Rigol DS1054Z noise.

[Wisnaeme]

Hi all,
I'm trying to work out of this noise is internal to my new DS1054Z or not.
I've got a signal that appears to be a sequence of five pulses that seem to repeat and are easily detected by the scope.

To remove external sources I'm tested at two location and switched everything off except the scope.
The signal is stronger near the fan of the DS1054Z.

Here's an example of the signal.
The blue trace is a shorted scope probe, using the flying ground lead.
The yellow trace is an audio feed from an audio mixer, just used as a test signal.

Image 32 is taken with a finger on the yellow lead showing the 50Hz AC pickup.

Looking for any advice on testing this further. Does you scope also have this issue?

Thanks,
Robert

[JPortici]

Try playing around with the scope, like enabling/disabling channels, changing scale so the relay switches and see if there is a change in the signal.

I certainly don't pick up anything like that from the scope (though the environment here is noisy enough that i pick up 100 MHz FM stations 12 km far)

[alsetalokin4017]

I'm not able to reproduce the issue here.

CH2 probe shorted with its own ground lead, positioned as close to the fan as I can get it. Tried it with probe set to 1x and 10x, without any difference.

Second shot below is 60Hz pickup on CH1, finger on probe. (USA mains pickup)

If you get the same thing when you are in some other location that you know to be EM-quiet... perhaps it's time to return the unit under the vendor's warranty.

[Fungus]

If it happens on two channels and you've tried it in two different buildings then it looks like a defective scope. Maybe a power supply issue or something.

If it's a new 'scope then send it back.

[Wisnaeme]

Another quick image.
Yellow Scope and probe are set to x10.
Bandwidth set to 20MHz.
Acquire mode to normal. Noise goes away only on Averaging x4 or higher but that's not really a solution.
This noise is across all channels.

The noise pulses are causing major issues with triggering as well.


Image is finger on the yellow probe.

Blue probe resting on desk at x1.

[alsetalokin4017]

Try the test with no probes plugged in to the scope. Also repeat with channels set to "ground" coupling.

If you still see that noise.... send it back.

[Wisnaeme]

Hi again,

Thanks for all the feedback. I've uploaded a couple of video's so you can see what's happening in real time.

https://youtu.be/d2tcsJuSjkk

https://youtu.be/96i-jOc-7S0

I've checked the earth/grounding. The scope screen/earths and screening casework inside are connected to the main earth/ground.
Selecting ground coupling stops all noise on the channel.
Changing from x10 to x1 on the probes changes very little in the signal detected.

I've rechecked with everything in the house switched off, routers, computers, lights, mobiles, everything. The only item powered on was the scope.

I think I give up with this. This device looks faulty to me. 

I'm going to link the dealer to the videos above to get their opinion.

Final thoughts?

Robert

[alsetalokin4017]

Ah.... wait a bit.

Your first video says that you have the Channel set to 5mV/div. But I see that the probe attenuation is also set to 10x.  On my scope it is not possible to set 5mV/div with a 10x setting, the lowest it will go is 10 mV/div.

Did you by chance "hack" your scope using the DSBA option code (500 uV/div) ?  The DS1054z is not compatible with this option. Before you contact your dealer I would suggest that you remove the hack and then reinstall the hack using a new key generated with the DSER option.


To remove the hack and restore the scope to "virgin" 50 MHz with no extra options,  you use the Rigol software, or telnet over a LAN connection, to send the SCPI command:

:SYSTem:OPTion:UNINSTall

exactly as written there.

Then repeat your test to see if the noise is still there.

Once you have a new key with the DSER option code, send the command

:SYSTem:OPTion:INSTall XXXXXXXXXXXXXXXXXXXXXXXXXXXX

where the "xxx" is your new key, with no hyphens, just the 28 characters.

Again, repeat the test to see if the noise is still present.

If you do decide to return your scope to the dealer you probably should reset it to "virgin", if that was the way it came to you.

[rstofer]

It seems that with the inputs set to ground coupling you had no noise.  So, the noise isn't coming from inside the scope or from the incoming power lines.  It really doesn't seem like a scope problem.  Then again, I don't really know where the input is grounded.  I would expect there to be a switch that selects the input from the DC blocking capacitor, the actual BNC input or ground.

It's difficult to get away from all sources of noise.

[rstofer]

The fact is, your body coupled the noise into the calibration point.  I don't know where the electrostatic field is coming from but your body is conducting it into the scope.  I doubt that there is anything wrong with the scope.  Particularly since you get no noise with Ground coupling.

[TurboTom]

The DS1000Z series only uses "digital grounding" , i.e, it disregards the output oft the ADC and displays the line at 0V. That's worth nothing regarding the noise problem.  I would simply try out the scope in a different location with a mains supply from a different  distribution point. If the noise is still there, return it  for a replacement.

Good luck,
Thomas

[cvanc]

Rotate the scope while you observe the fault.  If it changes to any significant degree it must be coming from outside the scope.  If it does not change it may or may not be outside the scope.

[alsetalokin4017]

The DS1000Z series only uses "digital grounding" , i.e, it disregards the output oft the ADC and displays the line at 0V. That's worth nothing regarding the noise problem.  I would simply try out the scope in a different location with a mains supply from a different  distribution point. If the noise is still there, return it  for a replacement.

Good luck,
Thomas

Ah, I didn't know that. Where did you find this information about "digital grounding"?
 
Perhaps that helps to explain how the scope is able to calculate and display a large RMS voltage measurement value even though the channel input is "ground coupled" and the display is showing a flat line at 0V.

Somehow this "feature" of the DS1000z series leaves me rather cold.

[vk6zgo]

The DS1000Z series only uses "digital grounding" , i.e, it disregards the output oft the ADC and displays the line at 0V. That's worth nothing regarding the noise problem.  I would simply try out the scope in a different location with a mains supply from a different  distribution point. If the noise is still there, return it  for a replacement.

Good luck,
Thomas

Ah, I didn't know that. Where did you find this information about "digital grounding"?
 
Perhaps that helps to explain how the scope is able to calculate and display a large RMS voltage measurement value even though the channel input is "ground coupled" and the display is showing a flat line at 0V.

Somehow this "feature" of the DS1000z series leaves me rather cold.

You must admit it is an elegant solution,if you don't anticipate the above "bugs".

Maybe the Engineer who designed this feature was a digital person & regarded the analog circuitry as a sort of "pipeline" feeding the signal to the ADC.
Or perhaps he was the "bunny" who had to come up with a "quick fix" after the analog guys forget to provide a proper
ground switch.

[pascal_sweden]

Maybe the Engineer who designed this feature was a digital person & regarded the analog circuitry as a sort of "pipeline" feeding the signal to the ADC.
Or perhaps he was the "bunny" who had to come up with a "quick fix" after the analog guys forget to provide a proper
ground switch.

Does that mean that the inputs of the Rigol scope can be made "floating" given that the analog guys forgot to provide a proper ground switch? Or did I misunderstand? What part is not properly grounded? The analog input of the ADC, or the digital output of the ADC?

[newbrain]

I had a similar problem with my 1054Z, traced in the end to a noisy (VERY noisy) USB hub power supply.
Even when used in a different room, on different mains circuit, you could still see it.

But, you say everything else has been disconnected, so...

Do you perchance live in proximity of an airport or military installation?
The cadence and repetition reminds me of some RADAR.

[Wisnaeme]

Hi all,
Back again. I've found it.
The noise is external and getting in everything. I borrowed another scope last night and dragged it home to test. Both had the same noise.

On closer investigation the "I unplugged everything" was proven correct, however there was a USB powerbank device that was switched on and remained so even with power removed from everything else. I sniffed it out using a long power cable and walking the property with the scope.
It just happens that my friend also has that same power bank (Generic Ebay) and his was also broadcasting a noise both out the USB ports and the charging cable. The rapid switching appears to be the power bank just topping off the capacitors as when power is drawn from the device the noise became much more constant as the switcher was working harder.

