Best 350MHz scope in a hackable world (Siglent SDS2104X Plus or Rigol MSO5072)

[Martin72]

So, if I may ask someone who's got access to both these scopes if he may try another configuration for a noise measurement:
TB 1µs, sensitivity set to the last "genuine" range, i.e. without digital magnification (I guess that's 2mV/div on the MSO), input impedance 1Mohm, 20MHz bandwidth limit enabled, BNC if possible capped off (even some aluminium foil wrapped aroud may do).
I'ld like to compare the following measurements: Sampling mode (1) normal and (2) peak (to understand if there's some additional "smoothing" going on in normal sampling mode). Measure functions: Vpp and STD (AC RMS) averaged. Maybe a few sensitivity comparisons would also be interesting, i.e. how the noise measurement changes when the input range is changed, maybe between 5mV/div and the lowest range available.

Before leaving home for visiting relatives, I´ve done a few measures.
TB 1µs, lowest vertical 1mV, 2mV, 5mV, 1mV Normal and 1mV Peak, further 1,2,5mV in normal mode but 10bit.
Input impedance 1M, no cap for bnc.

[Zlotnik]

10 bit and input noise defined it above all for me.

Even though the SDS2000X Plus appears to provide a lower noise floor than some of its competitors, it's an eight bit oscilloscope by hardware. Siglent uses two MXT2001 dual 1GSPS 8bit ADCs in its frontend that can be cascaded to provide 2GSPS when only two channels of the scope are in use.

Every marketing talk of 10bit or even more refers to things that happen in software (and should be specified as such). This doesn't mean that this option isn't useful. But for comparison purposes, one should always keep that in mind. There are only a few scopes in the entry or mid level range that actually contain ADCs that sample more accurately than eight bits -- a few Owons (utilizing the HMCAD1520 that can be internally reconfigured to provide 8/12/14 bit sampling depth, yet at reduced sampling rate - good choice for audio / sensor jobs) or the RTB2000 (Rohde&Schwarz) with its proprietary 10 bit full-speed ADC (up to 2.5GSPS).

BTW, over on the SDS2000X Plus thread, user maxspb69 posted a very interesting 1:1 comparison of the SDS2000X Plus and the RTB2000 input noise & ADC resolution. The Siglent’s front end really seems to be quite good, comparable to the much more expensive R&S.

If anything, I think the front end is almost too good, as you probably can’t make full use of it without a wider ADC as on the RTB.

[mawyatt]

I couldn't stop to reason about the input noise and finally, I simply had to add to my previous statements about this issue my latest thoughts:

If the (standard high-impedance) input of a scope is left open, it's resemling a resistance of 1Mohm in parallel with a capacitance of a few picofarads, say 15pF. This RC-combination is basically a low pass filter of first order with a 3db frequency of 67kHz (in this case), and not the 20MHz bandwidth that I used to estimate the thermal noise. At a frequency band of 67kHz, the thermal noise would be round about 33µV rms (surprise, surprise...), so if it's arranged that way, Siglent's noise figure may be quite accurate, provided they don't change that much with an open input.

I guess, in the next days someone will help out with a few tests - thanks so much for this  . I haven't got any of these instruments and as yet, don't see the need to change this (though TEA tries to convince me...).

One must also consider the noise model for this front end and the individual sources of noise. Changing the input source impedance and bandwidth effects input noise properties from the source R and any shot noise * R effects (although the shot noise is likely very small due to a FET input), but does not effect the bandwidth of the noise contributions behind the input since they don't "see" this bandwidth limit. The only way these noise sources are controlled is with post filtering just before the ADC like the DSO bandwidth control does.

So what I am trying to say it that the input bandwidth limit of the RC product only effects the noise contributions from the source R and not the subsequent amplifier stages.

Best and Happy Holidays,

[mawyatt]

If anything, I think the front end is almost too good, as you probably can’t make full use of it without a wider ADC as on the RTB.

Yes, as I've said before Siglent just got this DSO frond end design right, it's that good!! And begs the question for a new DSO version with a 10 or 12 bit core ADC

Best and Happy Holidays,

[Martin72]

And begs the question for a new DSO version with a 10 or 12 bit core ADC

You mean the thing with the pin-compatible up to 14bit ADC like GWInstek got ?
I´m not very deep into the technology ( but learn every day spending some time here), but the GW scope´s samplerate decrease dramatically low, when choosing the modes 10,12,14 bit.
Apart from this it will be of course a very nice thing, when siglent launch a 10 bit or more scope to a payable price.
And no, the SDS6000 line are not meant, as they are ways above the prices, hobbyists can pay.

For me, my limit is reached by paying 3000 bucks for private using a scope.

[mawyatt]

And begs the question for a new DSO version with a 10 or 12 bit core ADC

You mean the thing with the pin-compatible up to 14bit ADC like GWInstek got ?
I´m not very deep into the technology ( but learn every day spending some time here), but the GW scope´s samplerate decrease dramatically low, when choosing the modes 10,12,14 bit.
Apart from this it will be of course a very nice thing, when siglent launch a 10 bit or more scope to a payable price.
And no, the SDS6000 line are not meant, as they are ways above the prices, hobbyists can pay.

For me, my limit is reached by paying 3000 bucks for private using a scope.

At first, thought the SDS2000X Plus used the same ADC family as GWInstek but someone pointed out that's not the case. Since the 8, 12 and 14 bit versions of those ADCs are pin compatible that would have been nice!! Siglent will have to redesign the PCB as well as the firmware for a higher resolution ADC, so a more involved design task, and maybe not justifiable.

As mentioned the superb front end is already there in case Siglent is listening

Best & Happy Holidays,

[Martin72]

Hi Mike,

Do we really need 10 or more Bits ?
And then on a pricelevel, we can afford ?
Maybe a bad example, but last year we (company I´m working) bought a lecroy waverunner scope for 13500 bucks ( 40% off... ).
And it´s great, can do much more we ever need - And got 8 bit resolution..
Actually we got the siglent 2K+ with it´s 10 bit mode plus additonally up to 3 bits Eres function.
Should we really need more, with suitable price for us hobbyists ?
I think, when siglent will build scopes with a native resolution of 10 bits or more, they won´t be cheaper as the R&S Scopes, which already got it.
Native 10bit or more scope for less then 3000, 2000 bucks are a wet dream.
Combined with the excellent noise behaviour of the sds2k+ and the accurate timebase, even more wet.
Today...

[mawyatt]

Hi Martin,

Agree, 10 native bits with a good ENOB is likely all that's needed, but the enticing aspect of the ADC GWINstek is using in the 12, 14 bit "enhancement" in hardware. I can't say if this is better than doing the "enhancement" in software like Siglent and others are doing, but sure is enticing 

Think of what the core Siglent FFT would be like with hardware baseline 12 or 14 bit data, it's already pretty good as it stands now

Edit: Recall almost 15 years ago Fugi had developed a 8 bit ~56GPSP ADC in 65nm CMOS, a year later it was ~65GSPS, and a couple years later much much faster. This 56GSPS ADC had a core of ~320 capacitive ratioed SAR ADCs operating @ ~180MSPS, and each ADC had ~400 correction DACs, so 320 SAR Sub-ADCS and 400 DACs per ADC, and two ADCs per chip! These were for the wide bandwidth fiber-optic receivers, and had ENOBs ~6 bits, and error correction took place on-chip with the massive array of DACs correcting errors at the source rather than post processing.

If you follow the camera sensors then the sensor has each pixel column or row with a SAR ADC, usually 14~16 bits. A 36MP camera sensor has 6000~8000 individual ADCs. IC design guidelines are totally orthogonal to PCB design guidelines, and a totally different mindset all-together!

I've been out of the loop for the past couple years since retiring, but a little imagination will give an idea of what's going on today with 10, 7, & 5nm CMOS technology. BTW watch for a new type ADC called Non-Uniform Sampling emerging from the research labs of 10 years ago, the input analog waveform produces it's own anti-aliasing filter in post processing

Best & Happy Holidays,

[kahuna0k]

I posted this to the wrong thread, linking here for reference.

[IAmBack]

So, if I may ask someone who's got access to both these scopes if he may try another configuration for a noise measurement:
TB 1µs, sensitivity set to the last "genuine" range, i.e. without digital magnification (I guess that's 2mV/div on the MSO), input impedance 1Mohm, 20MHz bandwidth limit enabled, BNC if possible capped off (even some aluminium foil wrapped aroud may do).
I'ld like to compare the following measurements: Sampling mode (1) normal and (2) peak (to understand if there's some additional "smoothing" going on in normal sampling mode). Measure functions: Vpp and STD (AC RMS) averaged. Maybe a few sensitivity comparisons would also be interesting, i.e. how the noise measurement changes when the input range is changed, maybe between 5mV/div and the lowest range available.

