Strange discrepancy in transistor measurents

[camurgo]

I have a bunch of PNP and NPN transistors here (CBC327-25, C559B, BC560b, BC549, CBC328-25, BC640, BC546b, C550C, BC547b, BSR50, C32725, C32716, C556B, among others).

I also have two Fluke multimeters (87-V and 289), and several Chinese brands (Peakmeter PM8236, Borbede BD-168B, Zotek ZT219, Elecall MK72).

While measuring base-emitter and base-collector resistances, consistently across all transistors I get around 0.5 MOhms with the flukes and around 6 MOhms with the Chinese brands.

Anyone knows why that is?

[iMo]

It could be the meters use a different current while measuring the R.

[Signal32]

It is not strange at all, it is very expected.
You are basically measuring a diode by measuring B-E of a transistor.
The meters are using different voltage to measure the resistance.
Put one meter in resistance mode and the other in voltage measure mode and connect the leads to see what voltage they are using.
So, using a higher voltage will end up showing less resistance, because of this:
http://www.nlreg.com/diode.htm

What the meter is doing is applying a voltage, and seeing how much current is flowing.

[David Hess]

Meters usually apply a current and measure the voltage.

PN junction "resistance" is strongly non-linear with current so different test currents result in different resistance measurements.  The voltage drop changes by 60 millivolts per decade change in current so for example:

600mV 10mA = 60 ohms
540mV 1mA = 540 ohms
480mV 100uA = 4.8 kohms
420mV 10uA = 42 kohms
360mV 1uA = 360 kohms

Some meters have a "low ohms" mode where they use a very low test current and low input range so they can measure resistance which are in parallel with PN junctions without activating the PN junction.

[bson]

It makes no sense whatsoever to measure non-constant resistances with a DMM.  It can be useful for semiconductors when you know what the exact current is, especially if the current can be adjusted.  For example if you're experimenting with biasing a diode as a switch, or measure channel resistance vs gate voltage on a MOSFET.

[David Hess]

It is not quite useless; it can test for shorts and opens and sometimes bad leakage.  With some practice, the emitter and collector can be distinguished.