Author Topic: Rigol DS2072 Review  (Read 35370 times)

0 Members and 1 Guest are viewing this topic.

Offline JimmySte

  • Contributor
  • Posts: 13
  • Country: au
Rigol DS2072 Review
« on: August 09, 2012, 08:20:17 am »
Hi all...
Due to the total lack of reviews of the DS2000 series of scopes online I thought I would give reviewing a crack... :)

/*Background
I'm currently a student studying Electrical Engineering and approaching the pointy end of my course,
I thought it was high time to "graduate" from my father's 20MHz analogue scope and get my own...
I started out looking online at the Rigol DS1052, also discovering the DS2072 but unwilling to spend the dough...
Knowing that the DS1052 was very liked in reviews, on this forum and I'd used them at university, all be the 200MHz version.
So I was tossing up between the Rigol and the Atten ADS1062CAL... I liked the Atten because of it's larger screen, having experienced that menus covered almost half of the screen.
I eventually pulled the trigger Olympic pun... :) and got the Atten, got home, turned it on... and it had a faulty power supply :( not to dissimilar to the issue playfsx on the formums has/had with their scope...



Called up the shop and took it back, they connected it up to a Rigol DG4102 and the interference continued...
So they were quite apologetic and were willing to replace it with another Atten, until I asked what they thought about the DS2072 and if they were willing to let me play with one...
First impression WOW, not just for me but the shop owner, it was the first time either of us has seen one operating, the quality of the display make the Atten's look like dirt...
And as they say the rest is history...

/*Review - I think...
The Rigol DS2072 is a 2 channel DSO 2GSa/s clocking in at 70MHz, I got mine here in Australia at www.eyou.com.au for $839 + GST
They were a delight to buy from, as can be seen by the comments above...

It came double boxed with a English user manual, 2x 300MHz passive 1-10x probes, Australian power cable, USB cable and a CD
The first thing I noticed about the device was it's weight at 4.8kg, it's solid, double the Atten's 2.4kg.
Second was the quality of the screen at 8" 800x480 its massave and very clear to read.
The scope is a pleasure to use, the GUI is straight forward and with enough button presses the basics can easily be nutted out...
And for the more advanced aspects the user manual is quite comprehensive

And now for some images :)

An example of the low noise level of the scope


Connecting the channel 1 probe up to the 1kHz test point and pressing AUTO I got


Displaying all of the waveforms summaries


Some summaries at the bottom


Raising edge of the square wave


FFT of the Square wave


Other math operations available


The scope has an internal recording ability to take snap shots of a waveform and review them one at a time


I connectedthe TX from an Arduino to the scope and used it's ability to decode RS232


Although unfortunately these functions have a limited life and require more $$$ to keep them

And the time counts down weather you are using the function or not...

So that's about all... the DS2072 is a wonderful scope and money well spent :)
« Last Edit: August 09, 2012, 11:45:47 pm by JimmySte »
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3706
  • Country: sg
    • T4P
Re: Rigol DS2072 Review
« Reply #1 on: August 09, 2012, 02:25:05 pm »
Nice, i haven't really seen snapshots of internal operations before
but 1 thing is striking me,
What's this mythical "Attan" ADS1062CAL?
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #2 on: August 09, 2012, 04:04:05 pm »
Hi JimmySte

Seems that You have very nice scope there...
From my side any DS2000 series infos are more than welcome...

As for me 
If You could investigate bit more FFT functionality :D,
Please run again Your FFT test but with dBVrms and markers ON.
Not sure if You have Signal Generator to run some more tests but If You could run similar test as mine:  https://www.eevblog.com/forum/product-reviews-photos-and-discussion/fft-high-precision-mode-with-hameg-hmo-vs-rigol-ds2000-scopes/45/ we could hopefully have some comparison against DS1052E and bet that could be very interesting to many.

Rosendorfer


 

Offline prenato

  • Regular Contributor
  • *
  • Posts: 116
  • Country: us
    • paulorenato.com
Re: Rigol DS2072 Review
« Reply #3 on: August 09, 2012, 06:49:32 pm »
Hi Jimmy,

Thanks a lot for your review! As you know from my other post, I'm also considering buying this scope and your review is very helpful. It looks very good.
Is there any aspect that you think could be improved in the scope?

Paulo
Paulo
Professional Tinkerer,
www.paulorenato.com
 

Offline olsenn

  • Frequent Contributor
  • **
  • Posts: 993
Re: Rigol DS2072 Review
« Reply #4 on: August 09, 2012, 06:56:45 pm »
Seriously, I really don't understand why anyone would buy the DS2000 scopes? I'm not saying they're bad, but they're far more expenive than the DS1102E, which is almost as good, and for not much more than the DS2000-series, you can get a DS4000-series, which is FAR superior!

Am I missing something?
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #5 on: August 09, 2012, 07:27:21 pm »
Seriously, I really don't understand why anyone would buy the DS2000 scopes? I'm not saying they're bad, but they're far more expenive than the DS1102E, which is almost as good, and for not much more than the DS2000-series, you can get a DS4000-series, which is FAR superior!

Am I missing something?

Hi olsen

Well spoken......
But actually where from do You know DS2000 is almost as good or as some will tell as bad as DS1000 series, we are actually looking for any info on that. If You know something please share with us. 
Right now we do have nice looking marketing brochures full of features that makes DS1000 looks like toy in comparison and just few, early adopters reports.
And  regarding DS4000 being almost the same price... so where I can buy DS4000 for "not much more" than 900EUR or USD..?
So how much is actually "not much more".. 500-600-700-800 EUR or USD?

Rosendorfer
 

Offline prenato

  • Regular Contributor
  • *
  • Posts: 116
  • Country: us
    • paulorenato.com
Re: Rigol DS2072 Review
« Reply #6 on: August 09, 2012, 08:22:39 pm »
Seriously, I really don't understand why anyone would buy the DS2000 scopes? I'm not saying they're bad, but they're far more expenive than the DS1102E, which is almost as good, and for not much more than the DS2000-series, you can get a DS4000-series, which is FAR superior!

Am I missing something?

Hi Olssen,

The DS2000 series has a few significant advantages  over the DS1102E:
- 14M memory versus 1M
- 2GS/s versus 1Gs/s
- 500 uV/div instead of 2mV /div minimum vertical sensitivity
- Better refresh rate
- Much more advanced math functions and serial decode
- The build quality also seems  "a step above" going by the weight Jimmy quotes above:)

Whether this justifies 2x the price is another story, but there are some real advantages.

Paulo
Paulo
Professional Tinkerer,
www.paulorenato.com
 

Offline JimmySte

  • Contributor
  • Posts: 13
  • Country: au
Re: Rigol DS2072 Review
« Reply #7 on: August 10, 2012, 09:00:21 am »
Hi JimmySte

Seems that You have very nice scope there...
From my side any DS2000 series infos are more than welcome...

As for me 
If You could investigate bit more FFT functionality :D,
Please run again Your FFT test but with dBVrms and markers ON.
Not sure if You have Signal Generator to run some more tests but If You could run similar test as mine:  https://www.eevblog.com/forum/product-reviews-photos-and-discussion/fft-high-precision-mode-with-hameg-hmo-vs-rigol-ds2000-scopes/45/ we could hopefully have some comparison against DS1052E and bet that could be very interesting to many.

Rosendorfer

Hi Rosendorfer

Unfortunately I don't have a proper signal generator... could use my pc soundcard if I ground everything correctly

So I redid the FFT of the 1kHz square wave with the FFT set to dbV


And again with some coursers


And again with the FFT set to Hanning, as opposed to Rectangular


I hope this helps...
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 29883
  • Country: au
    • EEVblog
Re: Rigol DS2072 Review
« Reply #8 on: August 10, 2012, 10:18:21 am »
The DS2000 series has a few significant advantages  over the DS1102E:
- 14M memory versus 1M
- 2GS/s versus 1Gs/s
- 500 uV/div instead of 2mV /div minimum vertical sensitivity
- Better refresh rate
- Much more advanced math functions and serial decode
- The build quality also seems  "a step above" going by the weight Jimmy quotes above:)

I do believe it also has a high resolution mode like the Agilent et.al, and a corresponding lower noise floor.
And the huge 800x480 screen.

Dave.
 

Offline muvideo

  • Frequent Contributor
  • **
  • Posts: 393
  • Country: it
Re: Rigol DS2072 Review
« Reply #9 on: August 10, 2012, 10:54:15 am »
Hello, is it me or the probe seem undercompensated?
Fabio Eboli.
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #10 on: August 10, 2012, 12:35:28 pm »
Hi JimmySte

Thanks for the tests...
Having Sig Generator would be handy... Have a look at Siglent seems to be pretty decent for the money...
But If You could have a look at Your PC sound card output would be also very interesting, but as for me please do not play music, but just simple sin-wave... 8) and run FFT's...

And PLEASE !!! check program from dexter2048 : https://www.eevblog.com/forum/general-chat/using-rigol-ds1052e-for-spectrum-analysis/ . Big question if if it is working OK with DS2000 series files.
Well if You could check if it accepts wfs from famous but never used 12bit HR mode that would be also very interesting.

Rosendorfer
 

Offline JimmySte

  • Contributor
  • Posts: 13
  • Country: au
Re: Rigol DS2072 Review
« Reply #11 on: August 10, 2012, 03:02:36 pm »
Hi JimmySte

Thanks for the tests...
Having Sig Generator would be handy... Have a look at Siglent seems to be pretty decent for the money...
But If You could have a look at Your PC sound card output would be also very interesting, but as for me please do not play music, but just simple sin-wave... 8) and run FFT's...

And PLEASE !!! check program from dexter2048 : https://www.eevblog.com/forum/general-chat/using-rigol-ds1052e-for-spectrum-analysis/ . Big question if if it is working OK with DS2000 series files.
Well if You could check if it accepts wfs from famous but never used 12bit HR mode that would be also very interesting.

Rosendorfer

Hi Rosendorfer
I downloaded WFM Viewer and tried to open a saved waveform off my scope, unfortunately I got the following error message

The trigger is set to Auto and Edge so I don't know what the matter is...

And I don't understand what you mean by "12bit HR mode"??
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #12 on: August 11, 2012, 11:29:23 pm »
Hi JimmySte

Thanks for trying... as regarding wfm Viewer error seems that wfm format from DS2000 is different than DS1000. Maybe You could contact  dexter2048 regarding that and supply him few DS2000 wfm files.

Look at http://www.batronix.com/pdf/Rigol/Datasheet/DS2000_DataSheet_EN.pdf
At page 6 from DS2000 datsheet You will find remark of 12 bit HR mode.
I'm trying to get any info on that HR mode for some time as there is no much info on it in regular scopes reviews. My small theory is that most of 12 bit mode should be seen at FFT screen. If You could run few tests about it would be great.

Rosendorfer
 

Offline JimmySte

  • Contributor
  • Posts: 13
  • Country: au
Re: Rigol DS2072 Review
« Reply #13 on: August 12, 2012, 05:54:36 am »
Hi Rosendorfer

I got myself a signal generator and ran the 10kHz triangular FFT test


As for the High Res mode I found it...
Normal

HR
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #14 on: August 12, 2012, 09:44:03 pm »
Hi JimmySte

Interesting pictures....
And if You still have that generator PLEASE show FFT !!! from 12 HR mode.
There is rather little chances to see big differences between normal 8 bit and 12bit HR mode on 300pixels screen.
12 bit would require about 4000pixels screen resolution, maybe one day but not now.
So as I told there is my little theory that 12bit HR mode should bring improvement at noise floor level and dynamic (SFDR) of FFT measurements.
So if You could run simple sinewawe, that should give us more clear picture, at normal and HR mode FFT !!.
Interested to see if there are any improvements there.
Just as remark, as far as I know and tested myself there is no special magic in triangle being shown on the scope, seems that just nctnico for some reason have problems to do that on his Tek 7 series.
Anyway .......
FFT !!!!!!!! of simple sinewawe at normal and 12 bit HR mode, please!!  ;D

Rosendorfer
 

Offline JimmySte

  • Contributor
  • Posts: 13
  • Country: au
Re: Rigol DS2072 Review
« Reply #15 on: August 15, 2012, 11:30:32 am »
Hi Rosendorfer

The HR mode appears to have no effect on the FFT...
Normal


HR

 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #16 on: August 15, 2012, 11:08:09 pm »
Hi JimmySte

Well.... seems that HR mode is just marketing bs....

Bit unexpected...but not going argue with reality...
Thanks for doing the tests ... and have fun with new scope...still  "dream machine"...

Rosendorfer
 

Offline Gunb

  • Regular Contributor
  • *
  • Posts: 221
  • Country: de
Re: Rigol DS2072 Review
« Reply #17 on: August 17, 2012, 12:16:04 pm »
Hi there,

Hi JimmySte

Well.... seems that HR mode is just marketing bs....

