2008-10-26
Inductor and Capacitor Measurement Fixture
While I've built something similar in the past, this is a general purpose version for use with values of inductance and capacitance you tend to find around the bench. The circuit is almost exactly the Colpitts reference design in EMRFD (page 7.12, figure 7.24), except I changed a few of the component values a little.
I built the unit in a small diecast Aluminium box. Two polystyrene 1 nF and a NP0 12 pF capacitor were used for the oscillator. The calibration capacitor (~ 1 nF WIMA cap) is internal to the unit, with a switch used to activate it when needed. I measured it at 974 pF and have labelled the unit with this value. I only used 40 turns on a T68-2 core using 0.5 mm wire, but waxed them well to the core and ruggedly mounted the toroid in place. All the oscillator and buffer components were built point-to-point hanging off the banana jacks and switch posts. The resulting oscillator is extremely stable despite using a type-2 core, it sits for hours with < 1 Hz stability near 2.2 MHz.
I've written a calculator to speed usage of the unit. As I've mentioned previously, this is a problem I've wrestled with and failed to solve in the past, but with some dedication I finally got the problem out. The need to measure inductor true-inductance and distributed capacitance for other projects took me in a direction that turned out to be a much better approach, the solution was actually quite trivial and should have been apparent much earlier. BTW: This method measures the apparent inductance, which can be in error if the coil has significant distributed capacitance. You can utilise the internal calibration capacitance along with the distributed capacitance calculator to approximate the coil true inductance.
My unit has a calibrated inductance of roughly 10.6 uH and a capacitance of 514 pF - this compares very well with what was measured of the components out of the circuit and the expected strays. Comparisons with inductors and capacitors measured via different means suggests good accuracy and repeatability. I currently use a short piece of 4 mm wire as a "zero" inductance reference. I can measure the difference in its inductance based on how far I push it into the unit - the measurement is consistent with the inductance of a straight piece of wire - I was pretty impressed by this level of resolution. I'll be building a matching test jig that plugs into the banana sockets and enables SMD and other small devices to be measured and allow nulling with extremely short lengths of thick wire. I expect I'll use it more to measure inductance than capacitance as I already have a digital meter for capacitance but this unit's resolution is superior. It would not be extremely difficult to write some MCU code to implement the calibration (reed relay maybe?) and metering giving a direct-reading meter - might have a go at that in the future (Yes I know you can buy a kit, but where is the fun in that).
Update
I've since replaced the WIMA calibration capacitor with a carefully measured parallel combination of several ceramic NP0 capacitors totalling 1017 pF (and relabelled the unit). This has improved the calibration accuracy of the unit slightly.
I also added a pi LPF choke to the DC power inlet as I noticed a ~3 Hz shift in frequency when I choked the DC power lead externally with a ferrite core and could detect a small amount of RF floating on it with my wavemeter. With the internal filtering this is no longer apparent. (The filter is simply a few turns on a ferrite bead with 100 nF caps either side of it to the lug of the RCA plug.) The DC voltage from my bench PSU is stable enough, but an internal regulator might be advisable if you are going to make lengthy measurements following a single calibration from a battery supply. The pushing from the supply voltage isn't huge and would probably be lost in the experimental error for most measurements however.
4 comments.