Rigol DM3068 vs Fluke 87V disparity

Hi all:
I am reading a low duty cycle pulsed V (<1%) into a 12V batt and the Fluke 87V (both of them that I have, 1 brand new) reads it about 1 to 2% higher than the Rigol DM3068 (brand new).
Reading a steady battery DC source shows them all matched to 3 decimal places....so something is going on with the averaging/RMS readouts.

Any suggestions on resolving this?

Update:
On the DM3068 by selecting the 200V or 1000V range..the readings match the Flukes, on the DM3068 20V range (which happens to be the auto select in this case) ....they don't.

The discrepancy is undoubtedly due to the very high crest factor of the waveform you're measuring. Most DMMs can deal with high crest factor better if a higher meter range is used than you would need for a low crest factor waveform.

See: https://en.wikipedia.org/wiki/Crest_factor

It would appear to be related to the crest. As the battery charged and the int res. dropped so did the crest and the 20V scale is now just .03V off the 200V scale that agrees with the Fluke8&V at hi res.
I guess i'll just stay with the 200V scale to be confident of what's going on.

Mosaic said:

It would appear to be related to the crest. As the battery charged and the int res. dropped so did the crest and the 20V scale is now just .03V off the 200V scale that agrees with the Fluke8&V at hi res.
I guess i'll just stay with the 200V scale to be confident of what's going on.

Click to expand...

To be fair it's not something you should really attempt to be measuring, at least not with any expected degree of accuracy.