Measuring Current Draw from LED ring (WS2812B)

Using Arduino General Electronics
21

Grumpy Mike was incorrect in post #3 about your LED's being AC, They are PWM'd internally.

Also incorrect (post #6) to suggest you buy an expensive AC volt meter for the sole purpose of measuring an LED ring current consumption to a hobbyist. (Again, not AC)

The suggestion of the Oscilloscope was correct. You can usually find them relatively cheap and functional on ebay for a used one, and even some of the USB based digital ones are affordable and some of those can even analyze data busses like I2C, SPI and oneWire on top of acting as a digital storage scope.

However, If you connect a large value capacitor from the LED side of your current measuring resistor to GND, it will smooth out the current through the resistor of the Pulsed Current enabling you to get an accurate reading even if your using an analog volt meter. See attached schematic, Test Point 1 is where you will connect your negative lead from your meter, the positive side of the meter goes to the other side of the resistor.
While not perfect, it should suffice to get a decent reading at least for a few of the leds on full bright. Also I'm still correct that the chart DOES prove Ohm's Law! The bigger the current draw, the bigger the voltage drop across the resistor, and the lower the voltage at the side of the resistor the LEDs are connected to.

simple_current.jpg

22

Runaway Pancake, I was not implying that you said anything about the suitability of the resistor??
I was implying that others had told me that "my method of using a 1.2 ohm resistor and a multimeter is not suitable for this task."

However, I did imply that my results with my faulty measuring method did not match your statement that the current going through the LED ring at a lower value of brightness (0-254) would be no less that that at full brightness (255).

My measuring method is flawed. So I am not saying that you are wrong, only that my flawed measuring method is not consistent with your statement, which further supports the fact that I need to change something in order to get accurate measurements.

I might try rinkrides suggestion... and let you know how that turns out.

23

I am measuring the voltage drop on a 1.2ohm resistor in series...

Your multimeter should be able to directly measure current and in the current-mode it's internal resistance is probably less than 1 Ohm. But, digital meters are not good for voltages or currents that jump-around...

In a situation like yours, I'd just go by the specs & calculations (which is what I usually do), especially if you don't trust the readings your meter is giving you. Worst case, if the calculations aren't perfect and it turns-out the power supply can't handle it, you can upgrade the power supply.

While my multimeter would not measure high voltages... what else would it do to warrant its purchase that the oscilloscope and my present multimeter could not do ?

Usually, low voltages (millivolts) are the problem. A "random" oscilloscope won't necessarily measure higher voltages than a "random" multimeter. My 'scope at work goes down to 10mV per division and that's with a 10X probe, so it actually goes to 1mV/division.

The main purpose of an oscilloscope is for "looking at" waveforms/pulses. For example, you can see your CPU clock and you can see your address & data lines switching as long as the 'scope is fast enough). You can see if an audio waveform is [u]clipping[/u].

I use a 'scope every day at work, but I don't have one at home. There a few times I've wished I had one, but it's an expense (and more expensive if you want a good one) and it takes-up space and so far I've gotten-by without one.

At work, I mostly use it for troubleshooting digital circuits and I run a few routine tests that require me to check for pulses.

You can't read 8 or more address or data lines at once and the data usually looks jittery after the first (triggered) pulse so it's not "great" for digital data, but it's better than nothing... You can check for a data or address line that's stuck-low or stuck-high and you can usually tell if a two address or data lines are shorted together by the "funny looking" waveforms. I can also check for the presence of chip-enable signals, etc.

And, it's good for checking for noise on a power supply. And, you can check to see if a clock/oscillator is running at the right frequency. (My digital 'scope at work directly measures frequency, but you if yours doesn't you can measure the period and calculate frequency.)

24

ScottC:
While my multimeter would not measure high voltages... what else would it do to warrant its purchase that the oscilloscope and my present multimeter could not do ?

This was in response to MarkT's statement that I should invest in a multimeter with a 10A current range BEFORE investing in an Oscilloscope.

I am still curious about MarkT's statement as I too think that an Oscilloscope would be more useful overall for what I do (Electronics hobbyist)??

25

DVDdoug:
Your multimeter should be able to directly measure current and in the current-mode it's internal resistance is probably less than 1 Ohm. But, digital meters are not good for voltages or currents that jump-around...

In a situation like yours, I'd just go by the specs & calculations (which is what I usually do),

Thanks DVDdoug - originally I did try to directly measure the current - with similar findings. As the number of LEDs increased, the current reading per led got smaller and smaller. Then someone suggested measuring the voltage drop on a shunt resistor... the smallest value resistor I had was a 1.2 ohm resistor... and the rest of my progress is here on this forum...

I thought it was going to be a simple task - but clearly not.
Turned to the forum for help because I wanted to understand WHY I was seeing what I was seeing... and also understand how I could get an accurate measurement myself so that I could have confidence in any statements I make about these modules.

But yes - I had no problem "using" these modules, and after reading your response, I am really keen to get an Oscilloscope.....

26

Case 1: A WS2812 4X4.
Testing one "pixel", its R, G, B, separately, measured 22mA.
Pairs (in this same pixel) whether R+G, G+B, or R+B, measured 36mA.
R+B+G measured 50mA.

Case 2: A single WS2812.
R, G, B, separately, measured 14mA.
Any/Each pair of LEDs measured 27mA.
R+B+G measured 41mA.

The controller ICs inside each pull current (500uA ea.?), too.

Current always increased when the number of LEDs lit was increased.

These results as measured with Fluke 73 III, using its 300mA function.
VOM Volt-Ohm-Milliammeter

27

ScottC:
my faulty measuring method did not match your statement that the current going through the LED ring at a lower value of brightness (0-254) would be no less that that at full brightness (255).

The current is always the same when it is ON. It's on for longer or shorter periods depending how that's set, but when it is ON it is all there.

28

Ok - I get what you are saying.

29

Yes the total current always increased in my tests too.
In my tests, while the total current increased with every additional LED illuminated, the current per LED got smaller.

What current did you measure for 1,2,3,4 LEDs? In my case, the current draw for 4 LEDs was not 4 times that of 1 LED (with each LED illuminated at same brightness and colour).

30

I could re-write my sketch to have a dark interval and get a handle on the WS2812 housekeeping current.
And then turn on RED 255 on 1, 2, 3, 4, 8.

31

Done.
(RED 255)
Results:
none on, 8mA ( total housekeeping current, 16 WS2812s )
1 on, 21mA {21 - 8 = 13)
2 on, 34mA (34 - 8 = 26) 26 / 2 = 13
3 on, 47mA (47 - 8 = 39) 39 / 3 = 13
4 on, 61mA (61 - 8 = 53) 53 / 4 = 13
8 on, 113mA (113 - 8 = 105) 105 / 8 = 13

I was not being a real hard-### with the decimal - no lives were lost.