[answered] Why does Resistance decrease or increase as I'm probing this circuit?

[squiggy]

Hello,

I would like to understand this behavior: in measuring the resistance of R7 in the attached circuit, I observed the value reads differently depending on which "polarity" I place the probes. The "polarities" are marked as Config 1 and Config 2, where the '+' symbol indicates the input of the DMM and the '-' indicates the common/black input of the DMM. Furthermore, the resistance measurement changes over time, either increasing or decreasing, depending on the polarity.

+3.3V generated by a boost converter with an integrated bypass (schematic in 'dmm-resistance-boost-ckt.png'. When the boost is not used, +3V3 is fed by a CR2032 battery. In my measurement, no batteries are present.

The resistor is spec'd as +/- 5% and both readings fall within that tolerance. I observed this behavior with both 121GW and Rigol DM3058E (although there is a bit of difference in exact behavior) and I have no concern that either DMM is out of calibration (they both read a potentiometer sitting around just fine).

I also ruled out that I was applying pressure inconsistently on R27 (it is SMT 0402) by soldering some wires and using some clamps instead of sharp probes.

My best postulation for these observations is, assuming the DMMs are active as constant current sources while performing resistance measurements, that:
1. The measurement increasing or decreasing is caused by the capacitors charging or discharging
2. There is some internal path from Pin 9 (TP1) to GND or VCC within the MSP430 or another IC

Any help understanding this would be very much appreciated. My hope is to take away some 'best practices' for doing these types of measurements.

P.S.
1. I wasn't sure if this should go in the T&E section or Beginners - the example is somewhat specific but the understanding how to use the resistance measurement function properly seems like an important beginner skill
2. The reason I am measuring R7 is that I observed its value, surprisingly (and unsurprisingly not mentioned in documentation), influences the internal oscillator(DCO) and MCLK. In a small production run of this device, there was about a 25% failure rate where an output signal's timing failed specifications, and at this time I'm not convinced it's R7 causing the MCLK deviation.

[RoGeorge]

The resistance can not be measured correctly when the resistor is inserted in a circuit.  That is because the entire circuit will appear as yet another resistor in parallel with the one you wanted to measure.  The DMM doesn't know which is the resistor and which is the circuit.

To measure resistance correctly with a DMM, you'll have to unsolder at least one end of the resistor before measuring.

[jb79]

Hi,

as already stated: a correct measurement is only possible when you remove the resistor.

Regarding yourquestions:
1.) increasing/decreasing values of resistors connected to a circuit often come from capacitors within the circuit beeing charged/discharged by the measurement irself.
2.) Yes of cause. It's called protection diodes. Most pins of integrated circuits use them to protect the device itself from too high or too low voltages. Dave himself has done a video about that several years ago, that explains what's inside: https://youtu.be/2yFh7Vv0Paw?t=384

So when you put + to the reset pin and - to +3.3V you are actually measuring the 47k resistance in parallel with a diode, so some of the current will flow through that, causing the multimeter to see a lower voltage across the resistor resulting in a lower resistance shown on the display. Different readings with different multimeters are the result of different currents used for resistance measurement by that devices and the non-linear behaviour of the diode.

According to the clock: the resistance on the reset pin has nothing to do with the clock.
I suspect, you are using the internal clock, am I right? According to the datasheet, Page 28 the internal clock can between 4 and 20kHz at 3V, so "only" 25% is within spec!
If you need tighter tolerances in your project, you should consider to use another (external) clock source, if you are already using one, you should check that components (oscillator and capacitors), maybe your tolererance comes from poor quality or wrong component values there.

[squiggy]

Thanks for all your informative and straightforward explanations! I de-soldered one side and measured everything a second time correctly, and the resistors were effectively identical.

To close the thread, I also learned that for the chosen MSP430, the internal oscillator is only calibrated and guaranteed for specific frequencies (including but not limited to 1MHz and 8MHz).