Splitting an error signal into DC and High Freq.

[George Herold]

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the
cap., and using an instrument amp to read the voltage across the
resistor.


Error in----+---.
| |
R +--> Instrument
R -->High Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out
|
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

 

[Jim Thompson]

On Tue, 16 May 2017 09:07:58 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the
cap., and using an instrument amp to read the voltage across the
resistor.


Error in----+---.
| |
R +--> Instrument
R -->High Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out
|
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

Look up "cross-over networks".

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Thinking outside the box... producing elegant solutions.

"It is not in doing what you like, but in liking what you do that
is the secret of happiness." -James Barrie

 

[George Herold]

On Tuesday, May 16, 2017 at 12:14:09 PM UTC-4, Jim Thompson wrote:

On Tue, 16 May 2017 09:07:58 -0700 (PDT), George Herold
gherold@teachspin.com> wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the
cap., and using an instrument amp to read the voltage across the
resistor.


Error in----+---.
| |
R +--> Instrument
R -->High Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out
|
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

Look up "cross-over networks".

You mean like an LC thing? (I'd have to set the Q.)
I was thinking I could use an active filter. (S-V) and take
the HP and LP outputs. But don't I get more phase shift that
way?

George H.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Thinking outside the box... producing elegant solutions.

"It is not in doing what you like, but in liking what you do that
is the secret of happiness." -James Barrie

 

[Jim Thompson]

On Tue, 16 May 2017 09:43:59 -0700 (PDT), George Herold
<gherold@teachspin.com> wrote:

On Tuesday, May 16, 2017 at 12:14:09 PM UTC-4, Jim Thompson wrote:
On Tue, 16 May 2017 09:07:58 -0700 (PDT), George Herold
gherold@teachspin.com> wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the
cap., and using an instrument amp to read the voltage across the
resistor.


Error in----+---.
| |
R +--> Instrument
R -->High Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out
|
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

Look up "cross-over networks".
You mean like an LC thing? (I'd have to set the Q.)
I was thinking I could use an active filter. (S-V) and take
the HP and LP outputs. But don't I get more phase shift that
way?

No. Run the math.

George H.

...Jim Thompson
[snip]

You can easily make cross-overs with active filters.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Thinking outside the box... producing elegant solutions.

"It is not in doing what you like, but in liking what you do that
is the secret of happiness." -James Barrie

 

[George Herold]

On Tuesday, May 16, 2017 at 12:48:26 PM UTC-4, Jim Thompson wrote:

On Tue, 16 May 2017 09:43:59 -0700 (PDT), George Herold
gherold@teachspin.com> wrote:

On Tuesday, May 16, 2017 at 12:14:09 PM UTC-4, Jim Thompson wrote:
On Tue, 16 May 2017 09:07:58 -0700 (PDT), George Herold
gherold@teachspin.com> wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the
cap., and using an instrument amp to read the voltage across the
resistor.


Error in----+---.
| |
R +--> Instrument
R -->High Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out
|
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

Look up "cross-over networks".
You mean like an LC thing? (I'd have to set the Q.)
I was thinking I could use an active filter. (S-V) and take
the HP and LP outputs. But don't I get more phase shift that
way?

No. Run the math.
What, you want me to work for a living? :^)

Did I say the output is for a control loop?

A two pole filter has to have 180 degree phase shift across the
frequency band. RC has only 90 degree's. I don't know if that
matters for a control loop.. and what happens in the cross-over
region.

I did this ~10 years ago. I could go look in my notebook and (perhaps)
find it. It's easier just to do it again from scratch.

Is there anything wrong with my RC approach? (I've got it wired up,
I'll test it tomorrow.)

George H.

...Jim Thompson
[snip]

You can easily make cross-overs with active filters.

I may very well have used an active filter last time.

George H.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Thinking outside the box... producing elegant solutions.

"It is not in doing what you like, but in liking what you do that
is the secret of happiness." -James Barrie

 

[Tim Wescott]

On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV



--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!

 

[George Herold]

On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:

On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the x-over
freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to add
the BP too? (My 'scope can't sum three signals.)

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

George H.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!

 

[Tim Wescott]

On Tue, 16 May 2017 12:33:43 -0700, George Herold wrote:

On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High
Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. | | | ---
| | ---
'-' |
| | o-------------------------------o
LF out | |
| o-----------------o HF out |
|
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the
x-over freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to
add the BP too? (My 'scope can't sum three signals.)

2nd-order? Of course -- think about it:

LP: w^2 / (s^2 + 2*zeta*w*s + w^2)
HP: s^2 / (s^2 + 2*zeta*w*s + w^2)

LP + HP = (s^2 + w^2) / (etc.), which is a plain ol' notch filter.

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

Absent adjustability I belive my way is better (and they do sum to the
original signal).

But yes, if you need to tweak things, an instrumentation amp is probably
better than a dual pot.

--
www.wescottdesign.com

 

[rickman]

On 5/16/2017 3:33 PM, George Herold wrote:

On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the x-over
freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to add
the BP too? (My 'scope can't sum three signals.)

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

I assume you mean me? Thanks for the consideration, but in this context
those abbreviations were clear to me. I don't understand what you mean
about the LP + HP summing to zero though. I assume you mean they sum to
be the same as the input?

--

Rick C

 

[George Herold]

On Tuesday, May 16, 2017 at 4:17:37 PM UTC-4, Tim Wescott wrote:

On Tue, 16 May 2017 12:33:43 -0700, George Herold wrote:

On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High
Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. | | | ---
| | ---
'-' |
| | o-------------------------------o
LF out | |
| o-----------------o HF out |
|
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the
x-over freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to
add the BP too? (My 'scope can't sum three signals.)

2nd-order? Of course -- think about it:

LP: w^2 / (s^2 + 2*zeta*w*s + w^2)
HP: s^2 / (s^2 + 2*zeta*w*s + w^2)

LP + HP = (s^2 + w^2) / (etc.), which is a plain ol' notch filter.

Right, it made sense afterwards.
Hey, is this analogy correct?
So in an SV filter with HP, BP and LP output, can I think it as
a series LRC circuit, With HP the signal across an ideal inductor,
BP that across the resistor, and LP across an ideal capacitor.

That seems right.

George H.

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

Absent adjustability I belive my way is better (and they do sum to the
original signal).

But yes, if you need to tweak things, an instrumentation amp is probably
better than a dual pot.

--
www.wescottdesign.com

 

[George Herold]

On Tuesday, May 16, 2017 at 9:38:58 PM UTC-4, rickman wrote:

On 5/16/2017 3:33 PM, George Herold wrote:
On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the x-over
freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to add
the BP too? (My 'scope can't sum three signals.)

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

I assume you mean me? Thanks for the consideration, but in this context
those abbreviations were clear to me. I don't understand what you mean
about the LP + HP summing to zero though. I assume you mean they sum to
be the same as the input?

No, the sum is zero.. no signal. at "resonance" the LP and HP amplitudes
are equal, and the phase difference is 180 deg.

As Tim said I already knew this, 'cause you can make a notch by adding
LP and HP... but for me knowing (book-wise) and doing are two different
things.

George H.

--

Rick C

 

[rickman]

On 5/17/2017 9:21 AM, George Herold wrote:

On Tuesday, May 16, 2017 at 9:38:58 PM UTC-4, rickman wrote:
On 5/16/2017 3:33 PM, George Herold wrote:
On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the x-over
freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to add
the BP too? (My 'scope can't sum three signals.)

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

I assume you mean me? Thanks for the consideration, but in this context
those abbreviations were clear to me. I don't understand what you mean
about the LP + HP summing to zero though. I assume you mean they sum to
be the same as the input?
No, the sum is zero.. no signal. at "resonance" the LP and HP amplitudes
are equal, and the phase difference is 180 deg.

As Tim said I already knew this, 'cause you can make a notch by adding
LP and HP... but for me knowing (book-wise) and doing are two different
things.

Ok, I didn't get the resonance part.

--

Rick C

 

[George Herold]

On Wednesday, May 17, 2017 at 9:55:34 AM UTC-4, rickman wrote:

On 5/17/2017 9:21 AM, George Herold wrote:
On Tuesday, May 16, 2017 at 9:38:58 PM UTC-4, rickman wrote:
On 5/16/2017 3:33 PM, George Herold wrote:
On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the x-over
freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to add
the BP too? (My 'scope can't sum three signals.)

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

I assume you mean me? Thanks for the consideration, but in this context
those abbreviations were clear to me. I don't understand what you mean
about the LP + HP summing to zero though. I assume you mean they sum to
be the same as the input?
No, the sum is zero.. no signal. at "resonance" the LP and HP amplitudes
are equal, and the phase difference is 180 deg.

As Tim said I already knew this, 'cause you can make a notch by adding
LP and HP... but for me knowing (book-wise) and doing are two different
things.

Ok, I didn't get the resonance part.

--

Rick C

Resonance is where the BP output is maximum.

GH

 

[rickman]

On 5/17/2017 10:54 AM, George Herold wrote:

On Wednesday, May 17, 2017 at 9:55:34 AM UTC-4, rickman wrote:
On 5/17/2017 9:21 AM, George Herold wrote:
On Tuesday, May 16, 2017 at 9:38:58 PM UTC-4, rickman wrote:
On 5/16/2017 3:33 PM, George Herold wrote:
On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap., and
using an instrument amp to read the voltage across the resistor.


Error in----+---.
| |
R +--> Instrument R -->High Freq
out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. |
| | ---
| | ---
'-' |
| |
o-------------------------------o LF out
| |
| o-----------------o HF out
| |
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the x-over
freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP summed
to zero at the X-over. (I never knew that before.) I guess I'd have to add
the BP too? (My 'scope can't sum three signals.)

(BP = band pass, HP = high pass, LP = low pass)
('cause Rick will ask.)

I assume you mean me? Thanks for the consideration, but in this context
those abbreviations were clear to me. I don't understand what you mean
about the LP + HP summing to zero though. I assume you mean they sum to
be the same as the input?
No, the sum is zero.. no signal. at "resonance" the LP and HP amplitudes
are equal, and the phase difference is 180 deg.

As Tim said I already knew this, 'cause you can make a notch by adding
LP and HP... but for me knowing (book-wise) and doing are two different
things.

Ok, I didn't get the resonance part.

--

Rick C

Resonance is where the BP output is maximum.

I'm saying that I didn't catch the use of the term crossover to refer to
the point of equal attenuation of the two outputs. I was thinking of a
physical crossover circuit.

--

Rick C

 

[Tim Wescott]

On Wed, 17 May 2017 06:16:06 -0700, George Herold wrote:

On Tuesday, May 16, 2017 at 4:17:37 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 12:33:43 -0700, George Herold wrote:

On Tuesday, May 16, 2017 at 2:45:01 PM UTC-4, Tim Wescott wrote:
On Tue, 16 May 2017 09:07:58 -0700, George Herold wrote:

OK this is a bit of a silly question.
(But what the heck better than thawing tundra.)

I want to split an error signal into HF and LF parts.
I've wired this up as an RC filter, taking DC (LF) from the cap.,
and using an instrument amp to read the voltage across the
resistor.


Error in----+---.
| |
R +--> Instrument R -->High
Freq out.
R +--> Amp.(Ground ref.)
| |
+---+
+---Opamp buffer--> DC out |
C
C
C
|
GND

R = 300 to 10 k ohm. (300 is about most the opamp driving it can
supply.)
C = 2.2 uF. (About 10 Hz cross-over frequency at the low end.)

Is there anything wrong with this approach?

George H.

It'll need buffering. But no instrument amps. Equal C, equal R.

Vin o--------------o-------------.
| |
.-. | | | ---
| | ---
'-' |
| |
o-------------------------------o LF out |
|
| o-----------------o HF out |
|
| .-.
--- | |
--- | |
| '-'
| |
VVV VVV

Right I could have done that too. (I wanted an knob to change the
x-over freq.... no dual ganged pots.) And this way I'm pretty much
guaranteed that HF and LF will have the same time constant.

I screwed up the phase on the instrument amp the first time.

The sum of HF + LF was equal to the input. That's reassuring.

When I tired the same with a State Variable filter, the LP + HP
summed to zero at the X-over. (I never knew that before.) I guess
I'd have to add the BP too? (My 'scope can't sum three signals.)

2nd-order? Of course -- think about it:

LP: w^2 / (s^2 + 2*zeta*w*s + w^2)
HP: s^2 / (s^2 + 2*zeta*w*s + w^2)

LP + HP = (s^2 + w^2) / (etc.), which is a plain ol' notch filter.

Right, it made sense afterwards.
Hey, is this analogy correct?
So in an SV filter with HP, BP and LP output, can I think it as a series
LRC circuit, With HP the signal across an ideal inductor,
BP that across the resistor, and LP across an ideal capacitor.

That seems right.

At first flush yes -- I'd have to work it out to be sure.

Ditto for a current-driven parallel RLC with the signal being the
currents involved.


--
www.wescottdesign.com