Author Topic: Only 7 bit vertical resolution on scope?  (Read 3139 times)

0 Members and 1 Guest are viewing this topic.

Offline Circlotron

  • Super Contributor
  • ***
  • Posts: 2054
  • Country: au
Only 7 bit vertical resolution on scope?
« on: April 25, 2017, 03:56:44 am »
Say I want to do some measurements that go from zero to some positive value. Set the trace down to near the bottom of the screen so I can use the full screen height if necessary. Would that not mean though I am only going to get a 7 bit vertical resolution / 128 steps because the trace is not going to swing negative and so use the MS sign bit?

Is there some way of tricking the scope into using the full 8 bits for a positive only measurement?
 

Offline Octane

  • Regular Contributor
  • *
  • Posts: 101
  • Country: us
Re: Only 7 bit vertical resolution on scope?
« Reply #1 on: April 25, 2017, 04:35:22 am »
It is using all 8 bits. If you shift your waveform down horizontal it is done still in the analog domain in the frontend by offsetting the voltage before the ADC. I think usually the full scale of the ADC should be the 8 or 10 vertical divisions (however many your scope has). I think in some scopes there are two hidden divisions outside the top and bottom of the screen, but nevertheless the 256 steps are always scaled more or less to the screen. I hope this answer makes sense.

Michael
 

Offline bson

  • Supporter
  • ****
  • Posts: 1740
  • Country: us
Re: Only 7 bit vertical resolution on scope?
« Reply #2 on: April 25, 2017, 05:15:34 am »
In most cases you also have more than one sample per displayed time point, and they form a intensity graded vertical bar rather than a single point on the screen.  The measurement depends on what's measured, but when each time point has multiple samples you get better than 8 bit measurements in many cases - such as for rise time or RMS voltage.
 

Offline tmbinc

  • Regular Contributor
  • *
  • Posts: 228
Re: Only 7 bit vertical resolution on scope?
« Reply #3 on: April 25, 2017, 07:53:16 am »
There are two kind of offsets, a physical one (subtract DAC output before amplifier to ADC), and "display position" (simply shifts the waveform after acq).

For some scopes, it's the same - on the old TDS2xx for example, if you'd shift the waveform, it would change the physical offset. This way you'd always get the full dynamic range, even with a large offset. (I commonly measured ~1V of modulation on top of a 14V signal back then, so I would just move the waveform down).

For other scopes (most notably newer Tektronix ones), both are exposed individually. On the DPO7xxx-series you can press the position knob to switch between "position" and "offset". On some of the other scopes, you can only change position, but not offset using the knob (and you have to change the offset using the vertical menu).
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 11204
  • Country: 00
Re: Only 7 bit vertical resolution on scope?
« Reply #4 on: April 25, 2017, 01:02:59 pm »
Just imagine DAC '0' is at the bottom of the screen and DAC '255' is at the top*.

All the rest is done with amplifiers and offsets in the input.


(* or maybe it's the other way around in reality)
 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 4382
  • Country: ca
Re: Only 7 bit vertical resolution on scope?
« Reply #5 on: April 25, 2017, 01:56:59 pm »
No you do not see all 255 values, there is a margin on both top and bottom, 10%  or something on each side. Otherwise your screen top/bottom waveform would appear clipped.
To find out You can stretch a sinewave or other low slew signal across the screen and count ADC steps.
Facebook-free life and Rigol-free shack.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 11204
  • Country: 00
Re: Only 7 bit vertical resolution on scope?
« Reply #6 on: April 25, 2017, 02:04:51 pm »
No you do not see all 255 values, there is a margin on both top and bottom, 10%  or something on each side. Otherwise your screen top/bottom waveform would appear clipped.
To find out You can stretch a sinewave or other low slew signal across the screen and count ADC steps.

Easier to grab a screenfull of data over ethernet and dump the numbers.
 

Offline Keysight DanielBogdanoff

  • Frequent Contributor
  • **
  • Posts: 744
  • Country: us
  • ALL THE SCOPES!
    • Keysight Scopes YouTube channel
Re: Only 7 bit vertical resolution on scope?
« Reply #7 on: April 25, 2017, 08:40:31 pm »
For Keysight scopes, the ADC range is the full screen. All shifting, etc., is done in hardware before the ADC. So, shift the signal down and scale the signal to be full screen.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 11893
  • Country: us
  • DavidH
Re: Only 7 bit vertical resolution on scope?
« Reply #8 on: April 26, 2017, 10:41:12 am »
Say I want to do some measurements that go from zero to some positive value. Set the trace down to near the bottom of the screen so I can use the full screen height if necessary. Would that not mean though I am only going to get a 7 bit vertical resolution / 128 steps because the trace is not going to swing negative and so use the MS sign bit?

That is right; half of the digitizer resolution is used for negative signals.

They are not common but some "8 bit" DSOs are actually 9 bits in disguise which can be verified by checking the detailed specifications carefully.  Tektronix made a few modern TDS models like this.  These DSOs do provide 8 bits of resolution for positive only signals.

Quote
Is there some way of tricking the scope into using the full 8 bits for a positive only measurement?

Offset the input signal externally to start at a negative value or use the offset capability of the DSO if it has it.

In most cases you also have more than one sample per displayed time point, and they form a intensity graded vertical bar rather than a single point on the screen.  The measurement depends on what's measured, but when each time point has multiple samples you get better than 8 bit measurements in many cases - such as for rise time or RMS voltage.

This improves the resolution and reduces noise but does nothing for linearity so accuracy may not be improved even though precision is.

No you do not see all 255 values, there is a margin on both top and bottom, 10%  or something on each side. Otherwise your screen top/bottom waveform would appear clipped.
To find out You can stretch a sinewave or other low slew signal across the screen and count ADC steps.

This varies depending on the manufacturer.  Some, usually cheaper, clip at the display boundaries and some provide various amounts of over-range capability.  Good designs link this with the number of display points per division to prevent aliasing of the display and a performance advantage.
« Last Edit: April 26, 2017, 11:04:25 am by David Hess »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf