Is it ever unsafe to measure resistance?

[SkipMorrow]

I know that a voltmeter puts some voltage into a circuit when measuring resistance, so I was wondering if it is ever unsafe to measure resistance. Not unsafe as in dangerous to humans, but perhaps to the meter or to the circuit/component? What happens if you measure resistance across an energized component, for instance? What if the component is only designed for handling a very small amount of voltage? Could the meter exceed that and damage the component?

[Richard Crowley]

Unsafe to humans: Yes, if you attempt to measure a high-voltage (>50V) energized circuit.
Unsafe to the meter: Yes, if you try to measure (for example) a charged capacitor
Unsafe to the circuit: Yes, (for example) if the measurement voltage causes an unusally large signal spike.
Unsafe to the device under test: Yes, if it is a very delicate/sensitive component.
Unsafe to other connected components: Yes, it the test voltage/current causes a destructive abnormal signal and/or bias condition.

[alm]

In a well designed meter, measuring resistance in an energized circuit within the voltage ratings of the meter should not damage the meter. In a cheap meter, all bets are off, and it may very well get damaged by measuring resistance in an energized circuit.

Measuring resistance in an energized circuit is not very useful because the voltages that are present will likely mess up your resistance measurement. In addition, the current that the meter uses to measure resistance may affect the circuits in ways it could get damaged. For example, see this thread where the continuity (resistance) measurement most likely charged the gate of a power transistor so it would turn on and allow excessive current to flow, damaging an LED.

[David Hess]

If the maximum voltage in resistance mode exceeds the breakdown voltage of a part, then it can cause damage.  Digital multimeters usually limit the output voltage so that it is below the reverse breakdown voltage of a common base-emitter junction, about 5 volts minimum, while still remaining high enough to light an LED.  Tunnel diodes were sometimes damaged in this way.  Some small signal inductors and transformers rely on their core being magnetized at a specific level and a continuity test with too high of a current will destroy them.

[David Hess]

I think most DMMs will output less than 2mA, and you can't kill anything but MOSFET gates with such current.

2 milliamps is plenty to damage a bipolar transistor through base-emitter breakdown and it happens almost instantly.  It permanently lowers the hfe increasing bias current and noise.  This does not matter so much for RF transistors, switching transistors, and power transistors that operate with low hfe anyway but it ruins a differential input stage which is why operational amplifiers which lack a cascode input or wide differential input range (lateral PNP inputs) have input protection circuits limiting differential input voltage.

Various ohms meters that I have handy:

Tektronix DM501 10 Volts 1 Milliamp
Tektronix DM502 10 Volts 1 Milliamp
Tektronix DM501A  6 Volts 1 Milliamp
Beckman RMS225 0.6 Volts 0.8 Milliamps Ohms
Beckman RMS225 3.2 Volts 3.1 Milliamps Continuity
Tektronix DMM916 3.2 Volts 0.4 Milliamps Ohms and Continuity

Newer meters have a slightly higher test voltage so they can reliably test blue LEDs.

[David Hess]

2 milliamps is plenty to damage a bipolar transistor through base-emitter breakdown and it happens almost instantly.  It permanently lowers the hfe increasing bias current and noise.  This does not matter so much for RF transistors, switching transistors, and power transistors that operate with low hfe anyway but it ruins a differential input stage which is why operational amplifiers which lack a cascode input or wide differential input range (lateral PNP inputs) have input protection circuits limiting differential input voltage.

That's an interesting phenomenon. I will take an eye on it next time I deal with ultra high hFE transistors.

My own tests show that it can knock the hfe of a general purpose part like a 2N3904 down to 1/2 of its normal value or lower so it does not just apply to high hfe transistors like the 2N5089 series.  Where it matters most is if you are operating the transistor at a low collector current.

[amspire]

What if the component is only designed for handling a very small amount of voltage? Could the meter exceed that and damage the component?

Here is a direct quote from the Tektronix S4 Sampling Head manual:

DO NOT USE AN OHMMETER TO CHECK THE SAMPLING GATE DIODES.

These are very low voltage diodes with an extremely small geometry.

[amyk]

A related fact to the discussion: modern desktop or laptop CPUs can appear as a dead short across the power rail, because their transistors are so leaky --- they dissipate a few watts at idle, and have a supply voltage ~1V, corresponding to a resistance <1 ohm. This often confuses and surprises beginners.

[Vtile]

I think most DMMs will output less than 2mA, and you can't kill anything but MOSFET gates with such current.

2 milliamps is plenty to damage a bipolar transistor through base-emitter breakdown and it happens almost instantly.  It permanently lowers the hfe increasing bias current and noise.  This does not matter so much for RF transistors, switching transistors, and power transistors that operate with low hfe anyway but it ruins a differential input stage which is why operational amplifiers which lack a cascode input or wide differential input range (lateral PNP inputs) have input protection circuits limiting differential input voltage.

Various ohms meters that I have handy:

Tektronix DM501 10 Volts 1 Milliamp
Tektronix DM502 10 Volts 1 Milliamp
Tektronix DM501A  6 Volts 1 Milliamp
Beckman RMS225 0.6 Volts 0.8 Milliamps Ohms
Beckman RMS225 3.2 Volts 3.1 Milliamps Continuity
Tektronix DMM916 3.2 Volts 0.4 Milliamps Ohms and Continuity

Newer meters have a slightly higher test voltage so they can reliably test blue LEDs.

This might be a good topic in its own to technical/project or metrology section.

My old Meratronic Meratester (2nd half of 1970s) claims 25mV 30uW at ohms/continuity range. Which is 1.5mA with a quick calc. The extra low voltage is in-circuit measurement according to manual. Is the effect current or power related?

[Vtile]

My old Meratronic Meratester claims 25mV 30uW at ohms/continuity range. Which is 1.5mA with a quick calc. The extra low voltage is in-circuit measurement according to manual. Is the effect current or power related?

25mV can't turn even 14nm transistors on, and I think that should be safe enough. However, the DMM may have stray voltage caused by capacitive coupling or ESD or whatever, that may still be dangerous to MOSFET gates.
With true 25mV 1.2mA applied to anything, I can't see a reason why it can be destroyed.

Interesting, as it goes over my knowledge.

In case of Meratester one needs to use terms AMM or FET-Multitester though, instead of DMM.

[David Hess]

My old Meratronic Meratester (2nd half of 1970s) claims 25mV 30uW at ohms/continuity range. Which is 1.5mA with a quick calc. The extra low voltage is in-circuit measurement according to manual. Is the effect current or power related?

I did not include it in the above but the Tektronix DM501A and DMM916 include low voltage ohms modes which are intended to work without forward biasing a diode or transistor junction so they can measure resistances accurately.

I think the damage is logarithmicly proportional to current and time via hot carriers and is similar to ionizing radiation damage.  The junction can be repaired by high temperature annealing which is not something which is going to happen once the part is packaged.

[poorchava]

In my day job I'm currently working on a microohm meter that can output 200A/6V. CC source can output about 1.2kW.

Judging by the number of different things we set on fire during development (either because of test or just for fun) it quite often IS dangerous

[alm]

In what kind of application would you want to dump 1.2 kW in your DUT? Materials science? Power semiconductors?

[poorchava]

Power generation industry for example. Measuring resistance of terminal connections is also a very good way to check their quality. I've also heard that some clients use that sort of tests for verifying mechanical joints. I mean in this field a typical DUT is something of a size of a railroad track.

On the other hand, objects are often so large, that source cables are like 3m long each, so like half of the power is wasted in those .

[Vtile]

I don't know where I took that 30uW (from memory) for Meratester, but it somehow started to bother me today and I opened the manual and it states 1.5uW max. in (10X ohms range <10kOhms) Also the voltage is 24mV, makes ~62uA current output.  OT though.

Poorchavas case seems to be the other end. Odd meters from Poland. 

Many times the ohms ranges are pretty vulnerable in atleast older meter designs, even the better end equipment.

[poorchava]

http://www.sonel.pl/en/katalog-produktow/electric-power-industry/mmr-650.html (no cryptoadvertising intended)

this is the (much) much smaller version of what I'm working on. Here the source is only ~110W. This is not much of a feat. What is difficult is making it work with something like a primary winding on a 25MVA railroad transformer, which has some kilohenries of inductance. Also when measuring very small resistances in order to achieve any serious precision and accuracy all sorts of stuff must be taken into account when designing that stuff: thermoelectricity, piezoelectricity and several others.

On topic about safety: guess what happens when u force like 10A through a transformer that is 5x5x6m in size in order to measure winding resistance and then suddenly remove the test lead (hint: it's bright  )