Magnova oscilloscope

[nctnico]

If you refrained from all the name calling (which was unwarrented anyway) and just noted that at -3dB the signal level is much higher than what Martin72 showed, I would have just said: OK you are right.

Either way, I'll admit that in hindsight I triggered too quickly on seeing very tiny signal amplitudes in the screendumps.

Still, using max-hold closer to the noise floor is prone to errors. I picked some close frequencies (84MHz and 122.75 MHz) at high & low parts on the RTM3004's FFT noise floor (which is pretty flat) and fed in a -56dB signal. Using max-hold there is a 1 dB error between the levels and the levels are off as well in absolute sense. Using FFT averaging shows the correct levels at both frequencies (also verified using a measurement on the waveform). I use max-hold with a lot of caution and prefer to use other methods when possible especially when those methods cancel noise.

[tv84]

If you refrained from all the name calling (which was unwarrented anyway) and just noted that at -3dB the signal level is much higher than what Martin72 showed, I would have just said: OK you are right.

Don't read what is not there. I didn't see any level of "name calling" that could refrain you from the simple "OK you are right".

Glad that all is cleared and that the continuous FFT with peak hold is a pretty good/simple method for BW measurement.

[Geofrey]

I just received mine, and I am happy to report that my unit was nicely packaged in a bag and arrived perfectly clean. I ordered at the same time an extra Sensepeek SQ200 hands free probe, and it was not packaged in the outer shipping carton, but in the Magnova box. Batronix may pack the oscilloscopes when ordered, after loading the ordered bandwidth license. That's allowing to add small stuff in the box.

First impression is very positive. This is nice to use. the supplied probes and accessories are nice. The 1x-10x switch is quite firm, to the point that it's safer to use both hands to switch, but I like it that way, no chance to accidentally switch it.

Just 3 things :
I tend to accidentally touch the screen with my thumb when I turn the bottom knob, launching the keyboard pop-up.
There is no French (azerty) keyboard layout option, in fact only German and English options.
Last is probably a noob question, but I have a residual overshoot on the square 1khz side generator, both with the provided Testec probes and the Sensepeek ones after adjusting the compensation. Thanks to the nice big screen it looks big but it actually measures to 3% overshoot with the Testec and 4% with the Sensepeek. Sorry, no screenshot as I had no usb stick available, and not set-up LAN for it at work yet. There is 2 extra set screws on the Testec probe hidden under plastic caps, the probe manual says these are for adjusting this overshoot, but advise against messing with as these are factory adjusted. I'm not sure what is best to do here : leave as is, adjust on the scope 1KHz output, or adjust on a signal generator ?

[Martin72]

Hi Geofrey,
I would definitely leave it that way and definitely NOT touch the factory settings of the probes.
I just tried it myself and got just under 1% overshoot, you can live with that, you have to live with that.
For the fine adjustment, I would set the time and voltage so that you get one square wave from the signal on the screen and then carefully trim it again.

Martin

[Performa01]

Last is probably a noob question, but I have a residual overshoot on the square 1khz side generator, both with the provided Testec probes and the Sensepeek ones after adjusting the compensation. Thanks to the nice big screen it looks big but it actually measures to 3% overshoot with the Testec and 4% with the Sensepeek. Sorry, no screenshot as I had no usb stick available, and not set-up LAN for it at work yet. There is 2 extra set screws on the Testec probe hidden under plastic caps, the probe manual says these are for adjusting this overshoot, but advise against messing with as these are factory adjusted. I'm not sure what is best to do here : leave as is, adjust on the scope 1KHz output, or adjust on a signal generator ?

Passive high impedance probes are a very special topic. they have to be carefully matched to the scope input they are connected to.

What do we know about that input? 1 megohm in parallel to xx picofarad. This varies widely across the instruments, hence it has to be adjusted on the probes. This is why the so called LF-compensation is universal on every passive high impedance probe and oscilloscopes traditionaly provide a calibration signal (1 kHz square wave) for that.

Yet the detailed equivalent circuit of the input of an oscilloscope frontend is a bit more complicated than just a 1 meg resistor in parallel with some capacitance. There are parasitic inductors and capacitors, lossy transmission lines and dielectric losses, which affect the frequency response at high frequencies, hence also the under/overshoot at fast edges.

Passive high impedance probes need another compensation aka HF-compensation in order to provide optimal pulse response and signal fidelity. The supplied probes from any serious scope manufacturer are usually matched to the scope they are supplied with, hence do not provide any additional adjustments. But that has not always been the case and it does certainly not apply to random generic probes from third party manufacturers.

In the good old days, scopes were delivered with generic probes that did have a user adjutable HF-compensation. These scopes also provided a proper 1 MHz calibration signal with short rise time in addition to the traditional 1 kHz signal. All this has gone out of fashion by now.

Long story short: of course you need to adjust your generic probes to match your specific scope input. Since scope inputs differ from each other, there cannot be an optimal "factory adjustment". What can a manufacturer like Testec do? Choose the best selling oscilloscope (probably some Keysight Megazoom) that is likely to be used with a certain probe and adjust to that. That is unlikely to be a perfect match for a newcomer like the Magnova - or any other scope.

