REVIEW - Rigol DS2072 - First Impressions of the DS2000 series from Rigol

[marmad]

Sw ver: 00.00.01.00.02

I suggest an upgrade to 01.00.05 - but upgrade during bootup - NOT FROM GUI or you will lose the trial options. Also, don't forget: NO SELF-CALIBRATION for the moment - or you will also lose the trial options.

EDIT: Because of a memory reading bug discovered in FW 01.00.05 (but not in 01.00.02), I would now suggest waiting awhile before upgrading.

[EV]

I made the same test as with DS2202  for comparison also with Tektronix TDS3032. Sweep time is 0.1 sec and sweep is from 1MHz to 160MHz. Here is picture from Tektronix and the earlier picture from DS2202:

[Wim13]

@ EV,

Clearly some SWR on the Rigol. not completly 50 ohm
as on the Tek

[EV]

@ EV,

Clearly some SWR on the Rigol. not completly 50 ohm
as on the Tek

Yes, propably this termination with 50 ohm terminator using BNC T-adaptor is not very good. Tektronix has a real 50 ohm input impedance and needs no terminator.  It looks much better.

[EV]

I bought some Tektronix 011-0049-01 feed Thru 50 Ohm Terminators from eBay. When I get them, I test again this sweep with DS2202.

Edit: Rigol seem to have also these feed thru 50 ohm terminators.

[Wim13]

I bought some Tektronix 011-0049-01 feed Thru 50 Ohm Terminators from eBay. When I get them, I test again this sweep with DS2202.

Edit: Rigol seem to have also these feed thru 50 ohm terminators.

But i can not find any specifications of these from Rigol
You need at least calibrated or certified ones.

I have a several of these things, with and without certification.
and it is not easy to avoid any mismatches.

[Wim13]

Hallo,

Made two pictures, of FFT screen dump, with 300 Mhz on a Rigol 2072 ( 70 Mhz)

The first picture, with High resolution on, and the second on normal...
se the difference, two peaks suddenly disappears...

[Wim13]

Another two FFT pictures, with -70dBm signal at 100 Mhz and
the other with -70 dBm at 200 Mhz. scale is 5 dB.

So the noise floor here is about -95 Bm. The signal
is 70 uV and the scale is 500 uV /div

With 8 bit resolution, one single level is 16 uV on this scale.
so the 70 uV input only 4 or 5 levels of the 256 are used.

[EV]

Hallo,

Made two pictures, of FFT screen dump, with 300 Mhz on a Rigol 2072 ( 70 Mhz)

The first picture, with High resolution on, and the second on normal...
se the difference, two peaks suddenly disappears...

2 peaks more in normal mode? Also RMS is much bigger there. I am not familiar with FFT.

[Wim13]

Hallo,

Made two pictures, of FFT screen dump, with 300 Mhz on a Rigol 2072 ( 70 Mhz)

The first picture, with High resolution on, and the second on normal...
se the difference, two peaks suddenly disappears...

2 peaks more in normal mode? Also RMS is much bigger there. I am not familiar with FFT.

In high resolution , one should expect more detail then in normal....

[EV]

In high resolution , one should expect more detail then in normal....

Yes I thought so also. Does that bigger RMS in normal mode mean something?

[Wim13]

In high resolution , one should expect more detail then in normal....

Yes I thought so also. Does that bigger RMS in normal mode mean something?

The RMS is not reliable, to the book you may not use it in this ranges,
there are to less bits left , you have 8 bits for 8 divisions, 256 levels.
Thats 32 levels per division resolution for the RMS, it has to deal with statistics
to get more out of it.

[Wim13]

Here a FFT picture of the FM radio band at 88-108 Mhz,
just connected a VHF antenna on the input of the Rigol 2072.

You can clearly see the signals from 88 -108 Mhz in the air.

[zibadun]

You can clearly see the signals from 88 -108 Mhz in the air.

now we need to figure out how to hear them...  If there is a way to transfer samples to the PC in real time they can be demodulated and played back in something like GnuRadio. 

[larsie]

I pressed on and immediately pressed and released HELP (had to try several times).

I now have SW version 00.00.01 (and that's it... no .05) Does it mean it worked?

Is the self-calibration bug still in the v5 firmware? I assume yes, since you're warning me.

Edit: I managed to get the secret handshake now. Stupidly didn't realize I should go to system info afterwards. Thought it'd be a popup, but then I reread the forum, which I should have done anyway. So, it's not confirmed upgraded to 00.00.01.00.05.

Thanks.

Sw ver: 00.00.01.00.02

I suggest an upgrade to 01.00.05 - but upgrade during bootup - NOT FROM GUI or you will lose the trial options. Also, don't forget: NO SELF-CALIBRATION for the moment - or you will also lose the trial options.

