Benefits of going with all Siglent setup

[haloway13]

So much for me to unpack!!! So many responses! What a wonderful community!

So after reading up on MSO it does not look like I will go that route...

However, a combo Scope and AFG sounds like a good combination. What is the typical use case you have where the combo does not give you good enough amplitude?

What is you suggestion for a Combo Scope/AFG?

Thanks!
--Tim

[2N3055]

SDS2000x plus has AWG integrated, not a good one, but it is there and can do bode plot... no need for the DG811

IMHO the AWG is the biggest let-down of the SDS2000X+, which is otherwise a great scope for the price. The hardware seems pretty good with 14bit, 125MS/s and 50MHz BW, but the software implementation is just too basic and only really good for simple fixed (e.g. sine/square) waveforms. No modulation, no pulse train generator, no random noise generator, no noise-on-sine....

The AWG in the SDS2000X+ in its current form is not a replacement for a good signal generator and not up to the implementation in e.g. the R&S RTB2000. However, if Siglent makes a proper signal generator in a future firmware update, this scope is going to be really competitive!

The Bode analyzer functionality is good though (albeit a bit slow).

Even expensive Keysight 3000T built in AWG generator is very limited (only basic options, and very limited amplitude). It is most useful for FRA, and handy if you just need some small signal quickly, or you want to quickly capture something and than replay.
Most limiting is limited amplitude frankly...

So much for me to unpack!!! So many responses! What a wonderful community!

So after reading up on MSO it does not look like I will go that route...

However, a combo Scope and AFG sounds like a good combination. What is the typical use case you have where the combo does not give you good enough amplitude?

What is you suggestion for a Combo Scope/AFG?

Thanks!
--Tim

Limited amplitude is not limiting only when you use it for signals like audio amp repair and any low level signals.
If you use it only for that, than it's not limiting. But if you do only that, you don't need fully fledged signal generator, a DIY circuit with two opamp and  potentiometer is going to give clean sinewave, variable amplitude signal, for audio amp testing..

If you want to test mosfet switching, for instance, you won't be able to open it fully with 2.5V on the gate....... You pretty much need 10V at 50 Ohm to do that. So for that you need 20V P-P generator.

Idea of signal generator is that it is universal instrument to be used however your imagination would like it, limited by it's specs..

Other limiting factor is that having two synchronised channels is very important. There are many occasions when that is needed.

[tautech]

So much for me to unpack!!! So many responses! What a wonderful community!

So after reading up on MSO it does not look like I will go that route...

However, a combo Scope and AFG sounds like a good combination. What is the typical use case you have where the combo does not give you good enough amplitude?

What is you suggestion for a Combo Scope/AFG?

Thanks!
--Tim

The greatest versatility is always from a standalone 2ch AWG for feature set, amplitude and current source as many can supply 200mA that can also double as a DC source for small/simple projects.
If it can also be paired with a scope for Bode plot capability you get a useful tool set.

Sure 2 units hog more bench space however the easier usability is such it's better than an all in one scope/AWG.

[Howardlong]

Test conditions are two posts before in post #1743.

OK, thanks.

Without even going into the chosen setup it's obvious that:
a) Higher bandwidth scope = more noise.

You’ll see I had the 20MHz bandwidth limit apploed in all cases.

[Howardlong]

However, a combo Scope and AFG sounds like a good combination. What is the typical use case you have where the combo does not give you good enough amplitude?

What is you suggestion for a Combo Scope/AFG?
--Tim

Mostly for me it’s microcontroller ADCs with Vref over 2.5V, so, for example, a 2.5V AWG maximum often won’t test through the ADC’s full range.

The Rigols seems to be the best in terms of having 2 channels, and modulation and even AWG capabilities, but are limited to +/-2.5V. I’m not aware of any other scope AFG offering modulation capabilities, and very few offer dual channel.

Siglents top out at +/-3V and is single channel, and offers no AWG capability or modulation.

Keysights top out at +/-5V and are single channel on the 3000 series, dual channel on the 4000 and 6000, and do offer AWG capability but no modulation.

Tek MDOs are +/-5V and offer no AWG or modulati9n capability.

[nctnico]

However, a combo Scope and AFG sounds like a good combination. What is the typical use case you have where the combo does not give you good enough amplitude?

