First picture on EEVblog of the new R&S MXO4 series oscilloscope :)

[Domitronic]

I got the RTB2004-COM4 promo, so why not wait for the MXO4-COM4 package....

Well, looking at my budget i would need to wait for a MXO2-COM4 package 

[Messtechniker]

Just downloaded the 574 page manual. 
Will keep me out of trouble over the next few days.

[jjoonathan]

I am still trying to figure out the FFT architecture. From this video:


https://youtu.be/acE3d4TpiW4?t=46


it looks like changing the FFT settings also changes the time domain acquisition settings. Reducing the RBW 500Hz -> 400Hz -> 300Hz also reduces the sample rate from 1000 -> 833 -> 417 MS/s, which is what I would expect if the FFT were dependent on acquisition data (lower RBW => longer FFT window => longer acquisition) but not what I would expect from an independent FFT pathway.


The ability to have an independent FFT acquisition path would be very helpful for applications that require low-bandwidth-high-res time domain and low-bandwidth-high-frequency RF, e.g. looking at a power rail, spi bus, or VCO control at the same time as a corresponding 900MHz RF burst. On my bench, this is a very common situation. If the FFT must be calculated off acquisition data, the high fundamental frequency of the RF burst demands a high sample rate even though "morally" it has a lower information requirement driven by RBW+span. If the FFT and time domain have separate acquisition paths, these requirements are not in conflict, because the FPGA can send an efficient low-bandwidth-high-res time-domain trace alongside a low-bandwidth-high-frequency spectrogram, rather than a big high-bandwidth block of full-rate samples suitable for calculating both of them.


I want to believe that the MXO resolves this conflict with independent measurement pathways, but the video makes it look like the MXO has independent FFT in the more limited sense as the older R&S scopes, where you can trade off RBW against the number of stacked/overlapped FFTs, but you cannot have low-bandwidth-high-res time domain at the same time as low-bandwidth-high-frequency spectra unless you fill your acquisition memory with high-bandwidth samples. Could you clarify which definition of "independent" is in play here? Thanks. (EDIT: I made & pulled this same post yesterday, not wanting to cause out-of-hours work, but I was too slow and got an answer by PM anyway. Apologies.)

[Rich@RohdeScopesUSA]

I am still trying to figure out the FFT architecture. From this video:


https://youtu.be/acE3d4TpiW4?t=46


it looks like changing the FFT settings also changes the time domain acquisition settings. Reducing the RBW 500Hz -> 400Hz -> 300Hz also reduces the sample rate from 1000 -> 833 -> 417 MS/s, which is what I would expect if the FFT were dependent on acquisition data (lower RBW => longer FFT window => longer acquisition) but not what I would expect from an independent FFT pathway.


The ability to have an independent FFT acquisition path would be very helpful for applications that require a low time-domain sample rate but a high RF sample rate, e.g. looking at a low frequency power rail, spi bus, or VCO control at the same time as a corresponding 900MHz RF burst. On my bench, this is a very common situation. If the FFT must be calculated off acquisition data, the high fundamental frequency of the RF burst demands a high sample rate even though "morally" it has a lower information requirement driven by RBW+span. If the FFT and time domain have separate acquisition paths, these requirements are not in conflict, because the FPGA can send an efficient lowpassed time-domain trace alongside an efficient spectrogram, rather than a big inefficient block of full-rate samples suitable for calculating both of them. To the user, memory efficiency is visible as the maximum acquisition time when RF+DC pathways are active. If they are dependent, max acquisition time is milliseconds, but if they are independent, it can be many seconds.


I want to believe that the MXO resolves this conflict with independent measurement pathways, but the video makes it look like the MXO has independent FFT in the more limited sense as the older R&S scopes, where you can trade off RBW against the number of stacked/overlapped FFTs, but you cannot have low-sample-rate time domain signals at the same time as high-sample-rate FFTs. Could you clarify which definition of "independent" is in play here? Thanks. (EDIT: I made & pulled this same post yesterday, not wanting to cause out-of-hours work, but I was too slow and got an answer by PM anyway. Apologies.)

Good question - and I appreciate you not wanting to cause out-of-hours work    I PM'd our planner to get an answer - here's what they replied:

Historically, oscilloscope FFTs have had dependencies between frequency domain settings (CF, span, RBW) and time domain settings (RL, timebase).  If you changed the RL or timebase in the time domain this would cause a change to the FFT RBW and the converse was also true.  For the MXO 4 we removed these dependencies.  So you can set the timebase as you want and set CF/Span and RBW independently.  With that said, you are correct that adequate sample rate is still a requirement for bandwidth in either the time domain or frequency domain.  Nyquist still rules.  For MXO 4 the sample rate determines a common maximum bandwidth value for both time and frequency domain.

