+3V down to +1.1V

[DanF]

I'm sure we have someone who's also good at component level repair in this community (at least I'm hoping so)

I have an Asus X541UVK board with power issues. I noticed that while the 19V is present, 3V and 5V are missing. 3V is being dropped to 1.1V and at some areas it's fluctuating between 0V and 1.1V. DC step down IC is extremely hot, but I don't have the knowledge to conclude if the IC is the issue or if something else causing it. Probably a component on the same circuit that is short to ground can cause the IC to get extremely hot? 19V is definitely getting into the IC, but it's outputting 1.1V where it's supposed to output 3V.

You might ask if I could simply replace it and give it a shot, but ordering one would take a month to receive.

 

[Moltuae]

Gonna need more information, or a crystal ball

Do you have a schematic? Which power regulation/conversion IC is it using?

 

[DanF]

Gonna need more information, or a crystal ball

Do you have a schematic? Which power regulation/conversion IC is it using?

I have a boardview I can attach here.

The IC is: RT8249CGQW (Richtek)... PU8701 on the schematic.

 

[Moltuae]

I'd need a schematic rather than a board layout really.

Checking the IC's datasheet, it does have built-in protection and current limiting, so it's possible that's kicking in/out.

The current limit circuit employs a unique “valley” current sensing algorithm. If the magnitude of the current sense signal at PHASE is above the current limit threshold, the PWM is not allowed to initiate a new cycle. Thus, the current to the load exceeds the average output inductor current, the output voltage falls and eventually crosses the under-voltage protection threshold, inducing ICshutdown.

Try 'scoping the outputs and/or the 'PGOOD' status pin. If the PGOOD status is momentarily switching to the 'good' state and the outputs are repeatedly ramping up then shutting off, it would indicate that there's an excessive load elsewhere on the board, rather than a fault with the switched-mode power IC itself. You could also try disconnecting the power outputs from the rest of the circuit, if you can determine which components to remove (or which tracks to cut).

 

[DanF]

This is the only thing I could find so far.

 

[DanF]

It's true, PGOOD is going up and down.

Finding the tracks to isolate is tricky for me. You are referring to cutting out parts of the circuit by de-soldering jumpers right?

 

[Moltuae]

Yes, or components (... or you could cut tracks, as a last resort). But it's going to be difficult without a full schematic to find isolation points.

There's a good chance however that the circuit closely follows the Application Circuits on page 10/11 of the IC manufacturer's data sheet. In which case, if you can identify the equivalents to L1 and L2, you may be able to remove those and thus disconnect the load to enable you to test the outputs (albeit with some high frequency PWM noise present, due to the removal of the chokes and the low-pass filtering they provide, along with R12-R15 and the decoupling capacitors). If necessary, you could recreate the low-pass filter circuit externally.

It's potentially a lot of work/effort though. My background is electronics, but I stopped doing component level repairs decades ago because these days it usually costs more in labour to repair a board than it does to replace it.

Also, if PGOOD is oscillating, I'd say you can be 90% certain that the fault is not with the IC. Try 'scoping PGOOD (on a dual-channel oscilloscope) against an output to see if the output is ramping up to the full/correct voltage before PGOOD changes state.

 

[DanF]

Yeah I can find L1 and L2, but didn't quite understand what happens when I remove them (you went complex very quickly!) I've actually already replaced L2.

But, I'm only doing this for fun and possibly learn something new, cause it's not really worth the work. Having said this, I don't have an oscilloscope.

 

[Moltuae]

You should invest in a 'scope if you plan to do component-level repairs, even if it is only for fun. In fact, component-level repairs are much more fun with a 'scope because you can 'see' so much more. You'll learn more too because it's much easier to understand what's going on when you can view waveforms and trace signals through the circuit. For little more than the price of a multimeter you can pick up a basic oscilloscope, which should be sufficient for these kind of repairs.

To better understand what L1 and L2 do, you first need to understand how the IC generates and regulates the supply voltages ...

There are essentially 2 ways to create a step-down voltage regulator; you can use linear voltage regulation or switched-mode regulation.

In it's simplest form, a linear regulator uses 'controlled' variable resistance (in the form of a transistor), and feedback to keep the output voltage constant. The result is a smooth and linear output voltage but efficiency is very poor because the transistor has to dissipate the remaining power (simplified: Dissipated power = Amps x (Vin - Vout) ).

Instead, a switched-mode regulator rapidly turns the output on and off at a specific ratio (in this case using PWM) in order to generate the desired voltage through 'averaging'. So, for example, a 12v input switched rapidly on and off with equal durations (50% on, 50% off) would represent an average output voltage of 6v. Now of course to be usable, the switching pulses need to be removed from the output. This is where the ('low-pass') filter circuits and L1/L2 coils/chokes come in. There will be some amount of pre-filtering inside the IC but a relatively large choke and decoupling capacitor are required to really smooth out the output.

