I just came across an interesting “real world” example of multimeter burden voltage. I have a nifty USB voltage/amperage tester that I was using to measure the voltage/current/power draw of my iPhone 5s connected to one of those rechargeable battery power packs, and I kept seeing the phone, after some time, would report “accessory not supported” and stop charging. For the life of me, I couldn’t figure out why ((this is largely because I wasn’t paying attention to the meters all the time, so I didn’t notice the pattern right away)). The setup was running the battery pack into the USB tester, then to the phone. The USB Tester was plugged into my two Keithley 196’s, one measuring voltage, the other, current. Both were set to auto range ((you might see where this is going now, if you’re paying attention)). When you first plugin the phone, it would draw about 3/4 of an amp. It would then slowly draw less and less as the charge got higher and higher on the phone. Then at some point, it would stop charging.
Finally, I decided to watch when it happened, and then the burden voltage idea hit me. It would stop charging when the current usage dropped below .3A, and the meter would autorange down to the mA range. According to the Keithley 196 manual:
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.
The key part of all of that is “full scale”. So, basically, at the 3A input scale, with only 300mA going through, I’m only dropping 1/10th of the 300mV, or 30mV, which means the voltage to the iPhone is still 4.97V, well within the USB electrical spec (4.75V to 5.25V). At the 300mA range, though, I’m dropping that full 300mV, which drops me to 4.7V to the iPhone, and outside the spec ((Thanks to suicidaleggroll on the eevblog forum for pointing out the obvious.)). So, while not the typical burden voltage you get on some handheld meters that have a mV/mA rating (the Fluke 87V is rated at 1.8mV/mA on the 400mA scale), it’s still a good example of where a multimeter burden voltage can lead to some interesting issues that may not exactly be obvious at first glance (especially if you’re not paying attention to WHEN the issue occurs).
So, the long and short of it, don’t leave your current reading DMM on auto-range. Especially when you don’t care about that extra digit (do I really need to see µA being drawn?).