measuring battery drain

[alpha_geek]

Hi everyone

I am trying to decide on what battery(s) to use in my robot. The robot is built from an old PC, so I would like to measure the current the motherboard (with all peripherals) uses. I have been looking all over the net, but I can't find any info on how to measure the current a motherboard is using. There is plenty of info on how to measure currents, but I am not sure where to take the measurements from. Can anyone point me in the right direction?

[Soeren]

Hi,

There's lots of ways, but I think it would be the easiest to just make an adapter cable with mating parts of the power connector, to insert right before the motherboard.
This should be nothing more than wired in power resistors, one for each voltage. The resistors should be chosen so that they drop only around 0.1V or thereabouts at the expected max. drain.

Then you measure the actual voltage drop over each resistor and use Ohms law to find the current and that's it.

You have to use eg. a benchmarking program or similar, to make sure that it's working at max when measuring, to get realistic meaurements.

If, you think the +5V line will take say 15A max., calculate a resistor for 0.1V drop at 20A (to get some overhead).
That would be 5mOhm (0.005 Ohm), which will probably need to be made from resistance wire or bought as a shunt.
If you then measure eg. 80mV over that 5mOhm, then you calculate the current to be 0.08/0.005=16A

[Admin]

Dont forget the old fasioned multimeter, power supply, or current sensor, too . . .

And battery monitors . . .
http://www.societyofrobots.com/schematics_batterymonitor.shtml

[Soeren]

Hi,

Sure, they're included in "lot's of ways"
But if you want to measure all power lines of a PC without turning it on and off repetedly, I think a homemade cable with measuring resistors is the cheapest and easiest way to go...

- make cable
- insert cable
- boot PC and start program payload
- measure voltage drops
- turn off PC
- calculate current for each voltage

- A multimeter would be either risky or time consuming, if you're referring to using the current range and either hold the leads in place by crocks or connect it up properly (you still need the meter to measure the resistors I mentioned however, but it's so much easier to connect to them, since loosing contact woldn't mean a thing).
- A power supply is what's being measured, but on several lines.
- A current sensor, hey that's the resistors, but if you think off current transformers, they would be expensive compared to resistors, especially since it's DC lines, so linear Hall current transformers are needed.

[Admin]

- A power supply is what's being measured, but on several lines.

I meant the bench top power supplies that gives you current readings . . .

- A current sensor, hey that's the resistors, but if you think off current transformers, they would be expensive compared to resistors, especially since it's DC lines, so linear Hall current transformers are needed.

Yea I think the resistor method is the quickest way to do this (and best if you dont care for perfect accuracy). Make sure its a high power resistor or it will fry!!! Mmmmm fried resistors . . . magic smoke!

So the nicer way would be with current sensor IC's, which is what I meant. But thats for more permanent solutions, such as hooking up the output to an ADC of a microcontroller . . . Probably not what you want, but its an option. I only use them when my power draw has some high frequency cycle (changes a lot) and I need to know both peak and average power draw.

[hgordon]

There are some nice IC's that are designed for this function.  Do a search on "high side current sensor" or "high side current monitor".  You'll find chips such as Maxim 4372 and Zetex ZXCT1009.