Unboxing
Thank you for this first impression.
At last a rigol with low noise input...
Optically I don´t like it, but the rest seems to be nice.
And the price is very hot, no question.
They must have changed it on the HD model then, because that wasn't the case for previous Siglents I've looked at
The 6000A and the 2000HD got it from the beginning, on the SDS5000 it´s "new" with firmware update.
Why can't they just copycat the UI from R&S or Keysight, I don't know..
It seems they greatly value the somewhat videogame-ish appearance.
When you´re growing up with a (android) smartphone in your hand, this could be the result when you have to design the UI.
Why can't they just copycat the UI from R&S or Keysight, I don't know..
Some people dont like those UI's! Scopes are products that still have not settled on the "right" way to do a touchscreen.
Rigol have confirmed that you cannot turn off the Sin x/x interpolation.
Hello,
Maybe we can make a noise comparison between Rigol and Siglent 12 bit scope. For this purpose, 8 files (binary) per device would be useful.
Each with 50 Ohm and 1 MOhm
1 mV/div with 200MHz and 20 MHz with the largest possible memory and highest sampling rate.
And
1 V/div with 200MHz and 20 MHz with the largest possible memory and highest sampling rate.
The files would probably be quite large. But there are file hosters with free access and hopefully enough storage space.
Best regards
egonotto
5:50 in the video I totally missed that bias is actually different to position offset. You can actually DC bias the input to any level required which is pretty cool.
Maybe we can make a noise comparison between Rigol and Siglent 12 bit scope. For this purpose, 8 files (binary) per device would be useful.
Each with 50 Ohm and 1 MOhm
1 mV/div with 200MHz and 20 MHz with the largest possible memory and highest sampling rate.
And
1 V/div with 200MHz and 20 MHz with the largest possible memory and highest sampling rate.
The files would probably be quite large. But there are file hosters with free access and hopefully enough storage space.
Getting those now.
Maybe 1M memory is better as a standard?
Thanks! Can you confirm the LSB scaling? For RigolDS1-50ohm1mV20M.bin I am getting STD=19.8uV assuming full scale of 10 mV in 2^16 bits.
Also need to know sampling rate to calculate PSD.
....
Getting those now.
Maybe 1M memory is better as a standard?
With max memory an max sample rate than with Rigol one get 62.5 ms, With Siglent one get 100 ms.
This allows you to observe the noise over a large frequency range.
And you can try hires algorithms.
Best regards
egonotto
Thanks! Can you confirm the LSB scaling? For RigolDS1-50ohm1mV20M.bin I am getting STD=19.8uV assuming full scale of 10 mV in 2^16 bits.
Also need to know sampling rate to calculate PSD.
4GS/s for all files.
For actual LSB I don't know if there is a way to actually get this exact figure?
As I showed in my video there is some extra dynamic range outside of the displayed window. So it's at least 8 divisions + whatever extra there is. Maybe 10V full scale?
EDIT: Hang on, I think I know how to get this.
For 1 M Sample files it can't be 4GS/sec and 1 msec/div.
Actually checking your video, it seems sometime the horizontal scale is 10 divisions and sometimes it is 20 divisions.
But the spectra are looking nice... Just want to confirm scaling.
Thanks! Can you confirm the LSB scaling? For RigolDS1-50ohm1mV20M.bin I am getting STD=19.8uV assuming full scale of 10 mV in 2^16 bits.
Also need to know sampling rate to calculate PSD.
4GS/s for all files.
For actual LSB I don't know if there is a way to actually get this exact figure?
As I showed in my video there is some extra dynamic range outside of the displayed window. So it's at least 8 divisions + whatever extra there is. Maybe 10V full scale?
EDIT: Hang on, I think I know how to get this.
I have confirmed on the 100mV range that the dynamic signal acquisition range is about 960mVpp (assuming my Keysight 3000 sig gen is accurate.
