How measure power usage in a better way.

On 43oh blog there is this post
“How To Measure An Energia Application’s Power Usage with EnergyTrace”
http://43oh.com/2015/09/how-to-measure- … ergytrace/

EnergyTrace works by continuously measuring the energy supplied to the target as opposed to discrete measurement of voltage over a shunt resistor. This is done with a software controlled dc-dc converter that generates the power for the target. The pulses from the dc-dc converter are counted and the more pulses per unit time, the higher the current flow. Accordingly, even the most rapid events that contribute to current consumption are captured, although the sampling frequency is only on the order of 1 kHz.

Reading here
http://forum.43oh.com/topic/8913-using- … h-energia/

it seems that this method can be used with STM32 too and all chips that works at 3.3v

The board price is 15.99$ and shipping to Italy 7$
https://store.ti.com/msp-exp430fr5969.aspx

Energia, the Arduino-like IDE for TI chips, added some new API for LowPower like sleep (same as delay but in low power), suspend ad wakeup.

I’ve read those posts, but its not clear to me how this works.

It doesn’t look like the board you linked to is a power measuring device; Its just a TI microcontroller development board.

If you want to measure true power usage of an STM32 board, I think that building a variation of @ahull’s Pig-O-Scope would be a better proposition, i.e build something that measures current and then sum the current use over time.

The measurement of the actual current and voltage is especially important as people on this forum use a very wide variety of STM32 and GD32 based boards, and even the current drawn by a LED in a Blink sketch will not be the same from board to board as they all use different LEDs with different resistors in series

Depending upon how you view power, you may have learned the terminology differently:
http://hyperphysics.phy-astr.gsu.edu/hb … ow.html#c1

Essential to the measurement of electrical power (current turned into heat and EM radiation), would be the selection of a proper current shunt. Then the calculus of the situation is realized in measuring the instantaneous voltage drop of the precision shunt over time.

Practical guide to using AWG copper for shunt resistor:http://www.scienceshareware.com/bg-curr … toring.htm

Generally speaking, I use an old Fluke to measure the current of various uC states and then attempt to profile my software, “The Program”, into how long each state will be utilized. Then hack the program to invoke a state, measure the current, and move on to another state. It is best to put the profile (guessed / measured) into a spreadsheet so that as the program evolves you can adjust your module times. Then just sum the power over a full 100% use-cycle.

Ray

anyone know what’s inside one of these usb voltage/current monitors?
essentially can we hack it?

stephen

RogerClark wrote:It doesn’t look like the board you linked to is a power measuring device; Its just a TI microcontroller development board.

zmemw16 wrote:anyone know what’s inside one of these usb voltage/current monitors?
essentially can we hack it?

stephen

OK.

I guess the TI stuff is fine if you are using TI products.

Apart from that I suspect its probably easier just to use the INA219 etc
http://www.ebay.com.au/itm/SOT23-INA219 … 5d5bf8e081

As there is a lib for it by Adafruit

https://github.com/adafruit/Adafruit_INA219

I’ll probably order some of those modules from eBay or AliExpress, as they may some in handy.

If you want to experiment with current sensing,you might also like to take a look at the ACS712 also available on ebay. (And even if you are not interested in current sensing, Julian Ilett’s videos are worth a look).

Thanks Andy,

I have some of those hall effect based current sensors which I intend to use in for a mains power meter system, but I don’t have any of the low power DC modules.

I supposed its possible to remove the shunt from the ACS712 modules on eBay and then measure smaller currents, but I’ve not investigated the feasibility of that.

I use Adafruit’s USB current monitor board.. it has an analog signal corresponding to 1V = 1A, and is real time, not digital processed. So a scope on that signal shows current flow at the sub-millisecond level as the MCU varies processing. For example, I have an app that runs at 64Mhz with the PLL, but when it is time to do a lot of number crunching (integer and floating point), I briefly switch to 168MHz. That shows distinctly on the Saleae logic analyzer’s analog channel – where I can capture and review in time detail, for durations of seconds and many megasamples.

The initial quote seems to reflect a problem you often hit with digital designers with a less than deep analogue background.

They seem to be concerned about missing fast transient events due to sampling (which could be a valid concern but..), but there
solution to this is in effect a low pass filter that is integrating the power use for them (in the form, if I read it currently, of the
control side of a switchmode psu)…
That would of course work, as a SMPS works by energy transfer, although 1kHz transfer is very very slow.. I suspect 1khz is just their sampling not counting speed, and its a much higher (1Mhz+ with any luck) SMPS.. but the effect is the same.. integration.

However, you can create just as effective measurement using an RC filter, where the R is your sense resistor, and C is a power sense
resistor capacitance.. just tune the RC values to give you the required amount of low pass, and you will not miss anything, because in
effect you will be integrating the effects (you will also not see any fast transients, but you will not miss their contribution.. just the same
as the SMPS)..

So, just measure and possible add capacitance after your sense resistor for the required filtering, and job done, unless I am missing something major here?

I bought one of them and an MSP432 too that seems to have the same EnergyTrace tecnology and costs some dollars less.
The strange thing is that shipping costs were removed when I paid by PayPal.

The post and the video shows how to use with others TI chips that doesn’t have this tecnology on the board, it seems that the original board is used like a power source only. I will tell you.

The tecnology seems more accurate since it counts the pulses of the dc-dc converter, it is thought to measure these fast burst so I think is far more accurate than the shunt method. But I am not an elecronic engineeer and these are the words by the producer.

I will follow the video “Using EnergyTrace Technology without a Debug Session” that is here
http://www.ti.com/tool/EnergyTrace
I have downloaded CCS and said to TI that i will not use the product for military products and not sell to Cuba, North Corea and other country like these.

The board arrived last week but only today I had time to play with it.

The first link that explain HowTo measure refer to a page in a PDF and a Video. Boring.

It is more simple, I had to remove all jumpers that connect the MSP430 with the power and debug part, connect V+ and GND to the target, start CCS, go to Window/Preferences/Code ComposerStudio/Advanced Tools/Energy Trace and select USB1 in Connection, why this selection so deep buried I don’t know .
Then Push over EnergyTrace button and then, in the EnergyTrace Window select….. EnergyTrace and press Play.
But sometimes it seems that it doesn’t start recording. The default record session last only 10 seconds but you can change this.
PDF says to remove only GND, V+, RST and TST jumpers but I noticed that the onboard micro was alive until I removed RX and TX jumpers. I was unable to made it works with the MSP432 board.
The 430 board has a supercap and the example program run for hours taking internal temperature and recording it in the FRAM.

zoomx wrote:
<…>
The first link that explain HowTo measure refer to a page in a PDF and a Video. Boring.
<…>

I used a Firefox extension to download the video. It’s an mp4 video.