Is more resolution available on an oscilloscope when pulling data over USB?

[jrmymllr]

This one has me perplexed. I'm using a Tek MDO series scope and was surprised to see samples pulled over the TekVISA USB interface of around 100V peak on the 20V/div range. The signal is a sinewave, and it's definitely clipping on the built-in display which should be topping out at 80V with the reference set at the middle.

I thought there's no way this is possible because why wouldn't the manufacturer want to display the full resolution? Or is this a matter of exceeding the range and possibly not being in spec, which is why it clips on the display?

[Berni]

The actual hardware dividers and the display range might not always be 1:1 since the scope doesn't always have as many hardware gain settings in the AFE as it has ranges. There is also some headroom for calibration in software. ADCs and AFEs can also act funny in transitions into and out of clipping.

For example DMMs will often show more digits when read out trough SCPI versus what is on the display. They simply output the full unmodified raw values it is working with internally. The extra digits are likely useless, but the numbers don't have to be neatly formatted for human readability. It just minimises any rounding errors by skipping the rounding and just giving you the raw value, since if you are using SCPI that number is going to be interpreted by a computer, not a human. So any math the computer does on that number won't have rounding errors stack up.

For example on spectrum analyzers it is common to be able to get more points per sweep over SCPI than the built in display can show.

[electr_peter]

Not a full resolution of ADC is mapped to visible screen, giving some "headroom" above or below. IIRC Keysight X3000T maps 240 values (out of 256) to visible screen. Not sure how Tek does this, look for vertical range specs in datasheets.