Burden voltage?

[staze]

All,

Got a bit of a puzzler for ya all, and while typing out a post about it being burden voltage on a meter, I realized it may actually be the resistance change of the meter measuring current. So, here we go. And maybe this belongs in the test equipment section, but I figured this was more a question regarding Burden Voltage rather than the test equipment specifically.

Was measuring voltage/current/power draw of an iPhone over USB. Using a simple inline USB Tester (https://www.tindie.com/products/FriedCircuits/usb-tester-20/). Setup was two Keithley 196's, one for voltage, one for current. Both set to auto-range.

So, when plugged in at first, iPhone would draw about .75A, then as the SoC of the iPhone batt would increase, the draw would lessen. Eventually, the draw got to below .3A, and the meter would autorange to milliamp range, and the phone charge would stop saying "unsupported accessory". This could be fixed by turning off auto-range, and dropping it back to the 3A range.

Now, my thought is that this was burden voltage, but according to the Principals of Operation for the Keithley 196, and I quote:

The resistor current shunt network R28 is configured so that a full scale current input will result in a 300mV drop across the network on all current ranges. For DCA, this voltage is routed to the multiplexer through analog switch U24B. For ACA, the signal is routed through U21D X10 amplifier U28B. The amplifier signal then travels through analog switch U2lB and buffer U26B to the TRMS converter. The converted DC signal is then routed to the multiplexer.

So, I'm a bit confused by this. It sounds like there should be no real change due to the range changeover, but obviously there is. What's the deal?

[suicidaleggroll]

The current shunt network is sized so that a full scale current input will result in a 300mV drop.  When it's on the 3A scale, you're only running it at 1/4 of full scale, which means you'll only get a <75mV drop across the shunt (300mV @ 3A = 75mV @ .75A).  When it switches ranges at 0.3A, you're now running the full scale, and are getting the full 300mV drop (same as you would have got at 3A on the 3A scale).

[staze]

of course. So my supply is dropping to 4.7V or there abouts, and the phone is saying "something's wrong".

Hmm... that makes me wonder what the acceptable charging voltages are for the iphone. =P

[staze]

And a quick google later:

iPhone serires (all) are designed to charge with USB plug. It means, they are Power V-bus devices, and official USB specs will rate from 4.75V to 5.25V the voltage range, at max current output (500mA).

http://www.beyondlogic.org/usbnutshell/usb1.shtml

So yes, 300mV drop would put me outside that spec. Yay burden voltage!

[ejeffrey]

The actual voltage to the iphone may be even lower.  The port output might already be on the low end of nominal, including a voltage drop on a polyfuse or other current limiting device, and the cable has some voltage drop as well.

[staze]

Indeed. But, measuing the actual voltage AT the DUT is a big more burdensome *rimshot*

[Rick Law]

Mostly, I found DMM current measurement a waste of time.  I have a pack of 0.1ohm 3W 1% shunt.  I use just one when current is low or when voltage sensitivity of the DUT is high.  I use a pair in parallel if I want to minimize the Vdrop or when current is high.

Just over the weekend, I had to find out why my daughter's tablet is not charging.  I got a male and a female USB connector from my parts-box and quickly did a breakout...  With the shunts, I know at the 2A max the charger supplies, I am look at max Vdrop of merely 0.2V.  That is less than the typical variation across different USB ports.

I think a simple shunt may be a better choice for mA or A (not uA) - but of course there is the  "input protection issues".  For me, it would be well worth the trade off of lacking protection.

[free_electron]

the problem with autoranging is that they switch sense resistors around. this changes burden voltage as well as temporarily cuts the loop.

when doing precision current measurement a meter should always be set in fixed range, or a specifically designed machine should be use ( agilent makes special power sources that can measure with high dynamic range without perturbation of the loop )

[staze]

the problem with autoranging is that they switch sense resistors around. this changes burden voltage as well as temporarily cuts the loop.

when doing precision current measurement a meter should always be set in fixed range, or a specifically designed machine should be use ( agilent makes special power sources that can measure with high dynamic range without perturbation of the loop )

Yeah, this makes sense. And as noted, I switched out of auto, and back to the 3A range, and things worked again. Just never thought of it being an issue. *shrugs*