This segment of the road test will start exploring some of the applications I want to use this meter for - the real reasons I applied for the road test.
I design and assemble a lot of PCBs with surface mount components. Inevitably I end up with a variable number of random unused parts of unknown value. Capacitors in particular or small footprint resistors have no markings to indicate their value, so these parts become essentially useless. I have collected a small container with quite a number of such components with the vague intention to figure out a use for them. I went as far as to purchase a set of “tweezer” probes to measure surface mount components and it strikes me these probes might allow unknown components to be quickly sorted into standard value bins.
The probes did not come with a meter so this Keysight U1282AU1282A should be ideal for this application The large display can be read quickly and the capacitance function will allow those unmarked capacitors as well as resistors to be measured efficiently I think it will be interesting to demonstrate whether this application is quick easy and useful
Measuring and sorting SMT components is always going to be fiddly but with a tweezer probe and this meter and a little practice, it is feasible.
As you may know multilayer ceramic chip (MLCC) capacitors, like the ones I am measuring in this test, change capacitance significantly based on voltage. The following graph shows they can lose over 80% of their rated capacitance near their rated voltage. It is worse for dense capacitors so a 1uf 0603 will be affected more than a 1uf 1206. This will have a significant effect on filters and other circuits requiring stable capacitance. It may be worth placing such filters before your amplifiers.
The reason capacitance degrades with voltage is that the permittivity of Barium Titanate (ceramic dielectric) is a function of voltage – spontaneous local reversals of polarization at the crystal lattice level are increasingly inhibited by stronger electric fields. Polarization reversals increase permittivity and capacitance is directly proportional to permittivity, so the higher the voltage, the lower the capacitance.
To see where this meter is measuring on the degradation curve, I will monitor it with an oscilloscope. This may alter the measurement signal, but it should give some feel for the voltage level.
The scope trace tells us a number of things:
- The wave form is a triangle wave, which means the Keysight is using a constant current source and constant current sink to generate this frequency.
- The frequency is 99.1 Hz (implying the charge period is 5 ms), the capacitance is 21.46 nF and the voltage is about 850 mV; so the constant current is about 3.65 uA (because i=CV/t)
- The peak-to-peak voltage is 856 mV but the peak voltage is 1.18 V because the waveform is offset from ground.
The signal looks to be around 1 volt which is good because it is below where 16V MLCC capacitance falls off the cliff - as shown above. Most of my my capacitors are 16V or more, except maybe a couple of 10uf caps.
But here is what a 10 uF, 6.3 V 0805 MLCCapacitor looks like on the scope.
Note even at this low voltage the slopes of the triangle waveform are not exactly straight lines. This shows the capacitance is changing with voltage. The frequency is about 2.5 Hz.
Here is a high voltage 4.7 uF plastic film capacitor for comparison – its capacitance should be pretty stable with voltage:
Note the better triangle waveform. The frequency is about double the frequency of the 10 uF measurement, as you would expect with a constant current driver.
Capacitors are often required to be reasonably accurate for desired circuit performance - this exercise shows the designer needs to be aware of MLCC performance limitations and even capacitance meter characteristics in order to obtain predictable results.
Links to other installments of this road test:
component and DC voltage measurements exploration
waveform measurements exploration
applications exploration - component sorting
applications exploration - water level measurement
Top Comments