Bit unexpected...but not going argue with reality...
Thanks for doing the tests ... and have fun with new scope...still  "dream machine"...

Rosendorfer

No, that's wrong.

It's essential to understand what the HiRes-mode of DSOs means in detail and how it works.

Switching on the HiRes-mode means that the scope works with full sample rate, even if the displayed sample rate is lower.

A signal sampled with a lower sample rate than the scope offers means, that between two samples information gets lost. Switching on the HiRes-mode causes the scope to sample
with maximum sample rate always, even if the display shows a much lower sample rate --> higher horizontal accuracy!

Let's assume you sample with 5MSa/s but your scope has got a maximum sample rate of 2GSa/s. With HiRes activated the scope is sampling with 2GSa/s although it shows 5 MSa/s only. In this case 4 samples will be taken by the ADC per displayed sample. In detail the 4 samples will be summed up and divided by 4 and the result appears as a sample on screen.

Why that?

You get more information with lower sample rates because every displayed sample consists of the average of 4 real samples!

Where's the rub?

#1: It always depends on the ratio of the maximum sample rate of the scope and the chosen one. In case they are equal, i.e., the scope's display shows 2GSa/s and the maximum sample rate is also 2GSa/s you will not see any difference between normal mode and HiRes-mode.

#2: You get an averaged signal (if maximum sample != selected sample rate)! Let's say you search for an outlier, than it might get lost because the neighboured samples are much smaller and the average calculates the outlier away! In this case you would never use the HiRes-mode.

When I interprete the results above then it's exactly what I've expected. In both cases the signal is sampled with the maximum sample rate of the scope - what additional information should be there in this case? None!

Concluding: the limitations of an ADC can't be improved by software. Scopes with HiRes-mode offer the capability to take more samples into account of the displayed signal at lower sample rates. If the selected sample rate is equal to the maximum, there will be no difference.

For real accuracy an ADC with more resolution is mandatory. The HiRes-mode of scopes is only a mathematical approach using filter algorithms to use the maximum sample and take the maximum no. of samples into account than throwing them away at lower sample rates


Note: horizontal accuracy caused by the no. of ADC-bits != horizontal accuracy caused by filter algorithms

And: ERES != HiRes-mode. That's another way of calculating more accuracy.

Kind regards
Gunb


« Last Edit: August 17, 2012, 12:29:12 pm by Gunb »
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #18 on: August 17, 2012, 06:22:34 pm »
Hi Gunb

Thank You for sharing this informations.
There is no any real info on that, and lots of things I could find let me thing that HR mode is kind of multisampling  where ADC is exchanging time, as multiply measurements just takes longer for more data/"resolution" and lower noise floor. And that all leads to more SFDR that could be used at FFT scope measurements.

At Rigol DS2000 datasheets HR is described:

>>>>>>>>>>>>>>>>>>>>>>>>>

High Resolution

12 bit of resolution When ?5 ?s/div @ 1 GSa/s
(or ?10 ?s/div @ 500 MSa/s).

>>>>>>>>>>>>>>>>>>>>>>>>>

That what Rigol and Hameg puts in their dataset do not sounds like just switching to 2GSa/s constant.


Switching on the HiRes-mode means that the scope works with full sample rate, even if the displayed sample rate is lower.

A signal sampled with a lower sample rate than the scope offers means, that between two samples information gets lost. Switching on the HiRes-mode causes the scope to sample
with maximum sample rate always, even if the display shows a much lower sample rate --> higher horizontal accuracy!


Not that I would like to argue, far from that, but just to understand and learn.

Assume that we are using 10us/div timebase at HR mode so as You are saying @2GSa/s constant:
Lets do small calculation:
@2GSa/s ADAC is feeding the scope with 2 measurements per 1 ns or if my calculations are right  20 000(!!!) measurements per 10us or we get 160 000-200 000 (!!!) "horizontal data" to be squeezed at typical scope screen.
Actually DS2000 or DS4000 screens are like 800x480 so about 200 000 points of "horizontal data" is being squeezed at about 800pixels ....
Well... that just can be, we may have: "higher horizontal accuracy!" very true...
But there is question what for ..?  Do we need more "horizontal resolution"?
Actually proper question would be more than what..? And as far as I know even with normal mode ADAC is easily feeding scope with much more than needed "horizontal data" many times more than scope screen horizontal pixel count, specially at longer timebase, that is where HR can be used... bit counter-intuitive, one may say....

