oscilloscope (bandwidth & sampling rates)

[Sam Kaan]

although I know the bandwidth of a scope determines what kind of
circuit you can analyze as far as the circuit speed is concerned.
WHat I am unclear about is the bandwidth of the scope. Suppose you
have a scope that has a bandwidth of 100Mhz, for example this one:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2553989680&category=45005

This probably means the circuit you want to analyze, lets say a
computer operating at 100Mhz is all you can analyze. Now suppose the
same scope has very very high sampling rates, say 2Gs/s. Does this
mean that technically you can in fact analyze a circuit operating at
even higher frequency that the rated 100Mhz. Sorry someone told me
this and now I am confused.

I am trying to get my information all cleared up before proceeding to
purchase a used scope. The reason I like this one Tektronic 7603
(made back in 1970s) is because it can take a Logic Analyzer plugin.
So once u have that you basically have both an oscilloscope and a
logic analyzer all in the same unit.
Is it common to find oscilloscope with logic analyzer plug ins? What
about scope that can work as both an oscilloscope and a logic
analyzer?

 

[John Larkin]

although I know the bandwidth of a scope determines what kind of
circuit you can analyze as far as the circuit speed is concerned.
WHat I am unclear about is the bandwidth of the scope. Suppose you
have a scope that has a bandwidth of 100Mhz, for example this one:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2553989680&category=45005

This probably means the circuit you want to analyze, lets say a
computer operating at 100Mhz is all you can analyze. Now suppose the
same scope has very very high sampling rates, say 2Gs/s. Does this
mean that technically you can in fact analyze a circuit operating at
even higher frequency that the rated 100Mhz. Sorry someone told me
this and now I am confused.

I am trying to get my information all cleared up before proceeding to
purchase a used scope. The reason I like this one Tektronic 7603
(made back in 1970s) is because it can take a Logic Analyzer plugin.
So once u have that you basically have both an oscilloscope and a
logic analyzer all in the same unit.
Is it common to find oscilloscope with logic analyzer plug ins? What
about scope that can work as both an oscilloscope and a logic
analyzer?

A 7603 is an analog scope, so it doesn't have a sample rate.

Digital scopes have sample rates, but the rate may range anywhere from
2x the bandwidth (that's the theoretical Nyquist rate... practical
rates are more like 4x) for single-shot scopes to a tiny fraction of
the bandwidth (like, say, 1/100,000) for an equivalent-time sampler.
My 11801 is a 50 GHz scope with a max sample rate of 200 KHz. A
typical Tek lcd digital scope has a 100 MHz bandwidth and a 1 GHz
sample rate, but it's still just a 100 MHz scope.

The logic analyzers that plug into a 7603 are very limited, and not
useful except for very simple circuits. The best logic analyzer is the
one inside your head.

John

 

[Dr. Anton Squeegee]

My 11801 is a 50 GHz scope with a max sample rate of 200 KHz. A
typical Tek lcd digital scope has a 100 MHz bandwidth and a 1 GHz
sample rate, but it's still just a 100 MHz scope.

50 (fifty) Gigs? Are you absolutely sure about that? I recall
reading somewhere (probably an old Tek catalog) that the 11801 was good
to 4 GHz.

How'd you come up with 50? That's a heck of a bandwidth, even for
the highest-end spectrum analyzers.

 

[John Larkin]

50 (fifty) Gigs? Are you absolutely sure about that? I recall
reading somewhere (probably an old Tek catalog) that the 11801 was good
to 4 GHz.

How'd you come up with 50? That's a heck of a bandwidth, even for
the highest-end spectrum analyzers.

The bandwidth is determined by which sampling head you plug in. The
fastest for the 11801 was indeed 50 GHz. Ask google for details. My
scopes have heads that go to 12 and 20 GHz, which is plenty fast
enough for me. The 50G sampling heads only come single-channel, and
don't do TDR.

HP and Tek are up to about 70 GHz on their latest sampling scopes.

Picosecond Pulse Labs has a 100 GHz sampler, but so far I don't think
anybody has built a full scope with it.

