Post by keurslagerkurt on Apr 19, 2021 8:32:09 GMT
Hey all!
I thought I'd shed some light on my latest DIY aspirations since I had some (moderate) succes over this weekend! I'll try to document the project a bit here, and try to elaborate on my learning & thoughts process.
Learning more about the legendary 555 timer chip & the popular LM13700 OTA chip was on the wishlist, and while browsing through some projects from the fantastic Thomas Henry I stumbled upon the schematic for his 555 based VCO. Now there are quite some 555 based VCO online, mostly 'Atari Punkconsole' style noise machines (no disrespect!), but this one is quite a different beast. Watching some sound demo's made me really impressed by its sound & functionality. It has great tracking, big sounding 'hard sync' inputs (& fm inputs), a pretty massive saw, and something that would make me very happy in AE: an analog sinewave! This is the demo I watched:
The build is all standard parts and the 555-timer based core already runs on a single supply. So I figured that it might be possible to make an adapted circuit for AE! I started breaking down the circuit on paper, trying to break it down to just its triangle osc core circuit for a first test. I read up on the 555 chip in my copy of The Art Of Electronics, watched some video's on the LM13700 OTA and well... I had a lot of fun haha!
So here is the breakdown of the parts I left out to get to only the necessary parts for the triangle VCO.
Breakdown of the colors:
So that leaves us with the middle part: the 555 timer, one buffer and one OTA (=LM13700). Its a really simple and elegant circuit.
So to understand this basic circuit, you need to know one important basic thing about the 555's operation:
The voltage at the 'Trigger' (pin 2) and 'Threshold' (pin 6) pin controls the voltage at the 'Output' pin (pin 3). When the voltage at the Trigger pin drops below 1/3 of the supply, the Output pin swings to the positive supply rail. When the voltage at the 'Threshold' pin exceeds 2/3 of the supply, the output pin swings to the negative supply rail. That's all of it!
So the output pin can swing from the negative supply rail (for us: ground), to the positive rail (for us: 5V). This swing happens instantly. So this means we can get a square wave at our output pin. So we need a circuit to make a triangle wave out of this one! This is handled by the LM13700 pin. The LM13700 is a so-called 'operational transconductance amplifier', its a special type of 'opamp'-like chip that outputs a current instead of a voltage. It's a pretty hard to grasp IC (harder than opamps imo), but its use here is fairly simple: it will dump a steady flow of current on the 2200pF capacitor next to it.
Why we need a steady flow of current? Well, while a capacitor charges, it will build up a higher voltage and will actually try to 'resist' any incoming current harder and harder. So when you would tie eg a capacitor to a steady 5V, it would charge up, but not in linear fashion: while it is at 0V at the start, it will charge really fast (because you tie 5V to 0V). A bit later, while it is at eg 4.5V charged up, it will charge slower (because 5V & 4.5V is a smaller difference). This will give a logaritmic voltage curve. But we need a linear, straight line to make our triangle! So we don't need a steady voltage to charge the cap, we need a steady current that will charge the cap at the same rate at any time.
The last two things you need to know about the LM13700: it will compare its both inputs (+ and -), and depending on which one is bigger, it will dump a positive current, or it will 'suck' a negative current. So it can both charge and discharge the cap depending on its input pins. And the last thing: you can see a weird input at the bottom of the IC-pictogram, where Thomas Henry inputs the 1V/Oct signal. This input determines the gain of the LM13700, so depending on how much current you dump into that pin 1, the LM will output more or less current, making the triangle go up/down faster or slower. Handy for a VCO!
So the last question: why does it oscillate? It's the genius combination of the 555's threshold capabilities and the LM13700 steady current capabilities. Let's say we connect the circuit to power. We could suppose for a start, that the buffers output starts at 0V. So the Trigger pin is at 0V, which is less than 1/3 Vcc, which will make the 555 Output pin go to 5V. This will make the + pin of the LM13700 go up, higher than its - pin: it will start dumping a current on the capacitor, creating a linear slope up. This linear slope is fed back to the 555 via the buffer. So this slope up is actually connected, via the buffer, to both the Trigger & Threshold pin of the 555 timer. Once the ramp up reaching 2/3 of the Vcc, the Threshold of the 555 timer will start acting, making the output pin of the 555 timer go low.
