Hey guys, just a general question.
If you have a breadboard, a 5v 1a supply, and in a lot of your projects you need to have the left rails on the breadboard independent from the right rails, (isolated from noise and current drops, would back to back schottky diodes provide a good way of accomplishing this?


Thanks in advance

 

No. Use two independent supplies if you really need that isolation.
Also at least in my projects, I never had the need to use two isolated 5v supplies. Why? Do you have an optocoupler between them?

 

No I am working with audio circuits and need 88 isolated rails. Would be impractical to use 88 supplies. I need to understand how to isolate the supplies.
I guess a better question would be, how can I split 1x 5v 5a supply to have many isolate 5v rails? Lol I think I am asking a lot of people by asking this

 

You can somewhat isolate them with ferrite beads in both the +5V and common lines of each circuit, with many decoupling capacitors on each circuit.
Run lines from each circuit separately to the supply.

Why do you think you need such isolation?

 

I am building a big synthesizer, and using it as my motivation to learn electronics. I have the art of electronics, but im not far enough into it for this.

So my main issue is, and I dont want to give away my design concept on here because it is pretty insanely epic...
But lets just say, I am using cmos timers, and the issue I am encountering with just two oscillators, is changing the resistance on one oscillator can slightly change the resistance on another.

Now I am super amateur, so maybe there is a very easy fix that I dont know of of isolating the resistance in each part of a circuit.

But Im not advanced enough for that yet.

I am sorry to be so vague, I cant give much more info, but what I am making, once I am good enough, has the potential to pay the bills...

The bottom line, is resistance and noise would be ideal if they were isolated and I am too amateur to know how.

I heard of ferrite beads, especially when isolateing Agnd and Dgnd. Might give those a shot.

Does that make it clear as mud?

 

For a start, make sure each circuit has only one common 0V current path back to a central point so any current from one circuit cannot cause a voltage. change into another.
You probably do not need to have isolated supplies, just a well designed and filtered single power supply, decoupled to each circuit.
Have a search for "single point earth" and "earth loops".
And, a bread board is not the way to build such things. You will need a good circuit board layout.

 

I'm not building using a breadboard, I am prototyping lol!

Thanks for the tips!

 

changing the resistance on one oscillator can slightly change the resistance on another.

Without more detail I can't say for sure, but I can say for sure that is not the case. Frequency might be changing, but a fixed resistor is ... fixed. The problem is something else.

Electric pianos and synthesizers have been around a long time, and *no one* has ever has 88 isolated power supplies. Besides, all 88 signals have to come together at some point to form the output signal, so there goes any signal isolation. Better to clean up the power locally than to add 88 power transformers to the project.

Back to the secret circuit ... If you are using an oscillator circuit based on CMOS logic gates or inverters, know that the transition voltage (the input voltage level where the output changes state) changes with temperature, pressure (press on the chip and the frequency changes), and the actions of the other gates in the package. For example, if you are using a CD40106 schmitt hex inverter as six oscillators, adjusting one of them will have a small effect on the others.

ak

 

Its not a fixed resistor. When I change the resistance in one, the resistance changes in the other slightly.
For diagnostic purposes, I have oscillator 1 outputting to the left speaker osc 2, the opposite.
Frequency, in these timers, is altered by changing the resistance.

Even with the left speaker turned down, a change in the right oscillators frequency, will change the left oscillators frequency slightly. by 10-200hz depending on what freq I am at.
Here is the kicker, It happens when the oscillators are close to matching frequencies.

If I were to describe it in terms of water, its as if the frequency adheses " like a finger slowy coming out of a glass of water" to the harmonic range tone until it is no longer within one of the several resonant frequencies.

I know this is complicated crap.
All I can say is, if I connect the two chips separately on their own psu's, this doesnt happen.

I think i will try the ground as suggested above.
Seems like a right course of action.

 

Not wanting 88 psu's, didnt say that :S
Yes all 88 channels do come together, using an audio mixer that I will design at the end.
The mixer itself could be on its own psu.

 

But lets just say, I am using cmos timers, and the issue I am encountering with just two oscillators, is changing the resistance on one oscillator can slightly change the resistance on another.

Are both oscillators sharing one IC package?

 

No, two packages.

 

And no issue when on separate power supplies and routed into an audio mixer.

 

Run a test tone through it and use a scope to see what's actually happening. Check the DC lines too, are they still flat? If not then maybe you need more filtering. Be aware of what types of components you use; for example ceramic caps can pickup signals from nearby wires or traces so maybe electrolytic would be more appropriate, etc..

As mentioned above, with analog circuits the physical layout can be very important, so it may be difficult to troubleshoot on a breadboard.

 

You don't need multiple supplies. One supply will do, unipolar or bipolar, depending on your circuit requirements.
What you need is proper power supply filtering on your circuits.

 

You don't need multiple supplies. One supply will do, unipolar or bipolar, depending on your circuit requirements.
What you need is proper power supply filtering on your circuits.

Winner Winner Chicken Dinner.
Thanks boss

 

Whats wierd, is it only happens at octaves.
Which tells me its a harmonics issue.
Which maybe too complex for me right now.

 

Okay so thanks to mr software.
Yes it was a harmonics issue.
Yes it was solved by dropping a 4700uf cap between v+ and v- on each individual osc.

Thank you all for your help.

 

Great job! That's a big cap, you can likely get away with something much smaller. You might even get a better result with a smaller cap, as different size caps are more effective at different frequencies. Here is a great tutorial on youtube about it.

 

You can somewhat isolate them with ferrite beads in both the +5V and common lines of each circuit, with many decoupling capacitors on each circuit.
Run lines from each circuit separately to the supply.

Why do you think you need such isolation?

I'd use more inductance and bigger reservoirs for audio stuff.

Its the grounds that can tie you up in knots - do everything star point. Even ground planes may have one or two downsides.