Resistances measurement indicator on PCB

[kraka]

My question is about measurements i made on the resistors of a PCB, which is a component of my PC. All the readings I made in all resistors, one by one, agree with the labels on the resistors , except for two of them. In one the label on the resistor is 4700 Ω, but my measurement shows 40 Ω, and in the other, the label is again 4700 Ω, but my measurement shows 7.60 Ω.

The measurements were made with the resistors on the board.

The strange thing (at least for me who don't have the background as an electronics engineer) is that when I removed these two resistors from the board, and measured them, they gave the correct readings, i.e. 4700 Ω so it means that they are not broken.

After their detachment from the board, if I measure the resistance between the two empty pads of resistor, I find the same reading as when I was measuring, with the resistors on. It is as if the multi-meter bypasses the resistors and chooses a path with the least resistance. I know that current flows in the path with the least resistance, so is this a proof that there is a short circuit somewhere ?

My question is, whether this is an indication that there is a short circuit somewhere in the circuit. Near these resistors there is a PLD (programmable logical device) which is overheated. Is there any chance that the short circuit (if there is) is not in the PLD, but somewhere else ? And how can I locate it ? I don't have the schematic.




My question is about measurements i made on the resistors of a PCB, which is a component of my PC (it is a PCI card). All the readings I made in all resistors, one by one, agree with the labels on the resistors, except for two of them. In one the label on the resistor is 4700 Ω, but my measurement shows 40 Ω, and in the other, the label is again 4700 Ω, but my measurement shows 7.60 Ω.
The measurements were made with the resistors on the board.
The strange thing (at least for me who don't have the background as an electronics engineer) is that when I removed these two resistors from the board, and measured them, they gave the correct readings, i.e. 4700 Ω so it means that they are not broken.
After their detachment from the board, if I measure the resistance between the two empty pads of resistor, I find the same reading as when I was measuring, with the resistors on. It is as if the multi-meter bypasses the resistor and chooses a path with the least resistance. I know that current flows in the path with the least resistance, so is this a proof that there is a short circuit somewhere ?
My question is, whether this is an indication that there is a short circuit somewhere in the circuit. Near these resistors there is an IC which is overheated. Is there any chance that the short circuit (if there is) is not in the IC, but somewhere else ? And how can I locate it ? I don't have the schematic.

note : both of these resistors are connected to one of the I/O pins of the IC (not the same pin)

 

[KlausST]

Hi,

what you see is how electricity basically works.

In short: the test of resistors - built in a circuit - is not reliable.

Current flows at the paths they exist. And on a PCB a resistor almost never is "alone" .. there always are other paths for the current to flow.
And an Ohms meter does the Ohms measurement by forcing a current (or voltage) .. then measure back the resulting voltage (or current) and then use Ohm´s law to calculate the resistance.

--> So if there is a second path for a current to flow, then simply saying the shown value is not correct.

If you want to reliably want to measure a single part (resistor, capacitor, diode.....) yu have to separate at least one leg of your part. This prevents current to flow an unwanted path.

Klaus

 

[kraka]

This means that the resistor that i measured is connected to something else and the , lets say 10mA, that my multi-meter apples for the specific measurement "travels" not only through the resistor, but also through the other device or devices that the resistor is connected to.
From my readings, this "other" devices have a 7.6 Ω (for the one of the two resistors measurement). If the resistor and the "other" components are connected in parallel, it means that the total R is the inverse of the sum of the inverse of 4700 and the inverse of 7.6, which is 7.588 Ω. Then, the current that flows through the resistor during the measurement is 10mA multiply with the total R divided by R, which is 0.016 mA. If my multi-meter gives me the 7.6 Ω reading, it means that it receives a ~0.075 Volts through the two sides of the resistor.
Am i correct ?
But how this can help me to find a solution on my main question which is this :
"...My question is, whether this is an indication that there is a short circuit somewhere in the circuit. Near these resistors there is an IC which is overheated. Is there any chance that the short circuit (if there is) is not in the IC, but somewhere else ? And how can I locate it ? I don't have the schematic."

 

[barry]

You don’t even have a schematic and you expect somebody to tell you what’s wrong with your circuit? There could be a short ANYWHERE. It could be a bad IC. It could be a bad capacitor. Without more information, this is a waste of time.

 

[KlausST]

"...My question is, whether this is an indication that there is a short circuit somewhere in the circuit.

What is a short circuit?

You already realized that any other current path causes errors.

Is 7.5Ohms considered a short circuit?
or 0.75 Ohms?
or 0.075 Ohms?
or only truely 0.00000 pOhms? (I guess we both won´t experience supra conduction in real life.)

I can say: I built a heating application where the heating element can be down to 0.008 Ohms ... and NOT treated to be a short circuit.
Many others would treat 0.008 Ohms as short circuit.
thus the answer can only be: it depends.

Also mind that any measurement wire, any conductor, any thermocouple effect and so on will also cause error on the measurement result.

Klaus

 

[kraka]

You don’t even have a schematic and you expect somebody to tell you what’s wrong with your circuit?

I am not expecting. I hope.
But very few people can give hope. But surely they are the ones who are deeply knowledgeable on any particular subject.

--- Updated Oct 19, 2024 ---

What is a short circuit?

I am not here to answer but to make questions. But although it is not give any help to my initial question, i want to answer to this question.