So now we're down to a normal noise floor with some smaller switching noise still present.

I hate tracing these things out, but I guess that's all part of the job/hobby.

Thanks for all of your time/efforts testing at your end.

Robert

[TurboTom]

Ah, I didn't know that. Where did you find this information about "digital grounding"?
 
Perhaps that helps to explain how the scope is able to calculate and display a large RMS voltage measurement value even though the channel input is "ground coupled" and the display is showing a flat line at 0V.

Somehow this "feature" of the DS1000z series leaves me rather cold.

Okay, I re-checked that again and I seem to be somewhat in error here  -- the situation is a little more complex:

The analog front end of the DS1000Z series hasn't got a proper grounding switch/relay, the input stays connected to whatever source it is attached to (see Dave's reverse engineering thread: https://www.eevblog.com/forum/blog/eevblog-675-how-to-reverse-engineer-a-rigol-ds1054z/). Yet, there's still some noise displayed when sensitivity is wound up to maximum and input "grounded" via the vertical menu function. This noise is much less than with an "open" or externally terminated input. A closer look at the ADC that's presumably used in the instrument (Hittite HMCAD1511) shows that most of the attenuation/amplification is done with the (so-called) "digital gain control" which permits a 1:50 digital gain range (1:32 without missing codes) -- this matches very well with the 1:50 "hardware" attenuator in the front end (the relay you hear clicking between the 200mV/500mV (1x) range). There's one more section in the ADC that is relevant here and that's the input cross point switch (MUX Array). This is used to harness the four 250MSps channel inputs in the ADC to form four individual 250MSps, two 500MSps or one 1GSps channel(s).

It's not particularly mentioned but looking at the input matrix switching scheme http://www.analog.com/media/en/technical-documentation/data-sheets/hmcad1511.pdf, it should be possible to run the ADC with all the analog inputs disconnected which is in my opinion exactly what Rigol does to "emulate" a grounded input. The digital gain is still active but there's no connection between the ADC and the preamplifier. Hence, all internal noise generated or picked up by the analog input circuitry, will be camouflaged. The residual noise is just ADC and ADC interleaving noise (switch between single/multiple channels while observing a "grounded" input at maximum gain and 5ns timebase resolution to observe this).

To cut a long story short, the result is basically the same - you won't be able to tell if the analog front end amplifier is picking up the noise internally by setting input connection to "Ground".

Cheers,
Thomas

P.S: congratulations on spotting the source of the noise! It's quite amazing what crap some "el cheapo" power supplies are... 

[Fungus]

Somehow this "feature" of the DS1000z series leaves me rather cold.

So cold that you couldn't help bringing it into the middle of a completely unrelated thread?

Uhuh.

[Fungus]

Does that mean that the inputs of the Rigol scope can be made "floating" given that the analog guys forgot to provide a proper ground switch?

Yes. I can't see a "ground the input" switch anywhere on the schematic for the inputs.

This also neatly explains where the RMS 'bug' comes from (floating input into a sesitive op-amp), why this "bug" can be safely ignored, and why it's not going to affect anything else.

The fix would be to just display 0.00 on that channel instead of the actual input reading.

[Fungus]

I hate tracing these things out, but I guess that's all part of the job/hobby.

Yep.

It looked like that sort of switching noise, but:
a) Over 1V peak to peak is a lot.
b) When you said you tried it in "different locations" I thought you meant you tried it in two different buildings.

[Wisnaeme]

When you said you tried it in "different locations" I thought you meant you tried it in two different buildings.

Yup. Two buildings a mile apart. Just typical we had the same power bank in use at both locations.
My one has already been scrapped, torn apart and the cells removed for projects.

[Fungus]

When you said you tried it in "different locations" I thought you meant you tried it in two different buildings.

Yup. Two buildings a mile apart. Just typical we had the same power bank in use at both locations.

 

All part of the learning curve I guess.

[alsetalokin4017]

Somehow this "feature" of the DS1000z series leaves me rather cold.

So cold that you couldn't help bringing it into the middle of a completely unrelated thread?

Uhuh.

