Rigol HDO1000 and HDO4000 12bit oscilloscopes launched in China

[thm_w]

I'll get to the noise thing later, but the "I didn't measure pre/post hack but for CH1 I get 22uV (20MHz), 53uV (Max BW) on 1V/div." is probably at 1 mV/div?

oops, yeah 1mV/div.

thank you for your answer. I used a T-piece with a 50 Ohm termination. The generator is a SDG6022X which had a different motherboard after a repair, with a SDG6052 license. I repeated the measurement with only channel 3 active and a sine wave with 4 Vpp.

Interesting, well that gen is more than good enough so maybe mine is an issue. Still, can you do the FFT of a 50R short on channel 1 as I showed above and post here? 2Gs/s, span 0 to 1GHz.
You can try lower voltage ranges, but the difference in BW should not be huge between them.

[egonotto]

Hello,

I haven't really understood how to use FFT with the DHO1074 yet. The 200 MHz license is activated.
Here are a few noise measurements.

Now to the frequency response (50 Ohm):
With 500 mVpp I determined about 240 MHz.

1 kHz 264 mVpp

2 MHz 234 mVpp
100 MHz 220 mVpp
200 MHz 208 mVpp
240 MHz 184 mVpp
250 MHz 176 mVpp
300 MHz 164 mVpp
400 MHz 108 mVpp

Best regards
egonotto

[Martin72]

How good it would be if Rigol included the average mode.
You would have to write a petition...

[egonotto]

Hello,

you mean average for FFT?

How is the noise of the Siglent SDS2104X HD in the range of 1mV/div and 1 V/div at 1 ms/div and about 50 MS so that we can compare it with the Rigol?

Rigol 20 MHz with 50 Ohm termination:
1 mV/div: noise is about 23 uVrms
1 V/div: noise is about 9 mVrms
Rigol 200 MHz 50 Ohm termination:
1 mV/div: noise is about 103 uVrms
1 V/div: noise is about 11 mVrms

Best regards
egonotto

PS: But sometimes you get different values, for whatever reason.
As the pictures show.

[2N3055]

Hello,

you mean average for FFT?

How is the noise of the Siglent SDS2104X HD in the range of 1mV/div and 1 V/div at 1 ms/div and about 50 MS so that we can compare it with the Rigol?

Rigol 20 MHz with 50 Ohm termination:
1 mV/div: noise is about 23 uVrms
1 V/div: noise is about 9 mVrms
Rigol 200 MHz 50 Ohm termination:
1 mV/div: noise is about 103 uVrms
1 V/div: noise is about 11 mVrms

Best regards
egonotto

PS: But sometimes you get different values, for whatever reason.
As the pictures show.

8µ RMS and 300µV P-P at 1mV/div  is not right.....
Something very wrong there.


As requested, from SDS200X HD (1ms/div, 2GS/s, 20MPts, 50 Ω):

20MHz BW

1mv/div -  23 µV ACRMS
1V/div   -  4 mV ACRMS

200MHz BW

1mv/div - 48 µV ACRMS
1V/div   - 5,2 mV ACRMS

FULL BW (cca 550 MHz)

1mv/div - 74 µV ACRMS
1V/div   - 5,4 mV ACRMS


Like I said before, 2000XHD is better. Also, you can clearly see how at 1mV/div noise of preamp before AA filter is dominant, and at 1V/div ADC noise is...

But DHO1000 does decent job nevertheless, hardware vise. It is not best in class scope, but affordable one, and for the price range not bad at all.
But software really needs work...

[egonotto]

Hello,

thanks. I have switched the DHO1074 off and on again. Now I have a credible 23 mVrms.

Best regards
egonotto

[voltsandjolts]

Now I have a credible 23 mVrms.

mVrms?

[egonotto]

Hello,

it is 23 uVrms.

Best regards
egonotto

[2N3055]

Which means preamp is working really nice but ADC performance is not best in class for the BW.
Or ADC has much wider BW than needed.
Which would make sense if they used same ADC as in 800+MHz DHO4000...

[Fungus]

Hello,

I haven't really understood how to use FFT with the DHO1074 yet. The 200 MHz license is activated.
Here are a few noise measurements.

You don't use FFT to measure noise levels directly, use AC RMS measurement for that.

FFT can used to measure signal to noise ratios, or maybe to look a noise characteristics.

[Martin72]

Hello,

you mean average for FFT?

Yes.

[egonotto]

Hello,

now I have tried the 100 MHz license. BW is about 120 MHz.

Rigol 20 MHz with 50 Ohm termination:
1 mV/div: noise is about 21 uVrms
1 V/div: noise is about 4 mVrms
Rigol 120 MHz 50 Ohm termination:
1 mV/div: noise is about 58 uVrms
1 V/div: noise is about 8.5 mVrms

Best regards
egonotto

[egonotto]

Hello,

However, the calculations of DHO1000 are not trustworthy. You can also get a noise of 4 mVrms at 1 V/div instead of 8.5 mVrms as in the first measurement.

