Re: [Electronics_101] are step response and frequency response transfer functions the same?

Expand Messages
• The transfer functions should be exactly the same - if your system is linear. If you are getting different results, I d suspect: - System is not linear. (Try
Message 1 of 7 , Mar 10, 2013
• 0 Attachment
The transfer functions should be exactly the same - if your system is
linear. If you are getting different results, I'd suspect:

- System is not linear. (Try repeating the measurements using lower signal
levels.) If the system includes an op amp, one source of nonlinearity and
additional delay could be the limited slew rate of the op amp. It would
cause square waves to look like trapezoids, and sine waves to look somewhat
like triangle waves.

- Measurement or calculation errors.

For the step response measurement, I assume you are using a square wave as
a stimulus. Is the pulse width much longer than the expected lag?

> Hello
>
> I have been given a task where I need to work out the transfer function
> for a system firstly using the step response method then secondly using
> the frequency responce method. The system has a transfer function that is
> a 1st order lag.
>
> I have done both and have plotted the bode plots for the frequency
> response method.
>
> My question is should the transfer function for the step response method
> and the frequency responce method be roughly the same? As it currently
> stands my gain is the same and yet my time contast is different.
>
> The break point on amplitude ratio plot corresponds nicley my -45 degree
> point on my phase shift plots, so at least theese agree.
>

[Non-text portions of this message have been removed]
• Hi thanks for the response. The system is linear. For the step response the input isna syeo square with a pulse wider than the lag. I am happy that the results
Message 2 of 7 , Mar 10, 2013
• 0 Attachment
Hi thanks for the response.

The system is linear.

For the step response the input isna syeo square with a pulse wider than
the lag. I am happy that the results from the step respinse ia correct.

I am sure that the frequency response is incorect. The test board comprises
of a DC motor to wich a board comprising some amplifiers are conncted. We
applied a sinusoidal input and measures the volatge output from the tacho.

We used a 10volt peak to peak input. This nearly caused the measured output
volatge of the motor to clip {flat line} due to the motor reaching its top
speed . We also used a 5volt offset to keep the sinusoidal inout always
posotive and therby driving the motor forward. Perhaps we ahould have used
a smaller imout voltage or perhaps not used a DC off set.

I see your point regarding the op amps of wich there are at least 2 but the
output volathe from both the step and frequencu response looks normal with
no trapzoiding and no traingular sine waves.

On Mar 10, 2013 7:00 PM, "Jan Kok" <jan.kok.5y@...> wrote:

> **
>
>
> The transfer functions should be exactly the same - if your system is
> linear. If you are getting different results, I'd suspect:
>
> - System is not linear. (Try repeating the measurements using lower signal
> levels.) If the system includes an op amp, one source of nonlinearity and
> additional delay could be the limited slew rate of the op amp. It would
> cause square waves to look like trapezoids, and sine waves to look somewhat
> like triangle waves.
>
> - Measurement or calculation errors.
>
> For the step response measurement, I assume you are using a square wave as
> a stimulus. Is the pulse width much longer than the expected lag?
>
> On Sun, Mar 10, 2013 at 12:18 PM, scubadogct <scubadogct@...> wrote:
>
> > Hello
> >
> > I have been given a task where I need to work out the transfer function
> > for a system firstly using the step response method then secondly using
> > the frequency responce method. The system has a transfer function that is
> > a 1st order lag.
> >
> > I have done both and have plotted the bode plots for the frequency
> > response method.
> >
> > My question is should the transfer function for the step response method
> > and the frequency responce method be roughly the same? As it currently
> > stands my gain is the same and yet my time contast is different.
> >
> > The break point on amplitude ratio plot corresponds nicley my -45 degree
> > point on my phase shift plots, so at least theese agree.
> >
>
> [Non-text portions of this message have been removed]
>
>
>

[Non-text portions of this message have been removed]
• ?? you have mistake? my question is bark detector  :-) by luciano  :-) ________________________________ From: Charlie Taylor To:
Message 3 of 7 , Mar 10, 2013
• 0 Attachment
?? you have mistake? my question is bark detector  :-)

by luciano  :-)

________________________________
To: Electronics_101@yahoogroups.com
Sent: Sunday, March 10, 2013 9:16 PM
Subject: Re: [Electronics_101] are step response and frequency response transfer functions the same?

Hi thanks for the response.

The system is linear.

For the step response the input isna syeo square with a pulse wider than
the lag. I am happy that the results from the step respinse ia correct.

I am sure that the frequency response is incorect. The test board comprises
of a DC motor to wich a board comprising some amplifiers are conncted. We
applied a sinusoidal input and measures the volatge output from the tacho.

We used a 10volt peak to peak input. This nearly caused the measured output
volatge of the motor to clip {flat line} due to the motor reaching its top
speed . We also used a 5volt  offset to keep the sinusoidal inout always
posotive and therby driving the motor forward. Perhaps we ahould have used
a smaller imout voltage or perhaps not used a DC off set.

I see your point regarding the op amps of wich there are at least 2 but the
output volathe from both the step and frequencu response looks normal with
no trapzoiding and no traingular sine waves.

On Mar 10, 2013 7:00 PM, "Jan Kok" <jan.kok.5y@...> wrote:

> **
>
>
> The transfer functions should be exactly the same - if your system is
> linear. If you are getting different results, I'd suspect:
>
> - System is not linear. (Try repeating the measurements using lower signal
> levels.) If the system includes an op amp, one source of nonlinearity and
> additional delay could be the limited slew rate of the op amp. It would
> cause square waves to look like trapezoids, and sine waves to look somewhat
> like triangle waves.
>
> - Measurement or calculation errors.
>
> For the step response measurement, I assume you are using a square wave as
> a stimulus. Is the pulse width much longer than the expected lag?
>
> On Sun, Mar 10, 2013 at 12:18 PM, scubadogct <scubadogct@...> wrote:
>
> > Hello
> >
> > I have been given a task where I need to work out the transfer function
> > for a system firstly using the step response method then secondly using
> > the frequency responce method. The system has a transfer function that is
> > a 1st order lag.
> >
> > I have done both and have plotted the bode plots for the frequency
> > response method.
> >
> > My question is should the transfer function for the step response method
> > and the frequency responce method be roughly the same? As it currently
> > stands my gain is the same and yet my time contast is different.
> >
> > The break point on amplitude ratio plot corresponds nicley my -45 degree
> > point on my phase shift plots, so at least theese agree.
> >
>
> [Non-text portions of this message have been removed]
>

>

[Non-text portions of this message have been removed]

------------------------------------

[Non-text portions of this message have been removed]
• what has my question got to do with a bark detector? ... [Non-text portions of this message have been removed]
Message 4 of 7 , Mar 10, 2013
• 0 Attachment
what has my question got to do with a bark detector?
On Mar 10, 2013 8:34 PM, "Lucianopep Luciano" <lucianopep@...> wrote:

> **
>
>
> ?? you have mistake? my question is bark detector :-)
>
> by luciano :-)
>
> ________________________________
> To: Electronics_101@yahoogroups.com
> Sent: Sunday, March 10, 2013 9:16 PM
> Subject: Re: [Electronics_101] are step response and frequency response
> transfer functions the same?
>
> Hi thanks for the response.
>
> The system is linear.
>
> For the step response the input isna syeo square with a pulse wider than
> the lag. I am happy that the results from the step respinse ia correct.
>
> I am sure that the frequency response is incorect. The test board comprises
> of a DC motor to wich a board comprising some amplifiers are conncted. We
> applied a sinusoidal input and measures the volatge output from the tacho.
>
> We used a 10volt peak to peak input. This nearly caused the measured output
> volatge of the motor to clip {flat line} due to the motor reaching its top
> speed . We also used a 5volt offset to keep the sinusoidal inout always
> posotive and therby driving the motor forward. Perhaps we ahould have used
> a smaller imout voltage or perhaps not used a DC off set.
>
> I see your point regarding the op amps of wich there are at least 2 but the
> output volathe from both the step and frequencu response looks normal with
> no trapzoiding and no traingular sine waves.
>
> On Mar 10, 2013 7:00 PM, "Jan Kok" <jan.kok.5y@...> wrote:
>
> > **
> >
> >
> > The transfer functions should be exactly the same - if your system is
> > linear. If you are getting different results, I'd suspect:
> >
> > - System is not linear. (Try repeating the measurements using lower
> signal
> > levels.) If the system includes an op amp, one source of nonlinearity and
> > additional delay could be the limited slew rate of the op amp. It would
> > cause square waves to look like trapezoids, and sine waves to look
> somewhat
> > like triangle waves.
> >
> > - Measurement or calculation errors.
> >
> > For the step response measurement, I assume you are using a square wave
> as
> > a stimulus. Is the pulse width much longer than the expected lag?
> >
> > On Sun, Mar 10, 2013 at 12:18 PM, scubadogct <scubadogct@...>
> wrote:
> >
> > > Hello
> > >
> > > I have been given a task where I need to work out the transfer function
> > > for a system firstly using the step response method then secondly using
> > > the frequency responce method. The system has a transfer function that
> is
> > > a 1st order lag.
> > >
> > > I have done both and have plotted the bode plots for the frequency
> > > response method.
> > >
> > > My question is should the transfer function for the step response
> method
> > > and the frequency responce method be roughly the same? As it currently
> > > stands my gain is the same and yet my time contast is different.
> > >
> > > The break point on amplitude ratio plot corresponds nicley my -45
> degree
> > > point on my phase shift plots, so at least theese agree.
> > >
> >
> > [Non-text portions of this message have been removed]
> >
> >
> >
>
> [Non-text portions of this message have been removed]
>
> ------------------------------------
>
>
> [Non-text portions of this message have been removed]
>
>
>

[Non-text portions of this message have been removed]
• ... In theory. :-) ... I don t think the DC offset should cause problems. I would suggest you keep the 5V DC offset and try a smaller sine wave and square
Message 5 of 7 , Mar 10, 2013
• 0 Attachment
On Sun, Mar 10, 2013 at 2:16 PM, Charlie Taylor <scubadogct@...>wrote:

> Hi thanks for the response.
>
> The system is linear.
>

In theory. :-)

>
> For the step response the input isna syeo square with a pulse wider than
> the lag. I am happy that the results from the step respinse ia correct.
>
> I am sure that the frequency response is incorect. The test board comprises
> of a DC motor to wich a board comprising some amplifiers are conncted. We
> applied a sinusoidal input and measures the volatge output from the tacho.
>
> We used a 10volt peak to peak input. This nearly caused the measured output
> volatge of the motor to clip {flat line} due to the motor reaching its top
> speed . We also used a 5volt offset to keep the sinusoidal inout always
> posotive and therby driving the motor forward. Perhaps we ahould have used
> a smaller imout voltage or perhaps not used a DC off set.
>

I don't think the DC offset should cause problems.

I would suggest you keep the 5V DC offset and try a smaller sine wave and
square wave. (Say 5V P-P instead of 10V P-P.) If everything is linear, the
transfer functions you get should all be the same regardless of input
voltage.

You've probably got an op amp driving some transistors that drive the
motor. Do the transistors create a linear amplifier?

One problem might be that the transistor amplifier might not be designed to
actively pull the output voltage toward ground, if the motor is spinning
faster than the control signal is telling it to go. Thus the motor would
slow down more slowly than it speeds up. That would look like a lag.

You could replace the motor with a resistor and capacitor in series, where
R is the motor resistance and C is chosen to store the same amount of
energy (1/2 C V^2) as the kinetic energy in the motor's rotor at the same
voltage. The voltage across C would be proportional to the motor speed if
you had the motor in the circuit. Replacing the motor with the RC gets rid
of motor noise and eliminates the tachometer as sources of confusion.

>
> I see your point regarding the op amps of wich there are at least 2 but the
> output volathe from both the step and frequencu response looks normal with
> no trapzoiding and no traingular sine waves.
>

Have you looked at the tachometer output signal? I would expect it to look
like a low pass filtered version of the input signal. It should also look
like the capacitor voltage mentioned above.

[Non-text portions of this message have been removed]
• the responses should be the same. If they are not, then it quite likely that your either or both the step or sinusoidal excitation is too large causing the
Message 6 of 7 , Mar 11, 2013
• 0 Attachment
the responses should be the same. If they are not, then it quite likely that your either or both the step or sinusoidal excitation is too large causing the implementation to hit limits, step sizes of about 5% of rated quantities is a reasonable place to start.

>
> Hello
>
> I have been given a task where I need to work out the transfer function for a system firstly using the step response method then secondly using the frequency responce method. The system has a transfer function that is a 1st order lag.
>
> I have done both and have plotted the bode plots for the frequency response method.
>
> My question is should the transfer function for the step response method and the frequency responce method be roughly the same? As it currently stands my gain is the same and yet my time contast is different.
>
> The break point on amplitude ratio plot corresponds nicley my -45 degree point on my phase shift plots, so at least theese agree.
>
>
> Thanks
>
Your message has been successfully submitted and would be delivered to recipients shortly.