ECE4450 L24: State Variable Filters and the Oberheim SEM VCF (Analog Circuits for Music Synthesis)
Вставка
- Опубліковано 24 сер 2024
- Support this channel via a special purpose donation to the Georgia Tech Foundation (GTF210000920), earmarked for my work: • Support This Channel v...
This is a follow-up to the previous lecture on the theory of second-order filters: • ECE4450 L23: Second-Or...
I recorded this during the Spring 2021 offering of ECE4450: Analog Circuits for Music Synthesis, but this material will likely be appropriate for future offerings as well.
Is there anywhere we can give you donations? You're rapidly becoming one of, if not THE best synth DIY educators on the internet.
I'm working with the Powers that Be at my employer to coordinate a mechanism by which you can donate to Georgia Tech and have that donation earmarked to support this kind of work. Details coming soon.
I finally got it worked out! And it really helps with showing my superiors that there is a hunger for this kind of material. Do a UA-cam search on "lantertronics donation" (without the quotes.) I'll post the link a a separate comment since UA-cam seems to delete links sometimes.
Here is the link for the video with the details about donating: ua-cam.com/video/VBu-LST1p9c/v-deo.html
@@Lantertronics Awesome!
The diodes in SEM VCF are worth of analysis.
Definitely!
@@Lantertronics I'm using digital implementation of SV filter in my FPGA based synth and I would like to implement similar nonlinearity in the feedback path. It looks like the diodes are acting as exponential function and are somehow combined with the linear path in variable ratio. Interesting variation on the SV filter is also the Novation design: zeninstruments.blogspot.com/2017/09/novation-bass-station-schematic.html
Hey man listening to this again since I'm finishing the schematic for my version of it. Thank you!
I've got a Two Voice right here and yes indeed that's exactly how the mix and mod pots work. The frequency pots will blow your mind ... they are GEARED POTENTIOMETERS ... not a multiturn knob (EMS) on a pot or a multiturn pot!
That moment when just banging away in vcv rack leads me to "ah, that's what integration means". Been waaay too many years since any math or electronics class. Thanks for helping me start learning again. Found this video because I was curious what a SEM filter was.
You did a really good job explaining the structure and construction of a state variable filter.
Really helps with the theory behind building variable filters.
I was just covering state variable filters in a lecture yesterday in my Analog Filter Synthesis course. Great explanations!
Cool! Where are you teaching?
@@Lantertronics I teach at the University of Illinois at Chicago.
@@Lantertronics I sent you a "thank you " note via email.
Wow ❤
Thanks!
Haha, what gets me into synth DIY again after long pauses is, indeed, listening to certain Rush albums. Funny.
the mix pots did have center detent
As always, amazing content!
A little late in seeing this one. I've designed a few SV filters, but have never done one with the JFET buffers. Something I've wanted to try, and see how the character differs from my usual inverting op amp integrator types that I usually do.
“if you want to slam the poles into the imaginary axis”
Can I buy you a beer?
I don't script my lectures ahead of time (although I edit them heavily), so occasionally I surprise myself upon playback. :)
Actually, I'm working on a mechanism with my employer by which people can donate to Georgia Tech and earmark the donation to help fund more of this sort of work with my students and create more of this sort of content -- although I won't be able to buy a beer with it. ;)
Details will be forthcoming in a future video.
(Since I'm creating these lectures as part of my duties as a professor, I can't set up set up something like a personal Patreon and accept money directly -- Georgia Tech, like most universities, has policies about these things.)
Not a beer, but you can now support this channel a targeted donation to Georgia Tech earmarked for my work with my students. Do a search on "lantertronics donation" on UA-cam (without the quotes.)
Here's the link to the video with instructions for donating: ua-cam.com/video/VBu-LST1p9c/v-deo.html
Any amount appreciated, even the price of a beer. :)
About the thing where the feedback path from the first integrator is buffered by both a JFET and an op amp: Maybe the circuit in the bottom to choose Q and limit the oscillation would just load the JFET too much otherwise? The second JFET goes into a 100 k resistor for the feedback path, but the impedance into the first feedback path looks more like 25 k.
