Help with basic limiter/VCA at end of diy synth signal path

[floating-water]

I did some buffering with some tl072s on the input and out for the reverb, just treated it like a reverb pedal, seems to be working good. I need to get a small speaker for the build; I understand the smaller the ohms the more current it draws? For the tiny wattage of the lm386 are there any specs I should look out for like what does nominal rated power mean? Does it matter?

[PHOBoS]

[floating-water]

I have the 386-1 (of course i do) so would a 4ohm be too little/cause problems? Im running out of budget just want to get the right thing.

I've got this:
https://www.ebay.co.uk/itm/363076889353?ul_noapp=true

or this:
https://uk.rs-online.com/web/p/speaker-drivers/1218637

Going back to the synthesizer, everything's sounding great. I was wondering however, how would I switch between no fuzz-delay and just oscillators? No matter how much resistance I put between the outputs of the oscillators, any more than one osc going through the fuzz is inaudible. I have a main switch in the middle which is three way:

https://www.ebay.co.uk/itm/Guitar-toggle-switch-3-way-chrome-black-or-gold/390322262654?hash=item5ae105127e:g:FhsAAOSwC8pb5wRj

taken from a les paul style gutar. I can't seem to figure out how I could have:

switch in bottom position: off
switch in middle position: just osc's (i've got to this point)
switch in top position: osc with effect

Could I just wire the audio through the switch? What the proper way to do it?

[PHOBoS]

[floating-water]

Can't thank you enough for your help mate, ive learnt a ton. You're right about the switch but it will do for the 1st version of this project, there doesn't seem to any sound degradation.

[Ricko]

Hi.

Five ideas.

First, would adding a volume control get you a long way there? It can be just a single potentiometer.

Second, would a permanent cut in volume get you a long way there? So just a voltage divider (two resistors) at the outpit.

Third, I often put soft clipping on my modules. So this is like the anti-parallel diodes as above, but with blue or purple or green LEDs (highest voltage drops) preceded by a 1k or 2.2k resistor in series. You need the resistor otherwise the LEDs can blow, but getting much light out is not the deal. And the LEDs act like a clipping indicator if they do light up.

Fourth, it is possible to use a vactrol as a kind of passive limiter. I am putting out some boards for that, (fricko "tame") in November.

Fifth, if youbare a builder, consider thus fast limiter: https://sound-au.com/project67.htm

[floating-water]

Thanks for your info Ricko, I'm yet to attempt the limiting section just yet as still working on the noise making part; just a continuation of this, if anyone could help, is there a way to stop the oscillators affecting each others pitch, say, when one of them is at their highest on the potentiometer, it will affect the others. They are RA transistor oscillators

Another is how could I build a variable capacitance? would simply putting a variable resistor in the middle of two caps work? The two caps being the threshold of the minimum and maximum capacitance. It's for a kind of 'release' control on the trigger switches for the osc's

And the powering situation is causing some headaches, I've been breadboarding using a 9v batt but I want it to be mains powered, i' wired up a dc jack, did the diode/cap filtering but theres a load of noise. Is it most likely from the psu? Is there any procedures anyone takes for filtering noise in a basic 9v dc in?

Thx

[Ricko]

Hi. If you have several oscillators, and tuning one affects the others, your circuit has some coupling.

So it could be occuring through the power supply, in which case you add more decoupling capacitors as close each of the oscillators as possible (e.g. 1uF and a 10 nF capacitor from the +ve power to the 0V and probably the same from -ve to 0v if you use it).

Or it could be coupling on their outputs, if you have a simple mixer. Do your ocillators outputs go through op-amp buffers? Buffers are a good way to prevent interaction when outputs are mixed together.

For a variable capacitor, there is such a thing (a varactor, as used in old radios) but no synths have used them for maybe 60 years.

You cannot make a variable capacitor in the same way with two caps and a variable resistor. (Each end of the pot roration might be what you want, but in between it is not a simple response: probably more like a cross fade between the two capacitor values rather than a capacitor diminishing in value.)

[floating-water]

Thanks Rick, ye Im just using the passive resistor mixer atm, how do I go about mixing them with an op amp? I have some tl072s, I understand the inputs and outputs but how does one do any mixing with them? Do they not need dividing with resistors?

Another thing, is there a correct way to wire switches/buttons or 'keys' for the oscillators, do I need pull down resistors? And why are they necessary?

edit: ive hooked up a tl072 'correctly?' with the output and the inverted input connected and the osc going into the non inverted but the out put is basically nothing

Thanks

[Ricko]

A good way to apporoach problems is to be systematic.

So lets start with the theory that it is the passive mixing causing the interference. Are they just resistors or are they pots? If they are pots, does turning the knob change the severity of the problem?

If resistors, what value are tou using! You would want at least 10k. To ecperiment, at the passive mixer disconnect all of the oscillators except one (if there was one oscillator that was worst, keep that.) Now see if changing the other oscillators still has an effect. If it does not, then the mixer is probably the culprit.