What is you suggestion for a Combo Scope/AFG?
--Tim

Mostly for me it’s microcontroller ADCs with Vref over 2.5V, so, for example, a 2.5V AWG maximum often won’t test through the ADC’s full range.

The Rigols seems to be the best in terms of having 2 channels, and modulation and even AWG capabilities, but are limited to +/-2.5V. I’m not aware of any other scope AFG offering modulation capabilities, and very few offer dual channel.

Siglents top out at +/-3V and is single channel, and offers no AWG capability or modulation.

Keysights top out at +/-5V and are single channel on the 3000 series, dual channel on the 4000 and 6000, and do offer AWG capability but no modulation.

Tek MDOs are +/-5V and offer no AWG or modulati9n capability.

For completeness: the R&S RTM3000 offers modulation, arbitrary (32kpts) and noise. All in all it is a pretty complete generator. I do use it every now and then. Unfortunately there is no way to trigger on the internal generator though.

[2N3055]

However, a combo Scope and AFG sounds like a good combination. What is the typical use case you have where the combo does not give you good enough amplitude?

What is you suggestion for a Combo Scope/AFG?
--Tim

Mostly for me it’s microcontroller ADCs with Vref over 2.5V, so, for example, a 2.5V AWG maximum often won’t test through the ADC’s full range.

The Rigols seems to be the best in terms of having 2 channels, and modulation and even AWG capabilities, but are limited to +/-2.5V. I’m not aware of any other scope AFG offering modulation capabilities, and very few offer dual channel.

Siglents top out at +/-3V and is single channel, and offers no AWG capability or modulation.

Keysights top out at +/-5V and are single channel on the 3000 series, dual channel on the 4000 and 6000, and do offer AWG capability but no modulation.

Tek MDOs are +/-5V and offer no AWG or modulati9n capability.

For completeness: the R&S RTM3000 offers modulation, arbitrary (32kpts) and noise. All in all it is a pretty complete generator. I do use it every now and then. Unfortunately there is no way to trigger on the internal generator though.

Keysight 3000T also have modulation (AM,FM and FSK) but only on sinewave. Output is only 2.5V in 50 Ohm (5 V P-P only High-Z). IT has no sweeps. You do have logic level presets, add noise to signal and very easy and nicely made copy from channel capture and fairly useful editing of AWG.
All in all, useful in a pinch, but very limited.

I use it mostly for FRA and to capture some signal (from sensor for instance) and replay it, with ability to set amplitude and add noise... I use that on Picoscope too.. For low frequency signals and finer resolution I use 16 bit Pico.

That's for quick test, easy to simply capture real sensor and replay. If I need something more elaborate/defined, than I create AWG waveform for "real" AWG and load it to either Siglent or Rigol I have..

[haloway13]

Somehow after reading every post's suggestion I have arrived at the same list of Siglent pieces.

Siglent Technologies SPD3303X-E Triple Output Power Supply
Siglent SDS1104X-E 100Mhz digital oscilloscope 4 channels standard decoder
Siglent Technologies SDG1032X Arbitrary Waveform - Function Generator

What are the same comparable items with Rigol? My head is swimming with too many model numbers and features!

Thanks!
Tim

[haloway13]

I think these are the alternatives in Rigol:

Rigol DG1032Z 30MHz Arbitrary Waveform Function Generator   
Rigol DP832 195W Power Supply   
Rigol DS1054Z 50MHz 4-Ch Digital Oscilloscope

[Fungus]

You’ll see I had the 20MHz bandwidth limit apploed in all cases.

Yes, I saw that, but the statement still applies.

The dangerous part is extrapolating your result to when 20MHz limiter isn't enabled*. You're assuming the Rigol noise will increase in exactly the same way as the Siglent noise.

(*) ie. the Way the 'scope is generally used.

[Howardlong]

You’ll see I had the 20MHz bandwidth limit apploed in all cases.

Yes, I saw that, but the statement still applies.

The dangerous part is extrapolating your result to when 20MHz limiter isn't enabled*. You're assuming the Rigol noise will increase in exactly the same way as the Siglent noise.

(*) ie. the Way the 'scope is generally used.

But if I hadn't used the BW limiter, everyone would have complained that I wasn't comparing like for like. It wasn't to try to manipulate the results, quite the opposite in fact.

