Hi, I’m not here often, but if you don’t mind, I’d like to share a new open source noise analog synthesizer project.
This is a 4-voice fully analog synthesizer, the feature of which is 8 noise generators.
This is an open project and a link to full information for ordering such a kit and assembly will be at the end of the article.
Further in the article I will talk about its structure, operation and experience of designing and manufacturing such or similar device within Russia in September 2024.
Below, I have provided a block diagram with the structure of the synthesizer (I hope it can be considered)
Since we essentially have 4 very similar blocks on board, we will consider the diagram using one of them as an example.
In fact, each voice is a separate subtractive synthesizer with its own separate settings.
Instead of the usual VCO, I decided to use two generators for each voice:
- White noise generator
- Metallic noise generator
By mixing them together and changing the frequency of the metallic noise, you can get a unique sound base for creating atmospheric or percussive effects.
I used a fairly classic white noise generator circuit based on the avalanche breakdown of a NPN transistor.
For correct operation, it is necessary to select a transistor, as a rule, noisy transistors are found in cheap kits from Alika.
In the circuit, you can see two of them, this is due only to the additional landing place on the board, for different transistor cases.
To create metallic noise, I used 6 Schmidt triggers (CD40106), with a resistor in the feedback and a capacitor to ground.
Using the feedback resistor, the frequency of each of the 6 generators is set and then summed up.
To calculate the frequency, I used the golden section rule, but of course, due to the error of the parts, there can be no talk of accuracy and these values are very conditional.
To change the height of these generators, I modulate the CD40106 by power.
To control this parameter by CV, I used one NPN transistor shunted to ground, this solution turned out to be quite simple and effective.
To evenly blend these two generators I used a crossfade on one op amp.
Next, the mixed signal from the two generators is sent to the low-pass filter (Low Pass VCF). It is implemented on the transconductive operational amplifier LM13700.
This is a fairly popular circuit based on the KORG MS20 filter, but with a single-polar power supply.
It is worth paying attention to the LEDs installed at the midpoint.
I noticed that the circuit operation heavily depends on the voltage drop on these diodes.
I usually use red 3mm LEDs, but on low-quality parts the spread is from 1.62v to 1.75v, after some measurements and experiments, I was convinced that for correct operation, diodes with a drop of 1.8v are needed.
In the current circuit, I used orange diodes with a drop of 1.8v, it seemed that they produce a brighter sound and controlled operation of the circuit.
For CV control, I also used a NPN transistor, but with a pull-up to the power supply.
VCF A IN - the input is additional for connecting an external sound source to this voice.
For internal filter modulation, the circuit uses envelope generators and LFOs, which I will describe below.
Since we started with the audio path, the next one is the voltage controlled amplifier (VCA)
This is a fairly simple circuit using half of an LM13700, depending on the current on the control contact the gain changes, so we can control the volume using voltage.
The peculiarity of this circuit is the pull-up to ground of the output buffer, through the voice volume potentiometer.
Be careful when using this circuit in your devices, without a potentiometer at the output it will not work.
Separate envelope generators are used to control the filter (VCF) and volume (VCA).
Schematically, they are identical, except for the depth potentiometer (Depth) on the filter.
This is the simplest possible circuit on a dual op-amp.
The potentiometers limit the charge and discharge current of the capacitor C29/C32.
Its capacity is responsible for the operating time of the envelope, the greater the capacity, the longer the charge/discharge time.
At the output there is an op-amp in voltage follower mode.
A transistor for pumping current for uniform glow of the diode.
But there is a peculiarity.
The voices are grouped in twos and the second filter modulation (VCF) comes from the LFO.
In this LFO there are two waveforms, its selection is organized by the potentiometer, it mixes triangle and square.
In fact, this is not the entire audio path and then a group of two voices is sent to the delay.
This is a classic delay circuit on PT2399.
Of the features, I use the input buffer as a summator with two voices.
And also control of the delay speed by CV, it is also simply done through an NPN transistor, with a pull-up to ground.
Then, from two delays, the signal is mixed at the output through an active summer with the master volume.
As manual control of the envelopes, two logical sensor sensors are used.
They are connected to each other.
One is connected directly to the envelope inputs, the second one selects the solution of the first one (with or without fixation)
This is a simple capacitive sensor TTP223, the capacity of the capacitor can be adjusted to adjust its sensitivity.
We seem to have figured out the circuit diagram, let’s move on to the design.
