That elkayem project is what mine’s based on. But with a good number of changes (power supply, different op amp, different optoisolator, added protections on gate outputs, added fourth gate output, added LEDs…)
Looks good! How do you do your front panels?
3U size glass epoxy board is drilled. The design is printed on an adhesive sheet with a home printer.
The envelope follower is coming alone nicely. I’ve decided to not call it Envelope-O-Matic v2 (v1.0 is part of the analog vocoder I built some time ago) but Follow-O-Matic (the pressure from the marketing department to come up with something new was significant!).
Just in case: the yellow trace is the audio signal, the blue one the output of the envelope detector. I scaled them so that you can see they align quite well.
Depending on the clock rate of the sampling circuit and the smoothing applied to the output of the peak detector one can choose how snug the envelope is followed.
I’m building a double one, the first one is working well now, it is also producing gate signals (you can choose a threshold level and determine when the gate fires). Next thing to do is to get the 2nd one to behave as well.
This should be useful for making side chains and such.
This picture shows the gate which triggers depending on a threshold:
Looks like you are getting good results. Have you tried to make an “auto-wah” sounding patch with it yet?
I wonder if we could do a r2r ladder dac and get some good results. I wonder what bit depth would be necessary.
Forget R2R ladders made of discrete components…
To get any precision out of a R2R ladder, the resistors have to be manufactured together on the same substrate, and to get more than 8 bits of resolution, they have to be laser adjusted.
With 1% resistors, you can get a whole… 5 bits of resolution ! With a questionable linearity.
If you go 6 bits with 1% resistors, you can get non-monotonic behavior, that is, while incrementing the bit code, the output value can DROP at some points, usually at the transition from 011…11 to 1000…00.
Nope, but that is on my list of things to try. So far I’ve been checking the circuitry and looking at oscilloscope images.
Had a similar problem with my step divider. 2 “diode-AND” cascaded 4017, worked with a 555 but no joy with external clock sources. It wasn’t even clocking at all. A discrete schmitt-trigger before the clock input like in this schematic fixed it.
To be clear, i have used R2R as an 8-bit DAC with success. Im sure it not perfect, but my experience just doesnt align with “it doesnt work”.
For example, i used one with my abandoned eeprom sampler project and it worked just fine. Also considering this isnt audio but CV i would expect the resolution to matter less, not more. Could be very wronng though.
I built a DAC using R2R maybe 20 years ago and hooked it up to the printer port of my pc as a sound (output) interface. This worked nicely although the sound of course was a bit grainy.
Ran a 25’ hdmi and a usb cable from my desk thru the wall into my music room. Will now be possible to use daw to record much easier
That is a gorgeous setup. Do you happen to have plans for that case? I’d love to build one just like it.
Thank you!
Short answer is no there are no plans. I built as I went but I talk about it a little here with some general dimensions and pics:
And
That works for me! Thank you!
Today, a breadboard.
Tomorrow, the world! a MFOS WSG!
I decided to have a bash at using Fritzing, it’s a bit clunky and this layout is wrong; after reading the spec sheet, I thought I should be using pin 3 of the PJ-102AH 2.1mm barrel jack, not pin 2.
Here it is in reality. Maybe someone could clear up something for me. The red LED is spec’d at 1.8v to 2.1v 20mA
, so should I be calculating the resistor from the 9v supply, so 9 / 0.02 = 450R
, or from the voltage drop to get to the voltage the LED is spec’d at, so (9 - 2) / 0.02 = 350R
…? I went with the later and used three 1K in parallel. If I used two in parallel, or just the one, there was no magic smoke, just dimmer LEDs. I’m assuming it’s all about current draw, rather than voltage.
It’s been so long, I literally can’t tell you the last time I put components into a breadboard. Given my middle aged eyesight and that I now wear varifocal glasses, I shouldn’t have bought the transparent breadboards, as I can’t see the legs on the resistors, etc, when they’re over the board. I had to break out a magnifying glass to see where I was inserting everything…
That’s it. That’s minimum resistance to get maximum (rated) current for maximum (burn your eyeballs out) brightness, it’s fine to use a larger resistor for less light.
Spent some time in the garage today:
Lessons learned:
- I need something to sequence
- I need something to mix
- I need something to record
- I need to get good
Cheers
Less current tends to mean less risk for magic smoke
LEDs are current-driven devices – they generate light by plugging electrons into “holes” in the semiconductor material, turning most of the excess energy into photons instead of heat; the more electrons to play with, the more light you get, until the LED overheats. The resistor is there to limit the total current; without it the LED would happily accept all the current it can get, making itself look like a short to the power supply and frying itself in the process.
The current through a resistor depends on the voltage over the resistor, so you should remove the LED voltage drop (which is voltage over the LED, not the resistor) if you want to hit max current.
But as @analogoutput says you can skip that, and go for a higher resistance; most LEDs will light up at a tiny fraction of the max mA. 1k is common in 12 V circuits (~1 mA), and you can often go beyond that, especially if you have high-intensity LEDs.