Siglent SDS2000X Plus

[tv84]

Not necessarily related to this scope, but some comment: I don't get why these entry-level (but not cheapest) scopes don't have proper GPU included. There really isn't Zynq-like chip with Mali GPU? Or even they could slap some cheap mass-produced mobile SoC and link it up with FPGA with PCIe/USB3. Give it a HD screen, proper hardware accelerated UI that gets stable 60fps, maybe even offload all these Math features to compute shaders. But no, we instead end up getting CPU-blitted UI, so that leads to lowres screens, and any feature that have to be fast is hacked around by directly overlaying waveform from FPGA onto screen.

Here we go again...

Try to do the same for the same amount of $$$.

[Martin72]

1024x600 "Low-Res" Screen.....LOL.

[maginnovision]

If it's not high res the argument it's low makes sense to me. It's not even 720p so it's definitely not high res.

[Martin72]

For me who work with lecroy* scopes with vga resolution, it´s high-res...

*)Exept our two new ones with 1280x800

[maginnovision]

1280x800 with a 10" screen is what I'd call high res. If you push that resolution to/passed ~12-13" you lose me. It really all depends on many factors. The farther you are from it, the lower it can be(but scopes are usually close). Depending on the quality of the LCD resolution can seem worse than it is (it's mostly an opportunity to make it look worse). Compared to something like the 1054Z I bet the screen looks fantastic. In the end it's just 1 piece of the puzzle and I don't think it's going to be a deciding factor for most scope purchasers.

[Martin72]

To be honest, I give a s**t of the resolution  - Will you look a movie on it or watch simply "drawed lines and curves"....

[maginnovision]

That's the point. It's not required for a scope but it is a nice thing to have.

[Martin72]

Maybe therefore our new lecroy got "only" 1280x800, although they got 12.1" and 15.4" displays.
Well, the intel i5/7 cpu will not have any sweat handle this resolution with it´s own gpu area(they got intel 4-core i5/7 processors with 8GB ram....The performance in general is "wow!").

[thinkfat]

What makes a scope is the measurements and triggers. Not necessarily the fidelity or clarity of the display. The waveform display just gives you an idea what to look for.

[2N3055]

That's the point. It's not required for a scope but it is a nice thing to have.

No actually it's a horrible thing. It unnecessarily creates huge load on hardware without giving any benefit.
More than 96 DPI on the screen is not useful from normal viewing distances..

[maginnovision]

That's a ridiculous argument because those things aren't correlated. A nice display doesn't mean it's going to be slow and a crappy display doesn't mean it'll be snappy.

[2N3055]

That's a ridiculous argument because those things aren't correlated. A nice display doesn't mean it's going to be slow and a crappy display doesn't mean it'll be snappy.

Number of pixels does...

[maginnovision]

No it doesn't. If you put a 4k display and an arduino to control it then it'll be slow. If you put a 4k display with an appropriate SoC it won't be. You aren't required to mismatch your hardware.

[Martin72]

As you point it out right before, there is no need for a scope to have a high resolution display.
A rectangular waveform, sinusoidal, triangle wouldn´t look better on it.
So what could be the benefit having a high-res screen in real....

[maginnovision]

I already said what the benefit is. It's nice. Smoother graphics, smoother text, better graticule displays, being able to label everything accurately without taking up the whole screen. There are all sorts of things you can do by having a higher resolution. Just look at how they went for A1M instead of AC1M. That was likely a design choice limited by the display. They changed it because feedback told them users would rather have more descriptive label for the channel.

You don't need it, but there are things you can have with higher resolution displays you can't get with lower or with fewer trade offs compared to the lower resolution(DPI if you'd like, I'm assuming display size doesn't change). It's all a trade-off and I think Siglent did just fine but if I had a choice I would rather have a bit higher resolution than what they offer(1280x800 for a common resolution). As I said before I DON'T think it's a dealbreaker and it shouldn't be an issue for any of the SDS2000X Plus owners.

[TK]

Regardless of resolution or screen dimensions, Siglent needs to fix the waveform crippling effect when the menu is on the right side of the screen.  It does not make any sense to compress and expand the image horizontally all the time the menu appears and disappears.  A perfect sinewave looks like a stair case.

[tv84]

It does not make any sense to compress and expand the image horizontally all the time the menu appears and disappears.  A perfect sinewave looks like a stair case.

Agree. IMHO, it definitely looks cheap.