Coils and capacitors act in almost opposite ways, in that a capacitor will block DC and pass AC, while a coil will pass DC and block AC (the degree to which they pass/block being determined by their value). In the case of the application circuits here L1/L2 resist the AC (pulses) while allowing the DC through, then the large decoupling capacitors (C3/C17) pass (to ground) most of the remaining AC that gets through, further smoothing the output. If you remove both the chokes (and thus disconnect the capacitors too), you will likely see an output with a lot of pulse interference ... IF, that is, the IC doesn't decide to switch off the outputs because of the anomaly. If that happens you might need to reproduce the inductor/capacitor filter arrangement externally to fully test the IC.

 

[DanF]

Thanks for the time and explanation. Am I right to say that in this case the coils and/or caps are probably malfunctioning?

While I probably won't go further, because it got more complex than expected, it's definitely interesting to know and I got to learn something new.

 

[Moltuae]

Thanks for the time and explanation. Am I right to say that in this case the coils and/or caps are probably malfunctioning?

Unlikely I would say. Passive components such as those usually fail 'visibly' (burn marks, bulges, etc) and only usually if they're underrated/over-stressed. Without any further testing or info, and going by the oscillating 'PGOOD' (and past experience) alone, my guess is you're dealing with an overload. I would guess that the voltage regulator IC is working fine but that you have a short or failed component elsewhere on the board that is drawing too much current. You could try removing anything that is easily removable/unpluggable from the board, if you haven't done so already. Failing that, unless there are any visible signs of component failure, tracing an individual component failure on a large/complex board can be very time consuming.

 

[NviGate Systems]

Along with a scope, a good tool is a FLIR camera. Tells you almost instantly where a hotspot, aka where all the energy is going. That can help with issues like this. FLIR tells you what components are hot then you see if it's related to what's not working. Crude but effective.

The poor man's FLIR is 99% Isopropanol, pour it over the board, turn on and see where it evaps the quickest. That is if you are desperate. Depending on what's wrong it could help/harm.

 

[DanF]

Along with a scope, a good tool is a FLIR camera. Tells you almost instantly where a hotspot, aka where all the energy is going. That can help with issues like this. FLIR tells you what components are hot then you see if it's related to what's not working. Crude but effective.

The poor man's FLIR is 99% Isopropanol, pour it over the board, turn on and see where it evaps the quickest. That is if you are desperate. Depending on what's wrong it could help/harm.

I have a relative who has a CAT phone with FLIR built in, not sure how good it is, but I'll give it a shot.

In fact I've identified there is a problem with the step down IC because it was so hot, flux was literally boiling off.

Any recommended oscillscope? The ones I can find are the size of a multimeter. The bench ones are too expensive for me at this stage.

 

[Moltuae]

In fact I've identified there is a problem with the step down IC because it was so hot, flux was literally boiling off.

Any recommended oscillscope? The ones I can find are the size of a multimeter. The bench ones are too expensive for me at this stage.

The temperature of the step-down regulator IC does not necessarily indicate a failure. I suspect it's a symptom of the fault rather than the cause. If its inbuilt protection is working, it will be repeatedly cycling the outputs on and off (which your observation of the PGOOD status supports). Therefore each 'on' cycle will briefly be experiencing excessive current (before the IC shuts down again), causing it to heat up. Prolonged overheating could of course damage the IC but, according to the data sheet, it is capable of withstanding junction temperatures up to about 150°C.

Any oscilloscope is better than none so it really depends on your budget. I'd recommend dual channels at least, to enable comparing of 2 signals. PC-based 'scopes such as this can sometimes be a good way of getting a higher spec' 'scope for less money. Entry-level 2 channel bench 'scopes such as this are pretty cheap and should do the job (I haven't personally tested any of these by the way). Alternatively, you can get some good deals on used high-spec' 'scopes. You could try eBay, or look for somewhere that specialises in used test equipment. The specialist stores often stock a lot of ex-military equipment, which will usually have been serviced and calibrated frequently.

 

[Larry Sabo]

@DanF, you may find these videos by Electronic Repair School's YouTube channel helpful. I find his videos interesting because of his diagnostic techniques, even though his standards and repair techniques are often "dodgy."

 

[DanF]

@DanF, you may find these videos by Electronic Repair School's YouTube channel helpful. I find his videos interesting because of his diagnostic techniques, even though his standards and repair techniques are often "dodgy."

I am actually subscribed to him already I noticed that he sometimes takes "dodgy" shortcuts