alternate active all-pass

[John Larkin]

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.



--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[George Herold]

On Friday, April 26, 2019 at 12:01:19 PM UTC-4, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.

Aren't those the 'standard' all passes.
https://www.analog.com/media/en/training-seminars/tutorials/MT-202.pdf?doc=an-1328.pdf

I've used both. (but only ~audio freq.)

--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Tauno Voipio]

On 26.4.19 19:01, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.

It works fine, and IMHO, it is better to not feed spikes
into the circuit.

--

-TV

 

[bitrex]

On 4/26/19 12:30 PM, George Herold wrote:

On Friday, April 26, 2019 at 12:01:19 PM UTC-4, John Larkin wrote:
The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.
Aren't those the 'standard' all passes.
https://www.analog.com/media/en/training-seminars/tutorials/MT-202.pdf?doc=an-1328.pdf

I've used both. (but only ~audio freq.)

Off topic but...Analog Devices needs a bigger sign over at their Norwood
offices. they have like a five acre front lawn and a teeny lil sign
about four feet wide can't even read it from the road without 20/20
vision. What's that say...anal...what??

Yeah I realize New England tends to be understated but they could make
it a little huger.

 

[Phil Hobbs]

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.

Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[bitrex]

On 4/26/19 1:41 PM, bitrex wrote:

On 4/26/19 12:30 PM, George Herold wrote:
On Friday, April 26, 2019 at 12:01:19 PM UTC-4, John Larkin wrote:
The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.
Aren't those the 'standard' all passes.
https://www.analog.com/media/en/training-seminars/tutorials/MT-202.pdf?doc=an-1328.pdf


I've used both.  (but only ~audio freq.)

Off topic but...Analog Devices needs a bigger sign over at their Norwood
offices. they have like a five acre front lawn and a teeny lil sign
about four feet wide can't even read it from the road without 20/20
vision. What's that say...anal...what??

Yeah I realize New England tends to be understated but they could make
it a little huger.

Maybe they're worried passers-by will figure it's some kind of sex toy
company and complain.

 

[John Larkin]

On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs

We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[bitrex]

On 4/26/19 5:10 PM, John Larkin wrote:

On Fri, 26 Apr 2019 16:52:42 -0400, bitrex <user@example.net> wrote:

On 4/26/19 4:43 PM, John Larkin wrote:
On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.



has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

We decided to trash PCB rev A, and go for dumb and low risk on B. I
don't want a rev C.

https://www.dropbox.com/s/dghx81mb074ksvj/DSC03292.JPG?dl=0

There are lots of ways to get 90 degrees at one frequency.

https://www.dropbox.com/s/ienkb3moycl0cg8/Shifters.jpg?dl=0

I ran into a situation recently where I was into the process of
designing a Revision C of something when the client rang me up and told
me that they'd determined Revision B was quite satisfactory enough and
insisted on paying me my fee immediately.

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased labor
immediately and had a relaxing weekend with the girl friend instead.

Well it feels good to do math sometimes so I'll see what this idea looks
like out of curiosity. also she's away so I can get away with it. Heh
she probably worries I'm gonna go out on the town lol

 

[John Larkin]

On Fri, 26 Apr 2019 16:52:42 -0400, bitrex <user@example.net> wrote:

On 4/26/19 4:43 PM, John Larkin wrote:
On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.



has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

We decided to trash PCB rev A, and go for dumb and low risk on B. I
don't want a rev C.

https://www.dropbox.com/s/dghx81mb074ksvj/DSC03292.JPG?dl=0

There are lots of ways to get 90 degrees at one frequency.

https://www.dropbox.com/s/ienkb3moycl0cg8/Shifters.jpg?dl=0



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[bitrex]

On 4/26/19 4:43 PM, John Larkin wrote:

On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.

has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

 

[bitrex]

On 4/26/19 6:32 PM, bitrex wrote:

On 4/26/19 5:52 PM, Winfield Hill wrote:
bitrex wrote...

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased
labor immediately and had a relaxing weekend with the girl friend
instead.

Well it feels good to do math sometimes so I'll see what this idea
looks like out of curiosity. also she's away so I can get away
with it. Heh she probably worries I'm gonna go out on the town lol

  If you work on it and get her used to the idea that you
  need a little time for an outside affair, not to worry
  about it, you'll have plenty of free time for the lab.



I was going to make a rebuttal but I think I see your point

If I need some alone time I can always announce loudly as I walk into
the living room "Sweetie, I want to tell you all about the classic,
wonderful 807 beam power tetrode"

That works pretty good

 

[bitrex]

On 4/26/19 5:52 PM, Winfield Hill wrote:

bitrex wrote...

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased
labor immediately and had a relaxing weekend with the girl friend
instead.

Well it feels good to do math sometimes so I'll see what this idea
looks like out of curiosity. also she's away so I can get away
with it. Heh she probably worries I'm gonna go out on the town lol

If you work on it and get her used to the idea that you
need a little time for an outside affair, not to worry
about it, you'll have plenty of free time for the lab.