But as opposite we  possibly may wish more than typical  8 bit "vertical resolution"...
 With 8 bit ADAC usually 1 bit is lost in ADAC noise floor and we are left with 7 usable bits and it is just about 128 levels... Can we use/see more ..???
At 800x480 Rigol screen, that actually could be very handy. Instead interpolating about 127-180 pixels to 480 vertical screen pixels we cold use straight away without any tricks almost full 9 bits of real resolution, so we could use 10 bit ADAC.

Well actually not sure what to thing about ...
Not quite sold to HR being "more horizontal resolution" or 2GSa/s "constant on" thing or I just did not get it right...

Kind regards
Rosendorfer








 

Offline Gunb

  • Regular Contributor
  • *
  • Posts: 221
  • Country: de
Re: Rigol DS2072 Review
« Reply #19 on: August 17, 2012, 10:51:21 pm »
Hi Gunb

Thank You for sharing this informations.
There is no any real info on that, and lots of things I could find let me thing that HR mode is kind of multisampling  where ADC is exchanging time, as multiply measurements just takes longer for more data/"resolution" and lower noise floor. And that all leads to more SFDR that could be used at FFT scope measurements.


Yes, that seems to be right. But additionally to say the vertical resolution becomes mathematically higher - if possible at a certain timebase - because more real samples are used for the average =  the displayed sample.



At Rigol DS2000 datasheets HR is described:

>>>>>>>>>>>>>>>>>>>>>>>>>

High Resolution

12 bit of resolution When ?5 ?s/div @ 1 GSa/s
(or ?10 ?s/div @ 500 MSa/s).

>>>>>>>>>>>>>>>>>>>>>>>>>

That what Rigol and Hameg puts in their dataset do not sounds like just switching to 2GSa/s constant.


But that's the way how "Boxcar Averaging" works and that's what Hameg has mentioned in their manual. By the way it's easy to implement. Another proof is that they speak about "up to" 10Bits resolution. At a certain timebase you can see that the displayed sample rate goes down to a quarter of the highest sample rate when HiRes is pressed and the signal becomes "slower". At highest sample rate you don't need HiRes since you don't see a difference.



Assume that we are using 10us/div timebase at HR mode so as You are saying @2GSa/s constant:
Lets do small calculation:
@2GSa/s ADAC is feeding the scope with 2 measurements per 1 ns or if my calculations are right  20 000(!!!) measurements per 10us or we get 160 000-200 000 (!!!) "horizontal data" to be


OK.


squeezed at typical scope screen.
Actually DS2000 or DS4000 screens are like 800x480 so about 200 000 points of "horizontal data" is being squeezed at about 800pixels ....


No, there's nothing squeezed into 800pixels. You're mistake is that you do no not distinguish between the samples in the sample memory and the displayed samples that reside in display memory!
You're calculation is right but additionally you have to calculate 200.000 / 800 = 250. Not every sample in memory can be displayed because of the 800 pixels, only each 250th point is shown on screen at the timebase you've chosen for your example. What you will see is an interpolated sequence of every 250th sample, if you want to see more details you have to zoom in, as long as you see the real samples without interpolation. Disadvantage is that you can only see a small window of the complete signal, but without interpolation you'll see the real samples at small timebases.

That's a general restriction of digital scopes opposed to CRTs with their "infinite" tube resolution - there are simply no limiting pixels. But that has nothing to do with Hires-mode, that's the "problem" of all DSOs. As you've calculated you can see that you need more and more memory to keep the real samples for further investigation, for instance if you press RUN/STOP.



Well... that just can be, we may have: "higher horizontal accuracy!" very true...
But there is question what for ..?  Do we need more "horizontal resolution"?
Actually proper question would be more than what..? And as far as I know even with normal mode ADAC is easily feeding scope with much more than needed "horizontal data" many times more than scope screen horizontal pixel count, specially at longer timebase, that is where HR can be used... bit counter-intuitive, one may say....

