[time-nuts] measuring currents on USB powered devices

[Adrian Godwin]

A problem with inline measurement is that a suitable sense resistance for
measuring runtime current (perhaps at a few hundred mA) is not sensitive
enough for standby current (uA), and if chosen as a useful value for
standby current it drops far too much voltage during full power running.

An approach I've used is to bypass the higher resistance with a diode, then
use the measurements from that sense resistor only up to about 0.5V. If
above that value then the voltage over a smaller inline resistor is used
instead (perhaps via a second scope channel or voltmeter). Autoranging
voltmeters won't usually re-range fast enough to give useful measurements
of short on-periods and a scope, despite its low resolution, is often more
appropriate for such transient measurements.

The resulting 0.6 - 0.7V diode drop at higher currents is still too much
loss from a nominal 5V (unless subsequently regulated down to 3V3), but is
within the range that can be compensated by a bench supply's voltage sense
terminals, so a good quality 5V or less can be maintained while still
getting high speed measurements and a wide range of sensitivity, if the
current sense network is placed at the power supply output and the voltage
sense connected at the load.

I'm not sure why the raspberry pi is so intolerant of 5V variation given
that it does regulate down to 3v3 and lower onboard : presumably there is
some part of the system  (other than the USB ports) that requires the full
5V. Perhaps the ethernet controller ?


On Thu, Dec 5, 2019 at 3:23 AM Bob Albert via time-nuts <
time-nuts at lists.febo.com> wrote:

>  Bill, I have elected to do just that.  I ordered a few USB female
> connectors with tiny pcb and a male-to-male cable.  I will put a
> milliammeter in the dc line.  My total cost around two dollars and I use my
> own test equipment.
> Bob
>     On Wednesday, December 4, 2019, 05:01:20 PM PST, WB6BNQ <
> wb6bnq at cox.net> wrote:
>
>  Hi Jim,
>
>
> Whatever you do DO NOT bother with the el cheapo USB measuring devices.
> I bought a couple and I got to tell you they are, at best, a very crude
> indicator.  The calibration SUCKS big time.
>
>
> You would be better off using some USB breakout devices and use your own
> measurement system.
>
>
> Bill....WB6BNQ
>
>
> On 12/4/2019 8:56 AM, jimlux wrote:
> > Any recommendations or brickbats for devices that measure the power
> > drawn by USB powered devices.  I've been using a modified cable that
> > breaks out the red wire to run to a meter, but it would be nice to
> > have something that has a ADC and some sort of interface (USB?) that
> > would make it possible to log the power draw.
> >
> > I've got a bunch of wireless Beagleboards with GPS receivers and
> > RTL-SDR pods (to do phased array measurements) and I'm trying to come
> > up with better power information.
> >
> > (the recent WiFi NTP low power discussion prompts this).
> >
> > What would be great is if there were some USB "thru" pod that measured
> > this and reported the data over a USB interface (recognizing all sorts
> > of isolation issues that are possible), so I could plug the USB power
> > "to" the beagle board and the USB power to the RTL-SDR, and log them
> > both.
> >
> >
> > _______________________________________________
> > time-nuts mailing list -- time-nuts at lists.febo.com
> > To unsubscribe, go to
> > http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> > and follow the instructions there.
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>