buy a better oscilloscope than the Siglent SDS1104X-E

[uargo]

Hello, good afternoon. I wanted to ask you a question/query.
If you already had an oscilloscope (SIGLENT SDS1104X-E) and you were thinking of buying a better oscilloscope, one step further, which oscilloscope would you choose.

Stuff I already have:
* Riden RD6006P power supply
* Chinese linear power supply 30V 5A
* Siglent SDM3045X Multimeter
* UNI-T UT71E Multimeter
* Siglent SDG1032X signal generator
* Siglent SDS1104X-E Oscilloscope

Oscilloscope options I'm considering:
  * Siglent SDS2104X PLUS (8 bit, touch screen, 2GSa/s, 4 channels, hackable)
  * Rigol MSO5074 (8 bit, touch screen, 8GSa/s, 4 channels, hackable)
  * Rigol HDO1104 (12 bits, touch screen, 2GSa/s, 4 channels, hackable)

Suggestions of other oscilloscopes are accepted but for no more than €1,500, preferably touch screen, 4 channels, minimum 2gsa/s, if it has 10 or 12 bits better, and that can be hacked at 200Mhz or more

Thanks in advance

[Martin72]

Hi uargo,

* Siglent SDS2104X PLUS (8 bit, touch screen, 2GSa/s, 4 channels, hackable)
  * Rigol MSO5074 (8 bit, touch screen, 8GSa/s, 4 channels, hackable)
  * Rigol HDO1104 (12 bits, touch screen, 2GSa/s, 4 channels, hackable)

Exactly in this order...

In total, the siglent would be the "best" of the three, by the way it has a 10bit mode (with limited bandwidth 100Mhz).
The good old MSO5074 certainly has its charm and is over-featured, but has a slightly smaller screen, is still not bug-free and is not exactly low-noise.
The HDO has (in my opinion) currently only one advantage and that is the 12Bit resolution.
Otherwise, it is not yet mature, some features are missing.
I myself owned the siglent and the rigol MSO and had a HDO4024 on loan at the beginning of this year.

[tautech]

Oscilloscope options I'm considering:
  * Siglent SDS2104X PLUS (8 bit, touch screen, 2GSa/s, 4 channels, hackable)
  * Rigol MSO5074 (8 bit, touch screen, 8GSa/s, 4 channels, hackable)
  * Rigol HDO1104 (12 bits, touch screen, 2GSa/s, 4 channels, hackable)

Suggestions of other oscilloscopes are accepted but for no more than €1,500, preferably touch screen, 4 channels, minimum 2gsa/s, if it has 10 or 12 bits better, and that can be hacked at 200Mhz or more

Thanks in advance

Considering you already have SDG1032X and if you use the SDS1104X-E Bode plot feature another Siglent DSO might be a sound decision for their plug and play capability.

SDS2104X Plus is on worldwide promotion with a free BW upgrade to the 200 MHz model, SDS2204X Plus.
I have tested both those models to ~185 and 300 MHz BW.

If not in a hurry, the soon to be released 12 bit SDS1000X HD also might be of interest however it only offers 1 GSa/s.
https://www.eevblog.com/forum/testgear/siglent-sds1000x-hd-coming/

[uargo]

The MSO5074 has the problem of very exaggerated background noise.
The SDS2104X plus has 10bit mode, but does it really work well?

[uargo]

Oscilloscope options I'm considering:
  * Siglent SDS2104X PLUS (8 bit, touch screen, 2GSa/s, 4 channels, hackable)
  * Rigol MSO5074 (8 bit, touch screen, 8GSa/s, 4 channels, hackable)
  * Rigol HDO1104 (12 bits, touch screen, 2GSa/s, 4 channels, hackable)

Suggestions of other oscilloscopes are accepted but for no more than €1,500, preferably touch screen, 4 channels, minimum 2gsa/s, if it has 10 or 12 bits better, and that can be hacked at 200Mhz or more

Thanks in advance

Considering you already have SDG1032X and if you use the SDS1104X-E Bode plot feature another Siglent DSO might be a sound decision for their plug and play capability.

SDS2104X Plus is on worldwide promotion with a free BW upgrade to the 200 MHz model, SDS2204X Plus.
I have tested both those models to ~185 and 300 MHz BW.

If not in a hurry, the soon to be released 12 bit SDS1000X HD also might be of interest however it only offers 1 GSa/s.
https://www.eevblog.com/forum/testgear/siglent-sds1000x-hd-coming/

I didn't know anything about the SDS1000X HD, do you know the features, price and date?

[tautech]

I didn't know anything about the SDS1000X HD, do you know the features, price and date?

I have one. 

It's very much like SDS2000X HD with less features and only 200 MHz max.
No internal AWG and different LA/MSO which uses same module as SDS1104X-E.

Price is a guess but we expect under $1k. No release date is yet available but we expect before years end.

[Martin72]

The SDS2104X plus has 10bit mode, but does it really work well?

AFAIK yes, there are some demonstrations in the huge SDS2000X+ thread, but don´t ask me where exactly..
Another question is whether you really need 10, 12 or more bit resolution.

[coppice]

Hello, good afternoon. I wanted to ask you a question/query.
If you already had an oscilloscope (SIGLENT SDS1104X-E) and you were thinking of buying a better oscilloscope, one step further, which oscilloscope would you choose.

Define better. Some want more resolution. Some want more bandwidth. If you haven't specified I suspect you just want more expensive.

[tautech]

The SDS2104X plus has 10bit mode, but does it really work well?

AFAIK yes, there are some demonstrations in the huge SDS2000X+ thread, but don´t ask me where exactly..
Another question is whether you really need 10, 12 or more bit resolution.

In the first few pages of the SDS2000X Plus thread.
https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/msg2786952/#msg2786952

In screenshots there in the timebase panel we can see the 10 bit mode indicator displayed.

[Martin72]

Btw @uargo,

I´ve daily access to both scopes, SDS2104X+ and SDS1104X-E, just saying..

[uargo]

ok thanks everyone, I think the question then would be:
SDS2104X PLUS or SDS1104X HD?

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

[Martin72]

So the MSO5074 from rigol will be definitely out of the game...

[coppice]

So the MSO5074 from rigol will be definitely out of the game...

The MSO5074 is currently 810 pounds + VAT in the UK. The SDS2104X+ is 1050 pounds + VAT. If you can afford the higher price you get a better instrument. The market seems to be functioning properly.

[uargo]

but why do I want to buy the Rigol MSO5074 if it doesn't work for me later. The question is whether the Siglent SDS2104X PLUS will suffice, or better the Siglent SDS1000X HD or the Rigol HDO1000X HD.
Can someone who has these oscilloscopes advise me?

[2N3055]

But what does "better" mean in  this context?
You need to provide more detail of work to be performed to get more targeted advice.
Maybe you don't need a "better" scope at all?
What is a type of work you could not do that prompted you to want a "better" scope. Start with that.


Siglent SDS2104X Plus is a 500 MHz scope platform. Low noise, good for both high BW and small signals. MSO. Very good BODE plot.

Rigol MSO5000 is 350Mhz design without 50Ω inputs. It has 2 Ch AWG, 4 decodes. Higher noise and lower sensitivity. It is good for general purpose work, not ideal for fine analog work. For Arduino kind of work decent choice.

Both Siglent and Rigol 1000HD platforms are low BW, up to 200 MHz, so same as SDS1104X-E. They would be very good for fine analog work, but with BW limit.

So as @coppice already said, Siglent SDS2104X Plus is best compromise of all listed, with limited other info...
Even very expensive scopes will never be perfect for "everything". They all specialize something: BW, noise, math, screen size, decoding...That is why people end up with many scopes...

[Fungus]

but why do I want to buy the Rigol MSO5074 if it doesn't work for me later. The question is whether the Siglent SDS2104X PLUS will suffice, or better the Siglent SDS1000X HD or the Rigol HDO1000X HD.
Can someone who has these oscilloscopes advise me?

No, because if you don't know what you need then nobody does.

The only answer to that question is to buy the most expensive one and cross your fingers.

[markus_jlrb]

Dears,

beside all the pro and contra of the mentioned DSO's
I have obtained a SDS5034X this days due to the higher
sampling rate (5Gs) and the possibility to upgrade it to
a 750MHz BW. Price was a bargain due to a demo device
but with very less working hours.

I hope that later I will effort a HD version and could compare.

Markus

[uargo]

But what does "better" mean in  this context?
You need to provide more detail of work to be performed to get more targeted advice.
Maybe you don't need a "better" scope at all?
What is a type of work you could not do that prompted you to want a "better" scope. Start with that.


Siglent SDS2104X Plus is a 500 MHz scope platform. Low noise, good for both high BW and small signals. MSO. Very good BODE plot.

Rigol MSO5000 is 350Mhz design without 50Ω inputs. It has 2 Ch AWG, 4 decodes. Higher noise and lower sensitivity. It is good for general purpose work, not ideal for fine analog work. For Arduino kind of work decent choice.

Both Siglent and Rigol 1000HD platforms are low BW, up to 200 MHz, so same as SDS1104X-E. They would be very good for fine analog work, but with BW limit.

So as @coppice already said, Siglent SDS2104X Plus is best compromise of all listed, with limited other info...
Even very expensive scopes will never be perfect for "everything". They all specialize something: BW, noise, math, screen size, decoding...That is why people end up with many scopes...