[mawyatt]

Here's a print screen showing a vertical & horizontal minimum discrete steps. The delta X shows 0.5ns/step and the delta Y shows 8.33uv for minimum discrete step. Think this agrees with the 2GSPS ADC rate and vertical screen resolution (600), 8 and 10bit (oversampled) resolution. The first screen capture shows the Vectors (interpolated) and second shows Dots (actual ADC samples).

Hi. You write 8.33uV per discrete step and show that the cursors span TWO steps

That's because the step in between is not an ADC sample, it's interpolated. See the 2nd screen shot that shows the ADC samples as dots.

Best,

Mike

[rf-loop]

... 500uv/div, 10bit mode, and 50 ohms input termination. Zoom was enabled (maximum) ...

Here's a print screen showing a vertical & horizontal minimum discrete steps. The delta X shows 0.5ns/step and the delta Y shows 8.33uv for minimum discrete step.

No.

Here is part of your image again. I have marked ADC sample dots ETA:(for clarify to peoples who do not know SDS2000XP. Naturally not raw ADC samples but from 8 bit raw samples produced "10bit" resolution samples and fBW reduced to max 100 MHz) what I can find with my over best before date eyes.
Minimum step is roughly half you told.

[Vestom]

That's because the step in between is not an ADC sample, it's interpolated. See the 2nd screen shot that shows the ADC samples as dots.

Since you are running in 10 bit mode, all your steps are interpolated 

However, I did the exercise in a previous post and also found that the scope actually has 11 bits in 10 bits mode when running 500uV/div giving a max native resolution of 33uV and 4uV interpolated.

https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/msg3180170/#msg3180170

[Howardlong]

Regardless of resolution or screen dimensions, Siglent needs to fix the waveform crippling effect when the menu is on the right side of the screen.  It does not make any sense to compress and expand the image horizontally all the time the menu appears and disappears.  A perfect sinewave looks like a stair case.

I may be in a minority, but I much prefer this to covering up 20% of the waveform whenever a menu comes up, a la MSO5000.

I didn't understand your point "A perfect sinewave looks like a stair case" in this context?

[thinkfat]

Regardless of resolution or screen dimensions, Siglent needs to fix the waveform crippling effect when the menu is on the right side of the screen.  It does not make any sense to compress and expand the image horizontally all the time the menu appears and disappears.  A perfect sinewave looks like a stair case.

I may be in a minority, but I much prefer this to covering up 20% of the waveform whenever a menu comes up, a la MSO5000.

I didn't understand your point "A perfect sinewave looks like a stair case" in this context?

He's referring to the primitive scaling (column culling) applied to the waveform when the menu is open. The waveform is scaled horizontally by just omitting data points without filtering and that messes up the display.

[mawyatt]

... 500uv/div, 10bit mode, and 50 ohms input termination. Zoom was enabled (maximum) ...

Here's a print screen showing a vertical & horizontal minimum discrete steps. The delta X shows 0.5ns/step and the delta Y shows 8.33uv for minimum discrete step.

No.

Here is part of your image again. I have marked ADC sample dots what I can find with my over best before date eyes.
Minimum step is roughly half you told.

I didn't compare the dots on the left to the right, interesting!! These dots on the right, some are in between the dots on the left and it appears the scope can make a 8.33uv/2 step but not in direct sequence. Do you think this is due to the oversampling in the 10 bit mode?

I was impressed with 8.33uv/step, looks as if the scope can do 4.2uv in non-sequential sampling (maybe oversampling in 10 bit mode).

Best,

[milek7]

Here we go again...

Try to do the same for the same amount of $$$.

HW cost cannot be the problem there. You can buy board with beefy ARM for ~$100, add display for $50, so only $150 (and that's probably overestimating) more for much better display, scaling and UI performance.

I didn't understand your point "A perfect sinewave looks like a stair case" in this context?

It's quite jarring once you have seen it 
But I don't think they can fix it, except by cutting that area like on MSO5k.

[mawyatt]

That's because the step in between is not an ADC sample, it's interpolated. See the 2nd screen shot that shows the ADC samples as dots.

Since you are running in 10 bit mode, all your steps are interpolated 

However, I did the exercise in a previous post and also found that the scope actually has 11 bits in 10 bits mode when running 500uV/div giving a max native resolution of 33uV and 4uV interpolated.

https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/msg3180170/#msg3180170

See the reply above. I was referring to the steps that were introduced by the display processor and not a direct ADC sample measurement which the dots show. What's interesting, as mentioned above, is the 4.2uv steps apparently are not in sequence (in sequence 4.2uv steps are not ADC samples but interpolated for display in between 8.33uv steps which are actual ADC samples).

Either way, nice performance for a scope in this class and something I did not expect

Best,