Also, if the op amps are a bit slow, you would want to avoid to go through them for the feedback loop going through the second integrator, since the phase loss would tend to destabilize the loop, more so at higher cutoff frequencies. That should be less of a concern for the feedback loop from the first integrator, since the the phase around that loop is nominally -90 degrees, so far from instability (and it stabilizes the loop, if it's active), unlike the nominal -180 degrees of the second loop. But I don't know if the phase loss is a concern at all.
@@possible-realities That's a good theory! The 741 is certainly a slow op amp.
And I like your theory about loading the JFET...
Dear Aaron, what I can't figure out. How LPF (Low pass filter) can make it result if on the input it has BPF which has HPF on it's input. So we have series: HP = INPUT - LP; BP = Fbp(HP); LP = BP * Flp(BP); Where are Fbp and Flop are filters of BP and LP. The Chicken Eggs problem. I am getting crazy with simulation this circuits in the Altium. BTW Respect the engineers for 60s who designed all of this with pen and paper.
Just a bit of curiosity around that 10pF cap in the feedback path. (second path) Since the integrators are (granted small) capacitive loads, I wonder if there was a bit of parasitic oscillation that this helps remove? Some op amps are less prone to that, but not sure about the 741s, CA3080 OTAs, or maybe even the JFETs? I've had to remove some in somewhat similar scenarios, but I usually do it at any inverting amplifiers I may have in the design.
Huh... I honestly have no intuition about that 10 pF cap. At really high frequencies it's basically bypassing the 100K, which means you're getting *more* second order feedback, and it seems like that would potentially increase oscillation at high frequencies. Of course there may be some phase shift weirdness going on. So... beats me! :)
@@Lantertronics Ah, that makes sense. Maybe just a minor phase correction then. Thanks for the reply!
there is a lot of saturation with the filter...is that because of the J-fets?
I'm thinking it's mostly just from the OTAs themselves. That's just a guess though. One could breadboard it with JFETs and then try swapping in regular op amp buffers and compare.
13:10 I think you meant to say "integrate a current and turn it back into a voltage" here.
It looks like this video is missing from your ECE4450 playlist. I think it should be there as it's one of your most important lectures (why? 0:22)
Whoops! Thanks for catching that -- it was an oversight on my part. I just added it to the playlist.
In the case that the integrating OTA has a constant offset error, what keeps the integration cap from accumulating charge and saturating things?
Negative feedback to the rescue! :)
@@Lantertronics Ah true! And the feedback path that isn't the adjustable Q path is DC coupled, so it'll be fine no matter how the Q is adjusted. Thanks :D
aaa Preciso de um e ou , baixa potência, três circuitos independentes cada um SÓ com seus 2 ou mais amplificadores operacionais, em Classe D, com fonte SIMÉTRICA
I'd love some simpler circuits
I'd suggest checking out my ECE Design Fundamentals series: ua-cam.com/video/A720hgdE45M/v-deo.html
This video was clearly NOT made for beginners
Yup... these videos were primarily created for students in my class at Georgia Tech, which are mostly seniors in EE. So they do assume a EE background -- and the videos generally assume you've seen previous ones in the series: ua-cam.com/play/PLOunECWxELQS5bMdWo9VhmZtsCjhjYNcV.html
I feel like there's a lack of advanced engineering content on UA-cam, and I hope to fill that gap. If anyone is looking for a more tutorial introduction, I highly recommend checking out channels by "The AudioPhool" and "Moritz Klein."
@@Lantertronics I am no EE but I absolutely learned something from listening to your video twice.
Beginners should absolutely expose themselves to indecipherable information, eventually they will learn through context and then later on when they have a better understanding, go back to what they found confusing before and it will make sense.