R&S RTB2004 - adds an offset for some vertical divisions - an interesting bug

[javadesigner]

This is an interesting bug with the R&S RTB 2004   

1. Set any channel to 0.1: 1 attenuation (not 10:1 or 1:1 but custom 0.1: 1). Plugging in a probe is not necessary.

Edit: By the way, this attenuation isn't theoretical - it is necessary when using a current sensor probe that sends 10 mV for every 1 mA it reads.

2. Set Vertical position knob to control divisions (not offsets). Meaning, the knob moves the line by +/- divisions (upto +/- 5 maximum, since the
total divisions are 10).

The batshit crazy begins here 

3. Set Vertical scale to be 20 mV / div Move the vertical position knob down. The scope adds an offset after the line moves past 0.6 divisions and the display goes all wonky. 
4. Set Vertical scale to be 10 mV / div. Move the vertical position knob down. The scope adds an offset after the line moves past 1.4 divisions and the display goes all wonky. 
5. Set Vertical scale to be 5 mV / div . Move the vertical position knob down. The scope adds an offset after the line moves past 2.4 divisions and the display goes all wonky. 

And here is the kicker.

6. Set the Vertical scale to be less than 5 mV / div  - say 2 mV / div, 1 mV / div, etc. You can once again move the full +/- 5 divisions as expected and no offset is added!
7. Set the Vertical scale to more than 20 mV / div - say 40 mV or greater. You can once again move the full +/- 5 divisions as expected and no offset is added
 
Only for 5, 10  and 20 mV does the scope add offsets when you move the line by a few divisions. Offsets and divisions are totally not the same thing, so for the scope to be adding offests willy-nilly on its own - really sounds totally insane (especially since this does not happen for lower or higher values, only those three).

[javadesigner]

Uploaded a quick video file that shows this issue for 5 mv / div. You will see as you slide the line down, that it splits into two because the second dotted line is an offset value the scope is adding (wrongly). Again, this only happens for 5 - 20 mv / div, not for anything less or more (and probe attenuation set to 0.1: 1).

This dotted offset line should not appear at all since at no point I am entering any offset value.

https://www.dropbox.com/s/yrlm75xht3o4ejz/rtb2004-1.mov?dl=0

[javadesigner]

Here are another longer video I uploaded.

https://www.dropbox.com/s/5mthgm0f4n473z2/rtb2004-2.mov?dl=0


Pardon the shaky cam, was holding the phone with one hand. Video is simple. In sequence:

1. 50 mV / div, can move +/- 5 divisions no problem.
2. 20 mV / div, cannot move all divisions, scope adds offset dotted line
3. 10 mV / div, cannot move all divisions, scope adds offset dotted line
4. 5 mV / div, cannot move all divisions, scope adds offset dotted line
5. 2 mV, can move +/- 5 divisions, no problem.
6. 1 mV, can move +/- 5 divisions, no problem.

To reproduce:
- use custom 0.1 : 1 attenuation on any channel.
- No probe needs to be connected (this happens with both connected and unconnected probes)
- Use latest 2.3 firmware!

[javadesigner]

I heard that R&S support was able to replicate this and that they are working on a  fix (future firmware update).

Meanwhile, there may be workarounds, like setting the probe to be 1:1 and using Math functions to scale.

I have to admit, R&S support is responsive (via Rich@RS and this forum) and this is a very good feeling - thanks R&S! (I would still recommend R&S to anyone because of this support)..