But as opposite we  possibly may wish more than typical  8 bit "vertical resolution"...
 With 8 bit ADAC usually 1 bit is lost in ADAC noise floor and we are left with 7 usable bits and it is just about 128 levels... Can we use/see more ..???
At 800x480 Rigol screen, that actually could be very handy. Instead interpolating about 127-180 pixels to 480 vertical screen pixels we cold use straight away without any tricks almost full 9 bits of real resolution, so we could use 10 bit ADAC.

Well actually not sure what to thing about ...
Not quite sold to HR being "more horizontal resolution" or 2GSa/s "constant on" thing or I just did not get it right...

Kind regards
Rosendorfer


No, keep these things separated. The horizontal resolution has nothing to do with HiRes-mode.  You should differ between the sample memory and the display memory. The HiRes-mode relates at the sample memory first. In case your scope samples with 625MSa/s at a certain timebase and your ADC is able to sample with 2.5GSa/s  physically you get two options:

 #1: Sample with real 625MSa/s and store a sample every 1.6ns -> scale your y-axis with 8 Bit -> 256 steps
or
 #2: Sample with real 2.5Gs, add 4 samples and divide by 4, and store the result again every 1.6ns -> higher vertical resolution, not horizontal, but "slower" signal because of averaging. The effective sample rate is 2.5GSa/s / 4 = 625MSa/s again -> scale your y-axis with 10 Bit -> 1024 steps.

Now the timebase and screen resolution decides again which of these samples you will see on screen and be connected by interpolation. If you're timebase is small enough that 1.6ns can be resolved you will see every sample without interpolation.

By the way: the higher vertical resolution is caused by calculcated values that cannot be sampled so accurate because of the ADCs smallest resolution.

Kind regards
Gunb
« Last Edit: August 17, 2012, 11:32:12 pm by Gunb »
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #20 on: August 18, 2012, 10:44:14 am »
Dear Gunb

Thanks for all the info.... very appreciate You find time for explaining that.

Seems that we are now missing just one thing....
What actually all that technology can do for us ?

Any chance You could post any screens showing normal vs HR mode, so we could actually see something ?

Kind  regards
Rosendorfer
 

Offline PA4TIM

  • Super Contributor
  • ***
  • Posts: 1118
  • Country: nl
  • instruments are like rabbits, they multiply fast
    • PA4TIMs shelter for orphan measurement stuff
Re: Rigol DS2072 Review
« Reply #21 on: August 18, 2012, 01:14:41 pm »
It would be nice if this reviews could be accompanied by real tests instead of just display pictures of unknown sources.
For instance something as simple like a fast puls with known risetime and amplitude
Runt detections
Trigger tests on more complex signals
Bandwidth test on sinewave and pulse
Consistency of measurements
Channel separation
Input impedance vs frequency
Accuracy of timebase and amplitude
FTT from a signal of known amplitude and harmonics. So we see if intermodulation is caused by the scope, if harmonic amplitude is correct ect.

For instance on the first series of screenshots I see things I do not like. OK, nice pictures, nice resolution but it is an instrument for meaurements.
- the Tr of squarewave is 4 nS on the first squarewave picture. Then you zoom in and risetime changed. This seems to imply it has the same problem as the 1102. If I put a pulse on screen and measure Tr, then shift the trace up or down or left/ right the Tr changes. So accurate Tr meaurement is useless this way.  The same thing happend with amplitude.

- I believe it uses display memory for measurements. For instance on the 1102, press stop and change timebase and amplitude. Measurement information changes. If I do this on a Agilent, Hameg or the old B&W screened Tek TDS 210 ( or something like that, forgot the number) the measurement info is still correct

- i still think the noise is rather high ( fat traces) 500 uV/div is nice but not if the noise is allmost a dividion. I have a Tektronix plugin for my 547 and a TM5000 series diff- preamp plugin that both go sub-mV but the noise level is much lower.


10 mV/div, compare the trace noise to that with your pictures ( ok this scope is more expensive but without a compare, norm or standard a test is useless, what i think is good can be bad for someone else. Like a voltmeter showing 1.000,000 V. Nice but if the source in real is 1.000,200 V it is not so nice anymore  ;)

Most DSO's do not really measure under 10 mV/div this is just a sort of software zoom were they reconstruct a signal out of the noise by software.  I think the problem is in the analog part before the DACs. There the men get seperated from the boys. You only have 8 bit resolution.
See it like a multimeter made from an Arduino. Use the internal DAC or put a LTC1052 before it, ad a LTZ1000 and a 26 bit DAC as frontend. It can still be named an Arduino voltmeter but the measurement results are more then a bit different.

