Discrete-time systems

On 11/3/2012 10:36 PM, dbd wrote:
> On Saturday, November 3, 2012 7:30:29 AM UTC-7, Piergiorgio Sartor wrote:
>> On 2012-11-02 21:48, gyansorova@gmail.com wrote:
>>
>> [...]
> ...
>> What makes you think analogue system do not have delay?
>>
>> First of all, there is an upper limit, which is the speed
>> of "electrons" in circuit (far below the speed of light).
>> Second, you can always have a combination of analogue
>> components (LRC) causing a delay.
>>
>>
>>
>> In the end, analogue delay lines do exist.
>> ...
>> piergiorgio
>
> The delay in analog circuits is the time until the electrons start moving according to the signal, not 'how long does it take electrons to get there'. This is the propagation time of the electric field not the electrons. Speed of light in a vacuum, less in matter, as glen measured.
>
> Dale B. Dalrymple

Actually all filters have "delay", not just digital ones.  It is a 
consequence of frequency being related to time.  How can you distinguish 
any two frequencies instantaneously?  For that matter, how do you make 
*any* measurement instantaneously?  You don't.  All measurements and all 
computations take time no matter how they are done.

Rick

On 2012-11-04 20:54, rickman wrote:
[...]
> Actually all filters have "delay", not just digital ones.  It is a
> consequence of frequency being related to time.  How can you distinguish
> any two frequencies instantaneously?  For that matter, how do you make
> *any* measurement instantaneously?  You don't.  All measurements and all
> computations take time no matter how they are done.

This reminds me a story from the University.

Once a professor was asking if, given a tester to measure
*DC* voltage, its frequency response matters.

The student naive answer was: no, of course, it's DC!

Then the professor explained that when we connect the
tester to the measurement points, we feed its input with
a *step* signal, not a constant one.
If the tester has some low-pass filter in order to
filter our noise, depending on this one, it could take
time (delay?) before the readout reaches within the
precision (or accuracy, I always mix them up) of the
instrument.

bye,

-- 

piergiorgio

On Sunday, November 4, 2012 11:55:10 AM UTC-8, rickman wrote:
> On 11/3/2012 10:36 PM, dbd wrote:
> 
> > On Saturday, November 3, 2012 7:30:29 AM UTC-7, Piergiorgio Sartor wrote:

> > ...
> >> What makes you think analogue system do not have delay?
> >>
> >> First of all, there is an upper limit, which is the speed
> >> of "electrons" in circuit (far below the speed of light).
> >> ...
> >> piergiorgio
> 
> >
> 
> > The delay in analog circuits is the time until the electrons start moving according to the signal, not 'how long does it take electrons to get there'. This is the propagation time of the electric field not the electrons. Speed of light in a vacuum, less in matter, as glen measured.
> > Dale B. Dalrymple
>
> Actually all filters have "delay", not just digital ones.  It is a 
> consequence of frequency being related to time.  How can you distinguish 
> any two frequencies instantaneously?  For that matter, how do you make 
> *any* measurement instantaneously?  You don't.  All measurements and all 
> computations take time no matter how they are done.
>
> Rick

Propagation delay does not include measurement delay unless the measurement is part of the definition of the 'system'. The 'system' is a filter, not a filter plus a measurement.

Dale B. Dalrymple

On 11/4/2012 5:16 PM, dbd wrote:
> On Sunday, November 4, 2012 11:55:10 AM UTC-8, rickman wrote:
>> On 11/3/2012 10:36 PM, dbd wrote:
>>
>>> On Saturday, November 3, 2012 7:30:29 AM UTC-7, Piergiorgio Sartor wrote:
>
>>> ...
>>>> What makes you think analogue system do not have delay?
>>>>
>>>> First of all, there is an upper limit, which is the speed
>>>> of "electrons" in circuit (far below the speed of light).
>>>> ...
>>>> piergiorgio
>>
>>>
>>
>>> The delay in analog circuits is the time until the electrons start moving according to the signal, not 'how long does it take electrons to get there'. This is the propagation time of the electric field not the electrons. Speed of light in a vacuum, less in matter, as glen measured.
>>> Dale B. Dalrymple
>>
>> Actually all filters have "delay", not just digital ones.  It is a
>> consequence of frequency being related to time.  How can you distinguish
>> any two frequencies instantaneously?  For that matter, how do you make
>> *any* measurement instantaneously?  You don't.  All measurements and all
>> computations take time no matter how they are done.
>>
>> Rick
>
> Propagation delay does not include measurement delay unless the measurement is part of the definition of the 'system'. The 'system' is a filter, not a filter plus a measurement.
>
> Dale B. Dalrymple

My point is that the filter is making a measurement of the frequency of 
the signal in order to filter it.  That is why the delay of filters 
relate to the frequencies they filter, or more accurately, how closely 
they distinguish the frequencies of the signals, i.e. transition band 
steepness.  Or am I mistaken about this relationship?

