Author Topic: Short circuit on PCB (solved) - without schematics - question about my workflow  (Read 3048 times)

0 Members and 1 Guest are viewing this topic.

Offline ed1973Topic starter

  • Contributor
  • Posts: 11
  • Country: nl
Hello,

I am a mechanical engineer and only learned the basics of electronics (by average got 2 hours/week lessons in electronics).

I have been busy with a dead playstation 4, eventually found the root cause and the solution.
Because it cost me maybe 6 hours, and I am questioning : Could I have done this faster ?

This is what I have done:

1) Replaced power supply by new one -> same problem, no reaction on power on button, nor eject button

2) Measured output lines of power supply, 5V, but when connecting it to motherboard, is goes down to 0V.
Concluded, there is a component defective, or there is a short circuit on the motherboard

3) Took my voltage regulator and connected it to 5V input on the motherboard, took 5V and 50mA to be safe.
What I did see, was that the Amperage on my PC screen was fluctuating from 0A up to 2A and I heared a vibrating (maybe 2kHz) sound from the motherboard.
I listened very carefully and discovered that the component RT5069A (Power Management Multi-Channel IC (PMIC)) was making the noise, took my IR temperature measurement pencil and measured 40 degrees Celcius. Much higher then all other components.

4) Desoldered the IC, put 5V with 50mA on the mobo and problem was gone. Did not hear the sound.

5) Replaced the IC by a new one -> same problem

6) Compared the motherboard with a good working motherboard and found out, that on 1 output line of the RT5069A IC , was connected to ground on the defective motherboard, and not connected to ground on the good working identical motherboard.

7) Then I measured on both sides of the motherboard, to which points, the output line was connected and marked them all. Totally there were about 39 points.
I did not measure under the BGA chips.

8) Then I checked if there were single components in between that point and ground. I was searching for single components which were defective and shorted.
There were about 30 of them, of the 39 points.

9) Then I measured (Fluke 175) the resistence of those single components and started to desolder the components with the lowest resistence.

Eventually after desoldering 5 components (and of course soldering it back), I found that a smd capacitor was shorted.
Replaced that one with an identical one, from a scrap board, and everything was working.

This is the first time I went that far, I am thinking about, could I have done this faster ?
Maybe with the right equipment ?

Kind regards,

Ed
« Last Edit: September 25, 2018, 11:49:20 am by ed1973 »
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4694
  • Country: au
  • Question Everything... Except This Statement
My approach on a device like this with limited forward knowledge would likely be:

1. Try and power up PSU seperate from device, e.g. like how an ATX supply has a Power_On pin,
2. Measure the various rails to ground at the connector to see if there is a hard short,
3. Get out my variable lab supply and feed in some small amount of current. e.g. 100mA limited to 0.4V, and use my multi meter on mV range to narrow down where the short is, (the PCB traces drop voltage, so you can narrow down your search area by where the voltage for the positive rail is lowest, or on the ground traces where it is highest)

4. After that likely spray some isopropyl alcohol on the suspect area and see where it evaporated, (It doesn't always work, especially on switch mode boost transistors)
5. Hopefully I've found it and can lift the cap, take some guesses from where it is on its value, and throw in a new one.
 

Offline CJay

  • Super Contributor
  • ***
  • Posts: 4137
  • Country: gb
I think you did good.

Generally on a dead device that's making odd noises I'll check power rails first, the 'squeaking' noise you heard was probably a switch mode PSU shutting down under excessive load which is a pretty sure fire indicator of a short somewhere.

It's familiarity and practice, the 'right equipment' in many cases is just your brain and a multimeter.
 

Offline capt bullshot

  • Super Contributor
  • ***
  • Posts: 3033
  • Country: de
    • Mostly useless stuff, but nice to have: wunderkis.de
That was a quite systematic and good way, congrats on the successful repair.

With some more knowledge and equipment, there may be some shortcuts:
- Measure voltages and waveforms around the PMIC to identify the shorted rail while supplying 5V to the board, verify short by measuring resistance.
- Inject a somewhat higher current into the shorted rail, and try to find the shorted capacitor by measuring voltage along the tracks or use a thermal camera to find a warm spot (caused by the shorted capacitor)

Similar to Rerouters approach.
Safety devices hinder evolution
 

Offline ed1973Topic starter

  • Contributor
  • Posts: 11
  • Country: nl
Wauw, thank you a lot for your hint and tips. Definitely this will help me a lot.

