Bench multimeter current measurement (Keysight vs Rigol)

[Veketti]

Dear All,

Some time ago I bought my first bench multimeter Rigol DM3058E which is 5½ digit bench multimeter. I think some sort of test equipment bug bite me and I aquired more professional Keysight 34461A. I was comparing these two and was quite a bit amazed that when measuring currents with both I got following results at 12VDC from lab power and through 10kohm resistor.
Rigol 1.20655mA (Lab power supply Agilent 66321B showed 1.206mA).
Keysight 1.194697mA (Lab power suuply Agilent 66321B showed 1.195mA)
And 66321B with just the resistor 1.207mA.

Did I do something wrong, but is it normal that the more expensive Keysight adds so much series resistance to render the current reading useless? I'm amazed of the Rigol, as it basically didn't affect the reading.
Secondly I was wondering that the Keysight, being 6½ digit compared to Rigol 5½ show same amount of digits (when voltage is dual digit) as the Keysight shows 012.xxxx and Rigol 12.xxxx. So the Keysight loses one digit resolution by showing extra zero prefix.

[KungFuJosh]

I'm confused by your question. The Keysight reading is more accurate than the Rigol reading when compared to the lab PSU. How does that make it useless?

[DaJMasta]

A better check would be to put both meters in series with your source, then while the series resistance of the whole circuit is higher, the current through the two has to be identical, so you can see if they are actually measuring differently.

[Fungus]

Did I do something wrong, but is it normal that the more expensive Keysight adds so much series resistance to render the current reading useless? I'm amazed of the Rigol, as it basically didn't affect the reading.

The "burden voltage" (which is what you're seeing) will be listed in the manual for each - very easy to compare the specs without even measuring anything.

[Veketti]

Burden voltage is higher on Keysight than what it is for the cheaper Rigol so that's why it's affecting the reading more. Is higher burden voltage then sign of better multimeter or the opposite? Why isn't the multimeter taking the burden voltage into account when showing the reading? Or were you supposed to check the datasheet and manually calculate the real reading by taking the burden voltage into account?

[nctnico]

Burden voltage is higher on Keysight than what it is for the cheaper Rigol so that's why it's affecting the reading more. Is higher burden voltage then sign of better multimeter or the opposite? Why isn't the multimeter taking the burden voltage into account when showing the reading? Or were you supposed to check the datasheet and manually calculate the real reading by taking the burden voltage into account?

Yes. But you can use the 10A connection on the 34461A to have the lowest burden voltage. Personally I like to have accurate current readings on lab power supplies because DMMs always have a bit of burden voltage which then lowers the output voltage. Which you could compensate by using the sense lines from the PSU but that is getting messy on your bench.

[Kleinstein]

The meter is doing a current measurement. So no way to take the burden voltage drop into acount at the meter. The lab supply should take the drop at it's internal shunt into account and thus no voltage error from the current measurement there.

The cheaper meters often have some current ranges and may use the same shunt over an extended range. This can lead to ranges with rather and other with rather low burden voltage. A lower burden voltage is normally desirable, though a higher burden may allow for lower noise in the measurement in the range.
For the comparison 1.2 mA may still be in the 1 mA range (right at the edge of overrange) for the HP meter, while the Rigol meter has a 2 mA range. This alone make the direct comparison of the burden voltage tricky. The ranges are just different and thus not the worst case burden voltage at the same time.

[coromonadalix]

@klein ..   well explained    thks

I would add   it depends of the meter range too  and yess it depends on how they use the current part(s) for each range (s)

if you have 200,000k counts  or  some oldies like Keithley 196  who are 300,000 or 3,100,000  ....  etc ...  you see how  they will adapt to the display range in some ways

Now i'm using  precision shunts and or electronic current sensors / transducers with zeroing  on each start ... 

I'm not letting the meter / whatever brand or type taking the "juice" anymore, since on some, you have to dismantle it,  if the fuse is internal or on the rear ... but UN-moveable rack mount bolted on the floor    

And i would add,  i use the psu senses line near the DUT to compensate for the cable lenghts / drops etc ....  if they have it i always use them

[alm]

Everything else equal, a lower burden voltage is better. But also, if everything else is equal, the lower the burden voltage, the higher the noise. And the lower the burden voltage, the lower the accuracy. So pick your poison.