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

[uargo]

but why do I want to buy the Rigol MSO5074 if it doesn't work for me later. The question is whether the Siglent SDS2104X PLUS will suffice, or better the Siglent SDS1000X HD or the Rigol HDO1000X HD.
Can someone who has these oscilloscopes advise me?

No, because if you don't know what you need then nobody does.

The only answer to that question is to buy the most expensive one and cross your fingers.

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

[BillyO]

I first got the SDS1104X-E then bought the SDS2104X-P for improved bandwidth.

I still have both.  I have used the 10 bit mode on the SDS2104X-P and it does work well, but as mentioned limits the BW to 100MHz.

Maybe if you have more specific questions about the scope we could continue to answer them instead of just asking for a general opinion.  We all use our equipment differently because will all do different things so our "general opinion" will be heavily colored by our own specific needs and may be meaningless to you, or worse, cause you to make a bad decision.

I will make one comment with regards to sensitivity.  The SDS2104x-P's enhanced sensitivity is very nice to have, but it requires that you have a very "quiet" lab environment EMI wise.

[Fungus]

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

OK, so the Rigol MSO5074 can be eliminated, you're looking at one of the 10-bit or 12-bit 'scopes for that.

The Rigol HDO is the cheapest and is also the newest tech with really fast/modern user interface.



(nb. DHO1000 has identical functions...)

[nctnico]

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

An oscilloscope is the wrong tool for that. You'll need a true RMS DMM with a high enough bandwidth if you want to go cheap. Otherwise you'll need some kind of pre-amplifier to visualise noise that low using an oscilloscope. Probing will also be challenging.

[uargo]

I first got the SDS1104X-E then bought the SDS2104X-P for improved bandwidth.

I still have both.  I have used the 10 bit mode on the SDS2104X-P and it does work well, but as mentioned limits the BW to 100MHz.

Maybe if you have more specific questions about the scope we could continue to answer them instead of just asking for a general opinion.  We all use our equipment differently because will all do different things so our "general opinion" will be heavily colored by our own specific needs and may be meaningless to you, or worse, cause you to make a bad decision.

I will make one comment with regards to sensitivity.  The SDS2104x-P's enhanced sensitivity is very nice to have, but it requires that you have a very "quiet" lab environment EMI wise.

thanks I wanted to know how the 10 bit mode of the sds2104x plus works

[uargo]

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

OK, so the Rigol MSO5074 can be eliminated, you're looking at one of the 10-bit or 12-bit 'scopes for that.

The Rigol HDO is the cheapest and is also the newest tech with really fast/modern user interface.



(nb. DHO1000 has identical functions...)

thanks I'll look at the rigol too but they say they still have many bugs

[uargo]

I am designing an antenna rotor controller and I need to reduce the noise from the power supply and the reference voltage for the ADC readings. I need a fairly accurate oscilloscope to see the noise at 600uV or less

An oscilloscope is the wrong tool for that. You'll need a true RMS DMM with a high enough bandwidth if you want to go cheap. Otherwise you'll need some kind of pre-amplifier to visualise noise that low using an oscilloscope. Probing will also be challenging.

thanks for that I bought a SDM3045X It will arrive on Friday

[Fungus]

thanks I'll look at the rigol too but they say they still have many bugs

ALL oscilloscopes have bugs. They're very complex devices.

The question is will it do what you need? Even Dave's early release model seems to be able to measure noise.

[Martin72]

thanks I'll look at the rigol too but they say they still have many bugs

As Fungus wrote, every DSO has bugs.
But it depends on the type and amount...
If the scope hangs regularly, that's not a nice little bug (it may have been fixed, I had "my" rigol in february).
In addition, only rudimentary functions such as in FFT mode, a bodeplot is completely missing.
This can still change, if they are clever, this must also change.
Only if and when is always such a thing with rigol and the reason why I had once turned my back on the brand.

The MSO5074 is currently 810 pounds + VAT in the UK. The SDS2104X+ is 1050 pounds + VAT. If you can afford the higher price you get a better instrument. The market seems to be functioning properly.

The rigol DS/MSO 7000 series has the same noisy front end as the MSO5000 (I had myself confirmed by rigol) and cost 2600/3300€ each for the 100Mhz/4ch variant, so much for the topic more expensive is better.

[uargo]

thanks I'll look at the rigol too but they say they still have many bugs

ALL oscilloscopes have bugs. They're very complex devices.

The question is will it do what you need? Even Dave's early release model seems to be able to measure noise.

I already know it. but people comment that the firmware is still unpolished, that it is logical that all oscilloscopes have many bugs at the beginning and over time they are fixed

[uargo]

thanks I'll look at the rigol too but they say they still have many bugs

As Fungus wrote, every DSO has bugs.
But it depends on the type and amount...
If the scope hangs regularly, that's not a nice little bug (it may have been fixed, I had "my" rigol in february).
In addition, only rudimentary functions such as in FFT mode, a bodeplot is completely missing.
This can still change, if they are clever, this must also change.
Only if and when is always such a thing with rigol and the reason why I had once turned my back on the brand.

The MSO5074 is currently 810 pounds + VAT in the UK. The SDS2104X+ is 1050 pounds + VAT. If you can afford the higher price you get a better instrument. The market seems to be functioning properly.

The rigol DS/MSO 7000 series has the same noisy front end as the MSO5000 (I had myself confirmed by rigol) and cost 2600/3300€ each for the 100Mhz/4ch variant, so much for the topic more expensive is better.

I hope the rigol or the Siglent 1000 HD are very good, but it is still too early to confirm it, they look good, but it is necessary to wait a while, I think.
The Siglent SDS2104X PLUS convinces me more and more, the firmware is more established, if the 10-bit mode works for me to see the noise, it is an oscilloscope that I like

[KungFuJosh]

I have the SDM3045X and the SDS2504XP, and they're both great. My SDM3045X was upgraded to the SDM3055X-E, which you can also do to get 5.5 digits.

[Fungus]

I already know it. but people comment that the firmware is still unpolished, that it is logical that all oscilloscopes have many bugs at the beginning and over time they are fixed

Even so: If it does then things I need very well then I can overlook a few bugs in the CAN bus decoder, or whatever.

eg. The way the stats work and the overall user interface on these new Rigols could save hours per day if that's what you spend your time using.

I expect a big push from Rigol in the next few months when their new 800 series hits the market. I'm already saving up for mine.

[2N3055]

I already know it. but people comment that the firmware is still unpolished, that it is logical that all oscilloscopes have many bugs at the beginning and over time they are fixed

Even so: If it does then things I need very well then I can overlook a few bugs in the CAN bus decoder, or whatever.

eg. The way the stats work and the overall user interface on these new Rigols could save hours per day if that's what you spend your time using.

I expect a big push from Rigol in the next few months when their new 800 series hits the market. I'm already saving up for mine.

Well, other people are different.
For instance I will take simpler device, that is reliable and works well, over a device that was developed by marketing department and has impressive list of "features" in datasheet but none of it works as it should...

And praise how "new Rigols" have "good interface" is fact that they finally made touch screen interface. In fact the were late to the party, it was long due  on their part. And state of software is still what other manufacturers would call beta.. They will eventually get there, but their track record is that they do it slow.

[Performa01]

The Siglent SDS2104X PLUS convinces me more and more, the firmware is more established, if the 10-bit mode works for me to see the noise, it is an oscilloscope that I like

That's it. Do not believe everything posted here, like claims a DSO being the wrong tool for noise characterization. This might be true for some scopes with noisy frontends and/or lousy FFT implementation, but not for e.g. an SDS2000X Plus. Of course, an SDS2000X HD would be even better.

When analyzing noise, we don't want a single absolute number like from a DMM with its very limited bandwidth; we want to see the entire spectrum with accurate levels for every single frequency component. Such information also helps us to identify the source of the noise (so we might be able to do something about it).

Once again, instead of just claims and speculations, here is a practical demonstration (older measurement, not the latest firmware), where I happened to measure a weak 10 MHz signal and wanted to find out if the 10 bit mode can improve such measurements even though the FFT in itself provides a resolution enhancement already.

I've found a measurement of the -70 dBm level particularly, which is 71 µVrms or 200 µVpp respectively - far below 600 µV (whatever it should be, peak, p-p or rms).

First the measurement in 8 bit mode.

SDS2354X Plus_LVL_10MHz_1mV_-70dBm_8bit

Of course the signal is not visible in the time domain, because at about 600 MHz bandwidth, the oscilloscope's own noise is 76 µVrms or 685 µVpp. Yet the FFT shows the overall noise floor as well as the single signal quite well.

The frequency step is 119,2 Hz and with Flattop window the resulting RBW (Resolution Bandwidth) is ~450 Hz. I did not measure the noise floor back then, but from the graph we can estimate it to be about -124 dBm = 141 nVrms or 399 nVpp for a 450 Hz wide bin. From this, we can calculate the noise density to 141 nVrms / √450 Hz = 6.65 nV/√Hz. From this, we can calculate the total noise for any desired bandwidth.

Hint: The signal source for this particular test did not include noise deliberately, so this measurement shows the limits of the instrument in this particular configuration and setup. It means that external noise has to be stronger than about 10 nV/√Hz in order to be detectable. Still not too bad.

But even more important than the wideband noise floor is the identification and proper measurement of spurious signals, which can give a strong hint on the noise source.

In this example, with just 8 bits, we measured the 10 MHz signal as -69.8 dBm (72 µVrms, 205 µVpp), which results in an error of +0.2 dB or +2.33 % - not too bad for a microvolts signal level at 10 MHz - try that with any DMM!