John

 

[Mathew Orman]

although I know the bandwidth of a scope determines what kind of
circuit you can analyze as far as the circuit speed is concerned.
WHat I am unclear about is the bandwidth of the scope. Suppose you
have a scope that has a bandwidth of 100Mhz, for example this one:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2553989680&category=45005

This probably means the circuit you want to analyze, lets say a
computer operating at 100Mhz is all you can analyze. Now suppose the
same scope has very very high sampling rates, say 2Gs/s. Does this
mean that technically you can in fact analyze a circuit operating at
even higher frequency that the rated 100Mhz. Sorry someone told me
this and now I am confused.

I am trying to get my information all cleared up before proceeding to
purchase a used scope. The reason I like this one Tektronic 7603
(made back in 1970s) is because it can take a Logic Analyzer plugin.
So once u have that you basically have both an oscilloscope and a
logic analyzer all in the same unit.
Is it common to find oscilloscope with logic analyzer plug ins? What
about scope that can work as both an oscilloscope and a logic
analyzer?

High sampling rate gives the ability to capture single pulse events.
The 100MHz bandwidth means that you can still see 500MHz components
but it will be attenuated.

Sincerely,

Mathew Orman
www.ultra-faster-than-light.com
www.radio-faster-than-light.com

 

[jmc]

although I know the bandwidth of a scope determines what kind of
circuit you can analyze as far as the circuit speed is concerned.
WHat I am unclear about is the bandwidth of the scope. Suppose you
have a scope that has a bandwidth of 100Mhz, for example this one:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2553989680&category=45005

This probably means the circuit you want to analyze, lets say a
computer operating at 100Mhz is all you can analyze. Now suppose the
same scope has very very high sampling rates, say 2Gs/s. Does this
mean that technically you can in fact analyze a circuit operating at
even higher frequency that the rated 100Mhz. Sorry someone told me
this and now I am confused.

This is a resonse to a similar question that Mike Monet gave some time ago.
Note that you'll have to paste the link as one line into your browser
manually:-

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=3D56DEBD
..4697%40sneakemail.com&rnum=1&prev=/groups%3Fhl%3Den%26lr%3D%26ie%3DUTF-8%26
oe%3DUTF-8%26q%3Dmike%2Bsampling%2B%2Bsally%2Bsci.electronics.design


Well, the problem is a little more complicated.

You run into a problem sampling a 1 GHz signal at one sample per cycle. You
don't know which point on the waveform you are sampling, so a single data
point
per cycle doesn't convey useful information.

It's like watching a western on tv, where the camera frame rate happens to
match the rpm of the wagon wheels. Even though the horses are running as
fast
as they can, and a lot of dust is being kicked up, the wagon wheels may
appear
to be stationary, or even moving backwards. This is called aliasing. Here
are
some examples of aliasing problems in electrical waveforms:

http://euphoria.org/home/help/nyquist.html
http://www.cs.cf.ac.uk/Dave/Multimedia/node149.html
http://www.cage.curtin.edu.au/mechanical/info/vibrations/tut3.htm
http://ptolemy.eecs.berkeley.edu/eecs20/week13/nyquistShannon.html

Nyquist (and Shannon) showed the minimum number of samples needed to avoid
aliasing is two samples per cycle. With a 1 GHz sampler, the highest
frequency
component you can allow in the system is 500 MHz.

This means you need a very good filter to prevent strong tv stations between
500 MHz and 1 GHz from causing problems.

It turns out there is no physical filter than can provide linear phase up to
the cutoff frequency, and infinite rejection past it. But you knew that
already

So there's a compromise. The filter has to start rolling off before 500 MHz,
which means you start attenuating the highest harmonics of the signal you
are
interested in.

Fourier showed a complex signal can be broken into a fundamental plus odd
and
even harmonics. Good engineering practise recommends including all harmonics
at
least up to the fifth, which means the fundamental frequency in the waveform
must be 100 MHz or below.

Does this help answer your question?

73, Mike