Rick

On Sunday, November 4, 2012 4:35:28 PM UTC-8, rickman wrote:
...
> My point is that the filter is making a measurement of the frequency of 
> the signal in order to filter it.  That is why the delay of filters 
> relate to the frequencies they filter, or more accurately, how closely 
> they distinguish the frequencies of the signals, i.e. transition band 
> steepness.  Or am I mistaken about this relationship?
> 
> Rick

I think that you have a point that there is a relationship between frequency response and propagation delay. I don't think that it is related to a 'measurement'. For a filter that is linear time invariant, as we usually assume for passive systems, the filter characteristic is the same regardless of the values of the signal. There is no point in the filter that generates a measurement value and acts on it, although you can invent more complicated systems that do.

Dale B. Dalrymple

On 11/4/2012 12:28 AM, robert bristow-johnson wrote:

   ...

 > sounds like a sorta complicated "impedance".  even with silicon
> inside, it's a "passive" device.

Maybe, maybe not. I used to have a piece of what looked like 50-ohm 
co-ax with BNCs on both ends. It actually had a CMOS divide-by-4, a 
small capacitor, and a diode embedded in the inner insulator. The chip 
was powered by the signal source and its output fed the other end. With 
the cable connecting a square-wave generator to a scope. the trace 
amplitude followed the generator. So did the frequency if tou didn't 
notice that the displayed frequency was only a quarter of what the 
generator indicated it should be. It was the source of some fun, but was 
it passive?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

On 11/3/2012 4:02 PM, Piergiorgio Sartor wrote:
 > On 2012-11-03 20:14, glen herrmannsfeldt wrote:
 > [...]
 >> In an undergrad physics lab, we were measuring the
 >> properties of coax cables. The experiment was a little
 >> to teach about the physics, and more to teach about data
 >> analysis, but it is applicable here.
 > [...]
 >
 > Actually, this reminds me something.
 >
 > If I correctly recall, during World War II, someone was
 > designing FIR filters, using cables of a certain length
 > as delay elements and capacitors as "coefficients".
 >
 > It was a fully analogue FIR filter, not switched capacitors,
 > just delayed capacitors...

TV vestigial-sideband filters were designed using open and shorted co-ax 
stubs located along the length of the feed cable. Commercial (that is, 
high-powered) filters used waveguide stubs for the same purpose. I may 
still have my class notes and Smith charts somewhere.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

On Friday, November 2, 2012 10:14:49 PM UTC+13, gyans...@gmail.com wrote:
> Suppose you convert an analogue system with no time-delay to digital.
> 
> Say G(s) to G(z^-1) in backward shift format.
> 
> Intuitively, should the discrete-time version always have at least a 1 step delay
> 
> within it? ie can information travel from input to output instantaneously in the z-domain?

First of all I wasn't saying that information travels instantaneously in an analogue system. I have to distinguish what is normal rise-time due to a LTI system and that of a pure time-delay of at least 1 step sampling interval.
The latter has a nasty phase characteristic which will always be there nomatter what the sampling interval (though its effect will be less the higher the sampling rate) - there has to be a 1 step delay around a loop in the z-domain. The equivalent analogue would be exp(-sT) of course which is z^-1. 

On 11/4/2012 11:19 PM, dbd wrote:
> On Sunday, November 4, 2012 4:35:28 PM UTC-8, rickman wrote:
> ...
>> My point is that the filter is making a measurement of the frequency of
>> the signal in order to filter it.  That is why the delay of filters
>> relate to the frequencies they filter, or more accurately, how closely
>> they distinguish the frequencies of the signals, i.e. transition band
>> steepness.  Or am I mistaken about this relationship?
>>
>> Rick
>
> I think that you have a point that there is a relationship between frequency response and propagation delay. I don't think that it is related to a 'measurement'. For a filter that is linear time invariant, as we usually assume for passive systems, the filter characteristic is the same regardless of the values of the signal. There is no point in the filter that generates a measurement value and acts on it, although you can invent more complicated systems that do.
>
> Dale B. Dalrymple
>

I'm not sure exactly what your point is.  But the issue of delay in 
filters is fundamentally related to the process of measurement.  The 
filter does not need to produce an intermediate result that itself can 
be observed.

The delay in a filter is a fundamental result of the resolution of a 
property which is based on time.

Rick

On Monday, November 5, 2012 8:52:02 AM UTC-8, rickman wrote:
...
> I'm not sure exactly what your point is.  But the issue of delay in 
> filters is fundamentally related to the process of measurement.  The 
> filter does not need to produce an intermediate result that itself can 
> be observed.
> 
> The delay in a filter is a fundamental result of the resolution of a 
> property which is based on time.
>
> Rick

In engineering, a measurement is the determination of the ratio of a physical parameter to a standard unit for that parameter. Resolution is, for example:
 "The act or process of separating or reducing something into its constituent parts" http://www.thefreedictionary.com/resolution

I think that Ls, Rs , and Cs in simple passive networks are able to filter and delay perfectly well without performing any measurement or resolving anything into constituent parts. I think such filters work on the properties of Ls, Rs, and Cs and don't require any metaphysical hand waving about 'measurements' or 'resolution'. If you want to understand or predict performance, I think you should look at the values of the Ls, Rs, and Cs and not some 'fundamental' metaphysics. There are functional physical interpretations.

Dale B. Dalrymple