Now the same measurement in 10 bit mode.

SDS2354X Plus_LVL_10MHz_1mV_-70dBm_10bit

The signal is still not visible in the time domain, even though the input bandwidth is now limited to 100 MHz because of the 10 bit acquisition mode.

The noise floor is now at about -130 dBm = 71 nVrms or 200 nVpp for a 450 Hz wide bin. From this, we can calculate the noise density to 71 nVrms / √450 Hz = 3.34 nV/√Hz. That means, the noise floor has dropped dramatically by 6 dB – we have gained one bit of dynamic range, or in other words: the ENOB has been increased by one bit, just as expected – and up to 100 MHz there should hardly be any noise sources we could not accurately characterize with this.

Again, we can calculate the total noise for any desired bandwidth.

The measurement of the 10 MHz signal is now spot on as well. We can measure it as -70.015 dBm – an error of just -120 nVrms or -350 nVpp, equivalent to -0.015 dB or -0.17 % - no further comment necessary.

[nctnico]

The Siglent SDS2104X PLUS convinces me more and more, the firmware is more established, if the 10-bit mode works for me to see the noise, it is an oscilloscope that I like

That's it. Do not believe everything posted here, like claims a DSO being the wrong tool for noise characterization. This might be true for some scopes with noisy frontends and/or lousy FFT implementation, but not for e.g. an SDS2000X Plus. Of course, an SDS2000X HD would be even better.

When analyzing noise, we don't want a single absolute number like from a DMM with its very limited bandwidth; we want to see the entire spectrum with accurate levels for every single

If you look at OP's specific problem then you'll see it is about finding power supply noise in an analog circuit. Unless the OP has totally messed up the power supply decoupling, that means the problem is in the DC to several tens of kHz frequency range. You'd need a filter to make FFT usefull on an oscilloscope to prevent aliasing due to large sub-sampling factor. There are better tools out there compared to a general purpose oscilloscope. As I wrote before, probing also becomes an important factor so a device with floating or differential inputs will help to keep ground induced noise out.

[uargo]

I have the SDM3045X and the SDS2504XP, and they're both great. My SDM3045X was upgraded to the SDM3055X-E, which you can also do to get 5.5 digits.

Thanks, can you give me a link on how to convert the sdm3045x to sdm3055x-e?

[uargo]

I already know it. but people comment that the firmware is still unpolished, that it is logical that all oscilloscopes have many bugs at the beginning and over time they are fixed

Even so: If it does then things I need very well then I can overlook a few bugs in the CAN bus decoder, or whatever.

eg. The way the stats work and the overall user interface on these new Rigols could save hours per day if that's what you spend your time using.

I expect a big push from Rigol in the next few months when their new 800 series hits the market. I'm already saving up for mine.

I also hope to see a review of the new rigol and siglent 1000hd, I'm impatient

[uargo]

I already know it. but people comment that the firmware is still unpolished, that it is logical that all oscilloscopes have many bugs at the beginning and over time they are fixed

Even so: If it does then things I need very well then I can overlook a few bugs in the CAN bus decoder, or whatever.

eg. The way the stats work and the overall user interface on these new Rigols could save hours per day if that's what you spend your time using.

I expect a big push from Rigol in the next few months when their new 800 series hits the market. I'm already saving up for mine.

Well, other people are different.
For instance I will take simpler device, that is reliable and works well, over a device that was developed by marketing department and has impressive list of "features" in datasheet but none of it works as it should...

And praise how "new Rigols" have "good interface" is fact that they finally made touch screen interface. In fact the were late to the party, it was long due  on their part. And state of software is still what other manufacturers would call beta.. They will eventually get there, but their track record is that they do it slow.

+1 I think like you, even so they are interesting devices to keep track of them

[uargo]

The Siglent SDS2104X PLUS convinces me more and more, the firmware is more established, if the 10-bit mode works for me to see the noise, it is an oscilloscope that I like

That's it. Do not believe everything posted here, like claims a DSO being the wrong tool for noise characterization. This might be true for some scopes with noisy frontends and/or lousy FFT implementation, but not for e.g. an SDS2000X Plus. Of course, an SDS2000X HD would be even better.

When analyzing noise, we don't want a single absolute number like from a DMM with its very limited bandwidth; we want to see the entire spectrum with accurate levels for every single frequency component. Such information also helps us to identify the source of the noise (so we might be able to do something about it).

Once again, instead of just claims and speculations, here is a practical demonstration (older measurement, not the latest firmware), where I happened to measure a weak 10 MHz signal and wanted to find out if the 10 bit mode can improve such measurements even though the FFT in itself provides a resolution enhancement already.

I've found a measurement of the -70 dBm level particularly, which is 71 µVrms or 200 µVpp respectively - far below 600 µV (whatever it should be, peak, p-p or rms).

First the measurement in 8 bit mode.

SDS2354X Plus_LVL_10MHz_1mV_-70dBm_8bit

Of course the signal is not visible in the time domain, because at about 600 MHz bandwidth, the oscilloscope's own noise is 76 µVrms or 685 µVpp. Yet the FFT shows the overall noise floor as well as the single signal quite well.

The frequency step is 119,2 Hz and with Flattop window the resulting RBW (Resolution Bandwidth) is ~450 Hz. I did not measure the noise floor back then, but from the graph we can estimate it to be about -124 dBm = 141 nVrms or 399 nVpp for a 450 Hz wide bin. From this, we can calculate the noise density to 141 nVrms / √450 Hz = 6.65 nV/√Hz. From this, we can calculate the total noise for any desired bandwidth.

Hint: The signal source for this particular test did not include noise deliberately, so this measurement shows the limits of the instrument in this particular configuration and setup. It means that external noise has to be stronger than about 10 nV/√Hz in order to be detectable. Still not too bad.

But even more important than the wideband noise floor is the identification and proper measurement of spurious signals, which can give a strong hint on the noise source.

In this example, with just 8 bits, we measured the 10 MHz signal as -69.8 dBm (72 µVrms, 205 µVpp), which results in an error of +0.2 dB or +2.33 % - not too bad for a microvolts signal level at 10 MHz - try that with any DMM!


Now the same measurement in 10 bit mode.

SDS2354X Plus_LVL_10MHz_1mV_-70dBm_10bit

The signal is still not visible in the time domain, even though the input bandwidth is now limited to 100 MHz because of the 10 bit acquisition mode.

The noise floor is now at about -130 dBm = 71 nVrms or 200 nVpp for a 450 Hz wide bin. From this, we can calculate the noise density to 71 nVrms / √450 Hz = 3.34 nV/√Hz. That means, the noise floor has dropped dramatically by 6 dB – we have gained one bit of dynamic range, or in other words: the ENOB has been increased by one bit, just as expected – and up to 100 MHz there should hardly be any noise sources we could not accurately characterize with this.

Again, we can calculate the total noise for any desired bandwidth.

The measurement of the 10 MHz signal is now spot on as well. We can measure it as -70.015 dBm – an error of just -120 nVrms or -350 nVpp, equivalent to -0.015 dB or -0.17 % - no further comment necessary.

Thank you very much and great collaboration, summarizing an SDS2104X PLUS works, although an SDS2104HD would be better, and the 1000HD would need to see a review to see what background noise they have

[uargo]

The Siglent SDS2104X PLUS convinces me more and more, the firmware is more established, if the 10-bit mode works for me to see the noise, it is an oscilloscope that I like

That's it. Do not believe everything posted here, like claims a DSO being the wrong tool for noise characterization. This might be true for some scopes with noisy frontends and/or lousy FFT implementation, but not for e.g. an SDS2000X Plus. Of course, an SDS2000X HD would be even better.

When analyzing noise, we don't want a single absolute number like from a DMM with its very limited bandwidth; we want to see the entire spectrum with accurate levels for every single

If you look at OP's specific problem then you'll see it is about finding power supply noise in an analog circuit. Unless the OP has totally messed up the power supply decoupling, that means the problem is in the DC to several tens of kHz frequency range. You'd need a filter to make FFT usefull on an oscilloscope to prevent aliasing due to large sub-sampling factor. There are better tools out there compared to a general purpose oscilloscope. As I wrote before, probing also becomes an important factor so a device with floating or differential inputs will help to keep ground induced noise out.

Thanks, for that, apart from buying the SDM3045X, I wanted to see what oscilloscope could help me with this, and the answer seems to be the SDS2104X PLUS, the SDS2104 HD (out of my budget), and the 1000 HD series from Rigol and Siglent

[Performa01]

I also hope to see a review of the new rigol and siglent 1000hd, I'm impatient

Instead of being impatient, what about contributing to the community and do a thorough review yourself

Seriously though, there are plenty "reviews" out there. They will show you how to unpack the Rigol and some fanboys will praise it to no end, even though they've never touched one, just because it finally has an acceptable noise level and a user interface that doesnt look like a slot machine.
-----

Just to provide complete information on the SDS2000X Plus: aliasing isn't a problem even at low frequencies on a decent DSO with long FFT. For instance, we can analyze down to 100 Hz and still have 250 MSa/s for the FFT. Together with the 20 MHz bandwidth limiter and the 10 bit mode, which additionally limits the bandwidth at 100 MHz rather steeply, we don't have a problem with aliasing due to subsampling at all.

It is true though that general purpose oscilloscopes have a rather strong rise in noise below some 100 kHz, hence measuring very low noise will not be possible at low frequencies.