Once the output pin is low, the LM13700 will actually start sucking current from the capacitor, because its - pin is higher in voltage than its + pin (that one is at 0V now because of the output of the 555). So the LM13700 will start discharging the capacitor in a linear fashion. And you can see it coming: when this ramp downwards hits 1/3 of the Vcc, the trigger input from the 555 timer will make its output pin go high again! We got ourselves some oscillation. More specifically, a triangular oscillation between 1/3Vcc and 2/3Vcc, so for AE: between 1.66 and 3.33V!
I adapted some resistor values to make this work smoothly for the AE 5V, this is what I came up with:
Some extra notes:
- Its important to tie the reset pin to the Vcc (the 555 resets itself as soon as this pin is low)
- I connected pin 1 of the LM13700 to 5V with a 47k resistor. The bigger the resistor, the lower the current, the lower the LM's gain, the lower your oscillation frequency & note. If you want a semi-musical range, you can use something like a 200k resistor and the biggest pot in series with it to change the resistance upwards (I used 1M).
- You can tie the bottom of the 2200 pF cap just to ground. My drawing says 2.5V, but ground will work fine and is easier.
Next up:
- adapting the ramp for AE.
- adapting the sine wave circuitry for AE
- The hardest part: adapting the 1V/oct tracking
- hoping Thomas Henry likes my adaptation
All help, tips, questions, whatever are welcome as alway of course!
Cya
Kurt / Zeno
I thought I'd shed some light on my latest DIY aspirations since I had some (moderate) succes over this weekend! I'll try to document the project a bit here, and try to elaborate on my learning & thoughts process.
Learning more about the legendary 555 timer chip & the popular LM13700 OTA chip was on the wishlist, and while browsing through some projects from the fantastic Thomas Henry I stumbled upon the schematic for his 555 based VCO. Now there are quite some 555 based VCO online, mostly 'Atari Punkconsole' style noise machines (no disrespect!), but this one is quite a different beast. Watching some sound demo's made me really impressed by its sound & functionality. It has great tracking, big sounding 'hard sync' inputs (& fm inputs), a pretty massive saw, and something that would make me very happy in AE: an analog sinewave! This is the demo I watched:
The build is all standard parts and the 555-timer based core already runs on a single supply. So I figured that it might be possible to make an adapted circuit for AE! I started breaking down the circuit on paper, trying to break it down to just its triangle osc core circuit for a first test. I read up on the 555 chip in my copy of The Art Of Electronics, watched some video's on the LM13700 OTA and well... I had a lot of fun haha!
So here is the breakdown of the parts I left out to get to only the necessary parts for the triangle VCO.
Breakdown of the colors:
- Blue = 1V/oct circuit, doing linear to exponential conversion, FM, etc. Makes no sound in itself, pure utility
- Yellow = The 'Hard Sync' effect input. Restarts the wave when it gets a trigger. Also only 'utility'/effect.
- Green = The conversion from triangle to ramp output, has rescaling & mirroring. Not needed for just the triangle.
- Orange = rescaling of the triangle & passing through for the sine wave circuit on the second stage
So that leaves us with the middle part: the 555 timer, one buffer and one OTA (=LM13700). Its a really simple and elegant circuit.
So to understand this basic circuit, you need to know one important basic thing about the 555's operation:
The voltage at the 'Trigger' (pin 2) and 'Threshold' (pin 6) pin controls the voltage at the 'Output' pin (pin 3). When the voltage at the Trigger pin drops below 1/3 of the supply, the Output pin swings to the positive supply rail. When the voltage at the 'Threshold' pin exceeds 2/3 of the supply, the output pin swings to the negative supply rail. That's all of it!
