EEVblog #1218 - 1GHz Oscilloscope Spec Shootout

[EEVblog]

Spreadsheet spec and feature comparison of eight 1GHz bandwidth oscilloscopes
Also a look at 2nd hand scopes.
Siglent SDS5000
Rigol DS7000 (500MHz)
Rohde and Schwarz RTM3000
Keysight 3000X
Keysight 4000X
Tektronix Series 3 MDO
Tektronix Series 4 MSO
Lecroy Wavesurfer 3000

[tautech]

SDS5000X will have Bode plot added in coming firmware and PA added too as an option AFAIK.
Performa01 has made comment about this in a thread.

The Arb option is a 14 bit SAG1021 module as you've mentioned.

Oh, and as these run dual 5GSa/s ADC's, each has 250M of support for a total of 500M.

[chickenHeadKnob]

to put in an external 1080p hdmi out is maybe $20 or less worth of parts. inexcusable these days

[EEVblog]

SDS5000X will have Bode plot added in coming firmware and PA added too as an option AFAIK.

Yes, confirmed, I heard from them, being added around Q3

[EEVblog]

to put in an external 1080p hdmi out is maybe $20 or less worth of parts. inexcusable these days

One problem may be the data mapping system.
Keysight for example have a fixed 800x600 in the ASIC, and map samples to the screen all inside the ASIC.
So you'd have to have 1080p built into the mapping system from the get-go with a fast data path

[2N3055]

From what I saw while  testing it a bit, on R&S 3000 series, 10 bit is nice but not showstopper.. R&S has really nice low noise front end, that contributes more than 10 bit itself. Also it has lots of memory, but no search on basic serial protocols (I2C, SPI, UART), and for what little it has search for, search is only on current buffer, there is no search trough segments/historic buffers). So you have 50000 segments, and you have to go manually through them. Crazy. On Keysight 3000T you have 1000 segements max, but it can search and decode transparently trough it.. So R&S memory is great but useless because there are no tools to use it. On Keysight it works, memory is small but there are tools to use it to the max. Result: Keysight gives you 100% usability from 4Mpoints and 1000 segments, R&S gives you 0% of 400MPoints. Which is zero. Fail. A shame really, i wanted to buy it, but in the end, looks good on paper, but kept stumbling upon "really, it can't do that?" moments..
R&S3000 is great. 10Bit, good front end. But Keysight 3000T still holds its ground. There are many things R&S 3000 still can't do. Went and bought MSOX3000T. 6 months later, still would do the same.

Funny enough, Siglent 5000X starts to look better rounded scope than R&S for general use. They added search on analog events on segments on first FW update. There is more to come. Soon it will be more feature complete that R&S3000...
Interesting times.

[fanOfeeDIY]

I am totally behind updating the Oscilloscope chart at
https://www.eevblog.com/forum/testgear/digital-oscilloscope-comparison-chart

I really appreciate if anybody would add comments on the google sheet for the new Tek, Siglent, R&S models at the bellow link.
Any partial comments is fine, it is very beneficial.

https://docs.google.com/spreadsheets/d/1rtPqMAkNw2bSkqocuc7zcYkl57fsdPVlxzyqcJmbKIc/edit#gid=197124678

[branadic]

Hi Dave,

I'm missing Teledyne LeCroy HDO4104A(-MS), HDO6104A(-MS) and HDO8108A with 12bit ADC in your comparison.

-branadic-

[0xdeadbeef]

What's not really covered is trigger capabilities (e.g. complex/qualified triggers), (automatic) measurement possibilities (gating, measurement over the whole sample buffer vs. screen buffer etc.) and search/analysis features. These things are much more important than a peak detect or the waveform update rate in many applications.
Besides, it almost sounds in the video as if the Tek Series 4 would be the 1st scope to offer 12bit ADCs. This is not the case of course. E.g. there's the LeCroy HDO4000 line for quite some years which probably in the same price range as the Series 4, probably even cheaper.
Last but not least, even if a bit older, the WaveSurfer 510 seems to be missing. It's in the same price range as the Tek Series 3 or Keysight 4000X but offers 10GS/s and is a "real" LeCroy, not result of a Siglent cooperation as the WS3000(Z).
Anyway, personally I have high hopes for the Siglent SDS5000X. It's just s shame that the 350MHz model can't be updated to 1GHz and the 500MHz is a bit outside my limit. Still, it's the only scope in the contest with somewhat reasonable base prices and that comes with the basic decoders for free (without time limited discounts). I.e. it seems to be a bit unfair to to this comparison now, shortly after the R&S discount has started. Just a few weeks ago the bare naked 1GHz RTM3K would have been in the upper price range of this comparison with quite pricey options and not even the segmented memory for free (which is included even in the WS3000Z).