I just hate it, to give up. Always like to find the cause and solution.

Isn't there a method, to make the flow of electrons visible within the traces ? Measuring magnetic fields ?
Or is my idea too futuristic ?
 

Offline Samogon

  • Frequent Contributor
  • **
  • Posts: 457
  • Country: us
Good job!
Surr there are lot of tools used in repair shops, most of them are expensive, for instance ToneOhm from Polar has sound indicator of short so you can find shorted componet in few minutes. Then Huntron component tracer very usfull in locating abnormal component behavior visualizing its Volt/Amper characteristic.
Also just you hand can feel shorted or possibly shorted components by their temperature. Or simple duster spray can show where is most power dissipated.
 

Offline CJay

  • Super Contributor
  • ***
  • Posts: 4137
  • Country: gb
Good job!
Surr there are lot of tools used in repair shops, most of them are expensive, for instance ToneOhm from Polar has sound indicator of short so you can find shorted componet in few minutes. Then Huntron component tracer very usfull in locating abnormal component behavior visualizing its Volt/Amper characteristic.
Also just you hand can feel shorted or possibly shorted components by their temperature. Or simple duster spray can show where is most power dissipated.

Polar short locator is useful but expensive, I've seen very simple projects that replicate the function with not much more than a couple of op-amps.

Personally, I'd spend the money on a decent high resolution multimeter instead because it will be more generally useful.

Huntron Tracker is just an Octopus tester in a box, again expensive because it duplicates the function found on some analogue oscilloscopes like the Hameg ones. It's a trivially simple feature to add to a 'scope  though I do not know how well it would work on a digital 'scope.

I find the Octopus tester immensely useful when working on power supplies and other analogue equipment like amplifiers, radios etc.
« Last Edit: September 26, 2018, 09:10:18 am by CJay »
 

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4206
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
6 hrs to find a shorted decoupling cap? Not bad at all on such a complex product, especially without any schematics or layout diagrams.

My favourite technique is a bit more risky but also quicker, if you have the necessary equipment. Connect an external power supply to the defective rail, set it to the same voltage as the rail is supposed to have, and set the current limit to a low value. Scan the board with a thermal camera (eg. an E4, preferably with a close-up lens), and look for parts which start to get warmer as you increase the current limit.

A shorted decoupling cap can, if you're lucky, start to light up on the thermal image. No trial-and-error desoldering is required.

If you don't have a thermal camera, another (nasty!) option is to increase the current limit on the supply until the shorted component simply burns out. This doesn't qualify as 'science' in many peoples' books, but if the faulty equipment is scrap otherwise, there's nothing to lose by trying. More often than not, the component which fries is one that was dead anyway. (Make very sure to set the voltage correctly; don't be tempted to increase the voltage setting on your power supply to make the faulty component get hotter!)

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 5780
  • Country: ca
Quote

If you don't have a thermal camera, another (nasty!) option is to increase the current limit on the supply until the shorted component simply burns out. This doesn't qualify as 'science' in many peoples' books, but if the faulty equipment is scrap otherwise, there's nothing to lose by trying. More often than not, the component which fries is one that was dead anyway. (Make very sure to set the voltage correctly; don't be tempted to increase the voltage setting on your power supply to make the faulty component get hotter!)


This is a NO NO absolutely a NO NO,  seriously ??        I would never suggest that "nasty thing of yours" ...  you have to learn how to do trouble shooting, that's all, and take your time.

EX:  you have tantalum caps, tds tvs diodes who give an very good ohms short,  you will create more damage than needed, you may blow pcb traces and many other thing before you find or see you're "fried " parts goes up in flames   
 

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4206
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
I'd put this technique on the list of things you can do provided you understand what exactly it is you're doing, why, and what the risks are. If it's not something you'd be comfortable doing, then don't do it.

It may be appropriate depending on the time and resources available. Personally I wouldn't be inclined to spend what could turn out to be many hours removing and testing individual components, unless the faulty product were irreplaceable.

Given that my time has value, and I can order a new games console for next day delivery, the only repair methods that would make sense for me would be quick ones.


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf