Vichy VC99 (which is a great multimeter, btw) has both frequency and duty cycle measurements, so I decided to test my P60 dropins.
I hooked them directly to a 3xAAA holder (because it's easier to attach crocodiles to it than to a 18650 battery), and attached the multimeter leads in parallel.
Results (in the format of MODE PWMfreq/dutyCycle):
tMart Q5 #1: H no PWM / M 6.2kHz/40% / L 9kHz/9% / strobe 14.8Hz/50%
tMart Q5 #2: H no PWM / M 6.2kHz/40% / L 7kHz/9% / strobe 14.9Hz/50%
Manafont XML: H no PWM / M 123.5Hz/30% / L 126.6Hz/5%
Funny thing is, I couldn't measure the XML in low mode until I used my finger as an extra resistor in the battery/multimeter circuit. I've no idea why, but it was consistent... and I measured everything again with my finger in the circuit and got the same results.
Did I say I love this multimeter?
Two pictures of the setup (try to find which wire goes where... apparently the polarity affects duty cycle measurements [if I'd reverse the green/yellow leads it would read 70% instead of 30%] but not the frequency).
I had the same one a couple of weeks ago. It seems to measure a bit lower over 2Amp with the stock leads I'm afraid. I'm happy I now can use my crocodiles I got with my Turnigy charger. They seem way thicker than any leads I have. Haven't made the same measurements yet though.
I thought about the experiment a little bit, and was totally baffled.
It shouldn't have worked. What I was trying to measure was DC current change, but the multimeter measures frequency through voltage changes (it's parallel to the circuit), and since the flashlight is on and the driver just switches the LED on and off but keeps drawing a small current for itself, the voltage in the circuit is constant! (or so I thought)
Well, the DC voltage is indeed constant. (it decreases steadily, but does not fluctuate)
So I measured the AC voltage for the XML at the battery, and discovered: H 0.000v, M 0.120v, L 0.035v... Which means that the current ripples of the PWM process are measured as alternating current (with load - negative polarity, no load - positive polarity?, maybe that's why it's important for duty cycle readings), and that's what my multimeter actually measures. Neat!
The voltage itself is probably proportional to the duty cycle (linearly? non-linearly? I need another PWM mode to know that...) - maybe when reading AC it's trying to fit the square wave onto a theoretical sinusoid?
I guess the 0.120v at medium is strong enough for the meter to measure, but it probably has a noise filter at about 0.040v. When I used my finger as a resistor in the baterry-multimeter circuit, I probably created some extra load which translated to higher voltage. I tested this as well, and with my finger-resistor I measured about 0.050-0.060v for an instant, which was probably just barely enough for the meter to read the frequency.
All this may be total bollocks, my knowledge of frequency counters, AC and PWM is very lacking; please correct me if I'm wrong...
I was considering to buy a ddm capable of frequency measurement but i'm not sure i would really like to fiddle with all those cables. PWM actually never bothered me.
The VC99 DMM always appealed to me but i never found it under 22€. The reviews says it has junk internals but for the price difference of a genuine fluke i can live with that easily. I always wanted the capabilities VC99 provides.
I beg to disagree. Your analysis is well reasoned.
Whether it was a stroke of genius or simple serendipity to connect the leads across the battery to measure frequency/duty cycle it doesn't matter, you got a reasonable result. I wouldn't have anticipated enough signal to even attempt it this way.
This suggests that your finger wasn't acting as a resistor at all. Your body became an antenna, picking up some of the rf radiating from the battery wires.
There are two further experiments you could try.
Experiment 1: Check for inductive coupling between the circuit and the meter.
Clip the two leads of your meter together. Yes, as in short circuit. Bring those leads close to the battery wires. Do you get a frequency reading? If not, then coil the leads around a pencil and try again.
Experiment 2: Measure the pwm frequency/duty cycle directly using a light sensor.
Use a photo diode or silicon solar cell connected directly to your meter.
If you don't have either a photo diode or solar cell then any diode at all that has a transparent cover will work, such as an led. For more details refer to http://budgetlightforum.cz.cc/node/1853
Those counters aren't sensible enough to this experiment. Usually need some 'real' signal , in my Meterman 35XP is 2.5Vrms ...thought the solar panel could work but I suspect that his response is very slow to measure frequency...I'm back with the answer in a while.
Never mind, I was wrong ... The measure is possible because the the high resistance of aaa cells which make a great ripple , probably more than 2Vpp, I doubt could measure something with a big cell ...
Searching for a pocketable DMM too. UNI-T better than Xiole/Best brands? You can find them cheeper (10-13 on DX or ebay) but don't mind pay 5 bucks more for UNI-T if it's much better (still want a pocket one, I know that V97 is much better)
I'm not sure the Vichy brand is better than the UNI-T, it's just much more senisbly priced.
UNI-T have a very large selection of multimeters, scopes and the like, which all seem very good to my eyes.
There's a big fan of UNI-T in th eevblog forums (kiriakos, I think) and one model was even tested by Dave from eevblog and found to be not so horrible even though it's not a Fluke. So I guess this brand is considered okay even though it's Chinese.