[thm_w]

Hi Dave,
I'm missing Teledyne LeCroy HDO4104A(-MS), HDO6104A(-MS) and HDO8108A with 12bit ADC in your comparison.

$18,000, $20 ish?, and $32,000.
The first one maybe, but the second two are too expensive for the comparison.

What's not really covered is trigger capabilities (e.g. complex/qualified triggers), (automatic) measurement possibilities (gating, measurement over the whole sample buffer vs. screen buffer etc.) and search/analysis features. These things are much more important than a peak detect or the waveform update rate in many applications.

He did put zone trigger. But thats the kind of thing you'd need 20 more lines for on the sheet. Most of them have the basics.

SDS5000X: Edge, Slope, Pulse, Window, Runt, Interval, Dropout, Pattern, Qualified and Video (HDTV supported), zone (2 zones).
MSO7000: Edge, Pulse, Slope, Video, Pattern, Duration, Timeout, Runt, Window, Delay, Setup/Hold, Nth Edge, serial protocol trigger, zone (2 zones).
RTM3000: Edge, width, video, pattern, runt, rise time, fall time, serial bus, line, timeout.

[2N3055]

DSOX/MSOX3000T
Trigger:
Edge Trigger, Edge then Edge Trigger, Pulse Width Trigger, Pattern Trigger, OR Trigger, Rise/Fall Time Trigger, Near Field Communication (NFC) Trigger, Nth Edge Burst Trigger, Runt Trigger, Setup and Hold Trigger, Video Trigger, Zone Qualified PostTrigger (2 inclusion/exclusion zones)
Serial Trigger /decode:
I2C, SPI (2/3/4 wire), RS232/422/485/UART, CAN, CAN-dbc, CAN-FD (CAN-FD ISO), LIN, LIN symbolic, SENT, CXPI, FlexRay, MIL-STD 1553, ARINC 429, USB PD, I2S, User-definable Manchester, User-definable NRZ

[nctnico]

I think there is an error in the overview: the RTM3000 definitely has color grading and you can set it to multiple modes as well (including reverse brightness).

[Performa01]

Seems like some members here think that the price/performance ratio of the SDS5000X makes it an interesting option, so I thought I’d add my comments and try to clarify a few things:

Like any other Siglent X-series DSO, the SDS5000X does have color grade display mode of course.

The new Bode Plot II up to 120MHz with 3 channels, selective detector and >110dB dynamic has been introduced in the entry level SDS1004X-E for now and a further refined version (likely to at least 500MHz) will of course be ported into the SDS5000X.

Despite the fact that every vendor offers a PA (Power Analysis) option (Siglent has it too for SDS2000X), I still believe that not many will really miss/buy/use one. Nevertheless, we’ll get it for the SDS5000X at some point. Quality and feature set of these PA options vary quite a bit, and I think it’s a good thing that Siglent take their time and are willing to provide a good and comprehensive solution.

Comparison of Sample rate and memory depth between different brands can be misleading. The SDS5004X provides 2x5GSa/s and 2x250Mpts as a standard, so for example there is absolutely no difference to 4 channel Rigol DS7000 (with memory option) except when you stick to single channel use.

With regard to the triggers, Siglent might add some more in the future (like n^th edge, setup&hold), but much more importantly all triggers already support the “holdoff by event” feature. With this, you can’t only trigger on the n^th edge, but also n^th slope (risetime, falltime), n^th pulse, n^th window … and so on.

[EEVblog]

What's not really covered is trigger capabilities (e.g. complex/qualified triggers), (automatic) measurement possibilities (gating, measurement over the whole sample buffer vs. screen buffer etc.) and search/analysis features. These things are much more important than a peak detect or the waveform update rate in many applications.

