8vs 12-bit scopes in economy class - couple of questions...

[Hella_Wini22]

   1. Why do all of them have such shitty displays ?
      * What's the use of having 12-bit ADC if you can't display all that resolution greatness on a 1024x600 display ? 12-bit means 4096 levels.
And that gets to be displayed usually on a small part of the display. So why don't we see retina-style displays anywhere ?
      * At the time that we see insane big 500Hz displays for gaming crowd, why do all the scopes still have 60Hz models ? I there is one field that could use great, fast displays, that would be scopes, right ? All that famed "digital phosphor persistence" would look much better on such unit, not to mention many other subtle effects of old scopes one could reproduce.
     * Why none of them is using modern-day ARM units with beefy GPUs that could do wonders on scope displays and visualize in 3D many otherwise non-obvious signal properties ? All competitors go for simple marketing bragging rights for lowest hanging fruit they can find and call it a day. And usually even that seems to be just a marketing fluff.
  2.  If one is to do 12-bit properly , one needs much wider datapaths and much beefier ALUs, wider memory, tighter analog circuitry WRT noise etc. If one remains bottlenecked by those, what's the real effect of going to 12-bits?
  3. EEVBlog has said that 8-bits doesn't really make sense any more. Why would that be ?\
Given the significant (~2x) ratio in hardware needed, I wonder why he says that. 8-bit should always be significantly cheaper, all other things being equal.
  4. How much more useful is 12-bit over 8-bit for the most of hobbyistic to prosumer audience ? Scope is all about seeing/visualising of what is going on and much less about measuring.
  5. I guess FFT could be one field where it makes a difference. But even that is limited. FFT on a such scope is more like a toy, compared to real spectrum analyser/VNA. How many are using it and really couldn't do with it on an 8-bit scope ?

[Fungus]

Scopes aren't gaming PCs.

The thing about ADCs is that you never get all the bits. Usually the bottom bit is junk (at least) and there's always noise in the system. Really we're talking about 6.5 bit 'scopes vs. 9.5 bit 'scopes (very approximately) at a basic level.

EEVBlog has said that 8-bits doesn't really make sense any more, because... these are basically the same price as the old ones! Why bother with a lower-resolution, non-touch-screen, less functional model?

8-bit should always be significantly cheaper, all other things being equal

Would be true if they designed new 8-bit models using modern chipsets. But they don't.

How much more useful is 12-bit over 8-bit for the most of hobbyistic to prosumer audience ?

These 'scopes do have more pixels on screen than the old ones, and they're way nicer to use. Those old ones feel like antiques in comparison.

[Hella_Wini22]

The thing about ADCs is that you never get all the bits. Usually the bottom bit is junk (at least) and there's always noise in the system. Really we're talking about 6.5 bit 'scopes vs. 9.5 bit 'scopes (very approximately) at a basic level.

• Even so, the rest usually ain't pure noise and with oversampling one usually gets more of useful information even out of those.
• Even so, relative proportions still holds. SImple 1024x600 display has  roughly the space for 2 tracks in full res at 8 bit vertically.  New 12-bit scopes have mbarely any more vertcial resolution and yet their ADC resolution quadrupled. Plus they usually have oversampling tricks to get more- up to 16 bit res.

Would be true if they designed new 8-bit models using modern chipsets. But they don't.

Why ? As they age, their price depreciates. We can see this all over the field. THis is why FPGAs that mere moratls get to see, for example use "mature" processes.
Same with microcontrollers. And existing chips get cheaper and cheaper up to the point of obsolescence.

These 'scopes do have more pixels on screen than the old ones, and they're way nicer to use. Those old ones feel like antiques in comparison.

Jut before I was sternly warned that I should stick to the subject and that scopes are about testing, inspecting and service/repair, not about metrology or anything else.
And certainly not about pleasuring oneself. If I was seeking that, I'd be on OnlyFans or something like that.

So my question was, how USEFUL it actually is on a scope ?

SInce scopes are mostly about conveyuing VISUAL information to human operator, one would think that increased ADC resiolution would be closely coupled with matching increase in display resolution... 🙄

[Psi]

Put simply, you only use a small range of the total voltage range of the scope when looking at a signal.
So your 0 -> 2V signal showing on screen only covers a tiny range of the bits.

