What value is that capacitor you are about to use?
Some may take a quick look at the markings. For example, the one I’m looking at is marked “.0022 +/- 1%”
But what is the real capacitance?
That is a fair question these days. I recently built a kit that came from China–the Yosoo PIXIE QRP transceiver. (Watch for an upcoming “builder’s guide” on Amazon! Step-by-step instructions that will remind you of the manuals that came from a certain company in Benton Harbor. But I digress…)
I am in the habit of testing the components I can when building a kit. It is much easier to find a bad diode before you install it than once the circuit is complete but not working like it should. When I measured the 470 pF capacitors in the PIXIE kit, they all measured closer to 220 pF. Were they mismarked which made them available at bargain prices? After all, someone who orders a few thousand capacitors for manufacturing is likely to notice something that far out of spec. On the other hand, the typical kit builder is fairly likely to read the markings and install them. Because the values for most components in the PIXIE are not that critical, the transceiver is likely to work with maybe a slightly degraded level of performance.
But that did get me thinking about how I measure capacitors. Until a few years ago, most of my measurements were made on a Heathkit IB-5281 LCR bridge. It was not necessarily the most accurate of devices, but I could get fairly close and results tended to be consistent and repeatable.
Then I worked on a project where I needed something faster and more accurate. I ended up buying the Almost All Digital Electronics (AADE) L/C IIB. This was a popular meter that is no longer available and mine has seen a lot of use in the last decade.
But looking at the bench, I also have a Yosoo GM328 component tester, a Radio Shack 220-0075 True RMS DMM, and a DG8SAQ VNA that can all measure capacitance. Recently I have been using the Yosoo GM328 more than the AADE L/C IIB because it can also test diodes, transistors, inductors, resistors, and more. I also knew that the Yosoo GM328 was fairly accurate from some testing I had previously done. (If you are interested, you might want to check out the eBook Using the Yosoo GM328: a guide for radio and electronics experimenters by James McClanahan available on Amazon.)
I measured the capacitor I mentioned with all four devices. The AADE L/C IIB read 2.155 nF, the RS DMM read 2.228 nF, the Yosoo GM328 read 2.22 nF, and the DG8SAQ VNA read 2.21 nF.
A note about the difference between accuracy and precision is in order. If I measure something that is around 3.0 and then take the reciprocal (1 divided by that number), I get 0.333333… But that never ending string of 3s is misleading because the underlying values used to calculate it are only known to two or three digits. Some lab grade gear can measure capacitance out to several figures, but for most hobbyist grade gear you are probably better off assuming only two or three digits are meaningful (even if more are displayed).
But that aside, we still have a range of 2.155 nF to 2.228 nF. Years ago I had access to a lab with some high-end test equipment and I happen to know that this particular capacitor is actually somewhere between 2.21 nF and 2.22 nF (with the exact measured value varying slightly with frequency).
For now, let’s assume a value of 2.1215 nF which is midway between the two measured values if we toss out the high and low values. (That is also pretty close to the real value as previously measured on lab gear.)
The AADE L/C IIB work by having an internal L/C oscillator that the unknown component is placed in parallel with. The additional inductance or capacitance causes the frequency of oscillation to shift and by measuring the two frequencies (before and after), you can calculate the component value. Accuracy is going to largely depend on how accurately you can measure frequency. In this case, we are reading about 2.7% low which is respectable.
I’m not sure how the Radio Shack meter measures capacitance, but I suspect it is using a simple R/C oscillator that does not have a capacitor. It is rock stable at zero when no capacitor is installed which would be expected since the oscillator would not be oscillating in that situation. It looks like it is reading about 0.59% high which is more than respectable for a meter where the capacitance function was likely thrown in more as an afterthought!
I go into details on how the Yosoo GM328 measures capacitance in the book I mentioned earlier, but we can do a quick recap. First it discharges the capacitor, then it does a quick charge. If that “maxs out” things (i.e., fully charges the capacitor) the device knows the capacitance has a “small” value. If it does not “max out” it knows it has a “large” value. (Large and small are relative here and I’m going to gloss over the details.)
Once that is done, a second set of tests performs the actual measurement.
For a large capacitance, the device will discharge the capacitor and then charge it through a known resistance until it passes the voltage threshold on a comparator. Knowing the charging current, reference voltage, and charging time, you can calculate the capacitance.
For small capacitance, the device will discharge the capacitor and then charge it for a known period through a known resistance. At the end of that period, the voltage across the capacitor is measured. Knowing the charging current, charging time, and the ending voltage, you can calculate the capacitance.
This capacitor would be measured using the Yosoo GM328’s “small capacitance” method and the reading is within one digit (it only measures to three digits) and 0.22%. For a device costing under $30, that is extremely good accuracy! (Getting better than one digit is tough and since it stops at three digits, we can’t know if it internally calculated 2.2149 nF and then rounded to 2.21 nF, so calculating accuracy at this point is, to be candid, somewhat meaningless.)
The DG8SAQ VNA can measure the the reflected power (kind of like a really fancy SWR meter) and calculate the capacitance from that. One interesting feature of the VNA is that you get to choose the frequency used to measure the capacitance. This is more useful in radio frequency (RF) applications than at lower frequencies, but is worth noting. The accuracy of the VNA is basically the same as that of the Yosoo GM328 (with the same caveats).
The AADE L/C IIB was something around $75 as a kit if I remember right. The Radio Shack True RMS meter cost about the same new (except I picked mine up on closeout for around $10).
That leads to an interesting situation… The Yosoo GM328 is around $30 and the DG8SAQ VNA costs around $650. So the least expensive and the most expensive devices were also the most accurate.
(BTW, both the Yosoo GM 328 and the DG8SAQ VNA uses an Atmel ATmega 328 processor. The VNA also has other processors and requires a PC to run, so it isn’t an exact matchup from that perspective. But it is still kind of interesting knowing what is at the heart of both of them…)
So I will probably put the AADE L/C IIB to the side for now and do most of my quick measurements using the Yosoo GM328. For things that are more frequency dependent or where I want to double check the value, I will probably use the DG8SAQ VNA. Don’t take that as meaning the AADE L/C IIB (or the Radio Shack DMM) isn’t an excellent instrument. But if you are just starting out, the Yosoo GM328 is low priced, flexible, and (at least in my case) fairly accurate.
“With a soldering iron in one hand, a schematic in the other, and a puzzled look on his face…”