20/24pin ATX connector. So for example if you have 5V Standby and nothing else, the analyser is dead, it does not power on. Even worse you can have all voltages present apart from 12V and the analyser still will not power on - leading you to believe you have a dead PSU.

I recently wasted over two hours trying to fix a 'dead' PSU where I just could not find anything wrong, only to realise in the end it was working, apart from an open circuit track on the 12V output

So I decided to modify the Analyser so that it will power up and work if any voltage is coming out of the ATX PSU, whether it is 5VSB, 3.3V, 5V OR 12V

In fact it will now work with just 2.5V on any one voltage rail

The modification cost just €0.70 in parts and fits inside the existing housing.

If you have one of these analysers and would like to make it actually useful, then here is how I did it

https://www.youtube.com/watch?v=yw747J-9F1s

[stj]

interesting,
i have one of those and thought it was using the 5v standby rail - gonna have to look in it now.
i may also fit a switch on the control line to switch the main psu on and off.

[eccerr0r]

would make sense if it were using 5VSB, because ATX PSUs won't work at all if 5VSB is dead as it's needed for !PSON ... silly to run it off of 12V...

[sam_sam_sam]

Originally Posted by dicky96

Hi guys
I have one of these ATX PSU Analyzers/Analysers Tester

https://www.aliexpress.com/item/3280...73275478%22%7D

The analyzer is very good for testing ATX power supplies - if they work! But it's pretty useless for testing them if they faulty. Which kinda defeats the object.

The problem is caused by the fact the Analyser is powered from the 12V on the 20/24pin ATX connector. So for example if you have 5V Standby and nothing else, the analyser is dead, it does not power on. Even worse you can have all voltages present apart from 12V and the analyser still will not power on - leading you to believe you have a dead PSU.

I recently wasted over two hours trying to fix a 'dead' PSU where I just could not find anything wrong, only to realise in the end it was working, apart from an open circuit track on the 12V output

So I decided to modify the Analyser so that it will power up and work if any voltage is coming out of the ATX PSU, whether it is 5VSB, 3.3V, 5V OR 12V

In fact it will now work with just 2.5V on any one voltage rail

The modification cost just €0.70 in parts and fits inside the existing housing.

If you have one of these analysers and would like to make it actually useful, then here is how I did it

https://www.youtube.com/watch?v=yw747J-9F1s

Thanks for posting this information this will help me a lot

[eccerr0r]

If you have one stuck using 12V I'd suggest trying to just make it work of 5VSB. Probably a lot simpler than any other solution? Unfortunately I don't have such unit to see what the "simplest" "fix" would be...

[sam_sam_sam]

Originally Posted by eccerr0r

If you have one stuck using 12V I'd suggest trying to just make it work of 5VSB. Probably a lot simpler than any other solution? Unfortunately I don't have such unit to see what the "simplest" "fix" would be...

What about powering it with an external 5 volt power supply

I have a couple of them I will try doing this and let you how well it works doing it this way

[stj]

why?
if the 5v standby is missing then everything is dead and you have nothing to meter anyway.

[sam_sam_sam]

Originally Posted by stj

why?
if the 5v standby is missing then everything is dead and you have nothing to meter anyway.

All I am asking for now would it work if powered separately for now
Would there be any issues to be concerned about doing it this way

[stj]

well - i'll open one and take pictures for you.
i'm personally more interested if i can find room inside for a rocker switch.

[eccerr0r]

If you power it separately, the major drawback is that you'd need another outlet or a battery ...

Can't think of any reason not to power off of 5VSB other than if your meter draws more than 100mA - because ATX1.0 did not specify much current for 5VSB.

Hopefully bad cap ripple on 5VSB doesn't blow the tester...

[dicky96]

The way I modified it was all internal and only cost €0.76 in parts

So now it will power up as long as any voltage rail has at least 2.5V. That could be further reduced to about 2.15V if using Schottky diodes

I also considered the fact that if it would work from 5VSB then that would be good enough as it is not possible to have any other rails if 5VSB is missing...... or is it?

I can conceivably see a case where this would not be true - and I mention this in the video

Most ATX PSU I looked at have two windings, or taps, on the 5V STB transformer. The second winding/tap is typically 12V-24V with it's own rectifier diode and capacitor, and is used to power the PWM IC and associated transistor circuit that drives the base of the output transistors via a transformer

So if there was a fault where there was an open pcb circuit track or similar between the 5VSB transformer and the purple ATX wire, it is likely that the 12V-24V internal rail would still be there.

And seeing as this anaylser starts the ATX by physically connecting the green wire to ground, and needs no voltage on 5VSB to do that - it is at least possible you could have an ATX PSU that has no 5VSB but still powers up with all other voltages present when connected to the analyser.

[eccerr0r]

I really don't know why they used 12V to power initially, seems that you should be able to use 5VSB for nearly no additional cost, granted it will require it to be at least 3 or 4 volts for it to power whatever chips the tester uses. As said multiple times, ATX PSUs that don't have 5VSB working cannot turn on because that's what powers the !PSON - for both isolation and for keeping voltages within reasonable limits. Note that in the normal case, motherboards need 5VSB to power their switching circuitry, can't assert !PSON without any power.

It depends on the PSU design whether it uses two taps... isolation can happen with transformers or optoisolators.

[PeteS in CA]

If it will work from 5VSB, you could always have diode-isolated "battery back-up":

[stj]

not that brand though - no DuraLeak.

[sam_sam_sam]

Originally Posted by eccerr0r

If you power it separately, the major drawback is that you'd need another outlet or a battery ...

Can't think of any reason not to power off of 5VSB other than if your meter draws more than 100mA - because ATX1.0 did not specify much current for 5VSB.

Hopefully bad cap ripple on 5VSB doesn't blow the tester...

The reason I asked the question is if there is an issue with the standby circuit and provide more than 5 volts and you bypass the regulator because of low voltage rating for the meter parameters you might damage the tester

But in general I would agree with you that you could power it with the standby power supply output side but how do protect the tester from incorrect voltage from the standby circuit if has an issue

I would think that the meter module itself might not draw much current but the tester will put a load on the switching power supply output side of it

[eccerr0r]

As said yes a wayward 5VSB is a concern. Not like you completely remove the regulator, need to redesign it, perhaps with a shunt regulator to protect it. Have a LDO regulator there. Shouldn't be way too hard to protect against a wayward 5VSB.

[sam_sam_sam]

Would this regulator work

LP2950ACDT-5.0G

[dicky96]

Can I just mention guys that my internal mod actually costs less than the 9V battery, Duracell or otherwise

And you can't forget to turn it off

[sam_sam_sam]

Or would this work on the 5 standby circuit just in case the voltage has gone to hell and remove the regulator that is on the board now one note I am also going to put a switch on the power ON wire so I can control it coming ON

[eccerr0r]

Not sure if the TL431 is able to absorb the surge if 5VSB is considerably overvoltage... but yeah something like this, perhaps with a regular transistor as a shunt current amplifier.