600 µVpp would be equivalent to 212 µVrms or -60,46 dBm (I use dBm even though we are talking about voltages here, just because my measurements happen to be in dBm). Consequently, we can analyze noise as long as the FFT plot does not exceed that level. As we can see from the attached graph, even at just 1 kHz, the measured level is still -96 dBm for 450 Hz bandwidth – way below the -60.46 dBm calculated before. Only down at 200 Hz we reach a noise level of 900 µVrms.

But then again: if you need to look just for mains hum and its harmonics – and 600 µV would be a problem for your circuit alrady, then you actually need a dynamic signal analyzer (expensive) or a PicoScope 4262.

SDS2354X Plus_Noise_1M_BW20M_10bit

[Martin72]

...and a user interface that doesnt look like a slot machine.

 

Made my day.
And it´s true.

[KungFuJosh]

I have the SDM3045X and the SDS2504XP, and they're both great. My SDM3045X was upgraded to the SDM3055X-E, which you can also do to get 5.5 digits.

Thanks, can you give me a link on how to convert the sdm3045x to sdm3055x-e?

Check here: https://www.eevblog.com/forum/testgear/siglent-sdm3045x-enough-is-enough/msg3909188/#msg3909188

review that thread for other details. It's pretty easy.

[Fungus]

And praise how "new Rigols" have "good interface" is fact that they finally made touch screen interface. In fact the were late to the party, it was long due  on their part.

Look again, it's much more than "touch screen bolted onto legacy interface".

eg. Channel configuration:


Multi windowing:


Watch this video (on large touch screen via HDMI output):

[uargo]

I also hope to see a review of the new rigol and siglent 1000hd, I'm impatient

Instead of being impatient, what about contributing to the community and do a thorough review yourself

Seriously though, there are plenty "reviews" out there. They will show you how to unpack the Rigol and some fanboys will praise it to no end, even though they've never touched one, just because it finally has an acceptable noise level and a user interface that doesnt look like a slot machine.
-----

Just to provide complete information on the SDS2000X Plus: aliasing isn't a problem even at low frequencies on a decent DSO with long FFT. For instance, we can analyze down to 100 Hz and still have 250 MSa/s for the FFT. Together with the 20 MHz bandwidth limiter and the 10 bit mode, which additionally limits the bandwidth at 100 MHz rather steeply, we don't have a problem with aliasing due to subsampling at all.

It is true though that general purpose oscilloscopes have a rather strong rise in noise below some 100 kHz, hence measuring very low noise will not be possible at low frequencies.

600 µVpp would be equivalent to 212 µVrms or -60,46 dBm (I use dBm even though we are talking about voltages here, just because my measurements happen to be in dBm). Consequently, we can analyze noise as long as the FFT plot does not exceed that level. As we can see from the attached graph, even at just 1 kHz, the measured level is still -96 dBm for 450 Hz bandwidth – way below the -60.46 dBm calculated before. Only down at 200 Hz we reach a noise level of 900 µVrms.

But then again: if you need to look just for mains hum and its harmonics – and 600 µV would be a problem for your circuit alrady, then you actually need a dynamic signal analyzer (expensive) or a PicoScope 4262.

SDS2354X Plus_Noise_1M_BW20M_10bit

thanks

[uargo]

I have the SDM3045X and the SDS2504XP, and they're both great. My SDM3045X was upgraded to the SDM3055X-E, which you can also do to get 5.5 digits.

Thanks, can you give me a link on how to convert the sdm3045x to sdm3055x-e?

Check here: https://www.eevblog.com/forum/testgear/siglent-sdm3045x-enough-is-enough/msg3909188/#msg3909188

review that thread for other details. It's pretty easy.

thanks

[Martin72]

Look again, it's much more than "touch screen bolted onto legacy interface".

If you only watch videos and pictures, you might not get the right impression.
In reality, the creation and positioning of the windows was a difficult, annoying and not always as functional as intended matter.
I had also described this in the Rigol thread.
Nevertheless, it is the "best" user interface rigol has ever had, no question.
But as almost always, it is very well thought, but half-heartedly done.
This ultimately drove me into the arms of siglent.
What they offer is not super fancy, but it works as intended.
In practice, that makes all the difference.

[2N3055]

And praise how "new Rigols" have "good interface" is fact that they finally made touch screen interface. In fact the were late to the party, it was long due  on their part.

Look again, it's much more than "touch screen bolted onto legacy interface".

That is exactly what I said. Previous MSO5000/7000 was an awkward hybrid. But as Martin says, even the new one is far from being that good. Not because concept is wrong ( concept is kind of right this time) but you can literally see it is brand new and kind of thrown together in a  hurry. Bunch of stuff is missing etc.. You can see they were rushing it to the market.
But they are on the right track and if they finish it they might have good product.

I really wish they take it seriously and make good products. It is good for us to have choice..

Or maybe they will finish it 80% and then move to the new products in the future...
Only Rigol knows what are real plans and only time will tell.

[nctnico]

And praise how "new Rigols" have "good interface" is fact that they finally made touch screen interface. In fact the were late to the party, it was long due  on their part.

Look again, it's much more than "touch screen bolted onto legacy interface".

That is exactly what I said. Previous MSO5000/7000 was an awkward hybrid. But as Martin says, even the new one is far from being that good. Not because concept is wrong ( concept is kind of right this time) but you can literally see it is brand new and kind of thrown together in a  hurry. Bunch of stuff is missing etc.. You can see they were rushing it to the market.
But they are on the right track and if they finish it they might have good product.

I really wish they take it seriously and make good products. It is good for us to have choice..

Yes. Even if that means the price being on par with Keysight, R&S, Tektronix, Yokogawa, etc. Engineers cost serious money in China nowadays so realistically the Chinese can't make test equipment that really is on par with the A-brands cheaper than the A-brands can. And a lot of the problem is lack of knowledge as well. China is handing out large financial incentives for companies to bring leading edge technology into China (just to be clear: this is first hand information!).

[Martin72]

Even if that means the price being on par with Keysight, R&S, Tektronix, Yokogawa, etc.

Based on daily work and comparison with an "A" brand, I have to say that the Siglent scope performs very well.
If you want to include "more" than just the performance, you have to make sacrifices, that's clear.
There's the quality of the device's workmanship, the variety of functions and the comfort of the user interface, and above all the after-sales service.
If that would be on the same level as the other providers, there would be no significant price difference (and therefore no reason to buy them).
Privately, I have no problem with that, so I can disregard that and buy a "cheaper" scope.
At work, in the test field, I have nevertheless introduced the Siglent scopes, due to a simple consideration.
For the money, they perform exceedingly well and if there is something wrong with them, we just buy a new one instead of having it repaired.
What I don't like or find irritating, however, is the "general mood" here in the forum regarding scopes and their prices.
A scope for say 3000€/$ is still a cheap device today.
You should not fool yourself.

[Veteran68]

What I don't like or find irritating, however, is the "general mood" here in the forum regarding scopes and their prices.
A scope for say 3000€/$ is still a cheap device today.
You should not fool yourself.

What mood is that, exactly?

You have to remember this forum is made up of both hobbyists and those who do this for a living. I don't know what the mix is, but I suspect it leans more towards hobbyists than working professionals. In a professional lab, scopes and logic analyzers commonly run into the many tens of thousands of dollars. But outside of a job like that, or cheap(er) kits, back 30 years ago hobbyists had no ready access to that kind of equipment.

Thanks to sub-$500 oscilloscopes, now any semi-serious hobbyist or maker can afford a scope. Since those people are likely a majority of this forum, I'd say $3000 for a scope (while certainly cheap for commercial use) is still WAY out of reach of most of them. I certainly won't pay that for my personal use.

[BillyO]

For instance I will take simpler device, that is reliable and works well, over a device that was developed by marketing department and has impressive list of "features" in datasheet but none of it works as it should...

You mean like these: http://hantek.com/products/detail/17182

And these: http://www.fnirsi.com.cn/product/704340100322037760

 
 
 
 
 

[Performa01]

… so in my last posting I've shown the absolute worst case – and was curious if someone would jump at it. That's not happened, so here I'll show a more realistic use case.

In my last posting, I've used 1 meg input impedance and left the input open, which would be equivalent to a x10 probe at low frequencies. In any real power supply noise measurement we would either use a direct coax connection (total input impedance <50 ohms) or a x1 probe (total input impedance is a couple hundred ohms) – especially when looking for very low noise levels like 600 µV.

While the termination doesn't make a significant difference at higher frequencies when it comes to noise, it is still important at low frequencies. This time I use 50 ohms input impedance (the result would be the same if we choose 1 meg input impedance and use an external termination up to a few kiloohms).

Furthermore, there's no need to worry about aliasing, even when we lower the FFT-sample rate a bit. If the noise from the PS has high frequency content, then we need a wide FFT bandwidth hence also a high FFT sample rate anyway. If it's really just low frequency noise, then a low sample rate is sufficient. Aliasing would only fold down the scopes own noise, thus rising the intrinsic noise level of the measurement, but because of the strong 1/f characteristic it absolutely doesn't matter if a higher frequency part with <5 nV/√Hz gets added to the measurement. Consequently, I've now chosen 10 ms/div horizontal time base for a lower frequency limit of 10 Hz and a FFT sample rate of just 25 MSa/s.