Before leaving home for visiting relatives, I´ve done a few measures.
TB 1µs, lowest vertical 1mV, 2mV, 5mV, 1mV Normal and 1mV Peak, further 1,2,5mV in normal mode but 10bit.
Input impedance 1M, no cap for bnc.

I did those measurements too, but my results are different. Scope bought at the beginning of december.

For 1mV/div, 20M BW limit, TB 1µs, one channel active, 8bit normal mode: PK-PK is 200µV, Stdev 30µV, RMS depends on actual calibration (with "fresh" calibration it is about 40µV, but can be also as high as 300µV). Room temperature is about 25^oC (it may be important). Measurements done after more than 30 min. of warming-up. Turning input do "gnd" does not change results a lot.

Maybe it is worth to check it on other user's units?

[Martin72]

Hi,

Please make screenshots from...
Apart from this, I think this deviation is not a big worrying thing.
And you can try to cap the bnc with aluminiumfoil, in case your enviroment is "noisier" than mine here.

[KK1L]

A little off-topic, but nonetheless a very interesting read: @jjoonathan's write-up of his DS4000 fix, replacing two of Rigol's (supposedly) MXT2001 with ADC08D1000s.

I am impressed! Some people are really smart troubleshooters.

Smart and handy! Gifted talent. Respect!

Echo the bravo.  Regarding the PFA to look for a wirebond open it is often not the wire, but the bond. The wires are encapuslated and wont be moving even if they do break outright. Below is an example.


73 es God Bless de KK1L, Ron <><

[hpw]

I am also in the evaluation for a shoe box... so does Siglent plans to release soon new gear or updated gear for the SDS2xxx models?

Also not happy to loose a input just for triggering & also a pointer for the 16 port logic graph's, while would like to analyze propagation timing & delays.

hp

[tautech]

I am also in the evaluation for a shoe box... so does Siglent plans to release soon new gear or updated gear for the SDS2xxx models?

Not for a while as SDS2000X Plus models are very capable and popular.

Also not happy to loose a input just for triggering & ....

SDS2kX Plus has an Ext trigger BNC input on the rear which is limited to 1.5 VRMS or 7.5 VRMS depending on the internal divider selected. Most other Siglent DSO's have the same Ext trigger input.
Maybe that's sufficient sensitivity for your needs ?

also a pointer for the 16 port logic graph's, while would like to analyze propagation timing & delays.

What do you mean by this ? 

I have the Siglent STB3 test board and it can generate a wide range of signals so maybe there's some LA vs analog signal measurement you would like to see.

[hpw]

I am also in the evaluation for a shoe box... so does Siglent plans to release soon new gear or updated gear for the SDS2xxx models?

Not for a while as SDS2000X Plus models are very capable and popular.

Also not happy to loose a input just for triggering & ....

SDS2kX Plus has an Ext trigger BNC input on the rear which is limited to 1.5 VRMS or 7.5 VRMS depending on the internal divider selected. Most other Siglent DSO's have the same Ext trigger input.
Maybe that's sufficient sensitivity for your needs ?

also a pointer for the 16 port logic graph's, while would like to analyze propagation timing & delays.

What do you mean by this ? 

I have the Siglent STB3 test board and it can generate a wide range of signals so maybe there's some LA vs analog signal measurement you would like to see.

OK, if you have a ADC/DAC with various clocks & timing & various additional signals who relate to the master clock, I would like to measure & see how the edges are within the spec's. So within the graph, I would like to measure the delay/time or when the edge happen.

Hope it is now clear enough... a picture would be anyway better

[Martin72]

so does Siglent plans to release soon new gear or updated gear for the SDS2xxx models?

Why should they ?
SDS2K+ is "new", got several advantages in it´s priceclass against others, when you want some things "better", then grab the next higher range from them.
New gear in general is another thing but I guess, it will happen, when, in another pricerange.

[tautech]

also a pointer for the 16 port logic graph's, while would like to analyze propagation timing & delays.

What do you mean by this ? 

I have the Siglent STB3 test board and it can generate a wide range of signals so maybe there's some LA vs analog signal measurement you would like to see.

OK, if you have a ADC/DAC with various clocks & timing & various additional signals who relate to the master clock, I would like to measure & see how the edges are within the spec's. So within the graph, I would like to measure the delay/time or when the edge happen.