Meaning an 8 bit scope with 256 possible levels ends up very marginal once you consider the number of possible levels being used to represent your 0 - 2V signal.

It's why the voltage measurement on scopes is normally terrible and the noise makes it jump up/down by perhaps 0.1V.

That being said, normally you don't need accurate voltage readings on a scope. So a 8 bit scope doesn't usually cause any problems in 99% of cases.

[tautech]

Put simply, you only use a small range of the total voltage range of the scope when looking at a signal.
So your 0 -> 2V signal showing on screen only covers a tiny range of the bits.

Meaning an 8 bit scope with 256 possible levels ends up very marginal once you consider the number of possible levels being used to represent your 0 - 2V signal.

It's why the voltage measurement on scopes is normally terrible and the noise makes it jump up/down by perhaps 0.1V.

That being said, normally you don't need accurate voltage readings on a scope. So a 8 bit scope doesn't usually cause any problems in 99% of cases.

^^^^
This !

[nctnico]

Put simply, you only use a small range of the total voltage range of the scope when looking at a signal.
So your 0 -> 2V signal showing on screen only covers a tiny range of the bits.

Meaning an 8 bit scope with 256 possible levels ends up very marginal once you consider the number of possible levels being used to represent your 0 - 2V signal.

It's why the voltage measurement on scopes is normally terrible and the noise makes it jump up/down by perhaps 0.1V.

That being said, normally you don't need accurate voltage readings on a scope.

12 bit doesn't help to make a scope more accurate. If you look at the specs, you'll see the DC gain accuracy is typically between 1% and 2%. 12 bit doesn't improve that. What more bits does help with is showing more details of a signal. But to actually see those details, you'll need a bigger screen.

[Psi]

What more bits does help with is showing more details of a signal. But to actually see those details, you'll need a bigger screen.

To see it in the waveform itself yeah, but it's immediately obvious when using the measurement function to display the voltage on screen as text.
To me that is the main reason to get a scope with 10/12 bits. Not to see the bits displayed on screen as smoother vertical waveform resolution but to get better voltage measurements and measurement calculations

[mouz]

12 bit doesn't help to make a scope more accurate. If you look at the specs, you'll see the DC gain accuracy is typically between 1% and 2%. 12 bit doesn't improve that. What more bits does help with is showing more details of a signal. But to actually see those details, you'll need a bigger screen.

So basically a product like (example) 12-bit SDS814X HD with a 7in 1024 x 600 pixel screen and no HDMI output is mismatched. What resolution would you ideally need for a 12-bit scope. (720p) 1280×720  ? or FHD (1080p) 1920x1080  ?

[hpw]

As from R&S:

8 bit scope as using 1mV/Div is the LSB at 39uV. This means with 8bit ADC of maximal 256 digital steps,
we get 9.984mV on the ADC so it do not gets any overdraft (full scale).

On the 1mV/Div we have ~ 10mV full range, as 8 division grids on screen, we get about 25.6 LSB steps.

This means if we get to 12 or higher bits, zooming will show better figures. As shown on HD3 sheet.

[Fungus]

Would be true if they designed new 8-bit models using modern chipsets. But they don't.

Why ? As they age, their price depreciates. We can see this all over the field. THis is why FPGAs that mere moratls get to see, for example use "mature" processes.
Same with microcontrollers. And existing chips get cheaper and cheaper up to the point of obsolescence.

New designs are more integrated, use LESS chips.

So my question was, how USEFUL it actually is on a scope ?

The claim is that you'd have to be crazy to buy one of the old 8-bit 'scopes now that these are available.

That's 100% true. It would be true even if they were only 8 bits.

Do they measure a peak to peak voltage any better? Not really.

Do they show a lot more visible detail on screen? Yes.

[Phil1977]

Many reasons for 12-Bit:

- All 8-bit scopes I know have visible quantization error on screen, even on the small screens. All 12-bit-scopes I know show super clean lines.
- 12 bit give you much more resolution for doing maths, e.g. calculating the difference between two channels or a reference waveform and the channel
- An ADC works best if its lowest 1-2 bits are noise. So in a 8-bit scope you practically only have a little more than 6 bits real resolution per sample.

