Rigol DM3068 - Overvoltage bug

[salvodi89]

Hello everyone!
Today I noticed something strange with my DM3068...
Look at this:



The signal generator output is a sine wave 1kHz 100mVrms 1V offset. (Different frequence, same result)

The multimeter was set in manual range mode at 200mV.
Yes I tried to restart the multimeter but the value measured was the same.

I can confirm the offset value becouse I measure with my HP 3478A before the fuse blow, but I haven't a photo showing the measure!

[olsenn]

You are on DC voltage measurement mode, so the 0.98V reading in the bottom photo looks valid considering you have a 1V offset. I'm not sure if the DMM has a low-pass filter to eliminate the AC component in DC mode, but if not, you'd have essentially 100mV of noise.

The top photo seems to be measuring the AC voltage even though you are in DC mode as well. You say the frequencies were different? I believe the DM3068 has a bandwidth of 100kHz, but I could be wrong... not that this should affect the DC measurement.

[salvodi89]

You are on DC voltage measurement mode, so the 0.98V reading in the bottom photo looks valid considering you have a 1V offset. I'm not sure if the DMM has a low-pass filter to eliminate the AC component in DC mode, but if not, you'd have essentially 100mV of noise.

The DC offset value of 980mV was confermed by another multimeter!

[amspire]

I think the Rigol meter has a protection input clipping or clamping circuit or design quirk that limits the peak negative voltage  to - 100mV * 1.4142  + .1093183V = 250mV. This protection circuit can have the effect of offsetting the input voltage to keep the negative peak to -250mV. If this is so, it should correctly go to overvoltage on the 200mV range when the AC amplitude is below 35mV rms.

If you put pure overvoltage DC, even with the input clamping, you get -250mV on the 200mV range so it overranges. With the 100mV AC added, the input clamping shifts the voltage to the A/D so that the negative peak is -250mV, and the average then becomes -.1093138V.

This might be a protective feature, or some quirk of the A/D circuit.

If this is so, then what you have learnt is that this meter does not have much ability to handle noisy DC  - at least on the 200mV range. So the correct way to use this meter if you suspect noise may be on the DC is to always switch to the next range up and compare the results.

The HP meter probably has a great dynamic range in its input circuitry, so it behaves more predictably.

Richard.

[KedasProbe]

Hi salvodi89,

(I probably should not start a new thread for this)
Could you check two other (small) problems of the DM3068.
In the Youtube video review from thesignalpath of a year ago 2 things are said to be corrected in the next firmware.
- "The non-latching will be updated with firmware"
- Diode test, driving current through a green or blue led shows "OPEN" instead of the voltage.

Do you know if this was fixed/changed, do you know when they released the last firmware?

Thanks!

Edit: the new page 2-47 of the manual adds current source selection for diode and corresponding voltage range and clears that up. That's a cool thing.

[KedasProbe]

I checked it and I can confirm that the auto-range works on the displayed value, not on samples(?) in the measurement that are clipping. In DC range 2.0 it clips on 2.5 and increases DC range on 2.2 (leaving a margin)

So if you measure DC of a (DC+AC) signal with AC that is bigger than 25% of your range then you should not trust the auto-range of the DC setting and increase it manually.
If the AC signal is that big you could check AC first so you know the best range on DC but it's much faster to just increase the range and see if the value stayed the same.

Anyway I will ask them if they can improve their auto-range code for DC. (probably not but it can't hurt to ask )

P.S. Continuity test didn't change since the video review.