Scope bandwidth slope beyond max spec.

[gfmucci]

Using the Instek as my example:

I understand that beyond the 50MHz bandwidth spec of this scope that the downward slope of amplitude is quite gradual.  That is a plus.
I also understand that there are 70 MHz and 100 MHz unlockable options for this scope.

Question:  Will the amplitude slopes beyond these optional installed 70 and 100 bandwidth spec options have a steeper, sharper amplitude reduction than what it was with the base 50MHz model?  In other words, will these higher bandwidth options more sharply reduce the gradual amplitude slope of the 50 MHz model beyond their higher spec?

[David Hess]

The frequency response continues to decline , usually at about 6dB per octave, however the slope is not well defined and will have something between a Bessel and Gaussian response depending on the design margins.  Higher bandwidth oscilloscopes are more likely to have an aberrant or sharper controlled response because it is more difficult to maintain a Gaussian response at higher frequencies.

[gfmucci]

Can you tell me anything specifically about the amplitude slopes of the 50, 70, and 100 MHz Insteks?  Do the slopes drop off more quickly in the two higher bandwidth models?

[T3sl4co1l]

Also, if it's good, it'll have a Bessel or x dB Gaussian shape.

Bad old Teks (from the TDS2xx and other families I think??) just kind of did whatever, often peaking their response, presumably to hit risetime targets, and.. not much else?

Classic Teks, from the tube days, using distributed amplifiers, I think had extremely sharp cutoffs -- each stage was designed for a reasonably soft Bessel profile, but all together, plus the lumped-equivalent delay line, made for an aggressive asymptote -- on the upside, the Tek 585 gave you a flat 85MHz+ of bandwidth, in an almost/all vacuum tube design, back in 1959!

A more reasonable example from their mid age lines, my TDS460 -- the analog front end seems to be well behaved, dropping off gradually beyond the rated 350MHz.  Lots of custom (or at least specialized) ceramic+gold chips in there, the front end and ADCs.  I've observed signals (with stable trigger) over 650MHz with it.  (Main downside: the equivalent time sampling is preposterously slow on the lowest time/div settings.  Modern scopes are almost exclusively real-time, and this isn't a problem.)

Tim

[David Hess]

Can you tell me anything specifically about the amplitude slopes of the 50, 70, and 100 MHz Insteks?  Do the slopes drop off more quickly in the two higher bandwidth models?

For lower bandwidth oscilloscopes, the way it usually works is that the maximum bandwidth is the combined response of all of the various stages with some added equalization producing something approaching a -6dB per octave Gaussian response with the slope increasing at higher frequencies.  Then the lower bandwidth settings add a simple filter to produce a more accurate -6 dB per octave Bessel response.

If you really want to know, then it will be necessary to measure it.