What people usually do is have an op-amp buffer (eg a non-inverting buffer) at the saw output of each VCO. That isolates it from the effects of parallel circuits.

On the "pull down resistor" I expect the correct term to look up is voltage divider: two resistor to reduce a voltage that is too large. (A pull down resistor may look the same, but the name us used for a dufferent effect: to detetme what the signal is for "open collector" circuits or open switch circuits.)

[floating-water]

i'm using 47k resistors on each output from the oscillators. im hoping to use a centre negative 9v supply for this thing just fyi but have only ever tested on a 9v battery for now. don't op amps need a bi-polar supply? what powering situation would you suggest?

also when the oscillators are at their fullest/highest they stop making sound but assuming the current is still being drawn

i'm using either 10k or 50k pots on the oscillator frequency tune

[PHOBoS]

not necessarily a bi-polar supply but you will need a reference/bias voltage. With a bipolar supply you can use the GND for this
which can also be used by other ciruits to keep everything biased at the same voltage. Since you have a single supply you will
have to create a fake GND or reference voltage. if you only use it for the input(s) of an opamp the current draw is very low and
you can get away with just a voltage divider made with 2 resistors and a capacitor.

so something like this:


or this:

[floating-water]

Right so thats 9v going into the vcc, ground going to ground, output of voltage divider going into the non inverting inputs...

(and thats all the oscillators going into one inverting input on one of the op amps? i thought I'd need one op amp for each oscillator)

...and then the output of the first op amp going into the second op amp.
whats that cap and 1m resistor at the end? is it necessary?


edit: if i read that correctly, whats the limit to how many oscillators I can add to one op amp?

further edit: regarding power supply implementation; if im making a supply for a centre negative psu does the diode have to face a different way? can I just copy a power section from a pedal?

[floating-water]

ye that worked mate, amazing

[PHOBoS]

[floating-water]

Thanks so much for that info.

No doubt a stupid question but im on a time constraint, how do you recommend wiring the ground and power rails if you're hand-wiring it, also considering there is a starve section for oscillators. If the starve section just comes off the power rail, but last, will it affect the rest of the power?

[floating-water]

The oscillators now seem to rise in pitch slowly, is there something I should be aware of with opamp summing?

[PHOBoS]

Most important thing for power distribution is low resistance. Use wires that are thick enough, though you don't need a lot of current so they
don't have to be massive. Also if you have different circuits it's better to wire them to the power rails seperately than daisy chaining them,
so not from one circuit to the other.

[floating-water]

OK that makes sense for the power distribution, least resistance and separation. What are those things called that you can screw in or layer multiple connection points? I can't find the right thing on ebay. I want to say power terminals but it doesnt come up with what I've seen before.

The pitch issue resolved with proper power supply that finally arrived

[PHOBoS]

terminal blocks or screw terminals ?

[floating-water]

I think screw terminals, ordered. Another little thing if I were to use coins as conductive keys or switches whats the concept of that. I assume the coin needs to be tied to ground but would your finger also have to be touching ground when completing the circuit? Thanks

[PHOBoS]

For resistive touch control you need to bridge 2 contacts with one side usually connected to GND it can also be made very sensitive where you provide
the GND connection yourself. You can also use capactive touch sensors which you don't actually have to touch so there is no electrical connection with
your body. You can get those very cheap these days and they work pretty well (search for ttp223 touch sensor).

[floating-water]

Ah beautiful that looks pretty useful. After some experiments today i've just got them tied to pin 2 & 3 of the pitch pot and using them as extra pitch fluctuators I guess theyre just changing the resistance when u touch them.

Phobos befire we spoke about putting a reverb into an amplifier (lm386) and I''ve come to a point in the build where im figuring out the speaker and output jack wiring. I'm having an issue however, on the breadboard, its fine either with the speaker or with output jack but combined on the lm386 output...not so good, suddenly it distorts crazy. Is there protocol for wiring the outputs out? Also so that when you plug in an external speaker the internal one turns off?

edit: distortion seems to have gone now, is it a stereo jack that is required in order to cancel out the speaker when something is plugged in?

[PHOBoS]

Just a mono jack with a switch contact will do the trick. Most common are the ones that switch one or more of the pins,
which is what you need, there are also jacks available that have a completely isolated switch.
this page has some information which might help you a bit https://www.cuidevices.com/blog/understanding-audio-jack-switches-and-schematics

[floating-water]

I've completed the amp build but theres some what I believe to be ground hum for example, if I touch the switch that switches between the speaker and the headphone out (slight change in the design) you can hear your finger touching it. How do you usually sort this? Everything is connected to the same ground, are there things I need to be consider in the future to prevent this? Could I fix it quickly now? edit: it stops if i connect a crocodile clip to anyone of the grounds and touch the other end and as for the enclosure, its wood not metal