Hello,
I wonder if someone could suggest a good source of information, either in print form or online is fine, for understanding a particular negative feedback circuit as it applies to shaping an EQ response. In other words I would like to learn how it functions (I have a good idea of the basics) and what the mathematics are for determining component values and how to change them to achieve different requirements. This is the circuit I'm interested in:
I wonder if someone could suggest a good source of information, either in print form or online is fine, for understanding a particular negative feedback circuit as it applies to shaping an EQ response. In other words I would like to learn how it functions (I have a good idea of the basics) and what the mathematics are for determining component values and how to change them to achieve different requirements. This is the circuit I'm interested in:
Just Googling "active filter design" should turn up a lot of resources.
A couple of examples:
https://www.tinaja.com/ebooks/afcb.pdf
https://gctjaipur.wordpress.com/wp-content/uploads/2015/08/continuous-time-active-filter-design.pdf
Also, often it is assumed that the reader is familiar with the Laplace Transform. Some possible resources, if needed (basically, it is used to reduce problems involving differential equations into algebra problems):
https://www.sjsu.edu/me/docs/hsu-Chapter 6 Laplace transform.pdf
https://www.analog.com/media/en/technical-documentation/dsp-book/dsp_book_Ch32.pdf
A couple of examples:
https://www.tinaja.com/ebooks/afcb.pdf
https://gctjaipur.wordpress.com/wp-content/uploads/2015/08/continuous-time-active-filter-design.pdf
Also, often it is assumed that the reader is familiar with the Laplace Transform. Some possible resources, if needed (basically, it is used to reduce problems involving differential equations into algebra problems):
https://www.sjsu.edu/me/docs/hsu-Chapter 6 Laplace transform.pdf
https://www.analog.com/media/en/technical-documentation/dsp-book/dsp_book_Ch32.pdf
Last edited:
It's actually pretty simple ... in theory. As you've undoubtedly found if you read Lipshitz' paper, the math gets hairy in a hurry. But in theory the feedback network is "just" a voltage divider:
Pardon the mirror image for Z1. So now you can derive the divider ratio as:
Vout/Vin = Z2/(Z1+Z2)
Trouble arises because Z1 depends on frequency and that makes the math a bit complicated. It should be no trouble to play with the divider in a simulator and get some understanding that way, though. Keep in mind that the frequency response of the amplifier overall will be approx. 1/ß, so (Z1+Z2)/Z2, assuming you have sufficient forward gain.
If you're new to circuit simulation, I suggest learning QSpice. I'm thinking LTspice has become abandonware now that the author is with Qorvo.
Tom
Pardon the mirror image for Z1. So now you can derive the divider ratio as:
Vout/Vin = Z2/(Z1+Z2)
Trouble arises because Z1 depends on frequency and that makes the math a bit complicated. It should be no trouble to play with the divider in a simulator and get some understanding that way, though. Keep in mind that the frequency response of the amplifier overall will be approx. 1/ß, so (Z1+Z2)/Z2, assuming you have sufficient forward gain.
If you're new to circuit simulation, I suggest learning QSpice. I'm thinking LTspice has become abandonware now that the author is with Qorvo.
Tom
Thank you for these great suggestions. I've considered LTspice and QSpice for a long time and understand its potential. I have a Mac laptop and haven't found a clear answer to potential compatibility issues. I've read the Lipshitz paper a few times and must admit I get a little further with it each time. I also remember the article in Audio Amateur that used it to create tighter tolerance RIAA feedback values. Also thank you for those resource links, I will carefully absorb them.
The overall reason for this thread is that I have built a tape head preamp based on the Marantz 7 circuit that was inspired by a YouTube contributor who described how to convert the RIAA circuit into an IEC and NAB playback EQ. I built the unit with the ability to adjust the gain, input loading and low and high frequency parts of the curve. I have a calibration tape and can measure the playback frequency response and make adjustments. And actually the results are quite good compared to other tape preamps I have.
I sort of feel like I'm a fairly skilled builder and that a real engineer hands me a set of schematics and says, "here, build this". I'd like to know a bit more of what it is I'm building and the theory behind the results I'm getting. One example is when I change the pot for gain (the 510 ohm resistor) it effects not only the output level but it changes the way the bass response was matching the NAB curve. From Tom's post I believe I understand now that this resistor is part of a voltage divider. Thank you for that.
The overall reason for this thread is that I have built a tape head preamp based on the Marantz 7 circuit that was inspired by a YouTube contributor who described how to convert the RIAA circuit into an IEC and NAB playback EQ. I built the unit with the ability to adjust the gain, input loading and low and high frequency parts of the curve. I have a calibration tape and can measure the playback frequency response and make adjustments. And actually the results are quite good compared to other tape preamps I have.
I sort of feel like I'm a fairly skilled builder and that a real engineer hands me a set of schematics and says, "here, build this". I'd like to know a bit more of what it is I'm building and the theory behind the results I'm getting. One example is when I change the pot for gain (the 510 ohm resistor) it effects not only the output level but it changes the way the bass response was matching the NAB curve. From Tom's post I believe I understand now that this resistor is part of a voltage divider. Thank you for that.
Hi Johnny,
Thank you for posting that. I refurbished one of those a few years ago and learned quite a bit from working on it.
Thank you for posting that. I refurbished one of those a few years ago and learned quite a bit from working on it.
You basically have two routes:
- Install a virtual machine client like VirtualBox, vmWare Fusion, or Parallels. Then install Windoze and run QSpice in a virtual machine. This requires two licenses; one for the virtual machine client/hypervisor and one for Windoze installed in the virtual machine.
- Use something like Wineskin (now apparently replaced by Kegworks) and install the QSpice that way. I did that with TINA-TI and it works quite well. You can see the procedure here: https://neurochrome.com/pages/running-windows-programs-on-mac-for-free. I'd imagine the procedure is the same for Kegworks.
Anyway. That's a topic for its own thread.
Tom
Hi – that's a great looking build. A couple thoughts I think are worth sharing:
I'm a Mac user and I'm able to run LTSpice on my machine without issue. Below is the 'about' for the build I have installed, though I believe they're now up to 17.2.4 and you can download it here: https://www.analog.com/en/resources/design-tools-and-calculators/ltspice-simulator.html
As a tool for learning assistance, are you open to using AI? I've seen several threads on the subject in these forums, and it seems opinions vary about its utility and accuracy. Speaking from personal experience, ChatGPT 4o (the paid, $20/mo version) is a BARGAIN considering how quickly the tool has advanced my knowledge and accelerated my learning curve.
I'm new to the DIY electronics space and have found it amazingly useful, not just for understanding circuit function but also for advancing my understanding of electromagnetic theory in general. Additionally I've used it to advance my skills with the oscilloscope, turn CAD files into workable mechanical drawings for CNC, and learn unfamiliar software – LTSpice in particular, but also KiCAD and Fusion.
This comes with caveats, of course – it takes a lot of time, patience and critical thinking to prompt it effectively. Also, DEFINITELY test and confirm what it generates in the real world, and seek the advice of real experts with "real" intelligence. YMMV, but for me it's filling the knowledge gaps that make it possible to execute on ideas that I'd otherwise never be able to.
I'm a Mac user and I'm able to run LTSpice on my machine without issue. Below is the 'about' for the build I have installed, though I believe they're now up to 17.2.4 and you can download it here: https://www.analog.com/en/resources/design-tools-and-calculators/ltspice-simulator.html
As a tool for learning assistance, are you open to using AI? I've seen several threads on the subject in these forums, and it seems opinions vary about its utility and accuracy. Speaking from personal experience, ChatGPT 4o (the paid, $20/mo version) is a BARGAIN considering how quickly the tool has advanced my knowledge and accelerated my learning curve.
I'm new to the DIY electronics space and have found it amazingly useful, not just for understanding circuit function but also for advancing my understanding of electromagnetic theory in general. Additionally I've used it to advance my skills with the oscilloscope, turn CAD files into workable mechanical drawings for CNC, and learn unfamiliar software – LTSpice in particular, but also KiCAD and Fusion.
This comes with caveats, of course – it takes a lot of time, patience and critical thinking to prompt it effectively. Also, DEFINITELY test and confirm what it generates in the real world, and seek the advice of real experts with "real" intelligence. YMMV, but for me it's filling the knowledge gaps that make it possible to execute on ideas that I'd otherwise never be able to.
Thank you! It's funny you should mention AI. My second semi-obsolete hobby is stamp collecting. I started it as a kid and have recently taken it up again. In those days you had to use Scott's stamp catalogue to figure out where and when a stamp was issued and it took a while and lots of page turning over multiple volumes. But, now I just ask Google a question and depending upon how accurately I phrase the wording I get a really meaningful answer.
I worry about the provenance of the answer. Does it provide a reference?
"Too bad that I still can't make up my mind what I shouild do with it."
I know what you mean. I ended changing mine to a linear amplifier and connected it to an Advent 100A Dolby unit as a line stage bypassing the built in version:
I know what you mean. I ended changing mine to a linear amplifier and connected it to an Advent 100A Dolby unit as a line stage bypassing the built in version:
We all learn in different ways, I learn more by doing backed up with some theory & reading up on the subject. To my mind you can't beat prototyping a physical circuit then using a scope etc to test. That's me though, other folk as mentioned use Spice simulations to great effect.
Andy.
Andy.
It does.
Type something into Google. The AI summary is provided by Gemini, the Google AI thing. Click the "show more" and you'll get blurbs with a link to a reference.
Tom