EEVblog #1146 - New Low Cost Rigol MSO5000 Oscilloscope

[2N3055]

And then, IF after a year sales is dropping, they can still leak some hacking information to sell another 50.000 units for $100 profit each.

That assumes that competition does nothing. It is good that they think they can play with the big guys, but by all signs they are deeply mistaken. We'll see how it turns out.

I doubt anyone will go back to Rigol once they try R&S.

There a few thing Rigol missed. Hires etc.. Platform is few months old, so lots of debugging needs to be done.. But, it has craploads of sampling mem, it's hardware accelerated decodes and FFT, it can have 4 decodes at the same time...  Stupid 15000 € R&S RTM3000 doesn't even have search on I2C and SPI, and no search on 400 Ms of segmented memory. Talk about premium product.
Rigol 5000 pricing is crazy for now, but they will drop it if it doesn't sell..
And if you compare ds5000 to Keysight 2000 ...  If they make it stable enough, for MSO it could be very useful. Even for a bit more than 999 USD.

But at this point it is kinda funny product, together with DS7000 platform. No hires, no custom math (math is identical to DS1000Z right now), and no search or analysis on segmented memory.
If they add those, than it would be very interesting...

[David Hess]

Has anyone ever measured more than 200MHz without 50R? Sure, a good 1MR 10x or 100x probe can do it, but it comes with excessive capacitive loading on probe point, as well as signal quality issues due to the lossy coax (limits slew rate) and the ground loop (limits flatness).

I have gone up to 350MHz when either probing a terminated 50 ohm transmission line, which is how probe bandwidth is specified, and when using an inline coaxial to probe tip adapter but that presents the same 25 ohm source impedance.  500MHz is feasible under these conditions but I do not have an oscilloscope with high impedance inputs which is that fast.

An example of this is impromptu TDR measurements.  Some low-z digital logic also presents a 25 ohm or lower impedance and can be measured this way. (1)

A high-z passive probe under these conditions lacks the performance of a low-z probe but can tolerate greater overload.

(1) In some cases a special low-z "offset" probe is needed to make these measurements without upsetting the bias conditions but a high-z passive probe can do it without problems.

[0xdeadbeef]

50Ohm feedthrough resistors should be good to 500MHz or so. Did some measurements of a 20MHz (square wave) LVDS signals with (cheap) differential probes on a 600MHz Lecroy lately where I used feedthrough resistors instead of switching the inputs to 50Ohm (mainly since I was scared to forget to switch back to 1MOhm and then someone else could damage the inputs by applying >12V or so). Worked pretty well indeed.

[TK]

Looks like Keysight is releasing 4 channel 1000X series scopes:

https://www.eevblog.com/forum/testgear/4-channel-keysight-1000x-series-scope/msg1957309/#msg1957309

[Hydrawerk]

This is not confirmed yet. https://www.keysight.com/en/pcx-2759552/infiniivision-1000-x-series-oscilloscopes?nid=-32110.0&cc=CZ&lc=eng

[nctnico]

The one thing that I don't understand the most is the lack of 50R input. Say, you have an 8Gsps scope with custom ADC, custom ASIC and custom AFE. All are very nicely done and if it's hackable, for $999.
It looks super attractive, but the 50R killed it.

Are you serious? No 50 Ohm input mode on a 350MHz scope? 

[simone.pignatti]

Looks like Keysight is releasing 4 channel 1000X series scopes:

https://www.eevblog.com/forum/testgear/4-channel-keysight-1000x-series-scope/msg1957309/#msg1957309

Do you mean the 4ch version of the existing 2ch? That would be interesting since they sold none of the 2ch...

[supercilious]

Are you serious? No 50 Ohm input mode on a 350MHz scope? 

Theres no 50ohm on the R&S RTB2004 scopes either! I think that external feed-through terminators are okay up to ~300MHz, so its not a big deal here IMO.
Obviously on a 500Mhz+ scope its a deal-breaker though.

[TK]

Looks like Keysight is releasing 4 channel 1000X series scopes:

https://www.eevblog.com/forum/testgear/4-channel-keysight-1000x-series-scope/msg1957309/#msg1957309

Do you mean the 4ch version of the existing 2ch? That would be interesting since they sold none of the 2ch...

You wish... I bought one

[TurboTom]

Maybe we've just got to wait a year for the "MSO5000A" to appear on the scene, this time with 50Ohms mode and possibly probe power/sensing. The missing high res mode on both DS7000 and MSO5000 makes me wonder if Rigol "forgot" (or deemed it unneccessary) to implement the boxcar averager in their ASIC? If that's missing and the ASIC lacks the configurability of an FPGA, it will stay missing for the product life...

[iMo]

MSO5074 $999 fully hacked will be the winner in 2019.. They will sell it [Edit: kilo-]tons per day..

If they have to decide between selling 100.000 scopes for $100 profit each or only 5.000 units with $3.000 profit each, guess what they will do.
And then, IF after a year sales is dropping, they can still leak some hacking information to sell another 50.000 units for $100 profit each.

