Magnova oscilloscope

[Xyphro]

Out of pure interest: does Magnova use LMH6518 as frontend IC?

For context: https://www.eevblog.com/forum/repair/repair-of-rs-rtb2004-with-offsetprobem-of-toasted-lmh6518/

[Martin72]

Seen by chance and found to be very good:
A time bar that shows the progress of the acquisition in seconds.

[Xyphro]

Now that we're brainstorming:
A hires mode like on other scopes would be good. Meaning that it is not a fixed decimation factor, but one that scales also after having done a single shot capture depending on how much you zoom into the time axis. Right now it is basically a digital low pass filter only.

[KungFuJosh]

Seen by chance and found to be very good:
A time bar that shows the progress of the acquisition in seconds.

Yeah, it's pretty good. I mostly use that in the sense of: "oops, I see the time bar, I set the timescale too long!"

[Andre77]

Out of pure interest: does Magnova use LMH6518 as frontend IC?
For context: https://www.eevblog.com/forum/repair/repair-of-rs-rtb2004-with-offsetprobem-of-toasted-lmh6518/

The Magnova does not use the LMH6518.

[jusaca]

Is my understanding of RMS voltage wrong or is the magnova doing something weird here?
The input to the scope is a DC around 8V, but the RMS measurement shows something like 196mV...

I have the feeling the scope is not correctly calculating RMS when a DC offset for that channel is configured to be something other than 0V.

PS: Just realized how bad that dark blue is readable against the black background in screenshots, sorry about that. Will do better next time.

[2N3055]

Is my understanding of RMS voltage wrong or is the magnova doing something weird here?
The input to the scope is a DC around 8V, but the RMS measurement shows something like 196mV...

I have the feeling the scope is not correctly calculating RMS when a DC offset for that channel is configured to be something other than 0V.

PS: Just realized how bad that dark blue is readable against the black background in screenshots, sorry about that. Will do better next time.

You are measuring StDev that is AC RMS. There is separate true RMS(AC+DC RMS) measurement type.

[jusaca]

That are two separate measurements, I measure StDev at second slot, but RMS at the first slot.

[2N3055]

That are two separate measurements, I measure StDev at second slot, but RMS at the first slot.

Ahh, sorry it IS hard to see..

To verify how it works it is easiest to use 50% duty cycle squarewave.
First measure with no DC offset +-500mV (1Vpp)
then 0-1V (500mV offset). 

In first case measurement should be 500mV StDev and AC+DC RMS
In second 500mV Stdev, 707 mV AC+DC RMS

[jusaca]

Ah sorry, I don't mean a DC offset of the signal, but of the scope input channel.
I just played around a little bit more and the RMS values does weird things as soon as the channel reference is not on screen any more (for example when viewing the AC riding on a large DC with DC input coupling).
When I slowly increase the DC input offset on the scope to shift the trace around, the RMS values decreases as soon as the reference indicator is not on screen any more.

[2N3055]

Ah sorry, I don't mean a DC offset of the signal, but of the scope input channel.
I just played around a little bit more and the RMS values does weird things as soon as the channel reference is not on screen any more (for example when viewing the AC riding on a large DC with DC input coupling).
When I slowly increase the DC input offset on the scope to shift the trace around, the RMS values decreases as soon as the reference indicator is not on screen any more.

Now I seem to understand.
In order to verify anything a known signal needs to be measured.

Create 1 Vp-p 50% duty cycle squarewave with 2 V offset and feed it to scope.
On 200mV/div with 2V offset, ref will be below screen.
You should still read StDev of 500mV, and RMS should be 2V.

If it is not than something is wrong.

Make note that in this case, with fixed offset, all of the signal fits nicely on the screen and whole signal is in a ADC range + fixed offset was subtracted in analog frontend.

[Andre77]

I just played around a little bit more and the RMS values does weird things as soon as the channel reference is not on screen any more (for example when viewing the AC riding on a large DC with DC input coupling).
When I slowly increase the DC input offset on the scope to shift the trace around, the RMS values decreases as soon as the reference indicator is not on screen any more.

Thank you very much for your post! You are right - there is an error in the RMS measurement when the ground reference setting for the channel is moved outside the screen.

We overlooked this scenario in our tests. Sorry! It will be fixed in the next update.

[KungFuJosh]

I tried Sinisa's experiment above, and besides the Vrms bug, it's a little weird that the annotations don't match (update at the same time the measurement list does).

Thanks,
Josh

[rf-loop]

As most of peoples of course know but still good to repeat some times:
Also need take care about how to measure for get right result when signal horizontal position is "what ever".
In this example only one result is true for this continuous signal when the position of a symmetrical continuous signal in the measurement window on the time axis leads to power asymmetry. (how it need write in english... )

Stdev           (whole measurement window AC RMS)
RMS             (whole measurement window RMS)
Cycle Stdev  (one cycle  AC RMS)
Cycle RMS    (one cycle RMS)


For this 78.9 kHz continuous square wave, level 0 - 1V and with this kind of oscilloscope settings, only right result (for thermal power) is Cycle RMS
And naturally Cycle Stdev display nicely right value IF this signal is vertically symmetric -500mV to +500mV (DC removed)

So, with oscilloscope need think carefully how to select suitable measurement.
 

[Andre77]

I tried Sinisa's experiment above, and besides the Vrms bug, it's a little weird that the annotations don't match (update at the same time the measurement list does).

