Oscilloscope selection for Office

[yashrk]

Guys, we have new scopes from R&S, see a comparison below.
Which is more important deep memory or sampling rate?

Hi,

good comparision, but one short correction. The sampling rate of RTA and RTM is 5 GSa/s.

Markus

I have just shown sampling rate per channel not interleaved for all the scopes

[Markus@RohdeScopes]

Guys, we have new scopes from R&S, see a comparison below.
Which is more important deep memory or sampling rate?

Hi,

good comparision, but one short correction. The sampling rate of RTA and RTM is 5 GSa/s.

Markus

I have just shown sampling rate per channel not interleaved for all the scopes

Yes, this could be a fair way to compare sampling rate. But in this case the instruments from our competion also have 2.5 GSa/s per channel. You can send me a PN and I show you the parts from the data sheets.

Markus

[yashrk]

Guys, we have new scopes from R&S, see a comparison below.
Which is more important deep memory or sampling rate?

Hi,

good comparision, but one short correction. The sampling rate of RTA and RTM is 5 GSa/s.

Markus

I have just shown sampling rate per channel not interleaved for all the scopes

Yes, this could be a fair way to compare sampling rate. But in this case the instruments from our competion also have 2.5 GSa/s per channel. You can send me a PN and I show you the parts from the data sheets.

Markus

yup, my bad 

[nctnico]

Which is more important deep memory or sampling rate?

That depends on your application.

Yes, and as I mentioned before that I want to use it mainly for embedded application decoding serial lines (say spi, i2c, can) and some general electronics work, looking at analog sensor data, etc.

It will be helpful if you could give me examples where deep memory or sampling rate will be important.

I'd go for deep memory for generic microcontroller use. If push comes to shove then you can buy a decent >1GHz on the used market. More bandwidth usually means more noise (both acoustic from the fan and on the signals) as well. 100MHz to 200MHz is enough bandwidth.

[Rich@RohdeScopesUSA]

Guys, we have new scopes from R&S, see a comparison below.
Which is more important deep memory or sampling rate?

Hi,

good comparision, but one short correction. The sampling rate of RTA and RTM is 5 GSa/s.

Markus

I have just shown sampling rate per channel not interleaved for all the scopes

Yes, this could be a fair way to compare sampling rate. But in this case the instruments from our competion also have 2.5 GSa/s per channel. You can send me a PN and I show you the parts from the data sheets.

Markus

yup, my bad 

A few more small tweaks - the way you show memory currently in the table is per channel (which makes sense as Tek doesn't offer an interleaved option) - in this case, the Keysight's are 2M, not 4M.  You could also look at it as maximum memory depth available (allowing interleaving), which in that case it is correct for Keysight and Tek, but the R&S units would be 200M for the RTA4000, 80M for the RTM3000 and 200M for the RTE1000. 

The Keysight's and Tek scope also offer UART decode/triggering (didn't see those listed for them).

-Rich

[MarkL]

The MSO3000T is a very fast scope with a nice interface but I cannot justify the max 4Mpts of memory interleaved, it is just too small (@ 5GSample/s it is just 800us of data).

If your use cases are typical, you will care about that maybe 5% of the time.  Other scope manufacturers may brag about having a zillion gigasamples of acquisition memory, but extensive usability and performance compromises are often necessary in order to make that happen.

There are more issues with the way Keysight's ASIC works. For example the next to useless reference traces. If you pull the curtain up a bit further there are more problems to be found. I used to own a MSO7104A so I'm familiar with the limitations of the architecture.

What is wrong wit the reference traces?

You can't move, seperate or scale them. Most other DSOs have up to 4 seperate reference traces you can (at least) enable/disable. Moving, scaling and using the reference trace as input for math are very usefull additional features. On the Agilent MSO7104A you get a white line on screen and that's it and I doubt it is any different on the current Megazoom ASIC based scopes.

On later versions of Megazoom (at least on MSOX3104A), you can can scale the reference, offset it, and skew it in time.  You can also save it and load it.  But you can only display one of the two available reference memories at a time, and it's only 64kpts and not the full record length so zooming in on it is limited in resolution.  And still no math functions.

It's only useful as a visual aid for comparisons and not much else.  I'm not sure why they bothered to implement it except for the marketing check box.


EDIT: I took a quick look in the more recent Programmer Manuals.  For at least the X4000 and X3000T, the manual lists the reference waveform as a possible input for MATH functions.  So, presumably this issue has been addressed but I'd still want to see it in action.

[Keysight DanielBogdanoff]

1. I am confused between two scopes series of Keysight 3000 and the 4000 series except for the bigger size screen what is the difference? 

I'm a little late to the party, but the 3000T and 4000 X-Series are very similar with a few notable exceptions. Also, the 4000 X-Series generally runs ~15% more $.

The 3000T X-Series perks:
 - Display a math channel & FFT at the same time
 - optional 8 digit counter + totalizer (vs. 5-digit counter)
 - 1.6 ppm timebase accuracy (vs. 10)

The 4000 X-Series
 - Bigger screen 
 - optional dual channel WaveGen, larger output voltage
 - 4 Match channels (vs. 2)
 - 10 MHz reference in/out
 - LAN/VGA standard (vs. optional)
 - Can use 4 active probes without extra power supply (vs. 2)
 - 3 USB host ports (vs. 2)
 - 700 MHz probes included (vs. 500 MHz)

If you ever need 4 active probes at once (like for SMPS testing), the 4000 X-Series is a way better option.

Keep in mind that if you buy the Keysight 500 MHz scope the upgrade to 1 GHz is a replacement of the mainboard - there is no software/firmware upgrade.

It is a new board to go from 500 MHz to 1 GHz. But, buying the 500 MHz scope then upgrading to 1 GHz will cost you the same $ as just buying the 1 GHz from the beginning. So, the only cost is in lost lab time.

EDIT: I took a quick look in the more recent Programmer Manuals.  For at least the X4000 and X3000T, the manual lists the reference waveform as a possible input for MATH functions.  So, presumably this issue has been addressed but I'd still want to see it in action.

Confirmed

[yashrk]

Hey just got this scope to play with for a week 

Any one wants to test any thing about the scope ??
Or have any suggestions ?