-Rich

[jjoonathan]

Ok, so it sounds like the FFT settings are independent but constraints still propagate to acquisition settings (low RBW -> long acquisition, high frequency -> high sample rate). That's still a very useful capability -- I certainly appreciate it on my RTO -- but it still leaves the tradeoff where if I simultaneously want two low-bandwidth things (say, 1MHz of DC and a 1MHz sliver of RF) I still have to fill my acquisition memory with full-bandwidth samples (2.5GS/s rather than, say, 1*5MS/s + 1*5MS/s).

Even so, the new capabilities are beautiful and I look forward to the day when the platform refresh works up to MXO6 tier 

[pdenisowski]

The 18 short (~1 minute) videos showing different features of the MXO4 are now live on the Rohde and Schwarz YouTube channel

https://www.youtube.com/user/RohdeundSchwarz/videos

[Messtechniker]

Pity that the input BNCs on the MXO4 are not colour coded like on my R&S HMO1022.
Here LEDs next to the BNCs are used to colour code the active channels. Very useful IMO.

Moreover, on the MXO4 the channels are abbreviated as C1 ... C4
instead of the more usual CH1...CH4.

For us audio types it would be nice to know - just as a matter of interest - the true
distortion level of both generator and scope at 1 kHz.

The specs say: <0.1 % (-60dB) below 10 MHz. I expect the MXO4 to be much
better than that at 1 kHz.

Also a pity, that I do not have a use case for the MXO44 or even a MXO42 (+ generator option)
should a two channel version ever become available. Even a two channel MSOX4 with generator
option will probably be well over 5k Euros. Oh well.

My HMO1022 can do everything I need and also much more.

Toddles off to read more about this beast 

[egonotto]

The 18 short (~1 minute) videos showing different features of the MXO4 are now live on the Rohde and Schwarz YouTube channel

https://www.youtube.com/user/RohdeundSchwarz/videos

Hello pdenisowski,

in the thread:
"https://www.eevblog.com/forum/testgear/rigol-hdo1000-and-hdo4000-12bit-oscilloscopes-launched-in-china/" we want to analyze and compare the noise behavior of different 12 bit scopes in more detail.
It would be interesting if we also got data of the new impressive MXO 4.
Eight traces would be useful for this:

Maybe you should not use the first channel, because with my RTA4000 the display there also disturbs something.

All 8 measurements with 400 Mpoints (or 800 Mpoints) memory and 5 GSa/s sampling rate and open inputs:

1 mV/div 50 Ohm maximum bandwidth (200 MHz)
1 mV/div 50 Ohm 20 MHz bandwidth
1 mV/div 1 MOhm maximum bandwidth (200 MHz)
1 mV/div 1 MOhm 20 MHz bandwidth

1 V/div 50 Ohm maximum bandwidth (200 MHz)
1 V/div 50 Ohm 20 MHz bandwidth
1 V/div 1 MOhm maximum bandwidth (200 MHz)
1 V/div 1 MOhm 20 MHz bandwidth

Thanks in advance.

Best regards
egonotto

PS: In datasheet page 7
"Memory depth
 standard                       400 Mpoints with 4 active channel (single),
                                     400 Mpoints with 2 active channel (run)
R&S®MXO4-B108 option
                                     800 Mpoints with 2 active channel (single),
                                     800 Mpoints with 1 active channel (run)
"

Is this correct? I would guess that the memory is larger with 2 channels than with 4 channels with (single).

Is there is more detailed information on "The maximum available memory depth depends on the bit depth of the acquired data and, therefore, on the settings of the acquisition system, such as
decimation mode, waveform arithmetic or high definition mode"

[jjoonathan]

Not a 12 bit scope, but I took the same measurements on my 8 bit RTO with 20MHz and 200MHz HD filters:

https://www.dropbox.com/sh/e00tn9mtlpsd6gx/AABKMUdwkqgnRjF2VQDP6k_Oa?dl=0

(all traces are 10GS/s with filters to reduce bandwidth and improve vertical resolution)

Back-of-the-envelope says it starts at ENOB 7, the filters chop noise bandwidth by 20x, so ideally another ~2 bits and ENOB around 9, which is where most 12 bit scopes seem to land anyway. I'd love to see the apples-to-apples comparison of what it actually achieves, though -- I suspect the ENOB figures on the 12 bit scopes might be suffering from some kind of worst case and that the 3 least significant bits aren't typically complete rubbish.

[balnazzar]

I hate the shiny screen though, so... 

Buy a tablet matte screen protector. Then cut it.

They are made with different levels of "matteness".

[RBBVNL9]

Buy a tablet matte screen protector. Then cut it.

Over time, there are more and more instruments in my lab that have shiny screens, small or large. From a company called ViaScreens I bought their "Wacom Cintiq Pro 24 Matte Screen Protector". Not that I have that Wacom device, but it's among the largest protectors they have, and I think they even send you two if you order it. Works well with digitizer screens (after all, that's what the Wacom is about).