So yes, you need some clean 1 MHz square wave with rise time <5 ns (ideally, it should be faster than the scope rise time) and adjust the HF-compensation in order to get the best pulse shape, with minimal over- or undershoot.

Of course, if you don't have such a signal source available, then it would be better to leave the HF-compensation alone istead of messing with it.

[ralphrmartin]

The more I use this scope, the more I like it. Anyone want to buy my old Keysight 3024T-MSO? 

[Geofrey]

Thanks Performa01, that's very helpful.
The 1KHz square wave is quite slow at around 100ns rise time. The best I would have is a white DG992, with 8ns rise time and overshoot of its own, so unless I can rent or borrow better, not messing is probably what I will do.

[Martin72]

I wonder if this effort is really necessary – don't you get the same result as me?

[Geofrey]

I'll try to get screenshots tomorrow, but I definitely have more overshoot than you. Nothing crazy, but still.
At least I learnt something.

[Geofrey]

Good thing I told myself not to mess with that, who knows what it would have been otherwise...

Shot 1 is the Sensepeek probe, 2 is a Testec without touching the high speed compensation. I am pretty sure both were overshooting 1% more Friday. I have a screenshot in the internal memory of the scope, but somehow didn't found a way to copy or move it to a thumb drive, so I can't check nor share. I was measuring with both probes on 2 channels at the same time then, not sure if it made the difference.

Not too bad, so call it ok ? Naaah...
 
Shot 3 is from my "DG992", with a BNC cable to the scope, 50 Ohms termination set on the scope. The BNC cable was bought somewhere, some time ago, with the main criteria being price, so probably not the best cable possible. On the other hand, it is carrying a only 1MHz square signal, with 8ns rise time so it shouldn't have too much influence on the results I guess. There's some ringing, which seems typical of this generator and a bit of undershoot, which I am not sure if usual or not.

Shot 4 is with a Testec, connected with the supplied BNC adapter. This looks a lot like the user manual example for high speed overshoot. While I'm at it, shot 5 is in 1x gain. As expected, slower, exponential rise.

Shot 6 is the same setup, but with the HF compensations (2 screws for that) adjusted so that the waveform looks as square as possible. Shot 7 is the result on the 1KHz signal. Interestingly there the overshoot is higher.

Shot 8 is the Testec probe adjusted to exactly match CH2 on the generator, fed to the scope CH2 via the BNC cable. With 50 ohms termination, the peak peak value is reduced compared to the 1 Vpp setting. I changed the display zoom of CH2 to compensate. I found the matching impressive. Crossing channels on the generator and then on the scope did not change the matching. Is it a more faithful representation of the signal ? No idea. Can you match the behavior of a BNC cable 50 ohms terminated with this 10x probe ? Yes definitely.

Finally, shot 9 is the result on the 1KHz square. Not much of a difference compared to previous HF compensation fiddling in terms of overshoot, a bit slower rise time.

So what do I think I learned here ?
- The bigger your screen, the higher the resolution of the scope and the lower its noise, the more there is to see, both interesting and distracting. Knowledge is power, ignorance is bliss, all that things
- The overshoot on the scope 1KHz terminals seems to be really there, and not just a probing artifact. It's rise time is too slow for the overshoot to be affected by the HF compensation adjustments in a tangible way, and its shape is also unaffected by the usual compensation.
- The Testec probes are nice to use, with good provided tips and grounding options. There is 3 compensation screws which can set to match the response of a direct BNC 50 ohms connection. Not sure if the latter is something I want.

Also, I have 2 more usability remarks on the Magnova :
- As written above, I couldn't find a way to move files around folders, both within a drive and across drives. The manual wasn't helpful either. If someone knows how to do it, I would appreciate.
- When using long time bases (>2s/div) the data isn't displayed while being recorded, but only when the capture is complete. There is a handy count down displayed at the top of the screen showing the pre-buffer filling process before trigger and then the time to finish filling it, but when I'm working on signals <1Hz, I like seeing a real time trace filling the screen.

[woody]

Also, I have 2 more usability remarks on the Magnova :
- When using long time bases (>2s/div) the data isn't displayed while being recorded, but only when the capture is complete. There is a handy count down displayed at the top of the screen showing the pre-buffer filling process before trigger and then the time to finish filling it, but when I'm working on signals <1Hz, I like seeing a real time trace filling the screen.

Is roll mode (Acquire - Roll-mode) what you are looking for?

[Tomki]

- As written above, I couldn't find a way to move files around folders, both within a drive and across drives. The manual wasn't helpful either. If someone knows how to do it, I would appreciate.

There is no a file utility yet 🤓. This will be added in a future release…

[Sariel]

 have just received my Magnova scope, and I am super excited! The first thing I tried was the XY mode, which also has persistence.
This feature is very useful for many analog circuits I build.

