VOLTCRAFT DOV series

[Aldo22]

It seems there’s a 7.5 kHz interference on Channel 1 in addition to the noise. Is it just as bad on the other channels?
But for the price, it’s still a nice device.

As far as I know, you can also power it using a USB-C PD power bank.
It would be interesting to see if it makes a difference.

[egonotto]

It seems there’s a 7.5 kHz interference on Channel 1 in addition to the noise. Is it just as bad on the other channels?
But for the price, it’s still a nice device.

As far as I know, you can also power it using a USB-C PD power bank.
It would be interesting to see if it makes a difference.

Hello,

Even in that case, there are voltage converters. That's why I'm afraid the problem will persist.

Best regards,
egonotto

P.S.: I think Voelkner has 333 units.

[KungFuJosh]

P.S.: I think Voelkner has 333 units.

Did you add one to your collection yet?

[egonotto]

P.S.: I think Voelkner has 333 units.

Did you add one to your collection yet?

Hello,

I would only do that if the noise were much lower.
At 20 MHz and 1 mV/div, the noise would have to be well below 20 µVrms. Then there might be a chance I’d be tempted.

Best regards,
egonotto

[Filippo52]

"At 20 MHz and 1 mV/div, the noise should be well below 20 µVrms."
Is this a high-end (or at least medium-high) oscilloscope specification, or am I wrong?
Filippo
P.S.
My Rigol MHO954 (which, admittedly, has a reputation for being noisy) can only dream of that value under any measurement conditions, whether those used by Martin or those recommended by Rigol, even with the 20 MHz bandwidth limit set.

[Aldo22]

Something strange I noticed in a video of the Owon:
The ADS804 seems to refuse to measure any frequency higher than the specified bandwidth.
In the attached screenshot, it should be 100 MHz. The normal measurement and the counter only show “?”.
Just like with an analog scope, you have to count the divisions and do the math (or does the cursor at least work?).
Is that also the case with the DOV?

My even cheaper and now somewhat outdated 100 MHz Hantek measures all signals it can detect fairly accurately. In the screenshot, 300 MHz.

[Aldo22]

The “Hi-Res” mode is effectively useless.
It has no effect on the measurements, and I ultimately confirmed this with the Bodnar pulser.
The rise time remains the same regardless of whether the mode is set to normal or Hi-Res.

What do you expect it to do?

The manual says: "High Res: Reduce and improve the signal-to-noise ratio on aperiodic (single-shot) waveforms."

[2N3055]

The “Hi-Res” mode is effectively useless.
It has no effect on the measurements, and I ultimately confirmed this with the Bodnar pulser.
The rise time remains the same regardless of whether the mode is set to normal or Hi-Res.

What do you expect it to do?

"High Res: Reduce and improve the signal-to-noise ratio on aperiodic (single-shot) waveforms."

What do you mean? HiRes is well known method to low pass filter the signal. You get less BW and less noise (because all the noise above BW is suppressed). When you enable it, P-P and RMS noise should go visibly down, and signal should have limited BW (how much depends on HiRes depth).

Bottom line is that Martin is right to say that when you enable it, it should show visible results, as it does when it is well implemented.

[Aldo22]

What do you mean? HiRes is well known method to low pass filter the signal. You get less BW and less noise (because all the noise above BW is suppressed). When you enable it, P-P and RMS noise should go visibly down, and signal should have limited BW (how much depends on HiRes depth).

Bottom line is that Martin is right to say that when you enable it, it should show visible results, as it does when it is well implemented.

Thanks.
I'm only familiar with HR from my cheap Hantek, and on that device it doesn't affect the rise time.
It just makes the signal a little less noisy.
That's why I asked.
I guess this is slightly different.
Hantek writes: “High Resolution (HR): This mode uses a kind of ultra-sample technique to average the neighboring points of the sample
waveform to reduce the random noise on the input signal and generate much smoother waveform on the screen. This is generally
used when the sample rate of the digital converter is higher than the storage rate of the acquisition memory.
Note: "Average" and "HR" modes use different averaging methods. The former uses "Multi-sample Average" and the latter
uses "Single Sample Average".”

[2N3055]

What do you mean? HiRes is well known method to low pass filter the signal. You get less BW and less noise (because all the noise above BW is suppressed). When you enable it, P-P and RMS noise should go visibly down, and signal should have limited BW (how much depends on HiRes depth).

Bottom line is that Martin is right to say that when you enable it, it should show visible results, as it does when it is well implemented.

Thanks.
I'm only familiar with HR from my cheap Hantek, and on that device it doesn't affect the rise time.
It just makes the signal a little less noisy.
That's why I asked.
I guess this is slightly different.
Hantek writes: “High Resolution (HR): This mode uses a kind of ultra-sample technique to average the neighboring points of the sample
waveform to reduce the random noise on the input signal and generate much smoother waveform on the screen. This is generally
used when the sample rate of the digital converter is higher than the storage rate of the acquisition memory.
Note: "Average" and "HR" modes use different averaging methods. The former uses "Multi-sample Average" and the latter
uses "Single Sample Average".”

Aha. Thank you for excerpt from doc.

No, that is nothing different. That is exactly standard HiRes, albeit explained in a bit convoluted language (yes, I like puns!) and a bit confused wording.
Except it might not be working well, according to Martin testing.
I am sure he will test it thoroughly.

[Martin72]

Hi,

The Hi-Res mode I was familiar with up until now reduces the bandwidth.
As a result, the rise time should be longer, but since that didn't happen, I assumed there was a malfunction.

[tautech]

Hi,

The Hi-Res mode I was familiar with up until now reduces the bandwidth.
As a result, the rise time should be longer, but since that didn't happen, I assumed there was a malfunction.

As it should do when oversampling is engaged otherwise Nyquist gets challenged.

[Aldo22]

With the Hantek, HR visibly smooths the signal at higher time/div settings, while it seems to have no effect when I’m looking at fast rise times (<10ns/div).
Perhaps HR mode only works with certain settings?
Hantek states: “This (HR) is generally used when the sample rate of the digital converter is higher than the storage rate of the acquisition memory.”
Perhaps it’s similar with the Voltcraft/Owon?

[2N3055]

With the Hantek, HR visibly smooths the signal at higher time/div settings, while it seems to have no effect when I’m looking at fast rise times (<10ns/div).
Perhaps HR mode only works with certain settings?
Hantek states: “This (HR) is generally used when the sample rate of the digital converter is higher than the storage rate of the acquisition memory.”
Perhaps it’s similar with the Voltcraft/Owon?

There will definitely be timebases that will be too short and therefore have very few points in a capture, so you cannot average because you don't have enough.

Keysight MSOX 3000T ( and relatives), when set to HiRes, do not do HiRes until some 1µs/div.
Also HiRes decimates BW by some ratio to sampling rate. So as you go to longer timebases, as soon as your scope starts dropping sample rate, BW will go down too....
Less sample memory you use, that will happen sooner...
I really is simple in concept, but as scope changes several parameters during changing settings, it gets complicated fast.

On some newer Siglent scopes on recent FW you get effective BW displayed in CH when using ERES, so you now exactly what is going on. It is very nice detail.

[Aldo22]

I really is simple in concept, but as scope changes several parameters during changing settings, it gets complicated fast.

Yeah, that's how it looks.
All I can say is that the Hantek in HR mode doesn't slow down the rise time at 5 ns/div for example (Signal from NanoVNA).
So, at least for me, rise time is not an indication that the HR isn't working.

[egonotto]

Hello,

It depends on how the HR is implemented.

At 5 ns/div, there are only 5 points per division. It looks to me like Hantek has disabled the HR without displaying it.

Best regards,
egonotto

[Martin72]

Hi,

I was playing around a bit earlier, and the result is a few more pictures, which will be split up accordingly.

Intensity/Colorgrading
That turned out really well; I've seen much worse—Good!
With Colorgrading, you can't adjust the intensity (same like Rigol DHO).
But since it already looks good as is, it's not really necessary.

XY Mode
Nothing exciting—just standard.
The screen is a bit too small for my taste, but other than that... standard.

Decode
I quickly checked it with a UART signal from the demo board.
There are a total of decoder channels.
I’ll test the remaining decoders as well.
Since I’m going to show screenshots of all the scope’s menus anyway, there aren’t that many of them.

Then, as requested, I fed in a square wave signal again—first at 1 MHz, then 20Mhz and 50 MHz.
It looks good, and of course it depends on the generator.

Then there's another unpleasant issue that I had to deal with on at least two channels.
But I'll get to that in a moment.

[Aldo22]

Thanks. 

The 20MHz square wave looks really good for a 70MHz scope. Even the 50MHz signal is still recognizable as a square wave. Very good!
The intensity and color grading look great.

XY Mode
Nothing exciting—just standard.
The screen is a bit too small for my taste, but other than that... standard.

Apparently, there's a full-screen mode.

[Martin72]

Hi,

Indeed!

For that, the scope gets a star, because it allows you to select the channels for this purpose and isn't limited to just CH1/CH2.

To access the menu, you shouldn't click on “XY Mode” at the top of the screen; instead, you have to select “XY Mode” from the main menu. I hadn't noticed that, and it is a bit of a hassle.
But once you know that, it's fine.

[Martin72]

Now let's move on to the less pleasant aspects...
I noticed this by chance, and it's reproducible:
The signals look clean, but:
If you set the vertical scale to 200 mV/div or 2 V/div (10:1), you can see a ripple.
Not below or above that range, only at 200 mV, and this occurs on at least two channels; I think the other two are affected as well.
I’ll take a closer look at this in due course.

[Aldo22]

For that, the scope gets a star, because it allows you to select the channels for this purpose and isn't limited to just CH1/CH2.

Great!

Were you able to figure out if the axis-labels can be enabled in FFT mode?

I'm starting to like the scope. 

[Martin72]

The 20MHz square wave looks really good for a 70MHz scope. Even the 50MHz signal is still recognizable as a square wave. Very good!

That doesn't surprise me, because the rise time had already hinted that the scope has more than 70 MHz of bandwidth.
And lo and behold, much more...

The -3dB point is around 183 MHz, which is more than double—a nice bonus.
It's a shame that the frequency counter stops counting at 80 MHz.
I also took another request into account and looked at the memory/sample rate allocation.
By the way, you have to set the memory manually; there is no auto mode.
Also, the memory is not reset to the value originally set by the user if it was previously reduced by the system—for example, because another channel was turned on or off.
Regardless, the result matches the datasheet and confirms the presence of two ADCs.

[Aldo22]

 

Regardless, the result matches the datasheet and confirms the presence of two ADCs.

So that's true, then.
4 x 625 MS/s and 10 M are exactly twice the specs of the Rigol DHO804

[Martin72]

Were you able to figure out if the axis-labels can be enabled in FFT mode?

I tried it out, but unfortunately the “Label” function is really just a way to name the trace.

[Aldo22]

Were you able to figure out if the axis-labels can be enabled in FFT mode?

I tried it out, but unfortunately the “Label” function is really just a way to name the trace.

That's a bit of a shame, but at least there's the peak display and cursors.
I can live with that.