More than enough to cut out parts for all the shiny devices I have here ;-)

[Fungus]

You can probably get any size screen protector on Aliexpress.

I cut up phone screen protectors for all my multimeters, etc.

[Neurosurg]

Not a 12 bit scope, but I took the same measurements on my 8 bit RTO with 20MHz and 200MHz HD filters:

https://www.dropbox.com/sh/e00tn9mtlpsd6gx/AABKMUdwkqgnRjF2VQDP6k_Oa?dl=0

(all traces are 10GS/s with filters to reduce bandwidth and improve vertical resolution)

Back-of-the-envelope says it starts at ENOB 7, the filters chop noise bandwidth by 20x, so ideally another ~2 bits and ENOB around 9, which is where most 12 bit scopes seem to land anyway. I'd love to see the apples-to-apples comparison of what it actually achieves, though -- I suspect the ENOB figures on the 12 bit scopes might be suffering from some kind of worst case and that the 3 least significant bits aren't typically complete rubbish.

Jjonathan could you post a picture of yours R&S RTO1044 in HQ. In my opinon the RTE, RTO scopes looks much more professjonal and well doned

[jjoonathan]

Not a 12 bit scope, but I took the same measurements on my 8 bit RTO with 20MHz and 200MHz HD filters:

https://www.dropbox.com/sh/e00tn9mtlpsd6gx/AABKMUdwkqgnRjF2VQDP6k_Oa?dl=0

(all traces are 10GS/s with filters to reduce bandwidth and improve vertical resolution)

Back-of-the-envelope says it starts at ENOB 7, the filters chop noise bandwidth by 20x, so ideally another ~2 bits and ENOB around 9, which is where most 12 bit scopes seem to land anyway. I'd love to see the apples-to-apples comparison of what it actually achieves, though -- I suspect the ENOB figures on the 12 bit scopes might be suffering from some kind of worst case and that the 3 least significant bits aren't typically complete rubbish.

Jjonathan could you post a picture of yours R&S RTO1044 in HQ. In my opinon the RTE, RTO scopes looks much more professjonal and well doned

Here, I have performed a rigorous evaluation of the professional aspects of the scope, just for you 

[Martin72]

Now I know for what it´s good for..

[jjoonathan]

Ssh, don't tell marketing, or next time we will need to buy an option RTO-K666 to play doom 

[egonotto]

Hello,

what version of Doom is this? It looks like Doom 4.

Best regards
egonotto

[jjoonathan]

Doom Eternal, the one from 2020 with good gameplay. Highly recommended.

[Martin72]

Looking on the batronix site, the prices for the options are heavy ones.
Not under 3K/each (appx and incl. 19%VAT), bode plot, decodings, more memory and so on...
But one thing is interesting, you can upgrade the bandwith from 200Mhz up to the topmodel 1.5Ghz, so it´s always the same hardware.
Nice thing for hackers..
And interesting the second, buying the upgrade from 200Mhz to 1.5Ghz is NOT more expensive than buying directly the 1.5Ghz model.

[Fungus]

But one thing is interesting, you can upgrade the bandwith from 200Mhz up to the topmodel 1.5Ghz, so it´s always the same hardware.
Nice thing for hackers..

You probably need to go inside and solder JTAG wires to the PCB though, and working hacks are thin on the ground.

https://www.eevblog.com/forum/testgear/why-bandwidth-hacks-difficult-at-rs-devices/

And interesting the second, buying the upgrade from 200Mhz to 1.5Ghz is NOT more expensive than buying directly the 1.5Ghz model.

Are you sure the 1.5Ghz model doesn't have other stuff included as standard?

[goaty]

Are there any review videos coming up on the MXO4 ? Seems to me to be too quiet around that device - even if I cannot afford it, it would be interesting.

[Martin72]

Same here, around 10k makes it nearly (but not completely) impossible to buy in private.
But I would love to see "more" from it, it´s a very sexy scope from the first view..

[luudee]

Same here, around 10k makes it nearly (but not completely) impossible to buy in private.
But I would love to see "more" from it, it´s a very sexy scope from the first view..

Me too. If could get a fully loaded, top of the line  MXO4 for around $10K I would go for it immediately.

But it looks like a fully configured MXO4 right now would cost me $30K. That's a bit harsh.



Maybe R&S could do an introductory Christmas price ?!     R&S hint, hint ...     



luudee

[EEVblog]

Are there any review videos coming up on the MXO4 ? Seems to me to be too quiet around that device - even if I cannot afford it, it would be interesting.

I've been busy with other videos and scopes, and there was this 
It's being swapped tomorrow.

https://odysee.com/@eevblog:7/MXO4-Screw-Loose:2

[goaty]

I've been busy with other videos and scopes, and there was this 
It's being swapped tomorrow.

https://odysee.com/@eevblog:7/MXO4-Screw-Loose:2

Ah come on, really ? Would have made the "take it apaaaat" one screw faster.