Says the person who hasn't even loaded the latest firmware update into his miracle scope.

Read back in the thread. The issue arose when I suggested that the OP try testing with inputs ground-coupled. Was this wrong to suggest, or unrelated to tracking down a noise issue? Now, thanks to some good digging by TurboTom we know something more about how this scope does things internally, and how it is different from what some of us might expect. Is that wrong, or unrelated to tracking down a noise issue? Would you prefer that we did not have this information?

You are not helpful.  Perhaps you should put me on your ignore list, if you don't like reading my comments about "features" of the DS1054z.
 

[alsetalokin4017]

Hi all,
Back again. I've found it.
The noise is external and getting in everything. I borrowed another scope last night and dragged it home to test. Both had the same noise.

On closer investigation the "I unplugged everything" was proven correct, however there was a USB powerbank device that was switched on and remained so even with power removed from everything else. I sniffed it out using a long power cable and walking the property with the scope.
It just happens that my friend also has that same power bank (Generic Ebay) and his was also broadcasting a noise both out the USB ports and the charging cable. The rapid switching appears to be the power bank just topping off the capacitors as when power is drawn from the device the noise became much more constant as the switcher was working harder.

So now we're down to a normal noise floor with some smaller switching noise still present.

I hate tracing these things out, but I guess that's all part of the job/hobby.

Thanks for all of your time/efforts testing at your end.

Robert

Yes, congratulations on finding the problem.
 
However, my earlier comment about the 500 uV/div vertical setting (or 5mV/div with 10x probe setting) still holds.

[Wisnaeme]

However, my earlier comment about the 500 uV/div vertical setting (or 5mV/div with 10x probe setting) still holds.
[/quote]
I did take the updates back to DSER and not DSFR after resetting the licenses. This made no change to the noise and didn't seem to change anything other than the maximum sensitivity. I don't think I'll be going near those signal levels so nothing lost with this.

[vk6zgo]

Maybe the Engineer who designed this feature was a digital person & regarded the analog circuitry as a sort of "pipeline" feeding the signal to the ADC.
Or perhaps he was the "bunny" who had to come up with a "quick fix" after the analog guys forget to provide a proper
ground switch.

Does that mean that the inputs of the Rigol scope can be made "floating" given that the analog guys forgot to provide a proper ground switch? Or did I misunderstand? What part is not properly grounded? The analog input of the ADC, or the digital output of the ADC?

No!
One side of the input is always earthed.

The "earthing" switch referred to shorts the non-earthed side of the input to the earthed side.
This is commonly used to establish "zero volts" to set the zero line of the display to whatever graticule line is required.

Not having such a switch is no major drama,as a BNC connector with a very short connection between "inner & outer" can be made up,which will provide a real shorted inout.

[pascal_sweden]

The "earthing" switch referred to shorts the non-earthed side of the input to the earthed side.
This is commonly used to establish "zero volts" to set the zero line of the display to whatever graticule line is required.

Isn't this only relevant for analog oscilloscopes?

[timb]

The "earthing" switch referred to shorts the non-earthed side of the input to the earthed side.
This is commonly used to establish "zero volts" to set the zero line of the display to whatever graticule line is required.

Isn't this only relevant for analog oscilloscopes?

Why would it be? I regularly move the trace to different parts of the graticule on my DSO. (Though, instead of actually switching the input to ground, they could just artificially display a zero volt trace. No real difference there.)

The only reason you wouldn't need to set the trace to GND reference is if your DSO allows you to precisely place it. On my Tek, I can set a trace to, say, 4V (the center of the graticule being 0V, the first grid line above would be 1V/nVDiv, the first grid line below center being -1V/nVDiv, etc.) but I rarely do that, as it requires going through some menus, versus twirling a knob.

[ebastler]

The "earthing" switch referred to shorts the non-earthed side of the input to the earthed side.
This is commonly used to establish "zero volts" to set the zero line of the display to whatever graticule line is required.

Isn't this only relevant for analog oscilloscopes?

Why would it be? I regularly move the trace to different parts of the graticule on my DSO.