I was going to make a rebuttal but I think I see your point

 

[Winfield Hill]

bitrex wrote...

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased
labor immediately and had a relaxing weekend with the girl friend
instead.

Well it feels good to do math sometimes so I'll see what this idea
looks like out of curiosity. also she's away so I can get away
with it. Heh she probably worries I'm gonna go out on the town lol

If you work on it and get her used to the idea that you
need a little time for an outside affair, not to worry
about it, you'll have plenty of free time for the lab.


--
Thanks,
- Win

 

[George Herold]

On Friday, April 26, 2019 at 4:44:25 PM UTC-4, John Larkin wrote:

On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

Hey, so how about a few sections of RC phase sequence filter?
It's got a weird impedance vs. frequency.
I've only used it around audio (3 Hz to 3 kHz :^)
but I thought I saw things in the RF range.
At RF they call it by a different name.
(which I can't remember.)

George H.

 

[John Larkin]

On Fri, 26 Apr 2019 19:14:45 +0300, Tauno Voipio
<tauno.voipio@notused.fi.invalid> wrote:

On 26.4.19 19:01, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.


It works fine, and IMHO, it is better to not feed spikes
into the circuit.

I'd have to define the phase of a spike!


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[John Larkin]

On Fri, 26 Apr 2019 17:24:21 -0400, bitrex <user@example.net> wrote:

On 4/26/19 5:10 PM, John Larkin wrote:
On Fri, 26 Apr 2019 16:52:42 -0400, bitrex <user@example.net> wrote:

On 4/26/19 4:43 PM, John Larkin wrote:
On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.



has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

We decided to trash PCB rev A, and go for dumb and low risk on B. I
don't want a rev C.

https://www.dropbox.com/s/dghx81mb074ksvj/DSC03292.JPG?dl=0

There are lots of ways to get 90 degrees at one frequency.

https://www.dropbox.com/s/ienkb3moycl0cg8/Shifters.jpg?dl=0


I ran into a situation recently where I was into the process of
designing a Revision C of something when the client rang me up and told
me that they'd determined Revision B was quite satisfactory enough and
insisted on paying me my fee immediately.

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased labor
immediately and had a relaxing weekend with the girl friend instead.

We have one customer that, if we make a mistake, tries to pay us to
fix it!


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[bitrex]

On 4/26/19 10:01 PM, John Larkin wrote:

On Fri, 26 Apr 2019 17:24:21 -0400, bitrex <user@example.net> wrote:

On 4/26/19 5:10 PM, John Larkin wrote:
On Fri, 26 Apr 2019 16:52:42 -0400, bitrex <user@example.net> wrote:

On 4/26/19 4:43 PM, John Larkin wrote:
On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.



has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

We decided to trash PCB rev A, and go for dumb and low risk on B. I
don't want a rev C.

https://www.dropbox.com/s/dghx81mb074ksvj/DSC03292.JPG?dl=0

There are lots of ways to get 90 degrees at one frequency.

https://www.dropbox.com/s/ienkb3moycl0cg8/Shifters.jpg?dl=0


I ran into a situation recently where I was into the process of
designing a Revision C of something when the client rang me up and told
me that they'd determined Revision B was quite satisfactory enough and
insisted on paying me my fee immediately.

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased labor
immediately and had a relaxing weekend with the girl friend instead.


We have one customer that, if we make a mistake, tries to pay us to
fix it!

They must be used to working with lawyers

 

On Saturday, April 27, 2019 at 3:26:11 PM UTC+10, bitrex wrote:

On 4/26/19 10:01 PM, John Larkin wrote:
On Fri, 26 Apr 2019 17:24:21 -0400, bitrex <user@example.net> wrote:

On 4/26/19 5:10 PM, John Larkin wrote:
On Fri, 26 Apr 2019 16:52:42 -0400, bitrex <user@example.net> wrote:

On 4/26/19 4:43 PM, John Larkin wrote:
On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.



has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

We decided to trash PCB rev A, and go for dumb and low risk on B. I
don't want a rev C.

https://www.dropbox.com/s/dghx81mb074ksvj/DSC03292.JPG?dl=0

There are lots of ways to get 90 degrees at one frequency.

https://www.dropbox.com/s/ienkb3moycl0cg8/Shifters.jpg?dl=0


I ran into a situation recently where I was into the process of
designing a Revision C of something when the client rang me up and told
me that they'd determined Revision B was quite satisfactory enough and
insisted on paying me my fee immediately.

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased labor
immediately and had a relaxing weekend with the girl friend instead.

We have one customer that, if we make a mistake, tries to pay us to
fix it!

They must be used to working with lawyers

I've said some unkind things about John Larkin, but I've never accused him of being a lawyer.

--
Bill Sloman, Sydney

 

[John Larkin]

On Sat, 27 Apr 2019 01:26:07 -0400, bitrex <user@example.net> wrote:

On 4/26/19 10:01 PM, John Larkin wrote:
On Fri, 26 Apr 2019 17:24:21 -0400, bitrex <user@example.net> wrote:

On 4/26/19 5:10 PM, John Larkin wrote:
On Fri, 26 Apr 2019 16:52:42 -0400, bitrex <user@example.net> wrote:

On 4/26/19 4:43 PM, John Larkin wrote:
On Fri, 26 Apr 2019 13:35:07 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 4/26/19 12:01 PM, John Larkin wrote:

The classic opamp all-pass is in Williams 3,

https://www.dropbox.com/s/nhevg979w7gtxib/90_degree.JPG?dl=0

We're doing that but see some harmonic distortion at high frequency,
above 8 MHz in our case. This circuit inverts at low frequencies and
has unity gain at high frequency. Opamps have bad CMRR at high
frequencies, which is why Jim Williams said "always invert."

There could also be some problems from the finite output impedance of
each opamp driving the next stage, which sort of differentiates its
output. Our resistances are getting pretty low.

Wouldn't this work?

https://www.dropbox.com/s/jj9f3kp0svar7rw/Alt_All-Pass.JPG?dl=0

Circuit B is an inverter at high frequencies, unity gain at low. Gotta
try that. We'd just swap some parts on the board and redefine some
phase conventions.




Sure, B is the circuit I've always used for that. When I was a grad
student I built a 13-bit successive-approximation phase digitizer using
a SAR IC (AM2504), a 12-bit DAC, an all-pass phase shifter, a homemade
low-offset microvolt comparator, and a couple of flipflops to decide
which null to shoot for. Its input frequency was 60 MHz, and it ran at
50 kS/s with an accuracy of about 0.05 degree, which made me pretty
happy at the time. (It's the subject of my one and only instruments paper.)

The phase shifter was two of Circuit B in series, with MV209 varactors
with inductors in series and parallel. That linearized it pretty well,
at least enough so that a software linearization didn't lose precision.

Having one end of the varactor near ground makes it a lot easier to control.

Cheers

Phil Hobbs


We just had a meeting and decided that we'll dump the big all-pass,
which has been a hassle, with oscillations and distortion. We'll use a
dumb 90 degree phase shifter that works at one frequency. Our only
customer for the I/Q modulator so far works at fixed frequency, around
15 MHz.

That will be kinda like this:

https://www.dropbox.com/s/h37vfbv0cjt6ybc/90_deg_allpass_1.jpg?dl=0

We're driving an ADI analog multiplier that has diff inputs, so that
works nicely.

This is easier to explain to the customer. One path is 0 degrees and
one is 90.



has anyone thought about using an active impedance converter in standard
form #2 instead of the capacitor to ground? such as a negative inductor.
I think it would still be all-pass transfer function but offer
performance advantages in some use-cases. I'm doing the math on it...

We decided to trash PCB rev A, and go for dumb and low risk on B. I
don't want a rev C.

https://www.dropbox.com/s/dghx81mb074ksvj/DSC03292.JPG?dl=0

There are lots of ways to get 90 degrees at one frequency.

https://www.dropbox.com/s/ienkb3moycl0cg8/Shifters.jpg?dl=0


I ran into a situation recently where I was into the process of
designing a Revision C of something when the client rang me up and told
me that they'd determined Revision B was quite satisfactory enough and
insisted on paying me my fee immediately.

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased labor
immediately and had a relaxing weekend with the girl friend instead.


We have one customer that, if we make a mistake, tries to pay us to
fix it!



They must be used to working with lawyers

Not so much. They really value long-term relationships and think that
they should support "small companies" that do good work for them.

I think that we should admit and pay for our own mistakes. EMI
sensitivity is a borderline "mistake" but I thought that we should pay
to improve our own product.

https://www.dropbox.com/s/qjz9mfux9ucesw8/HTI_VME_Cage.jpg?dl=0



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Phil Hobbs]

On 4/26/19 5:52 PM, Winfield Hill wrote:

bitrex wrote...

The perfectionist-man part of me wanted to continue however the
business/family man part of me successfully overruled and I ceased
labor immediately and had a relaxing weekend with the girl friend
instead.

Well it feels good to do math sometimes so I'll see what this idea
looks like out of curiosity. also she's away so I can get away
with it. Heh she probably worries I'm gonna go out on the town lol

If you work on it and get her used to the idea that you
need a little time for an outside affair, not to worry
about it, you'll have plenty of free time for the lab.

Reminds me of an old joke.

Three friends, a doctor, a lawyer, and a physicist, were out having a
few drinks. The topic came up: would you rather have a wife or a mistress?

The doctor said, "A wife for me, all the way. I get called out at all
hours, even during holidays. A mistress would never put up with that."

The lawyer said, "Are you nuts? A wife can take you for all you've got.
A mistress for me any day."

The physicist said, "I'd like both."

The others replied, "Both? Why?"

The physicist replied, "Well, with both, if I tell the mistress that I'm
with the wife, and I tell the wife that I'm with the mistress, I can
spend my time at the lab!"

Boom-tsssh.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com