To answer this question one must make use of analogue thinking. In other words, he should not think of things as if they were autonomous and detached from the whole, but as parts of a greater whole. As well as he should not confuse his thinking with numbers. They say nothing: 1000000 divided by 1000000 gives 1, but 1 divided by 1 again gives 1. In this particular case, one could say that a short circuit is any situation in which any device (component) that participates in a circuit and is designed to pass through it, the current flowing throughout the circuit, is either bypassed, without being physically cut off from the rest of the circuit, but by creating a, prohibited by design, new road, or loses those properties for which the circuit designer chose it, so that it is as if it does not exist, even though the current passes through it.
Resistances of wires that used in circuits, are used as references like 0 V Ground used as a reference to rest of the measurements.
If a device or a simple resistor with designed resistance of 2000Ω, which is a part of an circuit with wires of resistance of 1000Ω , loose its property to have a resistance of 2000Ω and drop to 1000Ω, then what is the different between it and the wires ? Do we , then, have what we call "short" ?

 

[FvM]

I presume the thread is related to this previous discussion https://www.edaboard.com/threads/diagnose-bad-component-in-pci-audio-card.412458
We agreed that one or more ICs have been damaged by injecting overvoltage to PCI bus pin, showing up by excessive heating of these devices. Not quite sure what you are trying to achieve now with the described measurements, it's pretty clear that the ICs exposed to overvoltage need to be replaced.

The directly exposed IC pins may or may not show abnormal impedance, excessive current flow after damage can also occur through supply pins, the current flow out of supply rails can't be easily measured.

Regarding original question of in-circuit resistance measurement. It's possible in some cases by driving additional circuit nodes with a buffer to virtually isolate the device under test. You need respective measurement circuit and multiple probe needles contacting the board. It's relative unlikely that resistor value reduces due to overload, discontinuity or resistance increase may happen if a resistor is massively overloaded.

 

[KlausST]

Hi,

I am not here to answer but to make questions.

it was you who brought the phrase "short circuit" into this thread. But you missed to give any schematic, nor any definition:

so is this a proof that there is a short circuit somewhere

So - without any definition by you - how can we be able to answer your question?

I´m an electronics designer. And if a customer asks me to design a "short circuit detector" for him .. I surely ask him for a clear definiton of what he calls a "short circuit".
When I´m finished with my design, we test the device according HIS definition to find out whether my device is working like expected or not.

For sure - if you want to buy a truck ... and the vendor asks you wha you want to transport with it ... you can say: "I´m not here to answer questions."
It shows your attitude.

but by creating a, prohibited by design, new road,

Again: without schematic ... how can we know what path is "according design" or "prohibited by design"?
As an example from one of my circuits:
To measure currents up to 50A I used a shunt (resistor) in the range of milliohms. To be able to switch measurement_ranges I installed a 10:1 voltage divider with a total resistance of 1k Ohm. So if a smart guy measures across the 1k divider .. he will see only milliohms. How can anyone - without the according schematic - know if this is an erroneous short ciruit .. or if it is a situation according design?

Another example:
Many DVMs have a built in "continuity tester". The problem is, that there is no standard definition of what "continuity" means. Thus some DVMs beep (continuity) on a 47 Ohms resistor, others don´t. Now if one meter says there is continuity, but the other not ... it´s on you to validate the result. You, the one who does the measurement on the DUT. You, who has the information about the DUT. And although there is nobody to ask you about what you "define as continuity" .. you need to ask yourself. And you are free to tell yourself: "I am not here to answer but to make questions".

***
There are people able to say "please" and "thank you" .. and appreciate when others spend time for free to help them.
Others get added to my "ignore list". It saves my time .. and it prevents you to get annoyed by my lengthy explanations and questions..

Klaus

 

[kraka]

So - without any definition by you - how can we be able to answer your question?

It is very simple :
If you are not able to answer a question, then do not answer it.
Do not answer a question if you cannot. Why to answer ? Just for you to say that you have answer to one more question ?
If it is necessary for you (for some reason..i do not know), to answer then you could say :

"i cannot answer your question. Nobody can answer. Good bye."
It is only 1 line. Instead of this, you wrote 80 lines. It is not my fault that you waste your time like this. And i will not say "sorry" to somebody who want to waste his time for some reason...

--- Updated Oct 20, 2024 ---

No

I presume the thread is related to this previous discussion

No. It is about my NVDIA FX5500 Video Card.
But with all these new information (and knowledge) you gave me with your reply, you help me a lot.

--- Updated Oct 20, 2024 ---

Not quite sure what you are trying to achieve now with the described measurements,

I'm not trying to achieve anything. I don't need to prove to anyone, anything.

 

[barry]

It is very simple :
If you are not able to answer a question, then do not answer it.
Do not answer a question if you cannot. Why to answer ? Just for you to say that you have answer to one more question ?
If it is necessary for you (for some reason..i do not know), to answer then you could say :

"i cannot answer your question. Nobody can answer. Good bye."
It is only 1 line. Instead of this, you wrote 80 lines. It is not my fault that you waste your time like this. And i will not say "sorry" to somebody who want to waste his time for some reason...

--- Updated Oct 20, 2024 ---

No

No. It is about my NVDIA FX5500 Video Card.
But with all these new information (and knowledge) you gave me with your reply, you help me a lot.

--- Updated Oct 20, 2024 ---

I'm not trying to achieve anything. I don't need to prove to anyone, anything.

Wow, what a jerk.

 

[D.A.(Tony)Stewart]

Wow, what a jerk.

I agree but clearly an unexpected low resistance on board is a symptom of failure from anything connected. If the resistance is polarized it can mean that a protection diode is causing it, internal or otherwise.