SDS2354X Plus_Noise_50_10ms_20kHz_10bit

As we've seen before, 600 µVpp would be equivalent to 212 µVrms or -60,46 dBm and the OP could analyze noise and spurs in his scenario as long as the FFT plot does not exceed that level. As we can see from the attached graph, even at just 100 Hz, the measured level is still -91 dBm (this time for 45 Hz bandwidth), that leaves plenty headroom for analyzing a -60.46 dBm signal.

[Fungus]

Look again, it's much more than "touch screen bolted onto legacy interface".

If you only watch videos and pictures, you might not get the right impression.
In reality, the creation and positioning of the windows was a difficult, annoying and not always as functional as intended matter.

But it was there, it could do it. You could have a 3D view alongside a normal view.

This ultimately drove me into the arms of siglent.
What they offer is not super fancy, but it works as intended.

You'd have a point if they were both the same price, but they aren't.

If price isn't an issue, if you simply want the "best" then why stop at Siglent? R&S make some very nice devices.

Thanks to sub-$500 oscilloscopes, now any semi-serious hobbyist or maker can afford a scope. Since those people are likely a majority of this forum, I'd say $3000 for a scope (while certainly cheap for commercial use) is still WAY out of reach of most of them. I certainly won't pay that for my personal use.

Yep.

For the stated task, ie. measuring noise on a power supply, the Rigol HDO is perfectly capable.

[Fungus]

What mood is that, exactly?

The mood I usually see is a "BMW owners club" mood.

BMW owners justify spending more on their cars by going on and on about how fast they are and how they handle in the corners, how every other brand is somehow "compromised", etc. It all makes sense on paper but the simple fact is that most people just want to get around the place at normal speeds and carry shopping twice a week. A Ford Fiesta can do that perfectly and isn't missing any basic features like aircon or wheels.

Don't believe me? Siglent owners give the exact same excuses for not owning an R&S as BMW 3-series owners give for not owning a 5-series or 7-series.

Maybe when somebody asks "what 'scope should I buy?" we should start out by asking them what car they drive and answer accordingly. I predict these threads will be a lot shorter if we do that.

Thanks to sub-$500 oscilloscopes, now any semi-serious hobbyist or maker can afford a scope. Since those people are likely a majority of this forum, I'd say $3000 for a scope (while certainly cheap for commercial use) is still WAY out of reach of most of them. I certainly won't pay that for my personal use.

Yep.

[Performa01]

If price isn't an issue, if you simply want the "best" then why stop at Siglent? R&S make some very nice devices.

Oh yes ... "the  best". You sure have massive experience with R&S scopes. It would be very interesting how you manage to do the measurements I've shown above with just 128kpts FFT length? As always, I miss the demonstration and measurement results to back up your claims...

Yep.

For the stated task, ie. measuring noise on a power supply, the Rigol HDO is perfectly capable.

Here also ... have you ever tested one? If so, why not show us the results, so everyone can see this "perfect capability" with their own eyes? I do not doubt that it can do it, but seeing is believing...

[2N3055]

What mood is that, exactly?

The mood I usually see is a "BMW owners club" mood.

BMW owners justify spending more on their cars by going on and on about how fast they are and how they handle in the corners, how every other brand is somehow "compromised", etc. It all makes sense on paper but the simple fact is that most people just want to get around the place at normal speeds and carry shopping twice a week. A Ford Fiesta can do that perfectly and isn't missing any basic features like aircon or wheels.

Don't believe me? Siglent owners give the exact same excuses for not owning an R&S as BMW 3-series owners give for not owning a 5-series or 7-series.

Maybe when somebody asks "what 'scope should I buy?" we should start out by asking them what car they drive and answer accordingly. I predict these threads will be a lot shorter if we do that.

Thanks to sub-$500 oscilloscopes, now any semi-serious hobbyist or maker can afford a scope. Since those people are likely a majority of this forum, I'd say $3000 for a scope (while certainly cheap for commercial use) is still WAY out of reach of most of them. I certainly won't pay that for my personal use.

Yep.

Another bad car comparison.....

What are you tryin to say?  I honestly don't get it?

[Fungus]

Here also ... have you ever tested one? If so, why not show us the results, so everyone can see this "perfect capability" with their own eyes? I do not doubt that it can do it, but seeing is believing...

Dave's video was about pretty much nothing other than noise measurement:

[2N3055]

Here also ... have you ever tested one? If so, why not show us the results, so everyone can see this "perfect capability" with their own eyes? I do not doubt that it can do it, but seeing is believing...

Dave's video was about pretty much nothing other than noise measurement:

Errm... it is not?

As I said before, it seems Rigol uses decent ADC this time. And it seems to be (very?) decent noise levels, hardware vise.
But that is hardware "promise". With half baked FFT and unfinished software not there yet.

Problem in the real world is that measuring microvolt levels is not easy, and it is questionable if OP can make setup that will be better than the scope he has now... As Nico says, probing, grounding, shielding  etc is critical here.

It is unclear (even to OP) what he really needs. Saying "600 uV levels on power supply" without stating is it RMS or P-P or what BW is not really clear requirement. Also depending as to what this power rail is for maybe such low noise/ripple levels are not needed at all, i.e. devices powered also have it's own PSRR that might be sufficient for power not to be critical at all...

[mawyatt]

What mood is that, exactly?

The mood I usually see is a "BMW owners club" mood.

BMW owners justify spending more on their cars by going on and on about how fast they are and how they handle in the corners, how every other brand is somehow "compromised", etc. It all makes sense on paper but the simple fact is that most people just want to get around the place at normal speeds and carry shopping twice a week. A Ford Fiesta can do that perfectly and isn't missing any basic features like aircon or wheels.

Don't believe me? Siglent owners give the exact same excuses for not owning an R&S as BMW 3-series owners give for not owning a 5-series or 7-series.

Maybe when somebody asks "what 'scope should I buy?" we should start out by asking them what car they drive and answer accordingly. I predict these threads will be a lot shorter if we do that.

Thanks to sub-$500 oscilloscopes, now any semi-serious hobbyist or maker can afford a scope. Since those people are likely a majority of this forum, I'd say $3000 for a scope (while certainly cheap for commercial use) is still WAY out of reach of most of them. I certainly won't pay that for my personal use.

Yep.

Another bad car comparison.....

What are you tryin to say?  I honestly don't get it?

Maybe not os bad.

We Drive a pair of Siglent's and our MSOs (Max Speed Object) are a pair of Porsche 911s, one is a Twin Turbo (Black) for the faster waveform captures

Rigol has apparently developed a good custom new engine (12 Bit ADC) and new chassis (OS) but hasn't quite got the handling (UI) refined yet, whereas Siglent is using a GP engine, older chassis (OS) and has considerable refined handling.

So on the Track (actual use) the Rigol is fast on the straights (simple tasks), but struggles in the braking & corners (skilled tasks) due to the handling (UI) and suffers slower overall track times (user experience), whereas the Siglent is fast everywhere with the superior handling (UI) and sports an overall quicker lap time (user experience).

Best,

[Fungus]

Dave's video was about pretty much nothing other than noise

Errm... it is not?

Confirmed,  there's no arguing with Siglent/BMW owners.


Clue: Nobody's saying you're wrong.

[uargo]

Here also ... have you ever tested one? If so, why not show us the results, so everyone can see this "perfect capability" with their own eyes? I do not doubt that it can do it, but seeing is believing...

Dave's video was about pretty much nothing other than noise measurement:

Errm... it is not?

As I said before, it seems Rigol uses decent ADC this time. And it seems to be (very?) decent noise levels, hardware vise.
But that is hardware "promise". With half baked FFT and unfinished software not there yet.

Problem in the real world is that measuring microvolt levels is not easy, and it is questionable if OP can make setup that will be better than the scope he has now... As Nico says, probing, grounding, shielding  etc is critical here.

It is unclear (even to OP) what he really needs. Saying "600 uV levels on power supply" without stating is it RMS or P-P or what BW is not really clear requirement. Also depending as to what this power rail is for maybe such low noise/ripple levels are not needed at all, i.e. devices powered also have it's own PSRR that might be sufficient for power not to be critical at all...

600uVpp, I use a 2.5vdc CJ431 (Chinese TL431) voltage reference and the 12-bit adc from an STM32F401RCT6, so 2.5vdc/4096 steps = 0.0006v each step. but there is noise that I would like to reduce. and if I use a digital filter.
The source consists of:
1 - 12vdc input
2 - 12vdc input, dc-dc source with a Tps54331 and 5vdc output
3 - 5vdc input, linear source with Ams1117-3.3 with 3.3v output

STM32F401RCT6 and Cj431 voltage reference at 2.5v powered at 3.3vdc

[2N3055]

Dave's video was about pretty much nothing other than noise

Errm... it is not?

Confirmed,  there's no arguing with Siglent/BMW owners.

Clue: Nobody's saying you're wrong.

You are the king of arguing with irrelevant or nonconnected data.

Dave unpacked that scope for 23 minutes and then for 7 minutes has shown it has less waveform noise and better resolution that scopes with less resolution a larger noise. He could have shown comparison to several low noise 8 bit scopes (that includes several Siglents and a litlle Micsig) or compare it to other 12 bit low noise scopes.
But that is beyond point.

What would be interesting and "on point" for this topic would be showing how well Rigol DHO(HDO) would fare measuring a PSU noise.
Where you need an FFT, and high sensitivity, and low noise etc...

What he has shown is Rigol has nice packing, and unfinished product. That video is not a good recommendation. Rigol has to quickly debug and finish software and make a new video that shows it actually works..

