I have asked the poster earlier , but I did not get the answer.
1. What is the sample rate?
2. Do you apply FFT for all 250Msamples or you use signal stacking tecniques
3. Do you apply any FFT window function -e.g. Hamming, Tukey, Blackman
The reason why I asking is that I want to make apple to apple comparison.
It suspected that Rigol and Siglent used at least different FFT size. Lecroy data shows even higher difference in fft parameters. It looks to me that lecroy data uses 7...10 lower fft size. Confirmation/Discofirmtion from poster wail help me to understand such differences.
Different class of instrument Dave, you are comparing apples and oranges.
SDS2000X Plus is a 500 MHz design vs 200 MHz, it also has 2x 2GSa/s, 200 Mpts x2 mem depth, inbuilt AWG, MSO, 10x probe sense.....need I go on ?
Yes, but MathWizard seems excited by the 12bit architecture, but it put off by the price. The HDO1000 series is going to get you 12bit capability from $699.
Plus a huge touch screen, modern UI, etc.
All of which will have far more impact on daily use than 10x probe sense and a little bit more memory.
Gentlemen, are you sure to want 200Mpts ?
One file is 2.2Gb
what is the practical benefit?
Maybe you remember the cardgame "Quartett" from your childhood?
Hello,
please yes. With the maximum sampling rate, the devices can not dizzy. In addition, I plan to try out different smoothing methods with the data.
Best regards
egonotto
Hi,
After recognizing that Bin. format was really small (appx 400M vs 2.3GB) in comparison (thx to maxwell), I´ll repeat all records from yesterday in the later afternoon and upload it.
Meanwhile something to play for:
https://drive.google.com/drive/folders/1hoIWdt6K9MwJI5_JtyPjtvrhNytx3rWA?usp=sharing1mV 50R and FULL bandwith, then 1V and full bandwith.
Settings:
2GSa/s, 200Mpt, Full BW (500Mhz), 10ms/div.
Why is there such a big difference between the two diagrams on the Rigol HD4000? I think the curve of the Rigol is very smooth in the second diagram.
I agree. Something looks off here.
Why is there such a big difference between the two diagrams on the Rigol HD4000? I think the curve of the Rigol is very smooth in the second diagram.
I agree. Something looks off here.
The "noise" on the noise spectral density is just a function of data length and chosen frequency resolution. The same file can be analyzed to get smoother plot at high frequency or "noisier" plot that goes to lower frequency. You can see that I combined two different settings in the original plot. For comparison to Siglent I used even coarser frequency resolution for Rigol to match the frequency range of Siglent while still using all the 250M of Rigol's data. For Lecroy the file I have is only 100k points, that is why it has even more fluctuations.
The general property of the noise spectral density is that the average level does not depend on the number of points, sampling rate (if no aliasing) or on the frequency resolution. This is not the case with FFT power spectrum since the noise power in each bin depends on the width of the bin. On the other hand when looking at discrete peaks, it is better to use power spectrum.
I have asked the poster earlier , but I did not get the answer.
1. What is the sample rate?
2. Do you apply FFT for all 250Msamples or you use signal stacking tecniques
3. Do you apply any FFT window function -e.g. Hamming, Tukey, Blackman
The sampling rate can be found as twice the maximum frequency for each curve, I did not cut off the edge of the plot.
I mentioned that I used pwelch that breaks data into overlapping segments, and then averages resulting FFTs
The default window in pwelch is Hamming. It does not matter what is used because there are no strong peaks, I also checked rectangular window and Kaiser window. I subtract the mean to eliminate the spectral leakage from DC offset.
Here is a graph with the same settings for the HDO and SDS. But I only used the first megabyte from the HDO data, my PC can't handle the whole 250 MSamples. Attached also my Matlab code.
Peter
Thanks to sharing. Great job. Explanation, together with source code make the picture complete
I have asked the poster earlier , but I did not get the answer.
1. What is the sample rate?
2. Do you apply FFT for all 250Msamples or you use signal stacking tecniques
3. Do you apply any FFT window function -e.g. Hamming, Tukey, Blackman
The sampling rate can be found as twice the maximum frequency for each curve, I did not cut off the edge of the plot.
I mentioned that I used pwelch that breaks data into overlapping segments, and then averages resulting FFTs
The default window in pwelch is Hamming. It does not matter what is used because there are no strong peaks, I also checked rectangular window and Kaiser window. I subtract the mean to eliminate the spectral leakage from DC offset.
Thanks for the fast answer.
I think that the best way is to provide the source code as PeDre did few posts earlier.
I still dont know what is sown in your graphs
- What is the data size /total number of processed samples/ used in every scope?
- What is FFT length used by pwelch method?
- Did you use overlapping in pwelch methog? If yes how much
- how did you make coarser the frequency resolution for Rigol ?
Here is a graph with the same settings for the HDO and SDS. But I only used the first megabyte from the HDO data, my PC can't handle the whole 250 MSamples. Attached also my Matlab code.
Peter
Edit: The labeling is wrong, the SDS data also has a bandwidth of 200 MHz.