Either way, it's a pretty bad result for the Rigol. The first thing I noticed when using thus scope was the noise, and that was without the BW limiter: as you say that's how it's typically used. I didn't put the BW limit on to make the Rigol look bad, I just wanted a level playing field.

It's just not the result either of us wanted.

[nctnico]

You’ll see I had the 20MHz bandwidth limit apploed in all cases.

Yes, I saw that, but the statement still applies.

The dangerous part is extrapolating your result to when 20MHz limiter isn't enabled*. You're assuming the Rigol noise will increase in exactly the same way as the Siglent noise.

(*) ie. the Way the 'scope is generally used.

Actually I usually have the bandwidth limiter enabled on my oscilloscopes at appropriate bandwidths. It gives the cleanest signal for a given measurement.

[Fungus]

But if I hadn't used the BW limiter, everyone would have complained that I wasn't comparing like for like.

That's the problem: You aren't.

A person could easily need the Siglent for one type of job and the Rigol for another type of job. Neither is "best".

I didn't put the BW limit on to make the Rigol look bad, I just wanted a level playing field.

I'm not accusing you of trying to make it look bad (or of anything else in particular). The problem is that the Internet will now parrot "Rigol only has six bits" for all time.

Maybe some of the Rigol noise appears a bit further down the circuit path, after the 20Mhz limiting capacitor.

What noise levels do you see with the limiter off? That the figure would be interesting, too. Also: How much effect does the 20Mhz limiter have on each 'scope? Some of them might have a better limiter (ie. more than capacitor).

[nctnico]

Maybe some of the Rigol noise appears a bit further down the circuit path, after the 20Mhz limiting capacitor.

What noise levels do you see with the limiter off? That the figure would be interesting, too. Also: How much effect does the 20Mhz limiter have on each 'scope? Some of them might have a better limiter (ie. more than capacitor).

Don't be so hung up on why the Rigol performs worse. If the noise is bad with the 20MHz limiter it will be worse without the 20MHz limiter. Usually the 20MHz bandwidth limiter is a 1st order filter so don't expect big differences between various oscilloscopes. You'll have to live with the Rigol MSO 5k series having 6 bits only. I agree that this is dissapointing; usually the ENOB drops at much higher frequencies (hundreds of MHz). Having an ENOB of only 6 bits with the 20MHz bandwidth limiter on is pretty bad. Especially for today's standard. It seems that Rigol's own chipset has been a bridge too far and it is not the leap ahead they promise it to be.

[rvalente]

It seems that Rigol's own chipset has been a bridge too far and it is not the leap ahead they promise it to be.

Well, that's life, you get what you pay for. The MSO5000 is for me remarkable scope but i'd still prefer to own the SDS2000x+ because it seems more polished to me and frankly, I see no reason to have a 8gsa/s converter for a 350MHz signal, thats almost 23x (also, whats the point of having such high frequency and no 50ohm input, nahhh).

[colorado.rob]

I think these are the alternatives in Rigol:

Rigol DG1032Z 30MHz Arbitrary Waveform Function Generator   
Rigol DP832 195W Power Supply   
Rigol DS1054Z 50MHz 4-Ch Digital Oscilloscope

For the power supply, if I were buying one today, I would probably go with a GW Instek GPP model with the LAN option.

This is the 4-channel one.  You can find the 3-channel model a bit cheaper.
https://www.tequipment.net/Instek/GPP-4323/DC-Power-Supplies-/-Lab-Power-Supplies/

There's a good discussion about it here: https://www.eevblog.com/forum/testgear/seeking-psu-advice-rohde-schwarz-nge103b-com3-or-keysight-e36313a/msg3230970/#msg3230970

[ahope]

I think these are the alternatives in Rigol:

Rigol DG1032Z 30MHz Arbitrary Waveform Function Generator   
Rigol DP832 195W Power Supply   
Rigol DS1054Z 50MHz 4-Ch Digital Oscilloscope

For the power supply, if I were buying one today, I would probably go with a GW Instek GPP model with the LAN option.

This is the 4-channel one.  You can find the 3-channel model a bit cheaper.
https://www.tequipment.net/Instek/GPP-4323/DC-Power-Supplies-/-Lab-Power-Supplies/

There's a good discussion about it here: https://www.eevblog.com/forum/testgear/seeking-psu-advice-rohde-schwarz-nge103b-com3-or-keysight-e36313a/msg3230970/#msg3230970

Agree completely.
I am a big Siglent fanboy, but for the power supply I went with the GW Instek GPP-4323.
Four channels supply, two channels load, fantastic display (even better than 3055).
Remote software sucks though, Siglent scores higher in that regard.

