Rigol DS1054z with DG1062z question

[John321]

I just got the Rigol scope and function generator. 50Mhz scope and up to 25Mhz square wave. However, there seems to be a lot of vertical change between the square waves see attached. But then at other frequencies it seems ok. (see attachments) I have the newer firmware that should have fixed the jitter talked about in

Is this normal?

[Simon]

where is the scope triggering

[John321]

Rising edge about 0 volts. Some of the harmonics vary in amplitude, that seems odd. Like in pic 11 and 14.

[Simon]

I think this behaviour is normal it's due to the amount of waveform updates in a second so you are seeing multiple versions of the waveform which is not being repeated identically.

[John321]

About a 1500mV (750mV + 750mV) difference pp. Seems like a lot. And at some other near frequencies there is no deviation. Correction there are two different trigger levels depending on the picture, see upper right of o-scope screen.

[John321]

On the square wave the distortion disappears at 500kHz, 800, 1Mhz,1.6,2.0, 2.5, 4, 5, 8, 10, 12.5, 16, 20 and 25MHz.

However, the output of the arbitrary waveform is even worse it varies by 50% and the frequency is all over the place not just the amplitude. See attached.

Built in waveform Sinc.

[alsetalokin4017]

How are you connecting the scope and FG together?

[Wuerstchenhund]

I just got the Rigol scope and function generator. 50Mhz scope and up to 25Mhz square wave.

A 50MHz scope is not sufficient to capture a 25MHz square wave appropriately due to it's frequency components. To capture at least the 5th harmonic you'd need a bandwidth of 225MHz, i.e. a 250MHz scope, and to capture the 9th harmonic you're looking at 425MHz.

What you see on your 50MHz Rigol hasn't much to do with how the signal looks like in reality.

[John321]

A BNC cable from generator to scope.

The scope should act as a low pass filter. It shouldn't change the wave amplitudes each sweep. The disappearance of the problem at certain round frequencies makes me think this something to do with the generator. Also the problem goes below 1MHz (square wave), within the capacity of the scope.

[alsetalokin4017]

A BNC cable from generator to scope.

The scope should act as a low pass filter. It shouldn't change the wave amplitudes each sweep. The disappearance of the problem at certain round frequencies makes me think this something to do with the generator. Also the problem goes below 1MHz (square wave), within the capacity of the scope.

Well, the DS1054Z only has 1 megohm impedance inputs, and presumably the outputs of the FG are 50 ohm. So you should use a 50 ohm terminator, either a T + terminator or a 50 ohm feedthrough adapter terminator at the scope end of the BNC cable. I don't know if this will cure your jitter problem completely but it will make a significant difference in the accuracy of the displayed waveforms.

[Simon]

yes, the problem is reflections up the cable

[John321]

Ok I think that is the problem. Thanks!

[Tim F]

50ohm termination will help flatten the frequency response between the FG and scope, making the square wave squarer, but it's not causing the 'jitter' issue.

The difference in the harmonics (i.e. difference in ringing on the square wave) from one cycle to the next is due to the 200MSa/s sampling rate limitation of the FG. Because the DAC in the FG is non-oversampling you see the distortion caused by the 'stair-steps' (albeit smoothed out by some lowpass filtering) that occur at the transition of samples.
If you set a frequency which does not divide into 200MSa evenly, the sampling points change for every cycle so you see a change in the distortion of the waveform at every cycle. I.e. if you choose 18MHz, the FG reproduces the waveform with either 11 or 12 sample points, depending how the cycle fits into it's sampling time.
As you found, when you generate a frequency which divides perfectly into 200MSa/s, the problem goes away as sampling points remain exactly the same for every cycle. To prevent this you either need a FG with a significantly higher sample rate or an analog FG.

I think the problem you have with the Arb function is related - the FG simply does not have enough samples to reproduce high frequency waveforms without significant distortion. Realistically, a DG1062Z is only really useful up to 10-20MHz for sinewaves and a few MHz for more complicated waveforms if you want to avoid significant distortion.

[John321]

Great knowing why it has the jitter helps a lot. I was thinking it had to do with cable length and wave cancellation. But your explanation makes much more sense.