Attached is a picture I took while measuring some chaotic circuits and viewing the attractor on the XY display. Awesome!

I will be posting a full video on my LinkedIn account this upcoming Friday, showing many other XY views (I cannot upload long videos here).

[Tomki]

Also got a nice gadget to work in X-Y mode from Osziclock.net

[mawyatt]

Attached is a picture I took while measuring some chaotic circuits and viewing the attractor on the XY display. Awesome!

Nice display!!

Michael Kennedy's Chua Chaotic Circuit (dual op-amp) we presume.

https://people.eecs.berkeley.edu/~chua/papers/Kennedy93.pdf
https://people.eecs.berkeley.edu/~chua/papers/Kennedy93b.pdf

Best

[egonotto]

Hello,

how much memory does the Magnova have if segmented memory is used? Does it then have more than 320 Mpts?

Best regards
egonotto

[Martin72]

Channel Labeling:

I wanted to give the channels different names, so I went into the channel menu, selected Label.
That worked, see photo of display (I haven't yet figured out how to capture EVERYTHING, including open menus, with quicksave).
But then...you can't see anything.
But if you now “move” the channels (vertically), the created label appears briefly and then disappears again.
That's not ideal...

[woody]

I think you want Display - Grid - Marker label always visible?

[Martin72]

Thank you!
I wouldn't have thought of that spontaneously; I would have assumed something like that would be in the label menu...

[Martin72]

Now they are visible, albeit very, very small...

[Andre77]

how much memory does the Magnova have if segmented memory is used? Does it then have more than 320 Mpts?

There is no seperate "segmented memory" mode on the Magnova. All acquisitions take place based on the same mechanism.

That being said, the Magnova records up to 981.6 Mpts in total, up to 327.2 Mpts per recording. For example, up to 95 segments can be recorded at 10 Mpts (no matter if in History or "normal" mode). The exact amount depends on the memory depth per recording.

The "History" mode allows for inspection of all waveforms in memory. When initiating acquisitions from History mode, only one of the acquired waveforms will be displayed, leaving even more time for the acquisition itself. But the mechanism is not different to the general one.

[Martin72]

Decoding, here the UART protocol...
The Batronix Demoboard outputs Rx and Tx separately, so you need 2 channels to “see” both.
That's not a problem either, my siglent scopes, for example, display RX and TX simultaneously - in one decoder channel.
Of course, it may be that I didn't look in the manual properly again, but the way I see it, I need a decoder channel each on the Magnova to display RX and TX simultaneously.
That doesn't really matter, because the Magnova has four decoder channels.
But... now the table comes into play.
I can display the table for RX OR for TX, which I find a bit cumbersome.
Also, a larger separation between the lines would be more favorable to be able to read the individual values, so they just appear as on an “assembly line”.
Siglent provides a table that shows the values of RX and TX, each offset by the response, which automatically makes it easier to read.
However, the Magnova has a big advantage when it comes to character display.
The quotation marks around each character on the Siglent interfere with readability, which is not the case with the Magnova.

[Anthocyanina]

have just received my Magnova scope, and I am super excited! The first thing I tried was the XY mode, which also has persistence.
This feature is very useful for many analog circuits I build.

Attached is a picture I took while measuring some chaotic circuits and viewing the attractor on the XY display. Awesome!

I will be posting a full video on my LinkedIn account this upcoming Friday, showing many other XY views (I cannot upload long videos here).

(Attachment Link)

very nice! could you attach a screenshot of it as well? thank you!

[egonotto]

how much memory does the Magnova have if segmented memory is used? Does it then have more than 320 Mpts?

There is no seperate "segmented memory" mode on the Magnova. All acquisitions take place based on the same mechanism.

That being said, the Magnova records up to 981.6 Mpts in total, up to 327.2 Mpts per recording. For example, up to 95 segments can be recorded at 10 Mpts (no matter if in History or "normal" mode). The exact amount depends on the memory depth per recording.

The "History" mode allows for inspection of all waveforms in memory. When initiating acquisitions from History mode, only one of the acquired waveforms will be displayed, leaving even more time for the acquisition itself. But the mechanism is not different to the general one.

Hello,

thank you, does the Magnova have an internal memory in which waveforms can be saved? If so, how big is the memory?
Can the entire 981.6 Mpts be saved and transferred to a PC via LAN or USB?
If so, how long does the transfer to the PC take?

Best regards
egonotto

[Andre77]

thank you, does the Magnova have an internal memory in which waveforms can be saved? If so, how big is the memory?
Can the entire 981.6 Mpts be saved and transferred to a PC via LAN or USB?
If so, how long does the transfer to the PC take?
egonotto

Hello egonetto,

The internal memory size is 1.8 GB.

Various remote connections are already available and functional, but have not yet been officially released. A recent test via USBTMC has shown a transmission speed of 16 Mpts/s. However, we are still working on this and it may well be that it will become even faster. Perhaps also good to know: The measured values in the Magnova have a 16-bit resolution and are also transmitted in this resolution.

Best regards,
Andre