rayma,
You are reminding me of things I often forget and then have to start over from the beginning.
For a simple single pole low pass filter:
1 MHz Bandwidth has a rise time of 0.35 usec.
Bandwidth is the = -3dB point (0.707 of the low frequency voltage).
1 us delay at 20,000Hz is -7.2 Degrees phase.
Is that correct?
Thanks!
You are reminding me of things I often forget and then have to start over from the beginning.
For a simple single pole low pass filter:
1 MHz Bandwidth has a rise time of 0.35 usec.
Bandwidth is the = -3dB point (0.707 of the low frequency voltage).
1 us delay at 20,000Hz is -7.2 Degrees phase.
Is that correct?
Thanks!
Yes.
Oscilloscope manuals and Tek app notes cover such matters, along with filter design books.
I think the rise time rule of thumb assumes a linear phase characteristic, but it's close enough.
https://www.edn.com/rule-of-thumb-1-bandwidth-of-a-signal-from-its-rise-time/
Oscilloscope manuals and Tek app notes cover such matters, along with filter design books.
I think the rise time rule of thumb assumes a linear phase characteristic, but it's close enough.
https://www.edn.com/rule-of-thumb-1-bandwidth-of-a-signal-from-its-rise-time/
Last edited:
The constant that is used to calculate bandwidth and rise time for a hardware low pass filter is 0.35 for a Gaussian response.
In 1968 and 1969, I looked at some very good looking square waves on a sampling scope plug in, with the 10% to 90% rise time agreeing with the
bandwidth, and agreeing with the 0.35 gaussian constant.
No Ringing!
In about 2005 or so, I used a 50 Ohm square wave generator, terminated it with 50 Ohms (25 Ohms in parallel), to drive a 5 uF plastic cap.
The scope was not anything like a traditional sampling scope plug in, and was not like the new sampling scopes either.
I did that to show an engineer that the 0.35 constant, 10 % to 90% rise time, and -3dB bandwidth all came together at the same frequency.
No Ringing!
No change!
Of course, in all those years, the high frequency sampling scopes had new factors that were anything but 0.35 (more like 4.2 or somthing like that).
Different hardware and designs no longer met the single pole Gausian response (and yes, they had "ringing").
Yes, Ringing!
The best filter book ever written was a real 'filter bible'.
Written by a Russian Engineer, Zverev, and published by Westinghouse.
With that book, the mentorship of our chief engineer, the model shop, and another engineer's 8087 processor filter program,
I re-designed the 5 pole helical 10 MHz wide - 3dB BW @ 525MHz IF filter of the 2782 and 2784 spectrum analyzers.
Before my re-design, it was about 7MHz wide at -6dB, but needed wider bandwidth.
it had capacitive coupling from helical to helical, and had a sloped group delay. Good, but not what I wanted it to be.
I did not have the design time to alternate the helical to helical coupling . . . inductive coupling, capacitive coupling, inductive coupling, capacitive coupling. That would have made for a non sloped group delay.
Just my experience.
In 1968 and 1969, I looked at some very good looking square waves on a sampling scope plug in, with the 10% to 90% rise time agreeing with the
bandwidth, and agreeing with the 0.35 gaussian constant.
No Ringing!
In about 2005 or so, I used a 50 Ohm square wave generator, terminated it with 50 Ohms (25 Ohms in parallel), to drive a 5 uF plastic cap.
The scope was not anything like a traditional sampling scope plug in, and was not like the new sampling scopes either.
I did that to show an engineer that the 0.35 constant, 10 % to 90% rise time, and -3dB bandwidth all came together at the same frequency.
No Ringing!
No change!
Of course, in all those years, the high frequency sampling scopes had new factors that were anything but 0.35 (more like 4.2 or somthing like that).
Different hardware and designs no longer met the single pole Gausian response (and yes, they had "ringing").
Yes, Ringing!
The best filter book ever written was a real 'filter bible'.
Written by a Russian Engineer, Zverev, and published by Westinghouse.
With that book, the mentorship of our chief engineer, the model shop, and another engineer's 8087 processor filter program,
I re-designed the 5 pole helical 10 MHz wide - 3dB BW @ 525MHz IF filter of the 2782 and 2784 spectrum analyzers.
Before my re-design, it was about 7MHz wide at -6dB, but needed wider bandwidth.
it had capacitive coupling from helical to helical, and had a sloped group delay. Good, but not what I wanted it to be.
I did not have the design time to alternate the helical to helical coupling . . . inductive coupling, capacitive coupling, inductive coupling, capacitive coupling. That would have made for a non sloped group delay.
Just my experience.
Last edited:
Zintolo,
No, not quite the same.
The book I borrowed from our chief engineer, was written by Zverev (I am pretty sure, he was the only author).
It was about 10 inches by 8 inches, and 2-3/4 inches thick.
About 3 decades ago, I got a new boss, I wanted my own copy for work. But I could not justify the $230 new price, and $208 used price at the technical section of Powell's book store (PDX).
No, not quite the same.
The book I borrowed from our chief engineer, was written by Zverev (I am pretty sure, he was the only author).
It was about 10 inches by 8 inches, and 2-3/4 inches thick.
About 3 decades ago, I got a new boss, I wanted my own copy for work. But I could not justify the $230 new price, and $208 used price at the technical section of Powell's book store (PDX).
