ECE4450 L8: Voltage Controlled Oscillators: Sawtooth Cores (Analog Circuits for Music Synthesis)
ฝัง
- เผยแพร่เมื่อ 3 มี.ค. 2021
- Support this channel via a special purpose donation to the Georgia Tech Foundation (GTF210000920), earmarked for my work: • Support This Channel v...
MAJOR CORRECTION: The integrating capacitor should be 2.2 nF, not 4.7 nF; all of the calculations should then change accordingly. Also, I forgot to mention that the integrating capacitor should be robust to temperature changes; usually people recommend polystyrene, polypropylene, or C0G or NP0 ceramic (that's the good stuff, you don't want to use mediocre ceramics in that position).
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.
![คุณอาเรียโต๊ะข้างๆ พูดรัสเซียหวานใส่ซะหัวใจจะวาย ตอนที่ 2 [ซับไทย] 【Ani-One Thailand】](http://i.ytimg.com/vi/kbbfYakMY8k/mqdefault.jpg)
Thank you so much for this. I'm working on updating the ASM-1 VCO because the late Gene Stopp asked me to use it in everything I make, and It looks like it's derived from this. Your video helped me understand what I'm dealing with better.
Thought I'd mention a few of my favorite VCO integration op amps here, as they've given great results. It's quite a list. I use each one for different reasons. I'll list them all, but wanted to mention one specifically that's pretty amazing. (though recently it has doubled in price)
The AD8033. It has a FET input, super high slew rate, very very low bias current, wide supply range (compared to some) among other things. Pretty decent DC characteristics too.
Others that work quite well pretty much across the board for this task:
CA3140EZ
LF356
OPA132 (and even the OPA134 though it's more aimed at certain audio tasks I would say)
LF411 (I use this more in LFOs and slope generators, as its quite stable, low drift, etc. and also cheap)
OPA140
I saw a comment about possible substitutions below, so thought I'd mention these.
Also, for oscillators (I can't think of many, but Don Tilman's Trapezoid is one) that use the timing cap to ground instead of the inverting op amp integrator, something that's specced for capacitive loads might be better. Maybe something like the LM8261 (though there are much more common ones that are fine with this).
Hope that's somewhat helpful. Hard to elaborate further in a TH-cam comment. :D
Thank you for the class Aaron!
Great lecture! Subscribed.
I like the little tidbits you throw in about why one particular opamp vs. another, impedance considerations, etc; those details are very helpful for a beginner like me with little experience 😄
Bought the Chamberlin book! Thanks for the recommendation.
The JFet is really just being used as a voltage-controlled switch, when it turns-on it effectively shorts-out the integrator capacitor, the result is that the voltage across the integrator capacitor falls to zero and then the JFet gets turned off again, then the whole process of charging up the integrator cap starts again, rinse and repeat.
One like for that one lost dot😄. And ofcourse lecture is awesome 👍
I had a friend who liked 311 and we were all into progressive rock/metal/jazz. I love 80s synth and such. You should check out Porcupine Tree, anesthetized live, self titled song. Also the album Colors by Between the Buried and Me if you are into concept albums like Metropolis: Scenes from a Memory by Dream Theatre. Peace hah
Best thing since sliced bread
Thank you prof. Lanterman for this great video-lecture. Since i'm trying various simulation softwares to study the circuit, is there any replacement/substitute for CA3130? I can't find any LTSpice models for this. Thanks!
I have a hardcover copy of MAuP, will part with it only from my cold, dead hands.
In one sense it's "out of date" -- but no one has really written anything else like it!
@@Lantertronics well you have some spare time! It’s been superseded by a lot of technology but the fundamentals are there.
Thank you for this wonderful lecture Aaron, I appreciate all your lectures and the analysis of the external components, as I usually struggle to understand how they interact in DCOs and OTA VCFs. This is essentially a DCO core you're explaining, I've been studying the Roland and Korg DCO's and this analysis is excellent. They use NPN transistors instead of the JFET but the result is essentially the same. How would you turn it into a reverse saw or triangular waveform?
You can just run the waveform through an inverter to flip the saw direction. Turning saw into a triangle, you can center the waveform around DC, and then rectify it (take its absolute value), and shift and scale it again. I will talk about that in a future lecture.
Check out Electric Druid's VCO analysis: electricdruid.net/roland-juno-dcos/
And when you have the triangular waveform you can get a sine waveform with a pair of matched trancistors and an opamp.
I don't think that this qualifies as a digitally controller oscillator (DCO). In a DCO, the reset pulse is generated by a digital circuit, which typically divides down a high frequency clock from a master (crystal) oscillator. Here, the oscillator generates its own reset pulse, so I would say that it's a plain analog VCO.
@@possible-realities yes of course, the digital control part of it is missing, but the osc works the same way.
Someone needs to invent a ground symbol that looks good either way up.
Truth.
Question: How is the 270ns comparator time constant (and related RC component values) deterimined at design time for a 5V p-p waveform? Or is this something you can only determine empirically with a breadboard? Thanks for this video - it's well timed with a project I am currently working on and helped me understand this circuit way better.
No idea, actually. ;)
Next time I'm in my office I plan to see if I can find the original Electronotes article that VCO in Chamberlin's book is based on to see if I can find more insights.
@@Lantertronics I'm quite new to all this and I'm trying to test my knowledge a bit. But I think I have figured it out. The drain-source resistance of the FET is 25Ohm so that makes the RC time constant 117.5ns. 117*1.3 =270.25. After 2.3 times RC the capacitor is discharged about 90% which is good enough i guess.
The main timing capacitor *C* (4n7) and the *On* resistance of the JFET will have a time constant based on RC.