An other test. Take a squarewave, 6MHz or so, select around 10 mS/div. Then press stop and change the timebase so the squarewave is visible in detail. About 3 to 6 periods on the screen. On the Hameg (350 MHz) and Agilent ( 10 year old 200 MHz) the squarewave was still a nice correct signal. On the DS1102 ( 100 MHz) it was a mix of triangle and squares. The Tek too but not as extreme ( and in its defece, it is over 10 years old, 60 MHz and only 2.5Kp memory.
also look at amplitude and time measurement info. Correct on all scopes exceot the Rigol. That are things that , I think, are more important as nice boxes, manuals or screen colors.

But it is nice to see a overview of some features of an instrument. Also to see the inside, how well its build, but I only would like to see real tests too.
( about weight, i onces teared down a cheap redio. It looked good, felt massive, looked like real wood and aluminium. But it was useless as a radio. The cabinet was mdf with plastic woodlike folie, the thick massive looking aluminium was hollow thin plastic. So why the massive feel.  the weight was caused by a mystery box mounted inside on the back of radio. This box was filled with concrete......
 

« Last Edit: August 18, 2012, 01:24:27 pm by PA4TIM »
www.pa4tim.nl my collection measurement gear and experiments Also lots of info about network analyse
www.schneiderelectronicsrepair.nl  repair of test and calibration equipment
https://www.youtube.com/user/pa4tim my youtube channel
 

Offline grenert

  • Frequent Contributor
  • **
  • Posts: 446
Re: Rigol DS2072 Review
« Reply #22 on: August 18, 2012, 01:28:49 pm »
Seems that we are now missing just one thing....
What actually all that technology can do for us ?

Any chance You could post any screens showing normal vs HR mode, so we could actually see something ?

Sorry, I don't have this scope to provide you a picture, but hires mode will definitely lower your noise floor.  The image below shows the benefit of noise reduction in the Agilent 9000.

- i still think the noise is rather high ( fat traces) 500 uV/div is nice but not if the noise is allmost a dividion. I have a Tektronix plugin for my 547 and a TM5000 series diff- preamp plugin that both go sub-mV but the noise level is much lower.

I don't think you'll ever see a DSO have a noise level like that of an old analog scope, even a cheap one.  Really, 0.5 mV of noise is pretty good compared to other DSOs.
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: Rigol DS2072 Review
« Reply #23 on: August 18, 2012, 06:57:07 pm »
Hi Grenert

Interesting picture....
If I see it right HR mode on Agilent 9000 reduces noise at 10V/div and at 10mV/div and 5mV/div where we actually care about the noise it does nothing..??

Rosendorfer


 

Offline Gunb

  • Regular Contributor
  • *
  • Posts: 221
  • Country: de
Re: Rigol DS2072 Review
« Reply #24 on: August 19, 2012, 11:20:43 am »
Dear Gunb

Thanks for all the info.... very appreciate You find time for explaining that.

Seems that we are now missing just one thing....
What actually all that technology can do for us ?

Any chance You could post any screens showing normal vs HR mode, so we could actually see something ?

Kind  regards
Rosendorfer

Hi Rosendorfer,

an answer to your question (I've taken it from the net):

Boxcar Averaging: Boxcar averaging is a digital procedure for smoothing irregularities and enhancing the signal-to-noise ratio. It is assumed that the analog analytical signal varies only slowly with time and the average of a small number of adjacent points is a better measure of the signal than any of the individual points. In practice 2 to 50 points are averaged to generate a final point. This averaging is performed by a computer in real time, i.e., as the data is being collected. Its utility is limited for complex signals that change rapidly as a function of time.

I couldn't explain it better and it shows that the usage is limited to a special kind of signals: low signals with high-frequency interference and/or fast accidental irregularities.

And because of the calculated average the result might lead to a higher resolution than the ADC offers - quantization noise becomes lower.

But: it can never replace a real 10 bit ADC that always has got real 10 bit for any kind of signal.

Screenshots: I would say they won't show anything new compared to the ones that are already posted.

I believe to remember that you wanted to do some audio measurements. For that purpose I've bought a soundcard with a good SNR and PC applications - it has real 16 Bit and is much better suited for that purpose than the Hameg or the Rigol scope.


Kind regards
Gunb
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf