Changed to [EE]: as it is more appropriate.

> Me and a friend from my university are working on a PC-based oscilloscope.
> We were just wondering if anyone had any suggestionns on how deep
> the memory
> on such a thing should be etc...

As much as possible within any addressing restrictions you may have :)

I use a TEK TDS3000 series scope, and that has a (relatively) small buffer,
sufficient for one screen full at the timebase rate. When endeavouring to
look close in at a waveform by zooming in then there are distinct
limitations on how much magnification can be achieved without loosing the
clarity of the waveform.

A colleague has an Agilent scope which has much more capture memory  and
this allows one to zoom in on an aspect of a waveform without loosing the
clarity of the detail, making it a lot easier to see detail on glitch
problems and the like.

As a rule of thumb, I would say your capture memory should allow
magnification of 20x to 50x the screen size, based on my attempts at glitch
capture. However there may be other restrictions on what you can do that
could restrict this.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.