Modern digital scopes: real-time sampling or equivalent-time sampling

[David Hess]

I did not know that Phillips (or anybody else) made CCD based digital storage oscilloscopes.  That places the PM3320 in the same class as the Tektronix 2440 series of DSOs and neatly explains how they achieved 10 bits of resolution at 250 Msamples/second at that time.  The service manual says that the cycle time on the output of the CCD is 2.5 microseconds and the ADC is a Burr-Brown ADC803 which uses variable resolution successive approximation of 500ns 8 bits, 670ns 10 bits, and 1.5uS 12 bits.  The lowest end of current Tektronix models still use CCD sampling.

The PM3340 is a completely different beast; it is a *real* digital sampling oscilloscope with a 2 GHz bandwidth and 50 Gsample/second sequential equivalent time sample rate.  Amazingly they included delay lines which is unusual in a sampling oscilloscope but they make it much more useful.  I remember finding a reference saying that its real time sample rate is like 50 kSamples/second which is about right; that is actually fast enough to produce a real time display.

Phillips was not the only one using high resolution vector CRTs.  Many of the early Tektronix DSOs like the 7854 and 2232 use 1024x1024 on their 5:4 5 inch diagonal CRTs producing 260x325 dpi.  Some of the early HP DSOs with raster CRTs had doubled horizontal resolution.

I think LeCroy makes some stand alone real time DSOs with 12 bits of resolution but they are expensive.