The grid annotations and the measurement tables are rendered in independent threads. Since you record many waveforms per second, it is unlikely that they will refer to the same waveform.

However, once you stop the measurement, you can be sure that you are seeing the matching values of the last waveform.

[KungFuJosh]

I tried Sinisa's experiment above, and besides the Vrms bug, it's a little weird that the annotations don't match (update at the same time the measurement list does).

The grid annotations and the measurement tables are rendered in independent threads. Since you record many waveforms per second, it is unlikely that they will refer to the same waveform.

However, once you stop the measurement, you can be sure that you are seeing the matching values of the last waveform.

Wouldn't it be more efficient if the annotations were rendered from the same thread as their relevant measurement on the table? Why have separate threads?

Thanks,
Josh

[Andre77]

Wouldn't it be more efficient if the annotations were rendered from the same thread as their relevant measurement on the table? Why have separate threads?

I understand why this bothers you, even though it's only visible in a screenshot taken in Run Mode with multiple waveforms per frame.

We will discuss this internally and determine how to synchronize the display.

[john74]

I got my BMO350 also this week.

No issues with the packaging and first installation.

After playing around with the functions, especially with the math channels,
I think it would be a great feature, if we could use this also as trigger input.
(not sure if this even possible or how slow the reaction time would be :-)

I am also impressed about the functions and possibilities of the AWG.

The MSO is a 'must have' since the serial decoder is already implemented.
But here I would prefer another way to display the captured data.
Most protocols are using some type of termination byte to seperate junks of data.
So maybe we can have an opion to group the data (e.g. until <CR> or after 1ms IDLE time)

Also as 'how to visualize digital signals':
here a screenshot from a xilinx GUI:


Over all, a very good scope.
@Batronix: keep up with bug fixes and improvements to make even more customers.

[jusaca]

Now that you mentioned the math channels: Are there already any plans to provide more math functions?
We really miss some filters, especially low pass filters, as a math function.

Another thing which we missed this week in the lab: Having the measurements update in roll mode as well. I'm not entirely sure how they should be triggered and over which time interval they would best be implemented, but I'm sure something could be figured out.
Our application was: We measured the output of an amplifier for a strain gauge full bridge with quite a bit of noise coupling and wanted to view the trend chart of the RMS measurement. Unfortunately that is not possible in roll mode.
Maybe in roll mode the user can enter the time interval over which the measurement function works? Like: Newest 2s or something like that.

[KungFuJosh]

I recently got a trade for a 121GW meter. The previous owner did a great job calibrating DCV (thanks Scott!), but didn't have the tools to do ACV.

I tried using my SDG6052X to calibrate 5Vrms and below, but it simply didn't have enough resolution in the voltage control level to get within 7mV of the target.

Then I remembed the Magnova's AWG has more resolution. And it worked! My only complaint is that in the mVrms range, I could only adjust 2 decimal places, as the 3rd decimal was locked at 0.
Feature request: If possible, it would be awesome if that 3rd decimal place could be adjusted in the mV range.

Even with that limitation, I got some great results. Being able to adjust to 5 decimal places in the Vrms range made a huge difference!

For the test setup, the DAQ6510 is the calibrated reference, the AWG is connected directly to the 34465A, with 1 meter banana plug cables connecting the other meters.

Thanks,
Josh

[john74]

After some time with the magnova..here another thing that could be improved (from my point of view :-)

The horizontal and vertical scaling is overloaded and hard to read.
I have attached a screenshot with a slight offset from the center.

My suggestion for the horizontal axis:
1. get rid of the "s" in the units -> that's a redundant information
2. use the same unit for all the values -> change "800u" into "0.8m" if all other values are in "m"
3. center the values to the grid
4. add the values at the top of the grid or at the bottom, not on both edges
5. maybe mark only every second grid -> that should be enough

see the changed screenshot (made with paint)

the same goes for the vertical axis,
a value with 6 decimal places is just ridiculous
(e.g. -11.225276V)
the value should be rounded to "-11.225V"
(the offset also has only 3 decimal places)

and all values should be changed to "V" not "mV" in this case
("774.724"mV should be "0.774V")

that would be much cleaner in my opinion

[Xyphro]

Adding to that:
An snap function to snap in a magnetic way to more integer values when scrolling X or Y when using the touch screen would be cool. When trying to get trigger offset back to exactly 0s or y offset to 0V(for pure psychological reasons) or any other 1/10th  fraction relative to x/div or y/div it is quite hard to do so without falling back on using the rotary buttons.
This is the same for all other toucjscreen scopes I saw so far. Always wondered why that's never done. It could be a configuration option for those who don't want this.

[Fenstergucker]

This is the same for all other toucjscreen scopes I saw so far. Always wondered why that's never done. It could be a configuration option for those who don't want this.

The R&S RTB2000 has this snap function. However, it takes some practice to get the right speed when moving so that the measurement curves snap into place in the Y direction.

Peter

[Xyphro]

Oh, nice. Can imagine, that it also requires some experimentation play phase for such a thing to behave in a way that it feels "natural".

[Martin72]

With Siglent Scopes (and not only there), the axis labels can be:
-Fixed or floating
-Left-aligned, right-aligned, or centered
-Completely disabled...

That would indeed be a useful addition to Magnova.