[Teti]

Siglent fanboy here! 
I went for SDS1104XE scope, and SDG2082X sig gen as they outperform Rigol in the same class, allround much better product.
However I will go for Rigol DP832 bench PS as Siglent is not really 3 channel, more like 2 real channels with slapped on wall wart inside as extra. Will go for Siglent 5 1/2 DMM as much better display.
Benefit of having all siglent is that you can transfer stuff between devices etc. I don't use it at the moment, but I'm pretty sure this will be handy in the future!

[Howardlong]

It seems that Rigol's own chipset has been a bridge too far and it is not the leap ahead they promise it to be.

Well, that's life, you get what you pay for. The MSO5000 is for me remarkable scope but i'd still prefer to own the SDS2000x+ because it seems more polished to me and frankly, I see no reason to have a 8gsa/s converter for a 350MHz signal, thats almost 23x (also, whats the point of having such high frequency and no 50ohm input, nahhh).

That's pretty much my take, and I have both. As a bit of a Rigol fanboi myself, and having had a mediocre experience with the only bit of Siglent gear I had in the lab, I wasn't expecting much from the SDS2000X+, but it far out performed my expectations in terms of UI, functionality, performance and polish. On the other hand, while I had high hopes for it, I found the Rigol somewhat disappointing in many respects, the most egregious being the aforementioned noise and the UI sluggishness particularly when changing timebase.

Almost all of the Rigol deficiencies could be resolved with software updates, although I suspect the noise thing is endemic. I was glad to see that the dim screen reported by some on the Rigol was not present on mine. Bearing in mind the Rigol has been out for far longer than the Siglent, I'd have expected the Rigol to be more polished, but the exact opposite was true IME.

The three main negative takeaways for me on the Siglent were (a) the lack of decoder search functionality, (b) many functions are only available on the touch screen and (c) the memory management. None of these are a deal breaker IMHO. The Rigol will do decoder search, but only on the analogue channels for some bizarre reason. I don't have it on the bench right now, but almost all of the Rigol functionality can be accessed by the physical buttons & knobs, and the memory management seemed pretty reasonable to me.

[eeviking]

Somehow after reading every post's suggestion I have arrived at the same list of Siglent pieces.

Siglent Technologies SPD3303X-E Triple Output Power Supply
Siglent SDS1104X-E 100Mhz digital oscilloscope 4 channels standard decoder
Siglent Technologies SDG1032X Arbitrary Waveform - Function Generator

What are the same comparable items with Rigol? My head is swimming with too many model numbers and features!

Thanks!
Tim

Just to add some more water to the pool 

I would save some on the PSU and go for a SPD3303C
It's the same psu as the SPS3303X-E just without the fancy lcd and ethernet.
Over here it's 150€ cheaper and in my opinion better bang for the buck.

Then spend some more on the scope and upgrade to a SDS2102X Plus or SDS2104X Plus and use the built in generator for now.
I believe there is a DIY solution to cheaper MSO pods that maybe could "make up" for the missing channels on the SDS2102X Plus

If you really need a dedicated AWG I would go all in for a SDG2042X

Disclaimer:
I have SPD3303C, SDS1104X, SDG805 and kind of regret cheaping out on the scope and AWG but not the PSU.

[TurboTom]

...

Test conditions are two posts before in post #1743.



Effective bits:



My comparison of MSO5000 to SDS2000X+

I'm sorry but I dare to question @Howardlong's findings on the MSO5000 rms noise and thus the calculated equivalent resolution.

Just compare the peak-to-peak with rms noise figures at 500mV/div setting:

Vpp - 44.5mV
Vrms - 44.2mV

These two numbers are more or less identical which is mathematically impossible in case of an observation of AC signals. Vrms cannot be higher than Vp = 1/2 Vpp, which is the case if the signal has true square shape. And noise doesn't look that much like a square wave... On average, the crest factor for white noise can be assumed to be in the ballpark of 4, thus Vrms should be closse to 1/8 Vpp which matches the figures of the "big" Keysight and the TEK surprisingly well.

