How to locate high value short component

[harp]

It is not so important rather then discusion... for learning.

I have small box where put bad components in it, shorted diode, caps, burned resistor etc. I was trying to do some yt video about tracing shorts, so, I use one of these bad smd caps from box... it has few ohms, idealy for presentation... I put him on a known good pcb board, on 3.3v rail, parallel with other caps... but, after it cools down, the rail has 85 ohms... when I apply power, it took only 35mA and no voltage drop... so 0.1W...
Any idea, your experiance how to find 85R bad caps technique?

[stj]

only way i can think of is a low-ohms meter like a shorty to compare the resistance of the 3.3v rail at various points till you close in on the lowest reading

[harp]

I have not low-ohms meter to see if is capable to find 85R in parallel... maybe, but it must be very precise and wide range...
I try to inject volteges and see if any difference I can spot, on faulty cap or other. My dmm have 4000 count, so on 3V it show milivolts. Not a tiny difference for sure between all caps in rail while measurement cant see.
I hoping that the current path can be detected somehow, without modifiing pcb, so I inject pulsed 3V about 4kHz 75% and try searching some magnetic field with loop... but nothing again. Maybe different probe? Hall sensor?
I have not RF knowledge so I dont know how resistor will act in high frequency, with another resonant component or so, and can this be used to determine location or area anyway...

BTW, heating and cooling technique in this case I tend to avoid, becouse of unstable condition, eg, cap after threatman can be 100k... and potentialy catastrofic fail later.

[stj]

4000 count isnt great - what chip is in the meter? some are 6000 or even 10,000 count chips but they cripple lower models to justify the price differences.

i had an idea.
voltage injection wont work - on it's own.
voltage injection with current limiting to a few mA may cause a noticable voltdrop difference between different places.

[megaraider]

Well finding a short capacitor in a power rail is never an easy task unless a visual inspection reveals a cracked / smoked / burned one.
Less yet when you're trying to find it by measuring (the shorty project might be the most helpful here).
On the other end you have some well proved heuristic guess techniques, because a low resistance capacitor will heat way faster.
Thus:
1. 99% IPA spray over the board, power it up, the defective cap will make IPA evaporate faster;
2. Cooling spray, ...
3. Thermal camera
4. (...)

[petehall347]

i use my dmm on 200 ohms scale and see how long the beep lasts and also watch the reading then swap the probes .

[harp]

Stj, sorry for gap... I have not forget to feedback soon.

[harp]

Stj... it is Vici VC97 dmm what I have.
Your idea dont work. I have 3.260v supply and serial connected 300 ohm - that is feed to "shorted" board and give 1.666v on very input point, on various good capacitor in rail and on faulty one. Exactly same voltage 1.666v on all point.

Also, I inject 3.2v into rail, and monitor temperature of components with narrow low-mass ntc. Before connecting to power, all on pcb are measured 31.1 deg.C, and after connecting power, and wait a while, few minutes, there is no any difference between shorted cap or a goods one, all have same temperature 31.1, only main uC go some higher 31.9 deg.C and its surrounding fade off...
I dont have thermal camera, but I also doubth that the faulty capacitor (about 85 ohm) will be seen on it. Consumption on whole board is about 0.1W and most of the power go to main uC.

Regardless IPA...
There is no way to fine cotrol thickness of applayed ipa, and continuous shape, due to capilary effect and surface tension. On wery small power consumption, there is no any visual sign that for sure indicate suspicion for some location. While apply ipa in various thicknes and coverage, ipa evaporate unevenly where cooling surface area and interference with observation.
When you reach some power point who can be seen, ipa frst act like some small tension moves (see like light reflection moving from fluid surface) regardles to on/off power, but I am very far from that point.
The next level is to slightly faster evaporate then surrounding, but it takes even more power.
On yt is common to see detection with ipa where is disipated 1w or so on small component...

So for you who have that kind of equipment like liquid air, thermal camera and so on, just solder on scrap board 82 ohm resistor (must be smd for better cooling) on 3.3v rail and try detect it when is all cooled down. I am interesting for results and technique that work.

Peteh... I dont understand your method, it is not clear... can you explain how to it work?

[petehall347]

Peteh... I dont understand your method, it is not clear... can you explain how to it work?

thats a quick way of testing for open or shorted caps on the board . the more uf the longer the beep unless its constant beep of course . i dont use a capacitor meter if i take one out a new one goes in .