Rigol DM3068: Fried autoranger or bad design?

[spacehen]

Hi Guys,

I'm noticing some stange autoranging behaviour with my dm3068 multimeter.
If I put the DMM into ohms (with a NPLC of 1) and repeatedly measure a 21.6K Ohm resistor I get different ranges each time.

For example I might get a reading of 000.021M Ohm at times or 021.586 K Ohm..
It seems to jump to the mega ohm range quite commonly for this type of resistor value. It's even worse for a larger NPLC integration time...

The DMM is 2200000 counts so I'm not sure if this has anything to do with it.
Is this behaviour normal or should I send the multimeter back to rigol?

Some resistor values range correctly most of the time while others don't.

The reading itself is accurate but autoranging just doesn't work consistently.

[TurboTom]

I find the same behaviour on my DM3068 F/V 01.01.00.01.11.00.00 from Dec 23, 2022. I guess this relates to Rigol's general problem of unfinished firmwares and a certain reluctancy or incapability to sort things when / if they get public. If you're still within the "no questions asked return" period, I'ld say return it and get something else. The DM3068, despite being accurate and otherwise well made, should be mature by now and not showing these kinds of problems. Moreover, design-wise it's showing its age. Since it's not particularly cheap (unless you got it at a bargain), there may be other, more attractive choices available nowadays.

Happy Holidays!

[Wolfgang]

Buggy firmware, probably, despite its quite old. Never was a big seller and has SCPI bugs, too. I recommend getting a Keysight 3446X its a few hundred $ more expensive but has a graphics display, no bugs I know of and SCPI just works.

[TurboTom]

My favourite would be the Keithley DMM6500 but that's considerably more expensive than the Rigol... 

[Kleinstein]

The Sigilent SDM3065 is relatively similar, but already some graphics.

For the autoranging the 21.6 K could be just a boarderline case, an depending on noise / not immediate good contact could be in the transition region between 2 ranges. It is normal that some values can redult in 2 different ranges depending on what happend before. Some hysteresis band is needed to avoid frequent jumps between ranges. The is nothing gaines when noise or hum is enough to switch between 220 K and 0.220 M  in a fast sequence.

A 000.021 M reading still looks a bit odd as this is more than just 1 range up.

[TurboTom]

...
A 000.021 M reading still looks a bit odd as this is more than just 1 range up.

Yes, that also surprised me quite a bit. Funny thing is that after "cold" powering up the meter (mains switch at the rear), it autoranged perfectly well and very fast ten times or more in a row with my 21.77k resistor (15k and 6.8k in series). After I played a little with NPC and manual ranges and later returned the instrument to auto ranging, it failed to do so properly and showed exactly the same behaviour that @spacehen reported. I replaced the standard test leads with a shielded coax and clips at the end to eliminate external interference but didn't observe any difference. With the test resistor connected, shorting the clips and opening them again always resulted in proper autoranging, i.e. when "coming from below", the instrument correctly selected the 200kOhm range.

Didn't test with Kelvin probes, though, since they went missing recently... Guess it's time to tidy up the lab  .