To explain the measured figures, I state that all those numbers that deviated considerably from the 1/8 relation had some amount of DC offset superimposed. This is easily overlooked since the zero marker is such a tiny pip at the left border of the screen... The better method of evaluating the AC component of Vrms using internal measurements is just taking the standard deviation of Vpp. This eliminates the average (DC offset) by summing up geometrically the difference of the individual measurements minus the average, e voila -- AC RMS  .

The resulting equivalent resolution figures for some of the scopes will turn out much better if this calculation is done properly, but this wouldn't change the impression the reader now already memorized that the MSO5000 has got a front-end that's noisy like the Niagara falls... The apparent noise definitely differs, and Rigol for sure is among the noisier "party", but it's definitely not as bad as Howard evaluated in his table. Much of the apparent noise is also contributed to the way the traces are visualized, and Rigol's display engine traditionally emphasizes the extremal values and thus makes their scopes appear more noisy.

I'm not a Rigol or Siglent fanboy - I own gear of both companies and also had some "buyer's remorse" with both, but I think when publishing a comparison like this, we should at least make sure that the figures make sense from a mathematical point of view...

[nctnico]

Well spotted! It is however possible the Vrms calculation in the MSO5000 is wrong 

[2N3055]

So if I'm not wrong my calc for MSOX3104T would go like this, for those conditions:



and



so that gives:




Although ENOB is not single number, it is frequency/ENOB graph....

[Howardlong]

...

Test conditions are two posts before in post #1743.



Effective bits:



My comparison of MSO5000 to SDS2000X+

I'm sorry but I dare to question @Howardlong's findings on the MSO5000 rms noise and thus the calculated equivalent resolution.

Just compare the peak-to-peak with rms noise figures at 500mV/div setting:

Vpp - 44.5mV
Vrms - 44.2mV

These two numbers are more or less identical which is mathematically impossible in case of an observation of AC signals. Vrms cannot be higher than Vp = 1/2 Vpp, which is the case if the signal has true square shape. And noise doesn't look that much like a square wave... On average, the crest factor for white noise can be assumed to be in the ballpark of 4, thus Vrms should be closse to 1/8 Vpp which matches the figures of the "big" Keysight and the TEK surprisingly well.

To explain the measured figures, I state that all those numbers that deviated considerably from the 1/8 relation had some amount of DC offset superimposed. This is easily overlooked since the zero marker is such a tiny pip at the left border of the screen... The better method of evaluating the AC component of Vrms using internal measurements is just taking the standard deviation of Vpp. This eliminates the average (DC offset) by summing up geometrically the difference of the individual measurements minus the average, e voila -- AC RMS  .

The resulting equivalent resolution figures for some of the scopes will turn out much better if this calculation is done properly, but this wouldn't change the impression the reader now already memorized that the MSO5000 has got a front-end that's noisy like the Niagara falls... The apparent noise definitely differs, and Rigol for sure is among the noisier "party", but it's definitely not as bad as Howard evaluated in his table. Much of the apparent noise is also contributed to the way the traces are visualized, and Rigol's display engine traditionally emphasizes the extremal values and thus makes their scopes appear more noisy.

I'm not a Rigol or Siglent fanboy - I own gear of both companies and also had some "buyer's remorse" with both, but I think when publishing a comparison like this, we should at least make sure that the figures make sense from a mathematical point of view...

You're right to question them, as did I at the time.

The figures are from the scope's own measurements.

[TurboTom]

Howard -
on Rigol gear (I cannot tell for other brands), the RMS measurement always outputs AC+DC. And since the scope cannot distinguish between DC offset and DC signal component, and Rigol's frontends are notorious for considerable offset drift during warm-up, this effect can and will affect RMS readings considerably. Especially depending on where in the signal chain the offset drift happens, this may not be limited to the high sensitivity vertical ranges.

In order to retrieve a proper AC RMS reading to compare frontend noise, it's advisable to enable statistics on the Vpp measurement; the standard deviation will then contain the relevant figure.

Do you still have access to all the scopes that you listed in your table? It would be great if we could repeat the test with the DC offset component eliminated. If not, maybe other owners could possibly jump in. In this case, the input configuraion would have to be specified though (50R terminator, shorted, open but capped or just open).

Cheers,
Thomas