these lectures are basically mana from the physics gods. I am on chapter 2 (again) of the Horowitz book and almost everything here (augmented by Principles of Transistor Circuits - Amos& James and Electronic Music Circuits - Klein) makes sense - this is one of the best videos for bringing it all together. Currently I'm restoring a Roland JX-3P (two bad voices where the NPN transistors were smashed by the pitch bend assembly) and it happens to have the BA662 AND multiple JRC4558's and I've poured over those data sheets many times. There are a few videos on the hugely influential CEM3340, but it would be amazing given it's re-release to see how you analyze the design and the many ways it can be used - beyond what is in the data sheet. Thank you Dr. Lanterman!!!!
@@Lantertronics It's taken a while for it to sink in hence rereading it over and over. I can finally somewhat intuitively understand the individual parts of this VCO and other "voice" circuits. I have a Phy degree, but the way it was taught at the undergrad level, there was very little that bridged the E&M theory to actual components on a breadboard. So I've spent years in limbo trying to reconcile the two worlds, sadly.
I studied a degree in audio engineering and all I heard about was warmth and rock/pop music from the lecturers (they gave no recognition to edm or synthesis let alone be capible of technical descriptions). I studied electronics to a Cert IV level which was better but had no design aspect to it and mainly consisted of value solving 'textbook' circuits. You are a true Subject Matter Expert, thank you for this content. Exactly what I have been looking for about a decade!
I'm currently in the process of converting a non-working Yamaha CS15D Synthesizer over to a Modular Synthesizer system, I managed to find a complete service manual for the CS15D, which includes a complete set of circuit diagrams and adjustment procedures, the whole idea is to convert the circuits over to modular form, quite a few of the modules will be using the custom ICs that Yamaha used in the CS15D, such as the IGO0153 VCO, IGO0151 VCA, IGO0156 VCF, and IGO0159 ADSR Envelope Generator.
Awesome video once again. Given the simplicity of this VCA in today's standards you'd be able to recreate it for $20 or less with clone BA662's and some modern opamps. Simple but effective design
For the analysis of i_abc, unless I'm mistaken, there's a simple trick you can do to get a very rough ballpark number without analyzing the circuit fully. Assume Q3 or Q5 are dead shorts over C-E, as an extreme. Now, R22 and R27 control the maximum current. Divide the supply voltage by this resistance and you get for example i_abc < 12/6.8k = 1.8 mA as an absolute maximum limit. We will never see such a high current, so reasonably, in a practical circuit, i_abc < 1 mA.
Brilliant, I really enjoy your style of lecture. Wish I had had more fun lectures like this back in my days. Maybe I'd not have become an EE/CS student that only works on algorithms :P
Hi Dr. Lanterman, you mentioned at the beginning of the videos: "In the past couple lectures we have seen VCA.." but all I see as previous lectures are the history of synths. Did you post these videos somewhere else? (I'm an EE from UF building a synth for my senior design project and your lectures are pure gold) Thank you for sharing.
Here's the main ECE4450 playlist: after the history lectures, you'll see lectures about voltage controlled amplifiers: ruclips.net/video/mYk8r3QlNi8/видео.html
Also, there's a review of basic circuits and op amps that I make my synth students watch; that material is more general so I put it in a separate playlist (unfortunately RUclips doesn't seem to want to let me put a video in more than one playlist): ruclips.net/video/A720hgdE45M/видео.html
Great lecture, was very interesting and fun to watch how You tackle a design from the start. Interested to see somewhere in the future how all the transistor magic works. Thanks! I do have a more general question considering CV inputs like this in circuits. What is the plus side of using such a 'low' cutoff low pass filter? I get that it does not really matter as you should be plugging in DC or LFOs, but I love plugging in audio rate signals in CV stuff for interesting results. Like to get AM modulation with a VCA and two oscillators. It seems to me that a design like this only 'removes' an extra creative opportunity like that. But i suppose there is a good design reason for filtering out higher frequencies?
good questions, I am also interested in the answer. I have a guess that some components may behave in an unwanted manner at higher frequencies, but I don't really know
Minor thing but I think there is a mistake in the schematic? The bias resistor for Q4, below R25, should be V--, not ground. I don't think you will get anything going with Q5 otherwise and "linear" won't work.
@1m45, i used a simple 2SA733 in a linear V to I converter construction like found in the Roland Jupiter 8 and many others, i used a 1 resistor and trimmer. i have tested it and it works satisfying, tested it with scope. @9m55, i don't think it's experimentally, pretty much no control voltage in the input would go up to 15 volt power rail. many ADSR are 10Volt. blowing in 15 volt would produce a larger output than 10 Vpp at the output. @10m00, this trimmer aint really necessary, as the capacitors on the in and output would solve the problem and center the wave. as long as you use 1% or even better resistors. @12m19, you need that 10K resistor otherwise it would be a plain short circuit, using the buffer of the LM13700 and use a 33K resistor it will flatten the negative part of the wave, so the negative part of the wave isn't symmetric to the positive part, but smaller in amplitude.
@10m00 I disagree slightly. As you're saying, the inputs and outputs are AC coupled, so the trimmer is not there to adjust the absolute DC offset through the signal path. The trimmer is rather there so that reference level on the + and - inputs are identical to each other (or as close as possible to it). Specifically: a slight mismatch between + and - reference level would create a small DC offset going into the OTA. This DC offset would be multiplied by the control signal and so the control signal would be coupled into the signal path and give a small click or similar artifacts every time there's a sharp change in the control signal. Apparently, the Roland engineers thought this was important enough to address, perhaps even with 1% resistors.
@@Gameboygenius i have build this VCA less than 2 weeks ago using LM13700, but i didn't use the CV control circuit from Roland, but instead a simple 2SA733 transistor, resistor & trimmer. you don't need the "balance" trimmer. Mutable Instruments doesn't do either in her Ripples VCF design. mutable-instruments.net/modules/ripples/downloads/ripples_v90.pdf and she has researched this issue over and over. neither did she in the 4 pole mission. mutable-instruments.net/archive/shruthi/build/smr4mkii/ mutable-instruments.net/archive/schematics/Shruthi-Analog-SMR4-mkII-v02.pdf mutable-instruments.net/archive/documents/smr4mkII_analysis.pdf you don't have to worry about the mismatch .... the 10uF capacitors take care of that, they will center the wave again. it might be another story if you leave out the caps. if you really wanna be tedious, just get some 0.1% 1206, 0804 resistors, but 1% metal film resistors does the job just as good. i suspect even using normal 5% resistor would work, as long as you include the capacitors. .... and even if you would use plain 5% resistors, no caps, you probably not even hear it and still have lots of fun.
@@Gameboygenius but in this particular case, the capacitor after the mosFET is really needed. otherwise you would have a negative voltage on the output.
@@AnalogDude_ It may or may not be ok without the trimmer in practice. However, the fact that you're reiterating this "you don't have to worry about the mismatch .... the 10uF capacitors take care of that, they will center the wave again." tells me that you didn't read and take in what I said, and you're responding to something else.
@@AnalogDude_ Or how about this. High precision resistors are a modern luxury. In the Ripples BOM, the resistors in question are specified as, not 1%, but 0.1%. Unheard of in the 1970s when the Roland 130 VCA was designed. Roland probably sat down and did the math, and figured out that it was cheaper to have a factory drone trim the DC balance, than buying a resistor that had better tolerance from the factory. Today, SMD resistors are automatically calibrated in the factory by laser etching them to the specified tolerance.
hey professor Lanterman, nostalgia is a good thing, but i know a VCA schematic that will blow your mind, among a few others and beats pretty much anyone, Buchla, Moog, Roland, Korg, etc, etc. the establish brands. I'm pretty sure you will be amazed. that some small company can outsmart bigger things. have an email address so i can send it to you?
Oh, yeah, none of those old designs were particularly intended to be "hi-fi." Something like the SSI2164 would be cleaner, and if you really want a clean VCA, probably the chips by THAT Corp. would be the way to go.
@@Lantertronics the SSI2164 is probably cool if you want to squeeze 4 vca on a tiny board,, but i have build a (proper) VCF and this particular Roland VCA using lm13700, but simply linear 1 week ago, the LM13700 does a very good job, most people want even hear the difference. although it would be nice if Texas Instruments would make a CA3080 replacement, a single OTA without the darling buffer and like the CA3080 in a dip8 package ... or a version with identical FET like the 2SK30 Roland setup used instead of the darlington buffer. this hi pass filter i made, uses both, Roland FET configuration followed by the onboard darlington buffer, like ray Wilson did in his designs, it sounds awesome. ruclips.net/video/l_sAsZFENLo/видео.html i used 1nF / 1000pF WIMA 2.5% capacitors, i didn't measure it, but i guess the cut of point is around 4 or 5 kHz. in between. the images i have from my Tektronix 2212 scope from my experiments are good, i don't see any distortion in the wave forms, even uneducated unemployed people blind for brands screaming for the best of the best, accepted my build VCA.
I'm a musician and electrical engineering student. This lecture series is fantastic. Thanks for making it available!
same here. I am so grateful for this content
Ditto!:)
Thank you Mister Lanterman. All this lectures are gold for me.
these lectures are basically mana from the physics gods.
I am on chapter 2 (again) of the Horowitz book and almost everything here (augmented by Principles of Transistor Circuits - Amos& James and Electronic Music Circuits - Klein) makes sense - this is one of the best videos for bringing it all together.
Currently I'm restoring a Roland JX-3P (two bad voices where the NPN transistors were smashed by the pitch bend assembly) and it happens to have the BA662 AND multiple JRC4558's and I've poured over those data sheets many times.
There are a few videos on the hugely influential CEM3340, but it would be amazing given it's re-release to see how you analyze the design and the many ways it can be used - beyond what is in the data sheet.
Thank you Dr. Lanterman!!!!
If I could only keep one book, it would probably be Horowitz & Hill! :)
@@Lantertronics It's taken a while for it to sink in hence rereading it over and over. I can finally somewhat intuitively understand the individual parts of this VCO and other "voice" circuits. I have a Phy degree, but the way it was taught at the undergrad level, there was very little that bridged the E&M theory to actual components on a breadboard. So I've spent years in limbo trying to reconcile the two worlds, sadly.
I thought you would look really different in my head. Thank you so much for these downloads, you flippin rock as a lecturer!
These videos are absolute gold!
I studied a degree in audio engineering and all I heard about was warmth and rock/pop music from the lecturers (they gave no recognition to edm or synthesis let alone be capible of technical descriptions). I studied electronics to a Cert IV level which was better but had no design aspect to it and mainly consisted of value solving 'textbook' circuits. You are a true Subject Matter Expert, thank you for this content. Exactly what I have been looking for about a decade!
I'm currently in the process of converting a non-working Yamaha CS15D Synthesizer over to a Modular Synthesizer system, I managed to find a complete service manual for the CS15D, which includes a complete set of circuit diagrams and adjustment procedures, the whole idea is to convert the circuits over to modular form, quite a few of the modules will be using the custom ICs that Yamaha used in the CS15D, such as the IGO0153 VCO, IGO0151 VCA, IGO0156 VCF, and IGO0159 ADSR Envelope Generator.
That is awesome!
Awesome video once again. Given the simplicity of this VCA in today's standards you'd be able to recreate it for $20 or less with clone BA662's and some modern opamps. Simple but effective design
And I think it would work just fine (more or less) with an LM13700 or CA3080 (maybe with some tweaks).
Super !
Excellent video as usual. Many thanks Mr. Lanterman!
For the analysis of i_abc, unless I'm mistaken, there's a simple trick you can do to get a very rough ballpark number without analyzing the circuit fully. Assume Q3 or Q5 are dead shorts over C-E, as an extreme. Now, R22 and R27 control the maximum current. Divide the supply voltage by this resistance and you get for example i_abc < 12/6.8k = 1.8 mA as an absolute maximum limit. We will never see such a high current, so reasonably, in a practical circuit, i_abc < 1 mA.
Nice!
Brilliant, I really enjoy your style of lecture. Wish I had had more fun lectures like this back in my days. Maybe I'd not have become an EE/CS student that only works on algorithms :P
Hi Dr. Lanterman, you mentioned at the beginning of the videos: "In the past couple lectures we have seen VCA.." but all I see as previous lectures are the history of synths. Did you post these videos somewhere else?
(I'm an EE from UF building a synth for my senior design project and your lectures are pure gold) Thank you for sharing.
Here's the main ECE4450 playlist: after the history lectures, you'll see lectures about voltage controlled amplifiers: ruclips.net/video/mYk8r3QlNi8/видео.html
Also, there's a review of basic circuits and op amps that I make my synth students watch; that material is more general so I put it in a separate playlist (unfortunately RUclips doesn't seem to want to let me put a video in more than one playlist): ruclips.net/video/A720hgdE45M/видео.html
Great lecture, was very interesting and fun to watch how You tackle a design from the start. Interested to see somewhere in the future how all the transistor magic works. Thanks!
I do have a more general question considering CV inputs like this in circuits. What is the plus side of using such a 'low' cutoff low pass filter? I get that it does not really matter as you should be plugging in DC or LFOs, but I love plugging in audio rate signals in CV stuff for interesting results. Like to get AM modulation with a VCA and two oscillators. It seems to me that a design like this only 'removes' an extra creative opportunity like that. But i suppose there is a good design reason for filtering out higher frequencies?
good questions, I am also interested in the answer. I have a guess that some components may behave in an unwanted manner at higher frequencies, but I don't really know
Minor thing but I think there is a mistake in the schematic? The bias resistor for Q4, below R25, should be V--, not ground. I don't think you will get anything going with Q5 otherwise and "linear" won't work.
@1m45, i used a simple 2SA733 in a linear V to I converter construction like found in the Roland Jupiter 8 and many others, i used a 1 resistor and trimmer. i have tested it and it works satisfying, tested it with scope.
