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

[Carrington]

@Teneyes: This reminds me to the persistence effect. Although probably not have anything to do.

• 100MHz_Vec_2ch_trig0.png and 100MHz_dots_2ch_trig0.png overlay:

Note: Very nice the point clouds. 

[marmad]

Also got more noise, see picture with only 1 channel, and then enable channel 2,
the difference is clear, on 1 nS the sample rate is to low for two channels.
The performance is worse on 1 nS then on 2 nS.

@Wim: I think the problem in the images you posted has to do with the intensity-grading - not with noise or sample rate. Rigol seems to have fixed the intensity-grading issues in Delayed Sweep mode, but not those when two channels are turned on in Normal Sweep at time bases <= 50ns/div (I reported both problems to them via Drieg back in early June).

Look at this sequence of images made at 1ns/div: all of them have intensity set to 50% - except the last one, which is intensity at 0%. You can see that with two channels on in Delayed Sweep, the intensity is a pretty good match to the images with just one channel on. But in Normal Sweep, the intensity has to be set to 0% to get a closer match.

EDIT: I've added the following back into the bug list at the beginning:

2) The way the scope handles intensity grading at time bases <= 50ns/div when two channels are enabled with Normal Sweep is not the same as all other modes - it feels incongruous and incorrect.
[FW v.01.01.00.02 / FW v.02.01.00.03]

[Teneyes]

Intensity grade seems ok at slow scan rates, but not so good as scan rate increases
for both channels. Faster than 200nsec/div is limit
Does it seem that the min. persistence should vary for each condition and it does NOT?

1Ch alone seems OK

[EV]

Here are pictures abot this 2 channel trace problem. There are first 3 pairs of pictures with 1, 2 and 5 ns time base. Pictures with vectors and dots are from same sweep in normal single mode. 2 last pictures are from auto sweep.

[EV]

It looks that this sin(x)/(x) algoritm is not working very well if both channels are on.

[EV]

Yes, Sin(x)/X interpolation is outside the linear path. This can not be the correct way to do the algoritm. This is probably problem only with fast rising pulses and short time bases.  It works however fine if only one channel is on. It is enough for me.

@ EV   note that with 2CH there are less samples so the Sin(x)/X interpolation will make the curve go off the linear path
.
.

[marmad]

Yes, Sin(x)/X interpolation is outside the linear path.

Sin(x)/x is NOT a linear path - lots of information you can read about it online. 

"Unlike the narrow pointed triangle of linear interpolation, the window for SinX interpolation is a theoretically never ending damped sinewave. "

[EV]

OK, maybe it works correctly, but there are too few sample points for this fast rising pulse. Thanks for the picture! I am very lazy to read manuals and any other information. 

Sin(x)/x is NOT a linear path - lots of information you can read about it online. 

"Unlike the narrow pointed triangle of linear interpolation, the window for SinX interpolation is a theoretically never ending damped sinewave. "

[EV]

Here is for comparison the picture with same traces from my old TEK TDS3032. It is 300 MHz scope but the measured BW is over 400 MHz and it has 5 points per division as I told before.

[marmad]

OK, maybe it works correctly, but there are too few sample points for this fast rising pulse. Thanks for the picture! I am very lazy to read manuals and any other information. 

 

Well, the main thing to remember is that sin(x)/x is susceptible to errors if the original signal contains frequencies higher than the Nyquist frequency. At 1GSa/s (maximum with 2 channels on), that happens to be 500MHz. With the 300MHz enabled on a DS2000 HW v.1 model, 500MHz is around -12dB (according to Wim's chart) - so if you have questions about interpolation errors, you might also try it with the 100MHz BW filter on.

Here is for comparison the picture with same traces from my old TEK TDS3032. It is 300 MHz scope but the measured BW is over 400 MHz and it has 5 points per division as I told before.

Yes, I think your TDS3032 might be a TDS3052 by mistake     If your sampling rate is 5GSa/s, then the Nyquist frequency is 2GHz.

[EV]

Here are 2 pictures with BW limit 100 MHz. Deviation from dots path can still be seen.

[marmad]

Here are 2 pictures with BW limit 100 MHz. Deviation from dots path can still be seen.

Can you please post the following image again - taken with the DSO running and Wave Intensity = 0%?




And BTW, EV, why are your Rigol screen colors different than everybody else that posts images?

[EV]

Can you please post the following image again - taken with the DSO running and Wave Intensity = 0%?