So the output pin can swing from the negative supply rail (for us: ground), to the positive rail (for us: 5V). This swing happens instantly. So this means we can get a square wave at our output pin. So we need a circuit to make a triangle wave out of this one! This is handled by the LM13700 pin. The LM13700 is a so-called 'operational transconductance amplifier', its a special type of 'opamp'-like chip that outputs a current instead of a voltage. It's a pretty hard to grasp IC (harder than opamps imo), but its use here is fairly simple: it will dump a steady flow of current on the 2200pF capacitor next to it.
Why we need a steady flow of current? Well, while a capacitor charges, it will build up a higher voltage and will actually try to 'resist' any incoming current harder and harder. So when you would tie eg a capacitor to a steady 5V, it would charge up, but not in linear fashion: while it is at 0V at the start, it will charge really fast (because you tie 5V to 0V). A bit later, while it is at eg 4.5V charged up, it will charge slower (because 5V & 4.5V is a smaller difference). This will give a logaritmic voltage curve. But we need a linear, straight line to make our triangle! So we don't need a steady voltage to charge the cap, we need a steady current that will charge the cap at the same rate at any time.
The last two things you need to know about the LM13700: it will compare its both inputs (+ and -), and depending on which one is bigger, it will dump a positive current, or it will 'suck' a negative current. So it can both charge and discharge the cap depending on its input pins. And the last thing: you can see a weird input at the bottom of the IC-pictogram, where Thomas Henry inputs the 1V/Oct signal. This input determines the gain of the LM13700, so depending on how much current you dump into that pin 1, the LM will output more or less current, making the triangle go up/down faster or slower. Handy for a VCO!
So the last question: why does it oscillate? It's the genius combination of the 555's threshold capabilities and the LM13700 steady current capabilities. Let's say we connect the circuit to power. We could suppose for a start, that the buffers output starts at 0V. So the Trigger pin is at 0V, which is less than 1/3 Vcc, which will make the 555 Output pin go to 5V. This will make the + pin of the LM13700 go up, higher than its - pin: it will start dumping a current on the capacitor, creating a linear slope up. This linear slope is fed back to the 555 via the buffer. So this slope up is actually connected, via the buffer, to both the Trigger & Threshold pin of the 555 timer. Once the ramp up reaching 2/3 of the Vcc, the Threshold of the 555 timer will start acting, making the output pin of the 555 timer go low.
Once the output pin is low, the LM13700 will actually start sucking current from the capacitor, because its - pin is higher in voltage than its + pin (that one is at 0V now because of the output of the 555). So the LM13700 will start discharging the capacitor in a linear fashion. And you can see it coming: when this ramp downwards hits 1/3 of the Vcc, the trigger input from the 555 timer will make its output pin go high again! We got ourselves some oscillation. More specifically, a triangular oscillation between 1/3Vcc and 2/3Vcc, so for AE: between 1.66 and 3.33V!
I adapted some resistor values to make this work smoothly for the AE 5V, this is what I came up with:
Some extra notes:
- Its important to tie the reset pin to the Vcc (the 555 resets itself as soon as this pin is low)
- I connected pin 1 of the LM13700 to 5V with a 47k resistor. The bigger the resistor, the lower the current, the lower the LM's gain, the lower your oscillation frequency & note. If you want a semi-musical range, you can use something like a 200k resistor and the biggest pot in series with it to change the resistance upwards (I used 1M).
- You can tie the bottom of the 2200 pF cap just to ground. My drawing says 2.5V, but ground will work fine and is easier.
Next up:
- adapting the ramp for AE.
- adapting the sine wave circuitry for AE
- The hardest part: adapting the 1V/oct tracking
- hoping Thomas Henry likes my adaptation
All help, tips, questions, whatever are welcome as alway of course!
Cya
Kurt / Zeno