Rigol does $300-$350 profit per unit, I bet. With 300k (my guess) 5072' $909 (fully hacked) units sold in 2019 Rigol will do aprox $210mil revenue and aprox $100mil profit. With $2.7b global o'scopes market in 2019 I would be happy with such a result

[2N3055]

Maybe we've just got to wait a year for the "MSO5000A" to appear on the scene, this time with 50Ohms mode and possibly probe power/sensing. The missing high res mode on both DS7000 and MSO5000 makes me wonder if Rigol "forgot" (or deemed it unneccessary) to implement the boxcar averager in their ASIC? If that's missing and the ASIC lacks the configurability of an FPGA, it will stay missing for the product life...

That is exactly why i'm cautious about it. It is unclear at this point what of the shortcomings are hardcoded and what can be "fixed" or "upgraded" later.
I know they can add better math and search later.
I don't think they will be adding 50 Ohms to 5000 series. Apart from a bit more memory, 500MHz,  50 Ohm and active probe interface, 5000 has all the capability of 7000. 
Upgrading 5000 like you said would practically make it 350 MHz 7000 with standard memory...

[TurboTom]

Rigol also put the DS2000A on the arena and it basically included the 50 Ohms option as the only improvement over the DS2000. In my opinion, the MSO5000 clearly plays somewhere in between the DS2000A and the DS4000 (though both are mature platforms), IMO it hasn't got the "professional appearance" of the 4000. It's just a more "up to date" package with the bells and whistles the competitors offer since a few years. Also, bandwidth-wise it rather competes with the DS2000(A). I'm sure the hardware is capable of more than what's available right now bandwidth-wise, so who knows what Rigol's plans are? Anyway, it's interesting to watch... For my own part, I won't be buying yet another scope in the near future...at least not without getting rid of one of the existing specimen 

[2N3055]

Anyway, it's interesting to watch... For my own part, I won't be buying yet another scope in the near future...at least not without getting rid of one of the existing specimen 

Same here  , I need to get rid of one to even think of getting something new.. I don't really need one though.

[David Hess]

Theres no 50ohm on the R&S RTB2004 scopes either! I think that external feed-through terminators are okay up to ~300MHz, so its not a big deal here IMO.

Obviously on a 500Mhz+ scope its a deal-breaker though.

Based on the poor transient response results of various 1 GHz oscilloscopes shown in TSP's latest video, using an external 50 ohm feedthrough termination on a 500MHz vertical input is fine.

[JPortici]

The one thing that I don't understand the most is the lack of 50R input. Say, you have an 8Gsps scope with custom ADC, custom ASIC and custom AFE. All are very nicely done and if it's hackable, for $999.
It looks super attractive, but the 50R killed it.

digital zoom on 2mV and 1mV/div killed it for me. At least put ERES until the hack is available i see no reason to get one. Screw the memory and samplerate i'd rather go to the competition, especially if one needs more bandwidth. Used Lecroys in that price range are really attractive, then there's the RTB2000 which is also higher resolution et cetera.

I wonder how limited the decoders are (by design or artificially, to separate between the 5000 and 7000)

Oh, dave, i remember they put a number for overload recovery. could you test it ? does it signal overloading?

[nctnico]

Going through the datasheet it seems that the samplerate drops to 2Gs/s with 4 channels active. Also ERES is not necessary because the MSO5000 has signal filtering which IMHO -when implemented properly- is better compared to ERES because the cut-off frequency can be adjusted and remains constant regardless the actual samplerate.

[JPortici]

so there are filters? my bad. My understanding was that there is no form of filtering at all.

[Fungus]

so there are filters? my bad. My understanding was that there is no form of filtering at all.

The manual says this:


No detail provided there but it seems like they provide it instead of "hi-res" mode.

[2N3055]

so there are filters? my bad. My understanding was that there is no form of filtering at all.

The manual says this:


No detail provided there but it seems like they provide it instead of "hi-res" mode.

No, that is exactly what it says, antialiasing, to prevent aliasing (subsampling) artefacts in display . DS1000Z has that function too.

[2N3055]

Very good explanation of Hires vs filtering from Lecroy

http://teledynelecroy.com/doc/differences-between-eres-and-hires

[0xdeadbeef]

Batronix offers the MSO5000 now. MSO5072 starts at ~ €963 (including VAT). Logic probe not included.
And obviously there is no "free budle option" for the MSO5000 at this time.

[skander36]

Recomended prices from Rigol - https://www.rigol.eu/products/digital-oscilloscopes/MSO5000/
Simone (Batterly) offer the best price for now .

[dardosordi]

Dave, the forum link on the youtube video is broken.

[Fungus]

Recomended prices from Rigol - https://www.rigol.eu/products/digital-oscilloscopes/MSO5000/

809 Euros for the baseline model. They're going to make many millions after the hack appears. 

(we just need somebody to dump the ROM and everybody who knows the chip to look and find the license check algorithm)