a strange problem

Assuming it's a double layer PCB with vias... do you have gnd and vcc somewhere on an edge of the pcb? maybe there was a hangover from the plating process giving you a low impedance path between the layers, maybe even a few microns. you've blown this away when you applied power.

I've done similar with a home-brewed pcb where there was a remainder from the etching process. 12V @ 8amps did the trick - from short to open in 2us ;)

Heinz

Dear Heinz,

Thanks for sharing.

One more thing.

Will it be safe to use it now?

Might it occur intermittent issue in the future?

Regards

A capacitor round the wrong way??

Relay catcher diode round the wrong way??

--
Unless you know for sure what caused the fault and have repaired it
successfully, the answer is a resounding "NO".

Both cap or diode round the wrong way should have been noticeable by drawing a higher current (maybe for a few moments) and, eventually the occurance of magic smoke ;)

Heinz

Ohmmeter polarity?

John

:Hello.. : :I found a strange problem while working on a typical PCBA. : :It is a kind of pneumatic control PCBA. The PCBA uses 24V power supply :from which it draws current from 1 to 3A. : :Before its operation, normally we check the resistance measurement :between 24V and Gnd, which should be around 2 kilo ohms. : :For a particular PCBA, the measurement showed 30 ohms, we rejected it :for further analysis. : :After a week or two, I study the PCBA to determine what was causing :the problem. Accidentally I supplied 24V to it and found that it was :working normally. : :When I disconnect the powersupply and measure again, I found that it :showed higher than 2k ohms at that time. : :I visually checked it thoroughly to see if there were any defective :capacitor along the pcb traces but found no problem at all. : :I kept it for at least three weeks, I have tested it in operation for :more than 5 times, and until now there is no problem in the operation :of that particular PCBA. : :I wonder what could cause the resistance measurement to 30 ohms and :then higher than 2 k ohms after supplying voltage to it. : :Any idea? : :Regards

I'll bet you used a digital VOM to measure the resistance...

I have found a similar symptom when using my high quality digital VOM to measure resistance where there may be a multiplicity of components connected between the measuring points. The damned reading was way out from the expected value. Solution, I dragged out my old analog meter (AVO model 8) and now the value was as expected.

Some digital VOM's don't behave as they should.

measure

the

Thanks for sharing.

I also think that there is something wrong with my digital VOM.

On the other hand, i need a realistic and logical explanation for it so that I could put in my report to explain to everyone.

There has been a lot of capacitors installed along the power rails to keep them away from different ripple frequencies.

Accidental charge stored in capacitors could make the digital VOM readings to go wrong.

Correct me if I am wrong..

Regards

[concerning a PCB that had abnormal 30 ohm resistance across power until it was powered-up once]

I've seen hair-thin failed-to-etch blemishes that would account for this; the original PCB handled some high current, so might have been thicker copper than most, and those always take annoyingly long to etch to completion.

If it's not going to be a human-shock hazard, I'd say the board is fine. If there COULD be a shock hazard, more testing is required (HiPot type testing).

: :> I'll bet you used a digital VOM to measure the resistance... :>

:> I have found a similar symptom when using my high quality digital VOM to measure :> resistance where there may be a multiplicity of components connected between the :> measuring points. The damned reading was way out from the expected value. :> Solution, I dragged out my old analog meter (AVO model 8) and now the value was :> as expected. :>

:> Some digital VOM's don't behave as they should. : :Thanks for sharing. : :I also think that there is something wrong with my digital VOM. : :On the other hand, i need a realistic and logical explanation for it :so that I could put in my report to explain to everyone. : :There has been a lot of capacitors installed along the power rails to :keep them away from different ripple frequencies. : :Accidental charge stored in capacitors could make the digital VOM :readings to go wrong. : :Correct me if I am wrong.. : :Regards

That guess is as good as any. Without knowing the design details of every different type of digital VOM all one can conclude is that it happens for some unknown reason. If an analogue VOM shows the reading as expected then you can assume that the blame is with the digital VOM.

Go read John Larkin's reply and answer his question before going further. Forget the crap about the digital meter being wrong. You would be far better off learning what the meter is actually measuring and how it does so, than assuming it is somehow magically wrong because it is digital. In fact, unless it is broken, it doesn't read wrong - you just don't understand what it is showing you.

Ed

Good answer, Ed. I have been known to shout at trainees sometimes... "Look at the 'scope, what is it telling you?" :-(

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

If we knew what we were doing, it wouldn\'t be called research...
                    -- Albert Einstein

On Sun, 04 Jan 2009 21:26:38 GMT, ehsjr wrote:

:Myauk wrote: :> Hello.. :> :> I found a strange problem while working on a typical PCBA. :> :> It is a kind of pneumatic control PCBA. The PCBA uses 24V power supply :> from which it draws current from 1 to 3A. :> :> Before its operation, normally we check the resistance measurement :> between 24V and Gnd, which should be around 2 kilo ohms. :> :> For a particular PCBA, the measurement showed 30 ohms, we rejected it :> for further analysis. :> :> After a week or two, I study the PCBA to determine what was causing :> the problem. Accidentally I supplied 24V to it and found that it was :> working normally. :> :> When I disconnect the powersupply and measure again, I found that it :> showed higher than 2k ohms at that time. :> :> I visually checked it thoroughly to see if there were any defective :> capacitor along the pcb traces but found no problem at all. :> :> I kept it for at least three weeks, I have tested it in operation for :> more than 5 times, and until now there is no problem in the operation :> of that particular PCBA. :> :> I wonder what could cause the resistance measurement to 30 ohms and :> then higher than 2 k ohms after supplying voltage to it. :> :> Any idea? :> :> Regards : :Go read John Larkin's reply and answer his question before :going further. Forget the crap about the digital meter being :wrong. You would be far better off learning what the meter is :actually measuring and how it does so, than assuming it is :somehow magically wrong because it is digital. In fact, :unless it is broken, it doesn't read wrong - you just don't :understand what it is showing you. : :Ed

Well the evidence of what I found doesn't support your claim that it will only happen if the meter is broken. My meter (Gossen-Metrawatt Metrahit 25S) is fine and continues to be fine. The erroneous reading I got was obviously incorrect and I tried reversing the digital VOM polarity and eliminating every other possible meter fault as well without any difference to the erroneous reading.

I would ask the OP if the same digital VOM exhibited the same result with the specific unit under test every time he measured? And also whether he tried the same measurement with an analog meter on this unit? If he did, what was the result?

It is hard to argue with results proven by empirical method when all other logical explanations have been exhausted.

A bit of Googling turned up this

While this article refers to PC based plug-in multimeters one would imagine that the PC card uses a similar technique to measure resistance as do hand held or bench VOM's (ie. CC).

See the paragraph System Considerations For Resistance Measurements.

Agilent also recognise the problem (Hint 2)

Capacitance in parallel with the resistor being measured can cause erroneous offset measurement in a digital VOM.

The erroneous measurement may also be dependent upon the combination of resistance and capacitance involved and the particular VOM being used. If it is auto-ranging while endeavouring to come up with a valid reading may also cause it to produce an erroneous reading.

That's the problem when you try and use a multimeter ohm range to measure the "resistance" of an active circuit. It's not actually the "resistance" you are measuring, it's just what the multimeter thinks it is for a given test current. The figure will vary based on the type of multimeter you use, the probe polarity, previously charged capacitors, temperature, and the phase of the moon.

Dave.

measure

the

was

There is almost certainly nothing wrong your VOM. Go and measure your circuit with different brand VOM's and you'll almost certainly get 3 different values, then try the probes backwards and that'll really screw you up. Then measure just a resistor on its own with the same meters and they will all read the same value.

What you are trying to measure is non-linear and not purely resistive, that's your problem. This is not something a resistance range of a multimeter is designed to measure accurately or at all.

Dave.

that

Sorry Ross, but that's all bullshit. If you connect the DMM set to measure resistance of a circuit with a capacitor in parallel with a resistor and are mislead by what you see on the meter, that is *not* the meter's fault. It's yours. You seem to think that the meter should show you the value of the resistor, and that, if it doesn't, it's wrong. It's not wrong, you are, for not understanding what you are measuring and how it can affect the reading.

Suppose there is a resistor in the circuit and you connect your DMM, set to measure resistance, across it, not knowing that the circuit is applying a voltage across the resistor. Do you expect to get the correct value for the resistor displayed on the DMM? When it is not, do you blame the DMM for being wrong?

Ed

: :Sorry Ross, but that's all bullshit. If you connect the :DMM set to measure resistance of a circuit with a capacitor :in parallel with a resistor and are mislead by what you see :on the meter, that is *not* the meter's fault. It's yours. :You seem to think that the meter should show you the value :of the resistor, and that, if it doesn't, it's wrong. It's :not wrong, you are, for not understanding what you are :measuring and how it can affect the reading. : :Suppose there is a resistor in the circuit and you connect :your DMM, set to measure resistance, across it, not knowing :that the circuit is applying a voltage across the resistor. :Do you expect to get the correct value for the resistor :displayed on the DMM? When it is not, do you blame the :DMM for being wrong? : :Ed

What you are suggesting is that a resistance reading taken with a digital meter can not necessarily be relied upon. And where an unexpected reading does occur the user should completely analyse the system undermeasurement in order to determine why the expected resistance measurement is not being returned. That is just ludicrous.

A technician using a digital meter to measure a specific resistance combination should have no reason to suspect that there should be a significant difference in the result comapred to using an analog meter. The fact that there may be some parallel capacitance will have no effect on the analog meter reading (once the capacitance is charged) and the analog meter reading will be accurate.

A technician has every right to expect that a digital meter will also present an accurate reading without having to analyse whether or not any particular value of capacitance might be present to upset the reading. The fact is that digital multimeters, being sampling devices, can be upset by a certain combination of resistance and capacitance, but is it wrong to say that where an unexpected result occurs, the fault lies with the user because he has failed to analyse what might be upsetting the meter? Of course not. The digital meter is just a measuring tool the same as the analog meter and the technician should not have to be conversant with the specifics of the design of the two items in order to determine whether a reading is correct or not.

The user has every right to expect the same result (within reason) for the same measurement whether using an analog or digital meter, "particularly for resistance measurements". If the digital meter produces an unexpected result and the analog meter doesn't then where does the fault lie? The meter producing the unexpected result is obviously "wrong", despite any excuses, legitimate or not, which might be made to explain its erroneous measurement.