Hey folks,
My group and I are working on building a musical instrument using mostly analog circuits and are running into trouble with beat frequencies and sound quality. We're getting some pretty strong beat frequencies and would love some advice from an audio engineer on how we can combat the ****** sound we're getting. Here's some more about our system:
12 555timers in parallel for each frequency -> LM 386 op amp as a signal combiner -> bandpass filter -> (amplifier/filter (https://www.amazon.com/5V-12V-Ampli...words=arduino+amplifier&qid=1700168828&sr=8-3) chip -> 5W speaker. Currently, individual notes are working well but combining 2+ frequencies results in a large degradation in sound quality.
What could be the source of our terrible sound quality when playing multiple notes? I understand that we're amplifying twice, we also attempted to mix the individual signals together by wiring them together directly through 120k ohm resistors. Any thoughts?
Thanks in advance
My group and I are working on building a musical instrument using mostly analog circuits and are running into trouble with beat frequencies and sound quality. We're getting some pretty strong beat frequencies and would love some advice from an audio engineer on how we can combat the ****** sound we're getting. Here's some more about our system:
12 555timers in parallel for each frequency -> LM 386 op amp as a signal combiner -> bandpass filter -> (amplifier/filter (https://www.amazon.com/5V-12V-Ampli...words=arduino+amplifier&qid=1700168828&sr=8-3) chip -> 5W speaker. Currently, individual notes are working well but combining 2+ frequencies results in a large degradation in sound quality.
What could be the source of our terrible sound quality when playing multiple notes? I understand that we're amplifying twice, we also attempted to mix the individual signals together by wiring them together directly through 120k ohm resistors. Any thoughts?
Thanks in advance
Last edited by a moderator:
Beat (sum/difference) frequencies are typically generated when two or more tones are added and the result is passed through a sufficiently non-linear circuit. The solution would be to scale down the proportion of each of the tones by an amount enough to avoid driving the summing opamp into non-linearity.
I would suggest that you try reducing the proportion of each tone to 1/12 th of its current value, so that the opamp wouldn't saturate even when all the 12 tones are played together.
I would suggest that you try reducing the proportion of each tone to 1/12 th of its current value, so that the opamp wouldn't saturate even when all the 12 tones are played together.
An LM386 is a small audio power amplifier IC rather than an op-amp. How did you get it to work as a summing amplifier and filter?
Yes, Marcel has a point. It may be better to use a summing/mixing stage before the power amp as below. The output voltage is given as:
Vout = -Rf / Rin [V1 + V2 +V3]
Vout = -Rf / Rin [V1 + V2 +V3]
With NE555 you produce simple square waves. Square waves were the very first attempt building electronic organs. And definitely the poorest way to generate any sound. If you are really interested in high quality sound you may study the synthesizers of Bob Moog. But be warned - analogue implementations of that kind lead to a steep learning curve.
In 2015, I saw (among other things) analogue computers in the British National Museum of Computing and later that year, an early Moog synthesizer in the Dutch Teylers Museum. They looked remarkably similar to me.
First if the 555's are the standard sort, replace with CMOS versions such as the 7555, much better behaved to the supply rails, the original 555 crowbars the rail heavily and several on the same supply will almost certainly interact because of this. You should low-pass filter the harsh square waves at some point to avoid saturating the subsequent liear signal processing. RCRC filter with a ~40kHz rolloff might be a good start, 1k/3n3/1k/3n3 maybe?
Secondly a LM386 is not a low distortion opamp (especially in the ultrasonic range), which is what you need to combine signals, especially if there are ultrasonic components that can intermodulate - there are plenty of good audio opamps out there, what supply rail voltages do you have available?
Last edited:
Your circuit is clipping.
Clipping one square wave results in a square wave, but clipping the sum of two square waves results in intermodulation. Use the right amount of attenuation. Even better, don't use square waves - they sound awful.
Ed
Clipping one square wave results in a square wave, but clipping the sum of two square waves results in intermodulation. Use the right amount of attenuation. Even better, don't use square waves - they sound awful.
Ed