Do DMMs in resistance/diode mode damage components?

[pigrew]

(Please pardon my inability to search if this has already been asked)

Looking at repair videos on YouTube, I often see people probing random circuits with a DMM set in resistance mode, as a way to find shorts or check capacitor leakage in circuit. My understanding is that most multimeters will apply up to 10V in both the diode and resistance modes, and sometimes try to force up to a few milliamps through the circuit.

Nobody seems to worry about damaging components with overvoltages (most chips these days use 1.8V logic levels; 10V is scary). Why is this?

Or do people buy a Zener/avalanche diode to put on their DMM in order to limit the applied voltage? Could a Zener cause the auto-ranging of the resistance mode to oscillate, so it should only be used in fixed range?

I'm thinking about constructing a bananna->bananna feed-through for my DMM that contains a few Volt zener in order to limit the DMM's output voltage.

Thoughts?

[helius]

If you connect two DMMs together, it's easy enough to measure the test voltage used by each. In resistance mode, it is usually about 1 V.
Diode mode is higher because the forward voltage drop of certain diodes, especially LEDs, is higher than that. The test voltage there can be 3-4 V.
You can't measure capacitor leakage in-circuit without a phase-correlated instrument (an LCR meter).

[retiredcaps]

For the Fluke 87V, it can output up to 7.9V DC in diode or 600 ohm mode, but at 1mA.  The Fluke 87V is about the highest voltage that I know in a handheld multimeter.

[Chalcogenide]

All the ICs include ESD protection diodes that clamp the voltage to between GND -0.7V and VDD +0.7V. The current that a multimeter can provide is not enough to damage those diodes so testing ICs in this way is perfectly fine.

[tszaboo]

I did my share of in circuit measurements for QC reasons, and I never saw anything damaged by a multimeter that way, even when using a desktop multimeter. You should note, that to damage a component, you usually need to heat (pars of it) up, so it breaks. even 10V some mA is only a few mW, barely enough to heat up anything, including internal diodes and such. 0603 components are rated for 63mW. The DMM will not send enought power to it.
When you test something under power, that is a different story. You can permanently turn on a MOSFET for example, and break something, or latch up parts of the circuit. But I would not be worried about it.

[danadak]

Yes it is possible to destroy or change JFET characteristics by forward bias of small
signal transistors. The larger Id devices can tolerate larger Ig, but small signal devices
can incur damage. Refer to manufacturers datasheets.


Regards, Dana.

[Kleinstein]

The small test current can damage only a very few really sensitive parts. Even small FETs can usually stand a gate current of 1 mA.
The more likely way to damage parts by probing with the DMM is through ESD. Some internal nodes may not be ESD protected that well.

[David Hess]

Excessive current is rarely a problem because the currents are so low but reverse breakdown is a different matter.

Bipolar transistor beta is damaged by reverse breakdown of the base-emitter junction which in some cases can occur at only 3 volts.  Operational amplifiers with super-beta or limited differential range input stages are especially vulnerable to this which is why they usually include back to back diodes across their inputs.  High beta discrete transistors are also susceptible simply because they are used in circuits where high beta and low current noise is important.  Circuits using RF transistors and saturated switches will probably not even notice.

Tunnel diodes can be easily damaged by multimeter testing.

I am surprised the Fluke 87V test voltage is so high but I assume they increased it so people can test higher voltage LEDs.  Old analog meters that used a 22.5 volt battery were notorious for damaging semiconductors.  The test voltage on my relatively old Beckman RMS225 and Tektronix DM916 (contemporary to an early Fluke 97) is 3.2 volts, just enough for an LED of that age, and 0.5 volts in low voltage mode.  Test voltage on my even older DM502 bench meter is 11 volts, ouch.

Semiconductors are not the only thing which can be damaged by seemingly innocuous testing.  I remember reading a story which might have been from Robert Pease where incoming quality assurance was testing some kind of specialized small signal transformer and ruining them all by saturating the cores or something.

[pigrew]

I just checked my 34401A and it outputs about 7V on the diode and resistance modes. The manual suggests that it could potentially output 10V.

I often deal with thin-oxide (~8 nm) MOSFETs, and I know I'd destroy their gates with more than about 3V (In fact, I purposefully do this sometimes for reliability mesaurements).

My conclusion is that probing in-circuit with an ohmmeter is a bad thing to do with most multimeters, unless the voltage could be limited to perhaps 1 V. I'm just mystified why I've seen so many people do it on YouTube videos...

Perhaps it came from diagnosing non-solid state circuits. With electron valves, there wouldn't be any voltage-sensitive components (Ok... maybe the aforementioned transformer), and the valves would act as open circuits if they were not heated up.

Semiconductors are not the only thing which can be damaged by seemingly innocuous testing.  I remember reading a story which might have been from Robert Pease where incoming quality assurance was testing some kind of specialized small signal transformer and ruining them all by saturating the cores or something.

This could have magnetized the cores? Or maybe demagnetized them? Would a saturated/magnetized core act as a rectifier? I'm remembering using degaussing coils as a child after getting magnets too close to TVs....

[David Hess]

Semiconductors are not the only thing which can be damaged by seemingly innocuous testing.  I remember reading a story which might have been from Robert Pease where incoming quality assurance was testing some kind of specialized small signal transformer and ruining them all by saturating the cores or something.

This could have magnetized the cores? Or maybe demagnetized them? Would a saturated/magnetized core act as a rectifier? I'm remembering using degaussing coils as a child after getting magnets too close to TVs....

Robert Pease's book Troubleshooting Analog Circuits mentions the story starting on the end of page 35 but I think I read about it in one of his columns before his book was published.  Pease said that some magnetic cores rely on being manufactured at a specific point on their magnetization curve and saturation can screw that up.