[marmad]

Edit: I managed to get the secret handshake now. Stupidly didn't realize I should go to system info afterwards. Thought it'd be a popup, but then I reread the forum, which I should have done anyway. So, it's not confirmed upgraded to 00.00.01.00.05.

Great - enjoy!

[marmad]

I've had a couple of people PM me asking for a copy of the 01.00.05 firmware - but my DS2072 came with it installed, so I don't have one. Would someone who does be willing to post it - either here or in the software thread?  If someone has it, but is not comfortable publicly posting it - shoot me an email with it, and I'll be happy to do it. Just because Rigol is tight-fisted with the firmware, doesn't mean we have to be  And in fact, the more it spreads around, the better for everyone.  Perhaps it even pushes Rigol to speed up newer versions 

[EV]

Is it Valid to test at 160MHz Freg. when the DS2000 Scope is set at 100mS/div and ONLY 20Msa/s ??
Are you just showing aliasing?

Why it is not valid? What is aliasing (it is not in my dictionary)? I am testing the original probe and the way it is connected to scope. Tektronix scope is used for comparison because it has real 50 ohm input, but Rigol does not (it needs 50 ohm terminator).

[Teneyes]

Why it is not valid? What is aliasing (it is not in my dictionary)?

Here is a good video on aliasing


I think it is better to sweep 120-160 Mhz with DSO set for 2.0GSa/s  ie  500us/div   sweep time .5ms

[EV]

I think this aliasing does not influence to the shape of the curve and it can be corrected by long persistence time. Here are pictures with minimum and infinite persistence time.

[Wim13]

RIGOL 2000 Service Menu

I did not found any info on this yes on the forum, maybe somewhere else..?

But if you do this 4 button sequence, F7-F6-F7-Utility, under the trigger menu,
you get also the service menu, where you can do screen test, keyboard test,
and gaintest, and some other things i dont dare to touch.

Is there anybody , who already did some experience with this menu..??

[aghp]

Why it is not valid? What is aliasing (it is not in my dictionary)?

With Owon (and many many other oscilloscopes including also some Rigol lowest end models)I have made lot lot of lab tests.)

(here borrowed for example picture from Owon thread page 57 where is also more).

For this measurement can use slow speeds for BW tests. It is quite normal method. This may avoid also some possible other fast sweep problems)

There is very slow samplerate and still no need think aliasing (in this case) but wait a moment...

(aliasing is not known?? and using digital oscilloscope... perhaps this is good to upgrade asap for avoid mistakes in measurements... but no problem, it is easy to learn. it is one very very basic fundamental what need know and know what all it is and what kind of affect it have... in different waveforms in normal oscilloscope trace and FFT. )

You see owon image I have used very slow speed (10s/div) and tested with sweep from 1MHz to 250MHz and "samplerate" is 50 samples per second (becouse in this case only 10k sampling buffer selected)  This 50 samples/s  mean that ALL analog frequency components what ADC input can see and what are over 25Hz produce exactly sure aliasing (if you see it or not but it is there). This is basic fundamental what can not break!

But example in Owon behind this 50sample/second ADC circuit works still (in this case 1GSa/s) full speed.  For aliasing this do not matter and for alising it is 50 sample per second.

(ADC works full speed and systenm use only every 20000000's sample.)

How to do?

Turn sampling (acquire) mode to "peak" (it is used in this Owon test)

Now system select (inside this 1/50s window highest samples  (different scopes use different methods perhaps (of course this have also bad effect becouse now this sample time is where ever inside this "window" but in this case it do not mean anything) and  what happend... now you do not need put one penny for thinkin aliasing (in this case).

Of course if do some accurate tests also generator starting level need adjust with this frequency what want keep as reference level. Also it is good to use signal generator what have accurate flatness over used frequency band.

[EV]

Turn sampling (acquire) mode to "peak" (it is used in this Owon test)

Here is a new picture "peak detect" turned on. There is also under "acquire" "anti-aliasing" button which is now on. Persistence time is at minimum. There is no change on the curve compared to the earlier curves.

[marmad]

But if you do this 4 button sequence, F7-F6-F7-Utility, under the trigger menu,
you get also the service menu, where you can do screen test, keyboard test,
and gaintest, and some other things i dont dare to touch.

Is there anybody , who already did some experience with this menu..??

I've played around with the keyboard test - and looked at the other options. But I haven't screwed too much with it because I didn't want to risk messing up my factory calibration.

[aghp]

Here is a new picture "peak detect" turned on.

Now there on the forum no one can come and suspect if you show some aliasing...