I do that too. But my DSO (as, I believe, the vast majority of digital scopes) has channel indicators on the side of the screen, which show where the 0V level is. So I agree with pascal_sweden that switching the input to GND mode to get your bearings is a bit redundant with DSOs. (While in analog scopes it was the only quick way to establish where 0V are.)

[tautech]

The "earthing" switch referred to shorts the non-earthed side of the input to the earthed side.
This is commonly used to establish "zero volts" to set the zero line of the display to whatever graticule line is required.

Isn't this only relevant for analog oscilloscopes?

Why would it be? I regularly move the trace to different parts of the graticule on my DSO.

I do that too. But my DSO (as, I believe, the vast majority of digital scopes) has channel indicators on the side of the screen, which show where the 0V level is. So I agree with pascal_sweden that switching the input to GND mode to get your bearings is a bit redundant with DSOs. (While in analog scopes it was the only quick way to establish where 0V are.)

For those that don't want the display cluttered with on-screen measurements the need for grounding the input is necessary when setting the waveform zero level exactly to a graticule so that waveform amplitudes can be quickly observed.....a methodology for many that is a carry over from the days of no OSD on CRO's.

Not all use set methods of scope use and manufacturers know this.

[ebastler]

... my DSO (as, I believe, the vast majority of digital scopes) has channel indicators on the side of the screen, which show where the 0V level is. So I agree with pascal_sweden that switching the input to GND mode to get your bearings is a bit redundant with DSOs. (While in analog scopes it was the only quick way to establish where 0V are.)

For those that don't want the display cluttered with on-screen measurements the need for grounding the input is necessary when setting the waveform zero level exactly to a graticule so that waveform amplitudes can be quickly observed.....a methodology for many that is a carry over from the days of no OSD on CRO's.

Not all use set methods of scope use and manufacturers know this.

I was not talking about measurements, but about the channel indicators in the left margin of the screen (circled in the attached screenshot). These do the job for me: I can see, at any time and without the need for a GND switch, where the 0V level is, and can align it to a graticule  line if needed.

[alsetalokin4017]

Many owners of the DS1054z will have noticed that the displayed "zero volts baseline" rarely lines up exactly with the channel baseline indicators, and in some cases can be way off.

This is an uncorrectable DC offset and should be factored into any interpretation of Measurements that reference the channel baseline values. It also can move around, depending on whether you have run the self-cal routine at your present lab ambient temperature.



 

[tautech]

Many owners of the DS1054z will have noticed that the displayed "zero volts baseline" rarely lines up exactly with the channel baseline indicators, and in some cases can be way off.

This is an uncorrectable DC offset and should be factored into any interpretation of Measurements that reference the channel baseline values. It also can move around, depending on whether you have run the self-cal routine at your present lab ambient temperature.

Not at all uncommon for DSO's but not the end of the world either.
Self Cal after an appropriate warm up time should get any zero offset down below a mV, then if this amount matters you are really using the wrong tool for the job.

[ebastler]

Many owners of the DS1054z will have noticed that the displayed "zero volts baseline" rarely lines up exactly with the channel baseline indicators, and in some cases can be way off.

This is an uncorrectable DC offset and should be factored into any interpretation of Measurements that reference the channel baseline values. It also can move around, depending on whether you have run the self-cal routine at your present lab ambient temperature.

Not at all uncommon for DSO's but not the end of the world either.
Self Cal after an appropriate warm up time should get any zero offset down below a mV, then if this amount matters you are really using the wrong tool for the job.

Thanks, tautech. And the offset values alesatonkin's scope is showing are indeed all around 0.25 mV (assuming that all channels were set for 10x probes, as indicated in the screenshot for channel 4). Nothing to worry about, I would say.

[tautech]

Many owners of the DS1054z will have noticed that the displayed "zero volts baseline" rarely lines up exactly with the channel baseline indicators, and in some cases can be way off.

This is an uncorrectable DC offset and should be factored into any interpretation of Measurements that reference the channel baseline values. It also can move around, depending on whether you have run the self-cal routine at your present lab ambient temperature.

Not at all uncommon for DSO's but not the end of the world either.
Self Cal after an appropriate warm up time should get any zero offset down below a mV, then if this amount matters you are really using the wrong tool for the job.

Thanks, tautech. And the offset values alesatonkin's scope is showing are indeed all around 0.25 mV (assuming that all channels were set for 10x probes, as indicated in the screenshot for channel 4). Nothing to worry about, I would say.

Some might insist the display and onscreen measurements align at zero volts when directed to in the UI but then how is this being accomplished.......with a hard zero SW command? Then will subsequent measurements be correct after ignoring any internal noise in a DSO?
Let's post some screen shots with differing acquisition and detection settings to see if the zero level becomes better aligned.

[David Hess]

You must admit it is an elegant solution,if you don't anticipate the above "bugs".

Maybe the Engineer who designed this feature was a digital person & regarded the analog circuitry as a sort of "pipeline" feeding the signal to the ADC.

Or perhaps he was the "bunny" who had to come up with a "quick fix" after the analog guys forget to provide a proper ground switch.

Or marketing wanted the appearance of lower noise.  Doesn't the Rigol 2000A series do this as well?

It may be a safety issue for the oscilloscope.  In a proper design, ground coupling may be used to precharge the AC coupling capacitor preventing damage to the input circuits.

[Fungus]

Thanks, tautech. And the offset values alesatonkin's scope is showing are indeed all around 0.25 mV

??

More like a couple of mV. He has his 'scope set to 20mV/div.

According to the manual:

DC Offset Accuracy: ±0.1 div ±2 mV ±1% offset value

I'd say it's in spec.

Here's mine FWIW

[David Hess]

If you ground the DC coupled input externally, then you can see if the input bias current through the 1 megohm input shunt is causing any offset.

It seems odd that all of the offsets are in the same direction with about the same magnitude but as pointed out, without an physical analog ground coupling function, the automatic calibration may not be able to remove it.

[ebastler]

Thanks, tautech. And the offset values alesatonkin's scope is showing are indeed all around 0.25 mV

??
More like a couple of mV. He has his 'scope set to 20mV/div.

He has the scope configured for 10x probes. So that's 20mV/div at the (presumed) probe tip, but 2mV/div at the scope's BNC input. Hence, the offset at the actual scope input is well below a mV, right?

[Fungus]

He has the scope configured for 10x probes. So that's 20mV/div at the (presumed) probe tip, but 2mV/div at the scope's BNC input. Hence, the offset at the actual scope input is well below a mV, right?

Oh, I see what you mean.... yes, electrically it's 10x less.

[vk6zgo]

The "earthing" switch referred to shorts the non-earthed side of the input to the earthed side.
This is commonly used to establish "zero volts" to set the zero line of the display to whatever graticule line is required.

Isn't this only relevant for analog oscilloscopes?

An earthing switch of this nature is trivial to provide in analog 'scopes----just provide another position on the "input coupling" switch.
It gets messier with DSOs ,where the switch is normally a relay,driven ultimately by software.

As I said before,if there isn't an actual earthing switch,a  real "hard" earth can be provided by a shorted BNC connector.

If you put a real short across a DSO input,& your display says something other than zero volts ( in other words,an offset),you can see if the 'scope can auto calibrate that out,or just take a note of the offset for later reference.

The use of such is an example of a technique which seems to have been lost in this era of higher stability in test equipment,that of presenting the equipment with a known condition.

People do seem to assume that the accuracy of their equipment is good,& if all goes well that is OK,but in unforseen circumstances,it is good to know whether the DUT has a problem,or your test equipment is lying to you,

[Keysight DanielBogdanoff]

On closer investigation the "I unplugged everything" was proven correct, however there was a USB powerbank device that was switched on and remained so even with power removed from everything else. I sniffed it out using a long power cable and walking the property with the scope.

I'm imagining this (for better or for worse. Probably for worse.)

[nctnico]

You forgot to measure the noise from the brain of that grey haired dude you pushed away