This one?
https://www.amazon.com/Handbook-Filter-Synthesis-Anatol-Zverev/dp/0471749427
The 1967 first edition is free on the internet archive. Don't print it out, it's 586 pages.
https://ia803101.us.archive.org/20/items/HandbookOfFilterSynthesis/Handbook of Filter Synthesis.pdf
https://www.amazon.com/Handbook-Filter-Synthesis-Anatol-Zverev/dp/0471749427
The 1967 first edition is free on the internet archive. Don't print it out, it's 586 pages.
https://ia803101.us.archive.org/20/items/HandbookOfFilterSynthesis/Handbook of Filter Synthesis.pdf
Philwait,
In 1989, at MTTS in Long Beach, Calif., we introduced the Tektronix 2782 spectrum analyzer. I had already moved from engineering to Marketing Support.
HP and Tektronix gave each other an after-hours evening show-and-tell in our respective booths.
I met several of the HP engineers. One asked about our 3Hz analog resolution bandwidth, with which we were showing power supply sidebands on a 30GHz signal source.
I talked to one of the designers of the HP 8566, their flagship spectrum analyzer (the worlds best) until the Tektronix 2782.
If I remember correctly, I talked to Linkwitz (later of Audio fame). I am pretty sure he was one of the designers of the 8566, but I did not know of Linkwitz and audio at the time).
The 8566 had a 6GHz local oscillator.
The 2782 had an 18GHZ local oscillator. That allowed me to later set up a waveguide measurement of a 500GHz harmonic (properly identified by the ID function of the analyzer; and the 220GHz waveguide [is a brick wall high pass filter] prevented any of the 100MHz fundamental from coming through (measurable)).
Yes 1/2 TerraHertz.
In 1989, at MTTS in Long Beach, Calif., we introduced the Tektronix 2782 spectrum analyzer. I had already moved from engineering to Marketing Support.
HP and Tektronix gave each other an after-hours evening show-and-tell in our respective booths.
I met several of the HP engineers. One asked about our 3Hz analog resolution bandwidth, with which we were showing power supply sidebands on a 30GHz signal source.
I talked to one of the designers of the HP 8566, their flagship spectrum analyzer (the worlds best) until the Tektronix 2782.
If I remember correctly, I talked to Linkwitz (later of Audio fame). I am pretty sure he was one of the designers of the 8566, but I did not know of Linkwitz and audio at the time).
The 8566 had a 6GHz local oscillator.
The 2782 had an 18GHZ local oscillator. That allowed me to later set up a waveguide measurement of a 500GHz harmonic (properly identified by the ID function of the analyzer; and the 220GHz waveguide [is a brick wall high pass filter] prevented any of the 100MHz fundamental from coming through (measurable)).
Yes 1/2 TerraHertz.
rayma,
The old Zverev / Westinghouse book had a Red cover; almost looked like an old family bible; how appropriate.
I am not sure what any modern printing of the book will contain, versus the old book.
I was unable to view the one link, so do not know if it was the same, but with a publish date of 1967, I bet it was.
Tektronix had a running team for the annual Hood to Coast run. They called themselves the Mega Hurts (or Mega Hertz?).
The old Zverev / Westinghouse book had a Red cover; almost looked like an old family bible; how appropriate.
I am not sure what any modern printing of the book will contain, versus the old book.
I was unable to view the one link, so do not know if it was the same, but with a publish date of 1967, I bet it was.
Tektronix had a running team for the annual Hood to Coast run. They called themselves the Mega Hurts (or Mega Hertz?).
See if this shared Google drive Handbook of Filter Synthesis.pdf works. It doesn''t include the cover.
https://drive.google.com/file/d/1TDC2Pyaf4EgJVy285JreTsClVlEwTtWm/view?usp=sharing
https://drive.google.com/file/d/1TDC2Pyaf4EgJVy285JreTsClVlEwTtWm/view?usp=sharing
Sorry, I am not signing into Google.
You do not need to read those 500+ pages, just start reading the first few pages inside the front cover.
One of those first pages, before the table of contents, gives the credits to Zverev, the author, and Westinghouse the publisher (probably some publisher printed it on Westinghouse's behalf).
It has a very large section on different filter coefficient tables (Legendre, Bessel, Butterworth, etc.)
There are instructions on the Dishal method, and so much more that I can not even remember.
I more remember the process of how to set up to measure the unloaded Q of a helical, how to measure the loaded Q when you connect a wire to tap the coil and load it with 50 Ohms, the coupling from helical to helical versus the wall height from helical chamber to helical chamber, etc.
I used the tables for the Bessel filter, and some measurement results, and then did some preliminary calculations to enter into the software to run on the 8087.
As far as I can remember there was no mention of digital filters.
Obviously, a lot of the filter tables should apply.
Are you looking to purchase that book?
You do not need to read those 500+ pages, just start reading the first few pages inside the front cover.
One of those first pages, before the table of contents, gives the credits to Zverev, the author, and Westinghouse the publisher (probably some publisher printed it on Westinghouse's behalf).
It has a very large section on different filter coefficient tables (Legendre, Bessel, Butterworth, etc.)
There are instructions on the Dishal method, and so much more that I can not even remember.
I more remember the process of how to set up to measure the unloaded Q of a helical, how to measure the loaded Q when you connect a wire to tap the coil and load it with 50 Ohms, the coupling from helical to helical versus the wall height from helical chamber to helical chamber, etc.
I used the tables for the Bessel filter, and some measurement results, and then did some preliminary calculations to enter into the software to run on the 8087.
As far as I can remember there was no mention of digital filters.
Obviously, a lot of the filter tables should apply.
Are you looking to purchase that book?