Auto-scale vertical doesn't necessarily make the waveform fill the screen. Does this reduce the effective measurement resolution?
Autoscale and Autoset are different, and not all scopes do it the way you mention. Keysight Infiniium Scopes can separate each channel into its own grid, and each grid had its own scale, and is full-scale within that grid. Also waveforms can be overlayed, and each is full scale in the grid.
Yes this affects the resolution, here is an example on my DP3014 oscilloscope:
- I put the vertical scale on 1V / div- I apply a ± 5V signal- the signal exceeds the visible part of the screen which is 8 divisions, and therefore should cover 10 divisions.- If by GPIB I will see the values read, they barely reach the limit of an 8-bit converter, in this example + 125 / -126.- If I change the attenuator and go to 2V / div, the signal is fully visible on the screen, it occupies 5 divisions on the screen but the values read of the waveform are now half, so l equivalent of 7 bits.
A signal covering only one division will have the resolution of 4.68 bits: 2 ^ 4.68 = 25.6 for one division, therefore 256 for 10 divisions.
I'm looking for more information specifically on the auto-scaling algorithm.
The automatic vertical optimization is made to have the best possible resolution but on the visible part of the screen, generally 8 divisions (out of a total of 10). If you put a signal of ± 4V but you are on a scale of 2V / div for example, the AUTOSET will put it on 1V / div to have 8 divisions occupied by the signal.
However, if there is another channel activated, it will share the visible part of the screen for 2 channels with signals of ± 4V each, the scale will be 2V / div, one channel will occupy the 4 divisions of the the top, and the other the 4 divisions below.
For 4 channels, they will share via the AUTOSET automatic mode, 2 divisions each, giving a resolution of 5.68 bits per channel.
If you do some testing on your oscilloscope, you should be able to find the algorithm you are using for the vertical part.
You are absolutely right about the algorithm explained in my example, this method may be very different from one oscilloscope to another, I was only giving an example based on my DPO3014 and it is for this reason that I suggested to do some testing on its oscilloscope. As you explained, if each channel has its own grid, the resolution will be optimal, same thing if the traces are in overlay mode.
You are right, AUTOSCALE and AUTOSET are not the same function, I just wanted to say that when I use the AUTOSET function of my oscilloscope, the vertical scale is adjusted in the way I described, I should have be more explicit.
Retrieving data ...