@9m55, i don't think it's experimentally, pretty much no control voltage in the input would go up to 15 volt power rail. many ADSR are 10Volt.
blowing in 15 volt would produce a larger output than 10 Vpp at the output.
@10m00, this trimmer aint really necessary, as the capacitors on the in and output would solve the problem and center the wave. as long as you use 1% or even better resistors.
@12m19, you need that 10K resistor otherwise it would be a plain short circuit, using the buffer of the LM13700 and use a 33K resistor it will flatten the negative part of the wave, so the negative part of the wave isn't symmetric to the positive part, but smaller in amplitude.
@10m00 I disagree slightly. As you're saying, the inputs and outputs are AC coupled, so the trimmer is not there to adjust the absolute DC offset through the signal path. The trimmer is rather there so that reference level on the + and - inputs are identical to each other (or as close as possible to it). Specifically: a slight mismatch between + and - reference level would create a small DC offset going into the OTA. This DC offset would be multiplied by the control signal and so the control signal would be coupled into the signal path and give a small click or similar artifacts every time there's a sharp change in the control signal. Apparently, the Roland engineers thought this was important enough to address, perhaps even with 1% resistors.
@@Gameboygenius i have build this VCA less than 2 weeks ago using LM13700,
but i didn't use the CV control circuit from Roland, but instead a simple 2SA733 transistor, resistor & trimmer. you don't need the "balance" trimmer.
Mutable Instruments doesn't do either in her Ripples VCF design.
mutable-instruments.net/modules/ripples/downloads/ripples_v90.pdf and she has researched this issue over and over.
neither did she in the 4 pole mission.
mutable-instruments.net/archive/shruthi/build/smr4mkii/
mutable-instruments.net/archive/schematics/Shruthi-Analog-SMR4-mkII-v02.pdf
mutable-instruments.net/archive/documents/smr4mkII_analysis.pdf
you don't have to worry about the mismatch .... the 10uF capacitors take care of that, they will center the wave again.
it might be another story if you leave out the caps.
if you really wanna be tedious, just get some 0.1% 1206, 0804 resistors, but 1% metal film resistors does the job just as good. i suspect even using normal 5% resistor would work, as long as you include the capacitors.
.... and even if you would use plain 5% resistors, no caps, you probably not even hear it and still have lots of fun.
@@Gameboygenius but in this particular case, the capacitor after the mosFET is really needed. otherwise you would have a negative voltage on the output.
@@AnalogDude_ It may or may not be ok without the trimmer in practice. However, the fact that you're reiterating this "you don't have to worry about the mismatch .... the 10uF capacitors take care of that, they will center the wave again." tells me that you didn't read and take in what I said, and you're responding to something else.
@@AnalogDude_ Or how about this. High precision resistors are a modern luxury. In the Ripples BOM, the resistors in question are specified as, not 1%, but 0.1%. Unheard of in the 1970s when the Roland 130 VCA was designed. Roland probably sat down and did the math, and figured out that it was cheaper to have a factory drone trim the DC balance, than buying a resistor that had better tolerance from the factory. Today, SMD resistors are automatically calibrated in the factory by laser etching them to the specified tolerance.
hey professor Lanterman,
nostalgia is a good thing, but i know a VCA schematic that will blow your mind, among a few others and beats pretty much anyone, Buchla, Moog, Roland, Korg, etc, etc. the establish brands.
I'm pretty sure you will be amazed. that some small company can outsmart bigger things.
have an email address so i can send it to you?
Sure -- not hard to find my e-mail, just go here and click on e-mail: www.ece.gatech.edu/faculty-staff-directory/aaron-d-lanterman
Oh, yeah, none of those old designs were particularly intended to be "hi-fi." Something like the SSI2164 would be cleaner, and if you really want a clean VCA, probably the chips by THAT Corp. would be the way to go.
@@Lantertronics
the SSI2164 is probably cool if you want to squeeze 4 vca on a tiny board,, but i have build a (proper) VCF and this particular Roland VCA using lm13700, but simply linear 1 week ago, the LM13700 does a very good job, most people want even hear the difference. although it would be nice if Texas Instruments would make a CA3080 replacement, a single OTA without the darling buffer and like the CA3080 in a dip8 package ... or a version with identical FET like the 2SK30 Roland setup used instead of the darlington buffer.
this hi pass filter i made, uses both, Roland FET configuration followed by the onboard darlington buffer, like ray Wilson did in his designs, it sounds awesome.
ruclips.net/video/l_sAsZFENLo/видео.html
i used 1nF / 1000pF WIMA 2.5% capacitors, i didn't measure it, but i guess the cut of point is around 4 or 5 kHz. in between.
the images i have from my Tektronix 2212 scope from my experiments are good, i don't see any distortion in the wave forms, even uneducated unemployed people blind for brands screaming for the best of the best, accepted my build VCA.