Like Mike says, modern DSO are signal analysis machines. Rigol at this stage of development is an analog scope with some analysis options...
How well those develop with time will shape how I judge it's usability for the task..

[2N3055]

Here also ... have you ever tested one? If so, why not show us the results, so everyone can see this "perfect capability" with their own eyes? I do not doubt that it can do it, but seeing is believing...

Dave's video was about pretty much nothing other than noise measurement:

Errm... it is not?

As I said before, it seems Rigol uses decent ADC this time. And it seems to be (very?) decent noise levels, hardware vise.
But that is hardware "promise". With half baked FFT and unfinished software not there yet.

Problem in the real world is that measuring microvolt levels is not easy, and it is questionable if OP can make setup that will be better than the scope he has now... As Nico says, probing, grounding, shielding  etc is critical here.

It is unclear (even to OP) what he really needs. Saying "600 uV levels on power supply" without stating is it RMS or P-P or what BW is not really clear requirement. Also depending as to what this power rail is for maybe such low noise/ripple levels are not needed at all, i.e. devices powered also have it's own PSRR that might be sufficient for power not to be critical at all...

600uVpp, I use a 2.5vdc CJ431 (Chinese TL431) voltage reference and the 12-bit adc from an STM32F401RCT6, so 2.5vdc/4096 steps = 0.0006v each step. but there is noise that I would like to reduce. and if I use a digital filter.
The source consists of:
1 - 12vdc input
2 - 12vdc input, dc-dc source with a Tps54331 and 5vdc output
3 - 5vdc input, linear source with Ams1117-3.3 with 3.3v output

STM32F401RCT6 and Cj431 voltage reference at 2.5v powered at 3.3vdc

You will need a better reference.....

For interesting solution, lookup low noise designs based on old 723..  but you will need to power it with 12 V directly..
That would also be lower tempco...
 Use it as VDDA and ref...

[uargo]

Here also ... have you ever tested one? If so, why not show us the results, so everyone can see this "perfect capability" with their own eyes? I do not doubt that it can do it, but seeing is believing...

Dave's video was about pretty much nothing other than noise measurement:

Errm... it is not?

As I said before, it seems Rigol uses decent ADC this time. And it seems to be (very?) decent noise levels, hardware vise.
But that is hardware "promise". With half baked FFT and unfinished software not there yet.

Problem in the real world is that measuring microvolt levels is not easy, and it is questionable if OP can make setup that will be better than the scope he has now... As Nico says, probing, grounding, shielding  etc is critical here.

It is unclear (even to OP) what he really needs. Saying "600 uV levels on power supply" without stating is it RMS or P-P or what BW is not really clear requirement. Also depending as to what this power rail is for maybe such low noise/ripple levels are not needed at all, i.e. devices powered also have it's own PSRR that might be sufficient for power not to be critical at all...

600uVpp, I use a 2.5vdc CJ431 (Chinese TL431) voltage reference and the 12-bit adc from an STM32F401RCT6, so 2.5vdc/4096 steps = 0.0006v each step. but there is noise that I would like to reduce. and if I use a digital filter.
The source consists of:
1 - 12vdc input
2 - 12vdc input, dc-dc source with a Tps54331 and 5vdc output
3 - 5vdc input, linear source with Ams1117-3.3 with 3.3v output

STM32F401RCT6 and Cj431 voltage reference at 2.5v powered at 3.3vdc

You will need a better reference.....

For interesting solution, lookup low noise designs based on old 723..  but you will need to power it with 12 V directly..
That would also be lower tempco...
 Use it as VDDA and ref...

thanks

[Martin72]

What mood is that, exactly?

In any case, already no car club nonsense....

No, I sometimes have the feeling that some overestimate the measuring equipment for small money or expect too much from it, because you have to spend quite a lot on it in relation to other hobby expenses.

[tautech]

What mood is that, exactly?

In any case, already no car club nonsense....

No, I sometimes have the feeling that some overestimate the some measuring equipment for small money or expect too much from it, because you have to spend quite a lot on it in relation to other hobby expenses.

Fixed. 

[Veteran68]

What mood is that, exactly?

In any case, already no car club nonsense....

No, I sometimes have the feeling that some overestimate the measuring equipment for small money or expect too much from it, because you have to spend quite a lot on it in relation to other hobby expenses.

Fair enough. I can agree with that sentiment.

And I want to be clear, I'm a fan of nice toys. I'm not judging ANYONE for what they choose to spend their own money on as long as they're not hurting anyone else with it. I like BMW's, though I don't own one. I do however drive a 2021 Audi that was not inexpensive, so some would put me in the same boat. Point being, it's not like I'm a cheapskate who thinks hobbyists buying $3K scopes is stupid, it just doesn't make sense to me in my situation. Not that I can't afford it, just that I don't see it as a wise investment -- unlike a nice car, which others may rightly argue with. My cap would probably be in the ~$1500 range for a single piece of gear. Some will see that as cheap, and others will see it as exorbitant.

If someone gets as much enjoyment out of their high end scope that I get out of my car, more power to them, no judgment from me. But I can certainly understand it being way out of range of most hobbyists and why they might question the sanity of such an investment. Likewise, to your point, they should also understand that they're not going to get $3K capability/quality/support out of a $300 (or $50) scope , and they need to be okay with that too.

I have a considerable investment in my home lab, no question about it. I have a particular weakness for collecting multimeters and have spent a small fortune on my collection, and am not done. I'm eyeing another 6.5 digit bench meter, even though I have no real justification for it, other than I want an update to my 34401A. I just think about how many toys I can buy and play with for the cost of one $3K scope that I'll likely never reach the capability of.

[KungFuJosh]

$3K isn't much for a scope. But just for fun, look at the SDS2104X Plus. If you unlock all the options, and upgrade to 500MHz, the total cost is $6821. Or $6252 total if you buy the SDS2354XP and upgrade that with all options to 500MHz. That's $6k to $7k for what's not really considered a mid-level scope. It's a wonderful scope all the same.

[Veteran68]

$3K isn't much for a scope. But just for fun, look at the SDS2104X Plus. If you unlock all the options, and upgrade to 500MHz, the total cost is $6821. Or $6252 total if you buy the SDS2354XP and upgrade that with all options to 500MHz. That's $6k to $7k for what's not really considered a mid-level scope. It's a wonderful scope all the same.

Oh, I'm well aware. I've worked with scopes well into the 5 figures, but that was in a business/professional lab. I'm talking strictly hobby here. Maybe if electronics was my ONLY hobby, I'd be willing to drop $3K on a scope. But I have several hobbies, and all happen to be expensive...

[tautech]

$3K isn't much for a scope. But just for fun, look at the SDS2104X Plus. If you unlock all the options, and upgrade to 500MHz, the total cost is $6821. Or $6252 total if you buy the SDS2354XP and upgrade that with all options to 500MHz. That's $6k to $7k for what's not really considered a mid-level scope. It's a wonderful scope all the same.

Unlock = buy = make permanent whereas as standard you have all optional capability for 30 free trial uses.
BW upgrades have never had trial uses.

Yep SDS2kXP is a very capable DSO.

Yesterday a buddy/customer had issues with the timebase....self inflicted as he had never engaged the Reference position feature so increasing the timebase would shoot the H Pos setting off the display and exceed the max horizontal offset and flag a warning.
After he updated the FW after I discovered he never had  the same issue was present until the H Pos button was pressed to set it back to zero. Then no further problem...as expected, so guided him through setting the H Pos Ref position after which he was as happy as a dog with 2 tails. 

[uargo]

Taking into account that the Siglent SDS1104X-E has a background noise of approx =< 250uVrms

I HAVE MADE A TABLE WITH THE MAIN CHARACTERISTICS OF THE CONTENTS

         SIGLENT 2104X PLUS      SIGLENT 2104X HD      RIGOL DHO1104         SIGLENT SDS1104X HD               
         
BITS:         8 - 10 MODE(100MHZ)      12            12            12
CANALES:      4            4            4            4
FRECUENCIA:      100-350MHZ         100-500MHZ         70-200MHZ         70-200MHZ
SAMPLEO:      2GSA/S TOTAL         2GSA/S TOTAL         2GSA/S            1GSA/S
MEMORIA:      400MB  TOTAL         400MB TOTAL         100MB TOTAL         100MB TOTAL
SAMPLES:      120.000 - 500.000WFM/S      100.000-500.000WFM/S      50.000-1.500.000WFM/S      80.000-400.000WFM/S
GRADOS DE COLOR:   256            256            256            256
SERIAL PROTOCOL:   YES OPT.         YES OPT.         YES- OPT.         YES- OPT.
MIN VOLT:      80uVRMSNOISE 500uV/DIV           70uVRMSNOISE 500uV/DIV      66uVRMSNOISE 500uV/DIV      70uVRMSNOISE 500uV/DIV
MATEMATICAS:      2            2            4            4
FFT:         2MB            2MB            1MB            2MB
BODEPLOT:      YES OPT.         YES OPT.         NO            YES OPT.
POWER ANALYSIS:      YES OPT.         YES OPT.         NO            YES OPT.
DIGITAL:      16CHANNEL OPT.         16CHANNEL OPT.         NO            16CHANNEL OPT.
WAVEFORM GEN:      YES OPT.         YES OPT.         NO            YES OPT.
CONS:         LAN, USB(MASTER-DEVICE),   LAN, USB(MASTER-DEVICE),   LAN, USB(MASTER-DEVICE),HDMI   LAN, USB(MASTER-DEVICE),
         TRIGGER OUT, PASS/FAIL, WEB   TRIGGER OUT, PASS/FAIL, WEB   TRIGGER OUT, PASS/FAIL, WEB   TRIGGER OUT, PASS/FAIL, WEB
TFT:         10.1" TOUCH 1024X600      10.1" TOUCH 1024X600      10.1" TOUCH 1280X800      10.1" TOUCH 1024X600
HACKABLE:      YES            ?            YES            ?
PRECIO:         1.450€ AMAZON         3.546€ BATRONIX         1.523€ AMAZON (1.087€-70MHZ)   ?

