Keysight DSOX1102G strangely noisy

[dentaku]

A while ago I posted a screenshot from my Keysight DSOX1102G and someone mentioned how fuzzy the trace looked.
I've only been using this scope for audio related stuff with BW limit on and sometimes analyzing MIDI and SPI so the noise wasn't really evident. I had never even seen a digital scope before this one so everything looks noisy compared to my old Hameg analogue scope.

Last week I built a circuit that outputs a small 18mV signal and saw how horrible the signal is on my scope once I zoomed in. I don't know if it worked correctly when I first got it but I'm assuming I must have at least looked closely at the square wave signal coming from the Probe Comp output in the first few days I owned it.

SO, can anyone with a Keysight 1000X series scope show me what it's supposed to look like? I even brought my scope to another house to make sure it wasn't something originating from the houses mains power.

Look at all these spikes (you can see a repeating pattern if you zoom out) with a 10X probe and the vertical set to 10mV looking at the bottom of the square wave coming from the Probe Comp output.
The same noise happens with the probe connected to nothing or looking at a signal from a little DDS function generator and the smaller spikes go down if I touch my hand to the LCD of the scope but the bigger ones don't change.
The smaller spikes repeat at 520kHz and the bigger ones at about 71kHz but sometimes it's around 140kHz.

The strangest thing I've found about the big spikes is that if I turn on Cursors and move one, I get exactly the same big spike that follows the cursor just a few pixels to the right of it. I wonder if the screen on this scope or something in the front panel is messed up?

After lots of confusion with my scope's serial number not being in their system I eventually go one of the probes replaced because it was defective right out of the box but this is  quite a different situation.

[TheSteve]

Looks like the probe is just picking up some noise to me. Possibly from the scopes LCD. I don't think it is anything to worry about.

[blacksheeplogic]

After lots of confusion with my scope's serial number not being in their system

I think you will find that has now been fixed and all the effected scopes have now been entered.

[TK]

Have you tried with the probe set to 1:1?

[schopi68]

What noise-level do you have without any probe or even better with just an coaxial termination resistor instead of the probe?

[dentaku]

After lots of confusion with my scope's serial number not being in their system

I think you will find that has now been fixed and all the effected scopes have now been entered.

Yup, once I noticed that was fixed I contacted Keysight again and was able to get the probe replaced.

[dentaku]

Have you tried with the probe set to 1:1?

I get a few more spikes but the noise itself isn't any greater in amplitude.

[dentaku]

What noise-level do you have without any probe or even better with just an coaxial termination resistor instead of the probe?

It looks much better with no probe plugged into the scope and in fact it doesn't look too bad with the scope not connected to anything but with the "witches hat" removed.

Here's a screenshot of it with nothing connected to the BNCs.

[dentaku]

Here's another image. This time I'm looking at the probe comp output but with the horizontal zoomed out so you can see one of the corners.
The spikes are around 40mV so when I was looking at a circuit that was outputting an 18mV signal that's when I noticed how annoying it is.

The other mystery is how I get a similar spike that follows the orange Cursor markers when I move them sometimes?

[Keysight DanielBogdanoff]

That cursor thing is odd, I can't fathom how that could be related. Can you post a video or something so I can show support?

It does look like your probe is just picking up ambient noise, some tricks to help reduce it would be to 1- try a new environment and 2- try changing the orientation of the probe cables and see if that helps.

Can you also run an FFT and see if you can identify a noise frequency? Might shed some light on the situation.

[PA0PBZ]

That cursor thing is odd, I can't fathom how that could be related.

The only explanation I can think of is that he is picking up a lot of noise from the display for some odd reason.

Dentaku, does it get worse when you hold the probe close(r) to the display?

[JPortici]

Can you also run an FFT and see if you can identify a noise frequency? Might shed some light on the situation.

this is what i usually suggest, at my place for example there is a TON of interference from the whole FM radio band (which is obvious as soon as you make an FFT of the signal), it can make the 10mV/div and lower scales completely useless unless without some tricks and/or filtering.

[TK]

Here are some screen captures from my EDUX1002G.  I tried to replicate the exact conditions as used by OP.

Nothing connected to the BNC input


Detailed trace before rising


Trace with cursors

[dentaku]

Here are some screen captures from my EDUX1002G.  I tried to replicate the exact conditions as used by OP.

Thanks for the screenshots. I guess you get a bit less noise than me with nothing connected to the BNCs but it would vary depending on where you are in the world.

Did you think of checking  if you could see anything moving along with the cursor?

[TK]

I moved the cursor and could not see the ghost signal following it... The spikes I see on my scope seem to be pretty stable

[dentaku]

That cursor thing is odd, I can't fathom how that could be related. Can you post a video or something so I can show support?

It does look like your probe is just picking up ambient noise, some tricks to help reduce it would be to 1- try a new environment and 2- try changing the orientation of the probe cables and see if that helps.