I did this by feeding in a triange wave, capturing, and seeing where it clips. 970mVpp definitely clipped.
I assume the same on the 1V and 1mV ranges, so call it 9.60Vpp/4096 = 2.34mV/bit
For 1 M Sample files it can't be 4GS/sec and 1 msec/div.
Oops, it's 50MS/s on the 1M memory captures. 4GS/s on the 250M samples.
I have confirmed on the 100mV range that the dynamic signal acquisition range is about 960mVpp (assuming my Keysight 3000 sig gen is accurate.
I did this by feeding in a triange wave, capturing, and seeing where it clips. 970mVpp definitely clipped.
I assume the same on the 1V and 1mV ranges, so call it 9.60Vpp/4096 = 2.34mV/bit
It could be clipping before reaching full ADC range. Also the binary numbers are encoded in 16 bit unsigned integer. There should be some meta file saved along with binary data that would specify it. But 10 divisions full scale is probably fairly accurate.
I have confirmed on the 100mV range that the dynamic signal acquisition range is about 960mVpp (assuming my Keysight 3000 sig gen is accurate.
I did this by feeding in a triange wave, capturing, and seeing where it clips. 970mVpp definitely clipped.
I assume the same on the 1V and 1mV ranges, so call it 9.60Vpp/4096 = 2.34mV/bit
It could be clipping before reaching full ADC range. Also the binary numbers are encoded in 16 bit unsigned integer. There should be some meta file saved along with binary data that would specify it. But 10 divisions full scale is probably fairly accurate.
There is no extra file saved.
Yes it could be clipping before the ADC, but there is no easy way to know this?
I confirmed the 200mV range is the same FS ratio.
I have confirmed on the 100mV range that the dynamic signal acquisition range is about 960mVpp (assuming my Keysight 3000 sig gen is accurate.
I did this by feeding in a triange wave, capturing, and seeing where it clips. 970mVpp definitely clipped.
I assume the same on the 1V and 1mV ranges, so call it 9.60Vpp/4096 = 2.34mV/bit
It could be clipping before reaching full ADC range. Also the binary numbers are encoded in 16 bit unsigned integer. There should be some meta file saved along with binary data that would specify it. But 10 divisions full scale is probably fairly accurate.
The 1V range is different, 8.75Vpp. At 8.8V I can see the clipping.
Here is the noise spectrum using 250Msample files and assuming 10 div full scale. Looks very clean up to 200 MHz!
On 1 V/div (10V full scale) the white noise level is 240 nV/Hz^(1/2),
I think that corresponds to almost 10 ENOB. If I actually use proper equations instead of just "thinking", it comes out to 8.3 ENOB.
Hello maxwell3e10,
How do you make your graphics?
Best regards
egonotto
The noise looks very low.
1mv, 50ohm, 14bit mode.
1M poit fft, 0 ~ 2Mhz span, 100Msa/s.
If you show me the fft noise screen, I appreciate it.
There's no fft average view, right?
Quick video on getting the dynamic range:
It looks like it has a different amplifier path for 50 Ohm, giving lower noise at low frequency.
The ffts and plots are made with Matlab.
EDIT:
Actually, to be sure, was the scope input grounded externally when on 1 MOhm setting, Dave? Because it gives std=48 uV, instead of 19 uV. I think you mentioned in the video it does not make a difference when using on-screen rms measurements.
1M tests used an open input. If I used an external 50ohm termination then that's effectively the same as the 50 ohm mode, unless there is actually different paths for the 50ohm and 1M input which is usually
not the case for scopes.
So that graph you just posted in correct. I thought you wanted an open input?
Some scopes do have a different amplifier path, for example RTM3k and other higher-end scopes. One can tell if there is any difference in spectrum between 50 Ohm external short and internal 50 Ohm. It looks like there is not in this case, at least based on RMS measurements.
I am not sure how interesting is the spectrum with 1 MOhm open input. It might get better if one puts a shielding BNC cup (with no central pin) over the input.