Could you provide me with the matlab code used?
I might be able to run it on the full lenght data and send you the results.
I can read the data, but I am not sure what the vertical scaling is. If I read raw 16 bit numbers for the file 1V_50R_200Mhz.trc, their standard deviation is 6724 counts and their mean is -1515 counts. This should be scaled so it matches to the standard deviation and average reported by the scope. If possible, turn on full-waveform math.
If possible, turn on full-waveform math.
What do you mean with this ?
Some scopes calculate statistics based on reduced dataset to speed it up, can be anywhere from 1k to 1M points. I think a fancy scope would have an option to calculate statistics (like mean and std) on the whole datasets. But it's not a big deal, since std and mean should be roughly the same for a fraction of the dataset.
For me the filelength seems with appx 100Mb plausible, as the Memory is set to its maximum of 50Mpts.
But I can dive deeper in the menus of the HDO tomorrow..
The problem is not the file length, but vertical scaling. Without it there is no way to calculate the spectrum. One has to compare statistics on scope and in the file. Or you can record a real signal and measure its p-p value to compare with saved data. Or one can go through the documentation, it would specify the scaling or the format of the heater in the beginning of the binary file.
Hi,
LeCroy HDO6034A waveform files:
Settings: 10GSa/s, 50Mpts.
Hello,
thank you. There is a function for Matlab (to download) ReadLeCroyBinaryWaveform with the help of which you can read the files.
The files are 50000002 words long.
Best regards
egonotto
Well, this is a good example why RMS noise numbers don't tell the whole story. On the surface Lecroy HDO has 3 times less RMS noise for 1V/div scale and full bandwidth than Rigol (3mV vs 9 mV). But in reality, it is worse in many ways:
Much worse 1/f noise
More spurious noise peaks
Insufficiently steep 20 MHz low-pass filter on 1V/div scale
Brick wall 1 GHz filter - I can see why it would be there since these models go up to 1 GHz, but it means that claimed 10GS/sec speed is totally useless, just takes disk space.
So Rigol is looking pretty good compared to a nearly 10 times more expensive scope! They could still make it better if they implemented an additional sharp 200 MHz (or adjustable) digital filter.
Some scopes calculate statistics based on reduced dataset to speed it up, can be anywhere from 1k to 1M points. I think a fancy scope would have an option to calculate statistics (like mean and std) on the whole datasets. But it's not a big deal, since std and mean should be roughly the same for a fraction of the dataset.
Lecroys calculate on full data set by default.
I have contacted with batronix and the can´t sell the HDO1000 series. Rigol don´t allow them to do it because they sell siglent scopes too. A really dirty behavior from Rigol, a shame of company.
I have contacted with batronix and the can´t sell the HDO1000 series. Rigol don´t allow them to do it because they sell siglent scopes too. A really dirty behavior from Rigol, a shame of company.
By email? What exactly did they say? Can you post the email here?
I have contacted with batronix and the can´t sell the HDO1000 series. Rigol don´t allow them to do it because they sell siglent scopes too. A really dirty behavior from Rigol, a shame of company.
By email? What exactly did they say? Can you post the email here?
And the point you wanna make is ?
Manufacturers or resellers have every right to dictate who/what brands they sell.
I have contacted with batronix and the can´t sell the HDO1000 series. Rigol don´t allow them to do it because they sell siglent scopes too. A really dirty behavior from Rigol, a shame of company.
That is well within their rights but sounds weird.
What would they (Rigol) gain by it?
And the point you wanna make is ?
None. I just wanted to see the exact wording.
Batronix sells a lot of Rigol 'scopes, including the HDO4000.
https://www.batronix.com/shop/rigol/HDO4000.htmlIt's weird that they'd single out the HDO1000 for special treatment, especially as it's going to be a Siglent killer (assuming it performs anything like the HDO4000).
And the point you wanna make is ?
None. I just wanted to see the exact wording.
Batronix sells a lot of Rigol 'scopes, including the HDO4000.
https://www.batronix.com/shop/rigol/HDO4000.html
It's weird that they'd single out the HDO1000 for special treatment, especially as it's going to be a Siglent killer (assuming it performs anything like the HDO4000).
There’s 5 Rigol authorised resellers in Germany so if we are to believe ojete’s claim which other of those sellers that also sell other brands have also been forbid access to HDO1k ?
I contacted by email:
Hi,
thank you for your enquiry. We are distributor for Rigol AND Siglent. Therefore Rigol doesn`t allow us to sell the HDO series as long we are also working with Siglent. I am very sorry for no better news.
Phillip von Würzen
and my second email:
You are selling now the HDO4000 series. I think you mean the HDO1000 only. Is this right?
reply:
Hi,
correct. We are allowed to sell HDO4000 but not HDO1000.
Phillip von Würzen
I was very surprised of this, I'm interested in buying the HDO1000 but I don't like this kind of things from any company. For me this is a dirty trick and yes, Rigol can do whatever they want and choose the resellers they want and dictate them everything they want.