Hope it is now clear enough... a picture would be anyway better

Some digital channels ( 8 ) with 3 active signals with Ch4 connected to D0 showing skew between digital and analog signal paths. Triggering in ch4 with a zone trigger (off display) and digital decode engaged just for fun. X cursors on D0 and trigger point of Ch4 displaying 3.9ns propagation delay. Zoom mode used.

Now.....This is a SDS5054X but I would entirely expect SDS2000X Plus to produce exactly the same result as their firmwares are closely aligned.
Hopefully I collect more SDS2000X Plus tomorrow.

[hpw]

so does Siglent plans to release soon new gear or updated gear for the SDS2xxx models?

Why should they ?
SDS2K+ is "new", got several advantages in it´s priceclass against others, when you want some things "better", then grab the next higher range from them.
New gear in general is another thing but I guess, it will happen, when, in another pricerange.

Just asking, if some know if any new DSO are in the pipeline (like higher than 2 Gs or bits), I would wait.

As I see, the higher models 5xxx do not have the 10 bit feature, or I am wrong?

Are there any trigger Jitter spec's given for those models?

Hp

[Fungus]

Just asking, if some know if any new DSO are in the pipeline (like higher than 2 Gs or bits), I would wait.

Do you know that manufacturers deliberately don't tell people these things - to avoid them "waiting".

[tautech]

As I see, the higher models 5xxx do not have the 10 bit feature, or I am wrong?

Correct. SDS5000X only have 3 bits of ERES enhancement whereas SDS2000X Plus have the 10 bit mode and up to 3 bits of ERES and both can be applied together.

Are there any trigger Jitter spec's given for those models?

Yes of course, consult the datasheets for trigger jitter specs. Under 10ps for both models.

[hpw]

Currently reading the release notes on  V1.3.5R5 as 5. Fixed several bugs:

     b) 2-channel 100MHz model cannot be upgraded to 350MHz 

[Martin72]

Hm ?

[mawyatt]

Currently reading the release notes on  V1.3.5R5 as 5. Fixed several bugs:

     b) 2-channel 100MHz model cannot be upgraded to 350MHz 

Please don't tell that to my SDA2102X Plus, it has been "enhanced" with the help of folks here and now sports a measured input bandwidth of over 600MHz

Best,

[checksum]

Hi everyone - SDS2104X Plus is currently in stock at Labtronix in the UK.  www.labtronix.co.uk

Just bought one from your site today :-). Looking forwards to a delivery this week I hope.

[lastguy]

Thanks for all posts to refresh me, I don't own scope, but used scopes from 200MBW tek(DOS like GUI, poor waveform, Eth card costs me $668!) upto 50GHz Keysight. When I measure 16GHz differential signal I use 3.5mm CH1-CH3. I also tried 25GHz diff probe after calibration, yet it is a bit noisy - may be phase noise - to cause my measurement fail. Use 3.5mm/SMA I no need calibration.
BNC BW is 3GHz and well above 350-500M. Scope input has no divider but attenuator (except for OP gain change), for mv/Div ranges attenuators are bypassed or signal is directly passed.
Passive probe has its own capacitor compensation, traditional input cap is 13-15pf. That doesn't mean scope input cap. All low end scope (<=1G) uses simple relay to turn on 50ohm impedance if any, other than this, impedance matching (I mean signal path impedance) would be ignored, because path is short and is not 50ohm. The 50ohm is ended at the input to eliminate refection.
"Hires" as mfg stated is oversampling then "middle value filtering" (which is common is signal processing). "Average" is based on multiple trigging or samples average. Both can increase A/D EOB, which is also commonly used in SAR ADC.
I personally think the extra noise is more likely from ADC chip, yet shield input stage might help.
No doubt of 8GSa by interlaced sampling, or each ADC has 0.5ps delayed start pulse, I know FPGA can output calibrated delay line not sure whether can be so fine, I'm sure our DDR IP can, may be PCB/cable can, I use differential 3' cable pair with +-1ps tolerance. Many years ago HP 54601 scope uses "EQ" sampling, each capture can be shifted by 0.1ps, and real time sampling is only 10 or 100MHz (forgot)
5X is de facto std for sine wave, for example Keysight 3104T 1GBW has 5GSa, 16G PCIe signal needs 80GSa. Theoretically you can use Sinx/x and 2X+ sampling rate yet only when your signal is nice sine wave. 5X means each cycle you have only 5 samples, usually not enough for non-sine wave. So, 2GSa means you can use it for 400M signal, may be 500M sinewave.