Rigol DS1054Z cross-channel noise

[mbratch]

Hello

I apologize for the separate post on this topic. There are lots of posts regarding the Rigol DS1054Z and some of them discuss noise. But a lot of these noise discussions are sprinkled and buried within a broader topic, and I wanted to ask this particular noise scenario separately.

I just purchased (about 2 weeks ago) a new Rigol DS1054Z. It indicates board version 0.1.1 (I have not been able to find any information about the significance of this) and I have updated to the latest software (v00.04.05.02.00) and added the 100MHz "hack". I do not have the 500uV/div hack.

I was doing some simple/naive noise floor checking. With an open input, or with a probe connected and the ground clip connected to the tip, I get basically the same result which is ~700uV p-p noise. This looks very good, as expected. This is shown in the attached image and within expected behavior (no issues there).

In the same image, it also shows what the open channel looks like while I have an I2C trace in two other channels. What you see is the noise floor in channel 4 as I described above, but with large noise spikes (about 8mV p-p) that correspond to several of the I2C signal 5V to 0V transitions. These I2C signal transitions are about 10nS in width, so fairly abrupt.



Does anyone else observe this?

[Fungus]

Does anyone else observe this?

Yes.

It's mostly down to probing and grounding, but there's also an element of radio in there. When you connect a ground clip to the probe tip you're creating a radio antenna which will pick up the RF from your I2C circuit (try moving it closer/further away).

The display on my Micsig in front of right now is very similar. I'm coding an I2C function for a uC and it looks like this:


I'm probing by attaching the probes to two bits of wire stuck into a breadboard so all that ringing and "cross channel" spikes doesn't worry me at all. I know they're not really there in the circuit and there's nothing wrong with the 'scope. It's all down to lengths of wires, ground loops and stuff.

[mbratch]

It's mostly down to probing and grounding, but there's also an element of radio in there. When you connect a ground clip to the probe tip you're creating a radio antenna which will pick up the RF from your I2C circuit (try moving it closer/further away).

Thanks! I didn't think I had the probe that close to the circuit, but I went back and played with it a bit trying different "unconnected" scenarios and I could definitely make the spikes come, go, or change.

[marcus h]

I just purchased (about 2 weeks ago) a new Rigol DS1054Z. It indicates hardware version 0.1.1 (I have not been able to find any information about the significance of this) and I have updated to the latest software (v00.04.05.02.00) and added the 100MHz "hack". I do not have the 500uV/div hack.

I'm wondering about the hardware version of your scope.

When Dave reviewed the 1054Z in January 2015 and made his photos of the internals available one could see that it had a mainboard revision of "V01.01", dated 2013-05-04 -- both silkscreened on the board.

My 1054Z was purchased in December 2019 and it says it is 0.1.4 in both the system info screen as well as silkscreened on the mainboard: "V01.04", dated 2014-10-24

Could it be that you bought "old stock"? Hopefully at a good price?

[mbratch]

I'm wondering about the board version of your scope.

When Dave reviewed the 1054Z in January 2015 and made his photos of the internals available one could see that it had a mainboard revision of "V01.01", dated 2013-05-04 -- both silkscreened on the board.

My 1054Z was purchased in December 2019 and it says it is 0.1.4 in both the system info screen as well as silkscreened on the mainboard: "V01.04", dated 2014-10-24

Could it be that you bought "old stock"? Hopefully at a good price?

I think that's possible. It was open box but new. I did get a decent price (US$325). Everything was still factory tie wrapped and sealed in plastic. The scope itself had an older version of SW in it and all the features were marked "Trial". When I upgraded the firmware, they all became "Official" (but without the 100MHz of course).

It still has the factory seal on it, so I didn't want to open it. I might after a few months just to change the fan.

But back to the board version, it would be nice to know what the difference is between 0.1.1 and 0.1.4. I have not been able to find any information on that topic. I contacted Rigol with that question, and received the following brief note from an "Application Engineer" in Portland OR:

We don't have a list of changes to hardwareversions in the same way that we have changes to firmware. There are no hardware/firmware incompatibilities so any board version should be able to work with the newest firmware.

This wasn't a very satisfying response. Of course, the desired information must be available at Rigol, but the factory in Suzhou, Jiangsu, China probably doesn't provide that information to their App Engineers in the US I did ask the App Engineer if he would kindly inquire of the factory and he said he would. Very nice of him.

[TurboTom]

AFAIK it's all down to the arrangemnt of the MSO / digital inputs. The early H/W versions utilized a separate FPGA to implement the digital inputs while later on, this has been merged into the main FPGA, possibly changing the BOM item to a larger version.

Here's you find a visual comparison of the two H/W versions by @skander36: https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg1295400/#msg1295400

[mbratch]

AFAIK it's all down to the arrangemnt of the MSO / digital inputs. The early H/W versions utilized a separate FPGA to implement the digital inputs while later on, this has been merged into the main FPGA, possibly changing the BOM item to a larger version.

Here's you find a visual comparison of the two H/W versions by @skander36: https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg1295400/#msg1295400

That's a great thread, thanks!

I suppose the modifications from 0.1.1 to 0.1.4 may just be cost reductions/consolidations. I just spotted another thread listing versions in the range 0.2.x as well, with some opinions offered there.

I suspect the board version is probably of little concern.

[Someone]

Channel to Channel Isolation: DC to maximum bandwidth: >40 dB