[tautech]

FYI
All the Siglent scopes listed offer Bode plot capability as std.
Can confirm SDS1000X HD options can be unofficially enabled and for this model series just protocols I2C, SPI, UART, CAN, LIN decodes will be available.

[Performa01]

Taking into account that the Siglent SDS1104X-E has a background noise of approx 300uVrms

Where do you get that number from?

My Siglent SDS1104X-E has 56 µVrms noise for full bandwidth and 27.3 µVrms when 20 MHz bandwidth limiter is active (measured at 500 µV/div). Regarding noise, it is in a similar ballpark as the SDS2000X Plus, yet suffers more and stronger spurious signals.

I HAVE MADE A TABLE WITH THE MAIN CHARACTERISTICS OF THE CONTENTS

To the best of my knowledge, the basic serial triggers and protocols (UART, I2C, SPI, CAN LIN) are free on all Siglent models. Only the more specialized protocols are paid options.

Bode Plotter is a free option as well - you don't even need the AWG license to use the internal AWG for that.

In general, it's questionable what value such a table is. You look for a tool for a certain tasks - and the only thing that really counts is getting one that helps you accomplishing that task best and most effectively.
For example, you dont need a super high waveform update rate at all, but a powerful FFT can help a lot. You didn't even include the FFT in your table. Max. length? Speed? Tools? Yeah, most datasheets won't tell you that in detail anyway.

So if it just were to decide between the two Siglents, you can always look up the numerous demonstrations for both scopes in this forum to get an idea how powerful they are - also with their class-leading 2 Mpts FFT.

You can also listen to the ones who have evaluated the top HR model from Rigol (DHO4000) and tell you that its FFT is very basic and not impressive. And nobody has ever checked the accuracy of the measurements.
Or you listen to those who have never touched anything above a DS1054Z but still tell you how good the DHO800/900/1000 are. Maybe that's even true. But since you get only claims instead of hard facts and demonstrations, it remains a gamble.

[uargo]

Taking into account that the Siglent SDS1104X-E has a background noise of approx 300uVrms

Where do you get that number from?

My Siglent SDS1104X-E has 56 µVrms noise for full bandwidth and 27.3 µVrms when 20 MHz bandwidth limiter is active (measured at 500 µV/div). Regarding noise, it is in a similar ballpark as the SDS2000X Plus, yet suffers more and stronger spurious signals.

I HAVE MADE A TABLE WITH THE MAIN CHARACTERISTICS OF THE CONTENTS

To the best of my knowledge, the basic serial triggers and protocols (UART, I2C, SPI, CAN LIN) are free on all Siglent models. Only the more specialized protocols are paid options.

Bode Plotter is a free option as well - you don't even need the AWG license to use the internal AWG for that.

In general, it's questionable what value such a table is. You look for a tool for a certain tasks - and the only thing that really counts is getting one that helps you accomplishing that task best and most effectively.
For example, you dont need a super high waveform update rate at all, but a powerful FFT can help a lot. You didn't even include the FFT in your table. Max. length? Speed? Tools? Yeah, most datasheets won't tell you that in detail anyway.

So if it just were to decide between the two Siglents, you can always look up the numerous demonstrations for both scopes in this forum to get an idea how powerful they are - also with their class-leading 2 Mpts FFT.

You can also listen to the ones who have evaluated the top HR model from Rigol (DHO4000) and tell you that its FFT is very basic and not impressive. And nobody has ever checked the accuracy of the measurements.
Or you listen to those who have never touched anything above a DS1054Z but still tell you how good the DHO800/900/1000 are. Maybe that's even true. But since you get only claims instead of hard facts and demonstrations, it remains a gamble.

I have put the FFT in the table, look well, I have been researching on the internet to obtain the background noise of the SDS1104X-E and the only thing I have found is that it is:

Table 8 Noise Floor Limited Range for 2-Channel
Volt/Div        Stdev Limited Range
500 uv         ≤ 250 uv

link:https://int.siglent.com/upload_file/user/SDS1000X-E/SDS1000X-E_ServiceManual_SM0101E-E01A.pdf

and where do you get that it has 56uVrms ( better even than the $4,600 12-bit 350Mhz SDS2354X HD which has 70uVrms)

It is clear that my oscilloscope and yours, even though they are the same model, have different characteristics.

[Martin72]

better even than the $4,600 12-bit 350Mhz SDS2354X HD which has 70uVrms

65µVrms, look at the datasheet.
And: Noise rises more or less with the bandwith.
Plus: The difference between 56µVrms and 65µVrms is....nothing.
Plus, the second: You already have a SDS1104X-E, so you can measure it in real and now.
At the moment I let the SDS1104X-E warm up for it...

[rf-loop]

Here is one quite old image. At this time I own one Rigol 1074Z (mod to 1104Z)
In upper part of image (Rigol) there is also note about data used to measurements, today we know more also about this and it really use very extremely tiny intermediate buffer for measurements when Siglent use full acquisition data.

And then bottom part of image Siglent SDS1104X-E. Settings as same as possible what was used with Rigol for fair compare.
1mV/dif used because Rigol did not have 500uV/div. (also 1mV/div in Rigol is not true full resolution, it is some kind of production from 5mV/div)
Around 60uV AC RMS (Stdev) is quite good for 100+ MHz bandwidth with 1Mohm inputs.
Typically it is much better than service manual "go to repair" limit.

This is many years old..
Also I have measured several SDS1104X-E and I really can say that common value at 500uV/div is between 50 - 65uV/div  Stdev (AC RMS)

[Performa01]

I have put the FFT in the table, look well,

Well, yes, I've overlooked it in this messy table. FFT is normally characterized in points, not bytes and there are still some other parameters to consider.

Btw, your table contains a Siglent SDS1104X HD - a scope which doesn't exist outside China. Currently, nobody will be able to tell you what exactly we (the rest of the world) will be getting. However, I can assure you that it will be worth waiting for - if you really need a high resolution oscilloscope and don't mind a clumsy solution for the digital channels, like on the SDS1104X-E.

I have been researching on the internet to obtain the background noise of the SDS1104X-E and the only thing I have found is that it is:

Table 8 Noise Floor Limited Range for 2-Channel
Volt/Div        Stdev Limited Range
500 uv         ≤ 250 uv

link:https://int.siglent.com/upload_file/user/SDS1000X-E/SDS1000X-E_ServiceManual_SM0101E-E01A.pdf

This is a generic tolerance window for all SDS1000X-E series scopes, including the 200 MHz models, which have an actual bandwidth of ~240 MHz (SDS1104X-E: 110 MHz). The margins are wide; even though the SDS1000X-E are low noise like all contemporary Siglent scopes, Siglent quite obviously isn't willing to guarantee this for their bottom entry level scopes.

and where do you get that it has 56uVrms ( better even than the $4,600 12-bit 350Mhz SDS2354X HD which has 70uVrms)

Quite simple. I've measured it.

It is clear that my oscilloscope and yours, even though they are the same model, have different characteristics.

No. In theory, there could be slight differences, but since it's an easy task, many people have measured the noise of their SDS1000X-E and there never has been any significant discrepancy.
Hint: a 200 MHz device should show about 50% more noise, i.e. about 84 µVrms.

[Martin72]

At the moment I let the SDS1104X-E warm up for it...

Did it now, "yesterday" it was too late...
@uargo:
Important is to make a selfcal before, pic 1 shows the traces after switching the scope on.
Then warmup for appx 30min, then let the scope performing the selfcal, pic 2 show the result after.
Took channel 2, 500µV/div., 1ms/Div., no shielding at all, no averaging.

[Performa01]

At the moment I let the SDS1104X-E warm up for it...

Did it now, "yesterday" it was too late...

A few remarks:

• Even more important than the total bandwidth is the lower bandwidth limit, due to the 1/f noise characteristics. At 1 ms/div the lower bandwidt limit is ~70 Hz on an SDS1000X-E. This will of course yield significantly worse results than a measurement at e.g. 1 µs/div and ~70 kHz lower bandwidth limit. We can also take the current record length as a criterion (and 1.4 Mpts is a massive amount of data), but specifying the lower bandwidth limit is a much more robust statement.
• It has been mentioned over and over again: RMS is the wrong measurement for noise. If we use Stdev instead (on some DSOs it's called AC-RMS), then we need not worry about DC offsets and neither have to wait half an hour nor do a self-cal just because we want to measure noise. Apart from that, even with a fresh self-cal there will still be some remaining offset error which will be erroneously interpreted as noise when using the RMS measurement.
• Many scopes do not have 500 µV/div sensitivity (some even have only 4 or 5 mV/div and use digital zoom to fake higher sensitivity). So I recommend to use 1 mV/div throughout. It's a more fair comparison because of a near identical dynamic range for all contenders and it is visually better comparable as well.
• Your measurement appears to be in the right ballpark (61 µVrms) for full bandwidth, which could easily drop to 56 µVrms once you use the right measurement (Stdev).
• Your measurement with 20 MHz bandwidth limiter appears much too high (55 µVrms), where I have no other explanation than that is caused by an DC offset error. It would be nice to see a Stdev measurement to confirm my theory.
• Finally I want to repeat what rf-loop already mentioned: The Siglent always measures the entire record, hence 1.4 Mpts is this example. "Measurements" that  only use decimated (screen) data will just provide street numbers, because decimation alters the bandwidth.

[Martin72]

A few remarks:

 

Can repeat it with the wanted parameter.

[uargo]

better even than the $4,600 12-bit 350Mhz SDS2354X HD which has 70uVrms

65µVrms, look at the datasheet.
And: Noise rises more or less with the bandwith.
Plus: The difference between 56µVrms and 65µVrms is....nothing.
Plus, the second: You already have a SDS1104X-E, so you can measure it in real and now.
At the moment I let the SDS1104X-E warm up for it...

I put a link to the siglent document that supports what I say, can you do the same, please

[Martin72]

LOL...OK, look at the pic below.

Can repeat it with the wanted parameter.

All channels, history mode, 20Mhz bw limit, picking the lowest value.

(damn...500µV/div. ....)

[rf-loop]

Interesting.
How can I make sure that not only the screendata will be used for the values ? The history record ?

How you can be sure. Very Simply. Just believe what we tell about it. (but do not believe everyone --- you know....)

But if you do not trust, you can also easy evaluate it. 
 

You can somehow test if it use decimation for display is using example slow timebases and narrow  pulses. If it decimate before display there start missing some these pulses. Not happen with Siglent. But one vertical column can have many of these pulses overlaid on screen (example if burst include many pulses but whole burst is only one display pixel width. (all pulses overlayed in this column)).  But still measurement can see separate pulses. 


Some oscilloscope measure using screen mapped data and then  there is max resolution same as screen resolution.
Some oscilloscopes may use this screen (mapped) data for measure.
Some scopes may use some other kind of intermediate buffer for measurements. More resolution than display resolution but less than is in sample buffer (depending time base etc).
There is also some older Siglent models what use intermediate buffer for measurements (example SDS1000X models because there is not brute force for handle full resolution full data length measurements)

This SDS1104/1204X-E use full sample buffer for measurements (just this memory length what display info screen tell about current memory length).

I think if you do some simple tests it tell lot of about this.

Idea is just that input signal have more time domain events than  than display resolution is. You can easy find many kind of signals for test it.
Here one example...  no need just this... 
Simplest is 10MHz sine and scope 1ms/div and then measurements about +width, interval etc etc... scopes what use display buffer fails here totally.

Set 1ms/div.
Produce signal what have 50ns width pulses and pulses period 1.001us (1001ns) and risetime example 15ns
Set measure for Period  and +Width and risetime.

If oscilloscope use screen buffer for measurement... All these measurements fails totally.

If horizontal resolution is example 700 pixel and 14 div (as is in SDS1000X-E models)
If 1ms/div, then 1 div pixel resolution is 20us. (time axis 50 pixel/div)   

But in this example measurement display pulse width around 50ns and period around 1001ns and also risetime based to 1ns resolution (because sampling interval is 1ns and because it use full acquistion resolution.

Ok do it use also really full length of data. Yes. You can easy test it... make signal so that nearly start of sample buffer is one pulse and nearly end of buffer next pulse (example 13ms interval, 50ns pulses. Set trigger half div after display left border. Yes you get 13us interval so lenght used for measurement is least this , you can see next pulse half div before right border and as you can see resolution for interval is 1ns.

Try with Rigol 1000Z. Also you can try with some Megazoom Keysight.  (no need test, I can tell situation is just "game over")

But one note. Example in this case. There is not Interpolation between true sample points in this case. So resolution is now sample interval. 

Long time ago here in forum was some very long thread about this measurement resolution with different oscilloscopes and also something about testing it. In this thread there is lot of data and information.
Fast searching I didn't find it.

[Martin72]

Very interesting things, I need to read this in peace and find out more.
It's like a rabbit hole...

Actually, I wanted to repeat the measurements with 1mV/div. but I don't feel like it today, especially since the offset from the scope keeps "running away" from me.
At some point, that's no longer any fun.
But it's not mine, my HD is very easy to care for.

[uargo]

LOL...OK, look at the pic below.

Can repeat it with the wanted parameter.

All channels, history mode, 20Mhz bw limit, picking the lowest value.

(damn...500µV/div. ....)

the capture that you have put is at 50 ohm, of the SDS2104X HD model, THAT DATA OF THE SDS1104X-E does not exist published by Siglent. If you say that empirically it is that, then it will be, measuring the noise gives me unstable readings and strange things at those levels

[Martin72]

Hm ?

( better even than the $4,600 12-bit 350Mhz SDS2354X HD which has 70uVrms)

That´s what you´ve wrote, I don´t know from which datasheet you got this information because it´s 65µVrms.
You asked me where I have this information, I´ve posted a snip from the datasheet.

the capture that you have put is at 50 ohm, of the SDS2104X HD model,

No Sir, the SDS2104X HD have 50µVrms, look again..

[2N3055]

better even than the $4,600 12-bit 350Mhz SDS2354X HD which has 70uVrms

65µVrms, look at the datasheet.
And: Noise rises more or less with the bandwith.
Plus: The difference between 56µVrms and 65µVrms is....nothing.
Plus, the second: You already have a SDS1104X-E, so you can measure it in real and now.
At the moment I let the SDS1104X-E warm up for it...

I put a link to the siglent document that supports what I say, can you do the same, please

250 µV stdev is absolute limit (above which it is broken) for a 2CH model.

For 4ch model, at 1mV/div spec is "better than 25,55 dB SNR". There is formula below how it is calculated..
Which would be arround 150 µV stdev MAX. For it to be considered damaged....
Real noise will be half or better than that.

Martin posted real measurements from the scope...

[Performa01]

The noise of a DSO has been debated at all times and one of the earliest occasions where the Siglent SDS1104X-E has been looked at has been back in February 2018, when our fellow "Master of Noise", @maxwell3e10, had a look at the cheapest 12 bit scope, an Owon XDS3062A:

https://www.eevblog.com/forum/testgear/owon-xds3062a-input-noise-(good)-and-glitches-(bad)/msg1429292/#msg1429292

He posted a table showing the noise levels of all the scopes he had tested at that time and unsurprisingly the 12bit Owon won - particularly at lower sensitivities, where the granular ADC noise is the main contribution to the total noise.

You can see Tektronix being the king of noise, but also that a 12 bit ADC can be not much more than a marketing gag, if it is not properly calibrated and has high nonlinearities (take a close look at the 2nd screenshot in that post).

My measurements back then, including noise spectra:

https://www.eevblog.com/forum/testgear/owon-xds3062a-input-noise-(good)-and-glitches-(bad)/msg1432960/#msg1432960

Here we got an updated table (shortened version, quite obviously, nobody wanted to see the noisy Tek anymore):

https://www.eevblog.com/forum/testgear/owon-xds3062a-input-noise-(good)-and-glitches-(bad)/msg1433124/#msg1433124

And finally the spectral noise density for full BW as well as 20 MHz bandwidth:

https://www.eevblog.com/forum/testgear/owon-xds3062a-input-noise-(good)-and-glitches-(bad)/msg1433850/#msg1433850

EDIT: Oh - and have you noticed how nobody wanted to share data of the Rigol 1054Z, even though it was the most popular scope at that time? I guess I know the reason why...

[MarioBros69]

From someone with very little knowledge on these issues..., is the Siglent SDS1104X-E better than the Siglent SDS2202X-E 200MHz?
There is very little difference in price, but which one is more worth it?

[tautech]

From someone with very little knowledge on these issues..., is the Siglent SDS1104X-E better than the Siglent SDS2202X-E 200MHz?
There is very little difference in price, but which one is more worth it?

If you need the BW SDS2202X-E everyday as it is a 350 MHz design.
Very similar feature set to SDS1104X-E however the 4ch model offers more total capability if BW is not a concern.

[MarioBros69]

From someone with very little knowledge on these issues..., is the Siglent SDS1104X-E better than the Siglent SDS2202X-E 200MHz?
There is very little difference in price, but which one is more worth it?

If you need the BW SDS2202X-E everyday as it is a 350 MHz design.
Very similar feature set to SDS1104X-E however the 4ch model offers more total capability if BW is not a concern.

And about what is discussed in this thread, about noise, are the two the same?

Are both models hackable?

[MarioBros69]

Here is one quite old image. At this time I own one Rigol 1074Z (mod to 1104Z)
In upper part of image (Rigol) there is also note about data used to measurements, today we know more also about this and it really use very extremely tiny intermediate buffer for measurements when Siglent use full acquisition data.

And then bottom part of image Siglent SDS1104X-E. Settings as same as possible what was used with Rigol for fair compare.
1mV/dif used because Rigol did not have 500uV/div. (also 1mV/div in Rigol is not true full resolution, it is some kind of production from 5mV/div)
Around 60uV AC RMS (Stdev) is quite good for 100+ MHz bandwidth with 1Mohm inputs.
Typically it is much better than service manual "go to repair" limit.

This is many years old..
Also I have measured several SDS1104X-E and I really can say that common value at 500uV/div is between 50 - 65uV/div  Stdev (AC RMS)

Because the rigol shows the signal much thicker if the two are at 1mV/div?

Edit: Sorry..., I thought the same signal was being measured in both