1
\$\begingroup\$

I'm building a simple circuit with an Arduino Nano board, which will read signals from a Hall sensor and periodically send pulses to a small 5V solenoid. I'll drive it from 6xAA batteries in series, producing 7.2V. The on-time of the solenoid will be very short, and with a small duty cycle that will vary quite a bit (say 1% to 10%). I find here that the solenoid consumes 240mA at 5V if always on, and I guess the board and other bits-n-pieces would add another 50mA?

So, I'd like to be able to measure the current consumption over a set time period, such as a minute or five minutes - with an end goal of reducing the ongoing consumption (I'd like to use 6xAAA batteries, but AAs may be necessary, depending on how low I can push the current draw). However, with the current varying wildly over time, it is hard to get a static readout from my digital ammeter (it takes a good few seconds to settle) and they generally will change frequently (<0.1 sec).

A bit of research shows that I need an integrated ammeter, i.e. an ammeter that integrates over time as here. This question here is interesting, but answers recommend building such a device with an Arduino! I wonder, can such a thing be purchased inexpensively, ready-made?

\$\endgroup\$
4
  • \$\begingroup\$ Oscilloscope is what you actually need. \$\endgroup\$ Commented Mar 5, 2013 at 1:45
  • \$\begingroup\$ @Nick, thanks. I wondered if that was the case, and searched extensively for such a thing, but suitable resources came there none. Would it be possible to use a cheapie handheld scope for this purpose? I'm new to this hobby, and don't want to spend a fortune just yet :) \$\endgroup\$
    – halfer
    Commented Mar 5, 2013 at 8:43
  • \$\begingroup\$ (Happy to hear feedback as to the appropriateness of the question, incidentally. On StackOverflow, we like to see prior research, which hopefully my question above amply demonstrates). \$\endgroup\$
    – halfer
    Commented Mar 5, 2013 at 8:54
  • 1
    \$\begingroup\$ This does look like a good question to me, many many people find they need to measure their curent draw. \$\endgroup\$
    – Kortuk
    Commented Mar 5, 2013 at 14:19

2 Answers 2

2
\$\begingroup\$

I agree that the oscilloscope is a better solution for this sort of thing.

If you want to measure current consumption over periods of a few seconds or minutes, though, you could try building something using a coloumb counting IC like the DS2780. It may be possible to use it for something like this, given that measuring consumption accurately is it's primary task. Some creative repurposing may be involved, since they're designed to track battery energy status and therefore have registers and outputs and algorithms designed for that sort of use. Still, though, you'd need to use an arduino or such to interface with it anyway.

\$\endgroup\$
2
  • \$\begingroup\$ Alright, many thanks Chintalagiri. I'd have thought this would have been built into a multimeter of some kind, but both my research and your answer suggest not! I'll accept this in a day or two, to give others a chance to answer if they wish. \$\endgroup\$
    – halfer
    Commented Mar 5, 2013 at 10:39
  • \$\begingroup\$ There probably exist meters that measure this. I have never come across any that is inexpensive, though to be fair I've not looked very hard either. There do exist fairly high end power analysis equipment that also do this, along with other things. Such as programmable power supplys with built in average current readouts (usually fixed at 1 second or so) \$\endgroup\$ Commented Mar 5, 2013 at 11:05
1
\$\begingroup\$

I've often thought that I'd like to build a device for this purpose, since it does not seem like it should be difficult to produce something with finer resolution than many current-measurement ICs. Indeed, if one uses a switching supply in a flyback or boost configuration, the supply voltage and pulse length are relatively constant, and if the coil current drops to nothing on each cycle, then each pulse should transfer roughly the same number of electrons. If one doesn't need the output voltage to be absolutely stable, one could simply have a circuit which, 100,000 or so times per second, checks whether the output voltage has fallen below a threshold and, if so, kicks the coil with a pulse and bumps a counter. Calibrate the circuit by counting the number of pulses per second with a known load, and one should then be able to measure charge consumption on a much finer time scale than would be allowed by the charge-counter chips I've seen (e.g. on a circuit which wakes a processor once per second, it may be helpful to know how much current is used on individual wake-up cycles--a level of precision far beyond anything that coulomb-counter chips offer).

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for your thoughts. My use-case probably only requires sampling every 10ms or so - I've found the solenoid activates even with a 30ms pulse. So, not a particularly demanding requirement - I'd have thought the multimeter manufacturers would have made such a thing already :) \$\endgroup\$
    – halfer
    Commented Mar 6, 2013 at 10:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.