In general, there weren’t many questions about the diagram, but having had an unsuccessful experience of creating devices without prototypes (I’ll definitely write about this someday), I decided not to take risks and to start by assembling all the main units on a breadboard to check their functionality.
In parallel to this, I made edits to the already drawn schematic and the board being drawn on the fly, based on the corrections of the breadboard.
After making sure that everything works and transferring the edits, I returned to easyeda.
The schematic turned out to be quite large, but I divided it into equal blocks, for ease of reading.
There is no way to attach a PDF here, but I will definitely publish it as a separate file.
To reduce development time, I simultaneously prototyped on a breadboard, made changes to the circuit from it, and immediately updated the future board, which was partially ready by this time.
This device provides a “sandwich” format, so there are actually three boards.
The main board contains the entire audio path, control, and indication.
In fact, it is the entire synthesizer and can be used only by organizing control from external controllers.
The board-panel contains sensors and power supplies for them, I specifically decided to unload it as much as possible and transfer as many parts as possible to the main board.
For TTP223 sensors, a mandatory requirement is the absence of a ground plane under them, so it is not uniform across the entire board
And the bottom board, it is essentially a cover, but is intended as a board for extensions, mods or bug fixes.
It has a breadboard and pads for additional 3.5mm jacks for modifications and expansion of functionality.
After designing the boards and creating a Gerber file, I went in search of a manufacturer of printed circuit boards.
Within the framework of sanctions, cost, delivery times and the possibility of payment, I settled on the service:
Impressions from the service are quite mixed, a lot of approvals and requests to pay extra for the work, but if you refuse them, everything comes the same as it was in the order.
Ordering boards cost me $ 84, taking into account the discount.
I would like to separately note their own software for checking the correctness of the boards.
It shows controversial issues in great detail with a detailed explanation of the difficulties the Chinese will encounter when manufacturing each of these elements.
You can also get a short summary with examples in a pdf file.
I don’t see any point in describing the details, we are in different countries and the conditions are very different.
After about a month of waiting, the entire set of boards and components was received:
I did not order stencils and planned to do all the soldering with a regular soldering iron.
This is a fairly easy process when using convenient equipment.
For these purposes, I took:
- Tip type “K” (Axe)
- TAGS (flux, the choice was precisely on it because it is easily washed in warm water)
- Precise small tweezers
- Solder 0.6 mm with flux for soldering terminal elements (potentiometers / LEDs)
In general, this is a minimum set for more or less comfortable soldering of SMD components.
I also use a fairly large size “1206” in the design.
They are almost identical in size to, for example, terminal resistors for 0.125 W, so soldering them by hand is usually not difficult even for beginners.
As a rule, I start with microcircuits and transistors, and then move on to resistors, capacitors, etc.
Complete assembly by hand, with a regular soldering iron and tweezers, took two rather busy evenings.
The assembly itself turned out to be quite simple and comfortable.
Naturally, there were some difficulties, when I turned it on for the first time, I didn’t get any sound.
The output mixer/buffer refused to work.
Having assessed its circuit, it was decided that the problem was in the counterfeit op-amp.
Despite the inconvenient location, I still swept it away for a better quality one, unfortunately there were no changes.
Of course, I didn’t try to figure out the problem, but armed with a couple of 0.25W resistors, I made a passive adder for the output.
The solution turned out to be working, I plan to use the freed up space for a dip switch for different switching modes.
I also made a mistake with the choice of the landing place for the cable, I couldn’t find a suitable one in the quick access, so I decided to throw in wires, fortunately the step of the landing place is 2 mm and soldering the wire was not difficult.
As a result, I received a unique instrument with an unusual sound, for creating sound special effects, percussion and other sound design.
This time I included CV/GATE inputs on the top panel and I was right.
This instrument works great with various sequencers and external controllers, expanding the possibilities to a completely different level.
In the future, I will definitely record a video of the work and publish it separately.
You can open the entire project by clicking on the link, see the board, the diagram, and make your own adjustments.
NoiseLab by Carlo - Analog PolyNoise Synth
Other projects on GitHub:
https://github.com/EugeneCarlo
You can follow the news here:
- Blog in VK
- Blog in Telegram
Chat, learn more here:
- Chat “Online Assembly”
- Chat “Circle”
By the way, I often record circles with the operation of devices in the TG chat.
You can support not only with a like, but also with a coin here:
- Support with a coin.
If there are any errors, please pay attention to them, I will definitely fix them.
Love and music to everyone!