He did put zone trigger. But thats the kind of thing you'd need 20 more lines for on the sheet. Most of them have the basics.

Correct. Another 20 lines, and then you can have variation in capability within some of those as well.
It gets insane very quickly.

[EEVblog]

I.e. it seems to be a bit unfair to to this comparison now, shortly after the R&S discount has started. Just a few weeks ago the bare naked 1GHz RTM3K would have been in the upper price range of this comparison with quite pricey options and not even the segmented memory for free (which is included even in the WS3000Z).

It doesn't matter when you do the comparison, it's always going to be "unfair" to someone.

[0xdeadbeef]

It doesn't matter when you do the comparison, it's always going to be "unfair" to someone.

True, but at this very point, only the 1GHz model of the RTM3K is drastically discounted. Actually, while many of us dream of a >= 1GHz scope, very few hobbyists will be willing to invest >= 10k-20k USD/€. So a 500MHz shootout would have been probably more interesting for the target audience and also a logical step when including the MSO7000. And when comparing the 500MHz models right now, even the base model RTM3k-54 (without MSO option) costs more than twice as much as an SDS5054X.

Besides, the current Rigol discount for the DS7000/MSO7000 doesn't only include all options, but also a frequency upgrade for the 250MHz and 300MHz models.
https://www.rigol.eu/promos/
So buying a MSO7054 ($11299) instead of an MSO7034 ($8299) would make as much sense as buying the RTM3K base model right now.

[nctnico]

But how long will it take Rigol to fix the bugs? In my experience it is better to buy an instrument which works for real rather than one which looks nice on paper.

[0xdeadbeef]

I woudn't argue with that. IMHO, the DS7000/MSO7000 is still too expensive even considering free options and bandwidth discount. I only said the free bandwidth upgrade should be mentioned.

[Fungus]

But how long will it take Rigol to fix the bugs?

There's an update scheduled this month.

[Karel]

But how long will it take Rigol to fix the bugs?

There's an update scheduled this month.

The DS6000 series is still waiting for bugfixes and, yes, those bugs were reported (and confirmed) when it was still their "flagship"....
But they didn't care. Probably because they didn't sell a lot of them.

[Dubbie]

Dave mentioned something in the video about zone triggers. When I bought my scope I didn’t think I would get much use from them. Turns out I use them all the time. One thing they are great for is a really fast way to set up a trigger hold off. Just draw an exclude box to the left of your trigger point. The are tons of dumb tricks like that which can be done the “proper” way with other triggers, but when you’re in the flow you can do a ton with on-the-fly zone triggers.  Anyone else use them like this?

[NiHaoMike]

One problem may be the data mapping system.
Keysight for example have a fixed 800x600 in the ASIC, and map samples to the screen all inside the ASIC.
So you'd have to have 1080p built into the mapping system from the get-go with a fast data path

Limiting an ASIC to an ancient 800x600 seems to be a big overlook in future planning? If I were designing it, I would probably pick 1080p or 2.5K as the limit since LCDs with those resolutions are very common. Maybe even 4K to really be futureproof - ASICs are expensive to design but (relatively) cheap to manufacture.

[Dubbie]

That ASIC’s specs were probably set almost 15 years ago. 800x600 probably seemed extravagant at the time.

[thm_w]

Dave mentioned something in the video about zone triggers. When I bought my scope I didn’t think I would get much use from them. Turns out I use them all the time. One thing they are great for is a really fast way to set up a trigger hold off. Just draw an exclude box to the left of your trigger point. The are tons of dumb tricks like that which can be done the “proper” way with other triggers, but when you’re in the flow you can do a ton with on-the-fly zone triggers.  Anyone else use them like this?

Thats a good idea.
With the hold-off you are just dialing in a number which is not immediately obvious where it will occur, at least on my scope. They could make hold-off a bit more intuitive by showing a vertical line/arrow when setting it or allowing you to drag it around with your finger.

[NiHaoMike]

That ASIC’s specs were probably set almost 15 years ago. 800x600 probably seemed extravagant at the time.

1024x768 was a common resolution for laptops back then, with 1280x1024 being common on desktops and higher end laptops. Anything less was practically unheard of except on portable devices.