You want the comparator to be *on* long enough to discharge C, so you want the 15k ohm & 18pF at the comparator to have a similar time constant.
If that time constant is too long, *C* will completely discharge, and you'll get a flat spot at 0v before the ramp starts going up again.
Hi Aaron, jut a short question...i can not find numbers for the different lectures in the series. Is going through the playlist the only way to get them in the right order or did i just overlook the part number ?
When I was originally posting them I was sometimes posting them out of optimal viewing order (depending on what videos I finished editing first), and in some cases I wound up taking one video I was editing and splitting it up into two separate videos, so I often wasn't entirely sure what the final number of a video would be. I'll probably go back and add numbers at some point when I get a chance.
In the mean time, best thing to do is watch them in the order on the ECE4450 playlist:
th-cam.com/play/PLOunECWxELQS5bMdWo9VhmZtsCjhjYNcV.html
I like how fact that capacitor charge is not straight line but negative exponential form is ignored in all analog sawtooth/triangle waveform generators.
I'm not following how the charge on the capacitor would be exponential? It should just be a straight line if it's being fed with a constant current source and the op amp has low input current (like with a TL072).
@@Lantertronics Oh, you are right, of course, I've solved too many exercises about schemes and oscillators WITHOUT opamps at my course if basic electronics 20+ years ago. And, yes, one more note to myself: don't try to be smart at 4 at night.
Thank you for great videos!
Do you have an idea why this circuit uses a JFET insted of a MOSFET? I don't think Chamberlin explains why.
It acts like a reverse MOSFET. When you increase the gate voltage on a Jfet its resistance increases.
I would say that the JFET acts like a depletion mode MOSFET (except that you can never let the gate voltage go above the source or drain voltage, because of the gate-to-channel diode).
So I think the circuit would work basically the same if you exchanged the JFET for a depletion mode MOSFET. But I guess that it might be harder to find a suitable depletion mode MOSFET than a JFET? Or maybe JFETs have some other desirable properties for this circuit? Wikipedia says that JFETs have higher transconductance than MOSFETs, that seems like a good reason.
The circuit really wants to keep the leakage through the FET low when it's turned off, I guess that's why it's using a voltage swing between -15 V and 0 V for the JFET gate. I suppose that if you're gonna use that kind of voltage swing, a few volts of added threshold voltage that you would get with an enhancement mode MOSFET might not make such a big difference; you'd probably still want to drive the gate quite negative to reduce the leakage current? But with a MOSFET, you could use a comparator with a regular output and get an even wider voltage swing, not sure if that would be good or bad...
Anyway, the JFET's gate-to-channel diode doesn't seem to be a problem for this circuit, since the gate is always at = 0 V, so why not use a JFET, I guess? And even if you used a MOSFET, you couldn't let the oscillator have a much wider voltage swing, because the FET gate must have a much wider voltage swing than the oscillator to make sure that it's always on / off when it should be.
@@possible-realities I think Ian Fritz did some experiments using MOSFETs in this application.
Musical_Applications_of_Microprocessors-Charmberlin.pdf
says: 2700 pF as the timing cap, but you say 4700pF or 4nF7.
?witch one should i take¿
my provider doesn't have the 2700pF ( WIMA FKP2) TOL. 2.5% / POLYPROPYLEEN, but has WIMA nF7 FKP2 4700PF/63V TOL. 5% / POLYPROPYLEEN
Whoops! This is embarrassing -- I made an error on my schematic. It is indeed 2.7 nF. I made the slides for the Buchla 259 lecture (which I plan to tape and edit and post sometime this weekend), and that uses 4.7 nF. I created my sawtooth slides by taking the Buchla 259 circuit, stripping it down to just that op amp and capacitor, and then building up the rest of the circuit... but I forgot to change the capacitance value. In terms of the cap, you can use polystyrene, or nowadays people also use C0G or NP0 ceramic (which are high quality ceramics that are stable with temperature changes -- avoid mediocre ceramics).
@@Lantertronics thanx very much, i have watched the Buchla schematic often but mostly the timbre circuit and didn't notice, but indeed the Buchla has a 4nF7,
fluxmonkey.com/historicBuchla/buchlaFiles/Buchla_2590_2_200.jpg,
but i don't see a reset circuit, like most SAW core vco's have, but there is a feedback from the pulse "opamp" , is that the cause why it's a triangle?
like in LFO circuits? it seams the pulse is 4,29V, but nothing has giving for the triangle. strange "current mirror" for the buchla before the triangle core opamp.
.... and again thnx.
Bernie stopped selling Electronotes in Feb 2020, unfortunately.
Sigh... I was hoping he'd get somebody else to take over.
Is it in any way possible to get access to the article referred to here then? I believe it is Vol. 8, No. 62 pp 13-15 thats in question here.
@@havresylt There are a number of things from EN available for free from the Electronotes website, but the only thing from No 62 is page 12..... electronotes.netfirms.com/EN62pg12.pdf
witch one is best?
lm201 or lm301?
www.ti.com/lit/ds/symlink/lm301a-n-mil.pdf
Supply Voltage: lm201 = ±22Volts, lm301 = ±18Volts
my shop refers to lm201 as replacement for lm301
should i use TL431 Precision Programmable Reference for getting the +10V / -10V references
electronics.stackexchange.com/questions/398041/tl431-as-negative-voltage-reference
or simple use a zener and a non inverted / inverted opamp config?
www.ti.com/lit/ds/symlink/tl431.pdf
states: Sink-CurrentCapability:1 mA to 100 mA
does that mean the lt431 provides me with up to 100mA?