It is attached. I can not see the picture in your post, but I can get it from its url.

And BTW, EV, why are your Rigol screen colors different than everybody else that posts images?

I have added lightnig to the pictures. The attached picture is not manipulated.

[Carrington]

My results (Trigger mode normal):
- Vectors an Dots for one single pulse.
- Vectors an Dots for a pulse train.

[Carrington]

Now, (trigger auto) a pulse train, with wave intensity = 0% and 100%.

[Carrington]

I think it is due to the overlap of many waveforms slightly different each time, and the wave intensity processing. Like the persistence effect.


Note: Test now with average acquisition mode and only 2 averages.

[marmad]

I have added lightnig to the pictures. The attached picture is not manipulated.

Thanks! If you compare this 1ns/div image with the 100MHz BW limited 1ns/div image in your previous post, the BW limit seems to have had an effect.

[poida_pie]

BUG in latest f/w, 02.01

Delayed sweep malfunctions.
To reproduce:
Ch2 must be enabled, Ch1 optional
acquire:  7M samples, normal
trigger: ch1 or not important, Auto or Normal seems to have no effect
main timebase 500us/div, delayed t.b. 100 us/div or something
Then when you move the delayed t.b. towards 1 ns/div all seems well until you select 1 ns/div..
It locks up for about 10 seconds, no waveform update, no response to keys.
Then it comes good and lets you go back to 2ns/div

Of course this DS2072 is using the latest enable everything serial codes.

[marmad]

BUG in latest f/w, 02.01

Delayed sweep malfunctions.
To reproduce:
Ch2 must be enabled, Ch1 optional
acquire:  7M samples, normal
trigger: ch1 or not important, Auto or Normal seems to have no effect
main timebase 500us/div, delayed t.b. 100 us/div or something
Then when you move the delayed t.b. towards 1 ns/div all seems well until you select 1 ns/div..
It locks up for about 10 seconds, no waveform update, no response to keys.
Then it comes good and lets you go back to 2ns/div

I see what you're talking about, but as mentioned earlier in the thread in regards to problems with the 1ns time base:

It's another error (and there's a few of them) in an, as yet, officially unimplemented - and perhaps unfinished? - portion of firmware. As far as I know, Rigol has not started selling BW updates - but if/when someone on the forum has a DS2302A (or DS2302A-S), we can find out if the error(s) exist in the 'official' working version - and if so, report it to Rigol.

I suggest disabling 300MHz on non-A HW v.1 models. Not only is it buggy (I've had a couple of crashes from it), but it's not really perfectly implemented in the input stage.

[EV]

How is it possible to uninstall only BW option?

I suggest disabling 300MHz on non-A HW v.1 models. Not only is it buggy (I've had a couple of crashes from it), but it's not really perfectly implemented in the input stage.

[Fagear]

How is it possible to uninstall only BW option?

Do ":SYSTem:OPTion:UNINSTall" and then generate new keys without 300 MHz option (with 200 MHz for example) and install them.
It can be two keys: DSAZ (200 MHz, MEM, DC, AT) then DSEA (CAN).

[EV]

So there is no way to uninstall only one option! 

Do ":SYSTem:OPTion:UNINSTall" and then generate new keys without 300 MHz option (with 200 MHz for example) and install them.
It can be two keys: DSAZ (200 MHz, MEM, DC, AT) then DSEA (CAN).

[AndersAnd]

How is it possible to uninstall only BW option?

Do ":SYSTem:OPTion:UNINSTall" and then generate new keys without 300 MHz option (with 200 MHz for example) and install them.
It can be two keys: DSAZ (200 MHz, MEM, DC, AT) then DSEA (CAN).

Or simply use DSEZ instead as the only key, to combine CAN with all the options included in DSAZ (200 MHz, MEM, DC, AT) into a single key.
Read my post about this in the other topic here: https://www.eevblog.com/forum/testgear/sniffing-the-rigol's-internal-i2c-bus/msg356379/#msg356379

[marmad]

Just to let people know - I found still another bug in 300MHz option on my HW v.1 non-A DS2000:

At some point (I don't know when or how), the vertical trigger position got offset by a few nanoseconds (visible at any time base < ~50ns) - and nothing I tried could get it back in position. As soon as I removed the 300MHz option - it was back perfectly in the center.

[EV]

Are you going to do a separate list of these bugs? So we could check if they are in the next FW uodate.

Just to let people know - I found still another bug in 300MHz option on my HW v.1 non-A DS2000:
.
.