Why don't the manufacturers spend more on the display & GPU? I just assume they invest into the ADC and it´s very closely connected RAM. Most of the signal processing magic happens there for very good reasons. And if the pre-processing has been done there, you only need a reasonable but not a very fancy display unit.

[Fungus]

Example: The Rigol DS1054Z has 400 vertical pixels in the trace display area and an 8-bit ADC, at least one bit of which is mostly noise so really only 7 bits even on a good day.

Do you see why 12 bits might be better even though the screen resolution doesn't go up much...?

[2N3055]

Put simply, you only use a small range of the total voltage range of the scope when looking at a signal.
So your 0 -> 2V signal showing on screen only covers a tiny range of the bits.

Meaning an 8 bit scope with 256 possible levels ends up very marginal once you consider the number of possible levels being used to represent your 0 - 2V signal.

It's why the voltage measurement on scopes is normally terrible and the noise makes it jump up/down by perhaps 0.1V.

That being said, normally you don't need accurate voltage readings on a scope.

12 bit doesn't help to make a scope more accurate. If you look at the specs, you'll see the DC gain accuracy is typically between 1% and 2%. 12 bit doesn't improve that. What more bits does help with is showing more details of a signal. But to actually see those details, you'll need a bigger screen.

Or if you buy propper brand it can be 0.5% DC accuracy...

[Geoff-AU]

  4. How much more useful is 12-bit over 8-bit for the most of hobbyistic to prosumer audience ?

Not very.  I have an SDS2000X-HD at work and an SDS-2000X-Plus at home.  The Plus is more than adequate for all my needs.

  5. I guess FFT could be one field where it makes a difference. But even that is limited.

The FFT on my HD is just as clunky and horrible to work with as my Plus.  As you said, a real spectrum analyser is much better.  But for dealing with low frequencies, the scope will do.

[jasonquin]

Application processor costs. Drive a high-resolution display at high frame rate will put too much burden on the tiny Zynq, which is used in almost every entry level scope.

[noisyee]

   1. Why do all of them have such shitty displays ?

Bigger screen means more $, higher resolution means more processor power and more $, so how to be economy?

  2.  If one is to do 12-bit properly , one needs much wider datapaths and much beefier ALUs, wider memory, tighter analog circuitry WRT noise etc. If one remains bottlenecked by those, what's the real effect of going to 12-bits?

They do have wider data path, you can prove it through the measurement value step.
They do have more sensitive trigger (digital trigger), especially when triggering noisy signal.
They do have lower noise floor (compare to their 8-bit brother).
If they don't, just don't buy it.

  3. EEVBlog has said that 8-bits doesn't really make sense any more. Why would that be ?
Given the significant (~2x) ratio in hardware needed, I wonder why he says that. 8-bit should always be significantly cheaper, all other things being equal.

Actually not 2X, mainly 50% more (8-bit to 12 bit) on FPGA and memory. The retail price difference of the two may not be that significant.
If you can buy a 12-bit scope at (almost) 8-bit price, why bother to buy the old school one?

  4. How much more useful is 12-bit over 8-bit for the most of hobbyistic to prosumer audience ? Scope is all about seeing/visualising of what is going on and much less about measuring.

To be honest, not that much.
8-bit is enough, 12-bit is better. I won't say I can't live without 12-bit, but if I had choose, I won't use 8-bit.
Digital zoom capability on 12-bit is a killer feature, but it's not always needed.

  5. I guess FFT could be one field where it makes a difference. But even that is limited. FFT on a such scope is more like a toy, compared to real spectrum analyser/VNA. How many are using it and really couldn't do with it on an 8-bit scope ?

Not that much a difference, really.
But newer products has more powerful FFT function, that makes life easier.

[Fungus]

Many products are like this, eg. cars.

They could make new cars cheaper by constantly reworking the old ones and updating their production processes but they don't. They prefer to add features and keep the same price levels instead.

[KungFuJosh]

Many products are like this, eg. cars.

They could make new cars cheaper by constantly reworking the old ones and updating their production processes but they don't. They prefer to add features and keep the same price levels instead.

Bad example. They often keep the features the same, put lipstick on the pig, and then charge more. 😉