Best regards
egonotto

[Antonio90]

Could the test setup (moving probes or lights/appliances) influence the measurement?
I live close to military training grounds, and sometimes I get AM interference.

[egonotto]

Hello,

I don't think so, because you can often bring about this change in the measurement results by different operations (switch on the second channel, add measurement, switch off the second channel, delete the measurement), can bring about this change in the measurement results.

(This is translate with deepl)

Best regards
egonotto

[Antonio90]

Doesn't seem to be the case then. Activating a second channel could induce some crosstalk maybe?
But activating and deactivating measurements, I don't know what could be causing it.

[2N3055]

Doesn't seem to be the case then. Activating a second channel could induce some crosstalk maybe?
But activating and deactivating measurements, I don't know what could be causing it.

There is no crosstalk.
There is nothing connected to anything.
NO signals.
Just scope's own noise...

That is error in measurement software.
Or at least let's hope it is. Otherwise there is hardware problem.

[voltsandjolts]

Are the inputs shielded? Covered with tinfoil at least?

If 50 Ohm terminators fitted to the inputs are the noise measurements stable?

[egonotto]

Hello,

I suspect it's a software bug.

Best regards
egonotto

[2N3055]

That image that says 340µV P-P max, 280 µV P-P min does not have trace 0.3 div wide...
And it says 8µV RMS which, well, is not really possible...
That would be ratio of 30, not 6-8..

[TurboTom]

Hey egonotto -
I just scanned through your latest messages and I didn't find any information on the firmware version that you're running on your DHO1000. If it's a very early one, maybe try updating it to the most recent version and see if the flakey AC RMS measurement gets sorted. I don't remember to have seen anything like this on my DHO1000 which has the latest version (00.02.12) installed. Just an idea, though... 

[egonotto]

Hello,

when I got the DHO1074 it had firmware 00.02.04. But I had quickly set up 00.02.12

Best regards
egonotto

[Martin72]

I was also able to observe these phenomena with the DHO4204, and that was 10 months ago.
It remains to be feared that the "good old tradition" of rigol will also strike with the 12-bit models and that they will only do what is necessary to keep the box crash-free.
There is also the question of why the user has no leeway and things like interpolation and display mode are fixed.
Maybe it's better not to know that.

[core]

In the past, I've done some testing with the MSO5000, which still has the dots mode. In certain configurations, dot mode looks very clear and nice. But if you go to zoom in further, it starts to interpolate and put intermediate points using sinx/x interpolation. Siglent is more predictive, like "in the book".

I've come to terms with the fact that it's noisy and I don't use it below 20mV/div. I put it down to the fact that the input amplifier runs at high temperature and could be an explanation. It was the first oscilloscope I came across that had heat sinks on the input stages. Not very usual for me.

To get to 8GSa/s at a decent price I suspected that there were actually several DACs running in interleaved mode in the DAC chip. It could be for example 8x1GSa/s or even 16x500MSa/s combined in pairs of 2 and 4 respectively to cover the 4 channels. That would explain why it behaves so strangely in dots mode and why it doesn't have X interpolation mode. It's hard to align all these DAC's, so the software has to do something about it.

In this way I thought Rigol managed to get such a good price for the MSO5000.

The DHO also has the input stage operating at high temperature, but at the same time has low noise. I seem to be missing something here. Also the DAC they manufacture may be hiding some unknowns.

My main field of work is in power electronics, so I don't have expertise in high frequencies small signals. It's quite possible I'm wrong in the above 

[2N3055]

In the past, I've done some testing with the MSO5000, which still has the dots mode. In certain configurations, dot mode looks very clear and nice. But if you go to zoom in further, it starts to interpolate and put intermediate points using sinx/x interpolation. Siglent is more predictive, like "in the book".

I've come to terms with the fact that it's noisy and I don't use it below 20mV/div. I put it down to the fact that the input amplifier runs at high temperature and could be an explanation. It was the first oscilloscope I came across that had heat sinks on the input stages. Not very usual for me.

To get to 8GSa/s at a decent price I suspected that there were actually several DACs running in interleaved mode in the DAC chip. It could be for example 8x1GSa/s or even 16x500MSa/s combined in pairs of 2 and 4 respectively to cover the 4 channels. That would explain why it behaves so strangely in dots mode and why it doesn't have X interpolation mode. It's hard to align all these DAC's, so the software has to do something about it.

In this way I thought Rigol managed to get such a good price for the MSO5000.

The DHO also has the input stage operating at high temperature, but at the same time has low noise. I seem to be missing something here. Also the DAC they manufacture may be hiding some unknowns.

My main field of work is in power electronics, so I don't have expertise in high frequencies small signals. It's quite possible I'm wrong in the above 

You are not wrong, in theory.

MSO5000 had higher noise because it used both ADC and front end chip that are 4 GHz capable.
Many manufacturers use massively interleaved ADC (including Keysight), and there are many tricks of the trade there to achieve best performance.
From clock distribution, to buffering, to calibration...