Can you also run an FFT and see if you can identify a noise frequency? Might shed some light on the situation.

It IS quite weird. I did some more experiments and noticed that sometimes I even got two little ghost blips that move in opposite directions and that they don't move at the same speed as the cursor. Depending on how my horizontal is set they will move further across the screen than the cursor has moved.

Here's the best I could do by pointing my phone at the screen.
You can see two blips moving in opposite directions on the left and the cursor on the right.
10X probe connected to the probe comp.

https://1drv.ms/v/s!ArYn6kcDECISkRY7WS_erYNxklZ8
This .mp4 won't play in a browser for some reason so you'll probably have to download it first.

I know Dave did a video quite a while ago about how people keep saying that digital scopes are noisier than analogue ones even if it's not true. I was still surprized at what the trace looked like when I was trying to measure an 18mV signal without using BW limiting and high resolution mode.

It was the same at someone else's house (but in the same village) and moving the probe wires around makes the noise worse most of the time.

[schopi68]

Will the Noise change when you short the input with a very small wire (Bnc outer side to inner connector) - or better a BNC shorting cap?

[snoopy]

That cursor thing is odd, I can't fathom how that could be related. Can you post a video or something so I can show support?

It does look like your probe is just picking up ambient noise, some tricks to help reduce it would be to 1- try a new environment and 2- try changing the orientation of the probe cables and see if that helps.

Can you also run an FFT and see if you can identify a noise frequency? Might shed some light on the situation.

I've seen this problem before on other scopes and is most likely nothing to do with the scope itself. More likely a red herring rather than a real fault   It usually has something to do with the ground clip acting as an antenna or inductive pickup and injecting switching regulator noise etc into the probe or circuit under test. Sometimes out of band noise can be aliased back if you are using a low sampling rate which will create lots of phantom noise artifacts as well. Move the probe away to the back of the DUT or putting the metal lid back on the equipment will substantially reduced the effect. If you are making audio measurements etc try connecting a shielded coax cable directly from the output of the DUT to the scope will help as well

cheers

[JFJ]

Look at all these spikes (you can see a repeating pattern if you zoom out) with a 10X probe and the vertical set to 10mV looking at the bottom of the square wave coming from the Probe Comp output.

I believe that your oscilloscope's Probe Comp output signal is 2.5V peak-to-peak. If that is the case, then the input signal would be 250mV peak-to-peak (after being attenuated by the 10:1 probe). At 10mV per division, it would take at least 25 vertical divisions to display that signal properly. The 8-bit vertical resolution of your oscilloscope couldn't possibly capture 25 vertical divisions with any degree of detail. Hence, much of the noise that you're seeing is likely to have have been generated by a display processing algorithm, while trying to make sense of the overloaded input.

[Keysight DanielBogdanoff]

So the noise is almost certainly just environmental, but the spike moving with the cursors  . I've sent the clip to R&D and we'll see if we can figure out what's up with that. It reminds me of aliasing/folding in the frequency domain, but I've never seen anything like that in the time domain.

I'll keep you updated, thanks!

[Keysight DanielBogdanoff]

Just to be clear, the cursor + spike thing happens with the 1000 X-Series scope connected to a probe and nothing else?

[dentaku]

Just to be clear, the cursor + spike thing happens with the 1000 X-Series scope connected to a probe and nothing else?

Good question... I just set up my scope just like in the screenshot and it seems to only happen when I have the probe connected to the Probe Comp output and not when it's connected to nothing.
I can't even reproduce it with the probe connected to the Gen Out set to the same frequency and amplitude of square wave.

I just found it a strange curiosity, the noise you get when the scope is set to 10 or 20mV is what was most surprizing to me but from what people are saying, that's just normal.

[dentaku]

I did some experiments last night with my old Hameg HM605 scope and another signal generator and NOT an output from the Keysight and the same sequence of spikes happens on that one too if I turn the vertical scale all the way up so I guess I've learned that this house and probably the whole neighbourhood, because I brought it to another house, just has that type of noise.
Would this have something to do with our mains power?

Anyway...
If I wanted to measure a small signal, can I simply set my scope to BW limting and High Resolution and trust that my measurements are roughly accurate?
I know scopes aren't really designed for making extremely accurate voltage measurements but it's still the best instrument I own.

I'm attaching some screenshots of the circuit that made me see how much noise scopes pick up and one image of my analogue scope picking up the same noise I was seeing on my Keysight.

[Keysight DanielBogdanoff]

Anyway...
If I wanted to measure a small signal, can I simply set my scope to BW limting and High Resolution and trust that my measurements are roughly accurate?
I know scopes aren't really designed for making extremely accurate voltage measurements but it's still the best instrument I own.

Yes, you can. High resolution will help filter out some of the random noise, and BW limiting will help cut high frequency noise. You can also give averaging mode a try if you have repetitive signals. Using the probe in 1:1 mode will also add another effective bandwidth filter to your measurement, but could load your signals more.