Noise Flaws

Bring The Noise!

This post will be me looking at why the SS-30M is so infernally noisy.

Since I brought it inside and started fiddling with output levels a bit I've become far more distracted by the noise-floor of the SS-30M. With a quiet 'studio' space with high-quality monitor speakers and headphones, the noise is far more obvious. Maybe not unignorably bad, but certainly not very good.
 
So far I have identified four types of noise

1. Unidentified 'ticking'
2. The Orchestra section's bucket brigade delays
3. The 'swarm of bees' effect 
4. General hiss

I'll start analysing each one and return to this post as and when I identify the root causes or fixes for each.

MIDI Interface 

The MIDI interface is the major new component in the system. On that basis alone it's a prime suspect for all sorts of issues. However, the first thing I did was disconnect the power to the interface and none of the problems were even slightly better.

Just a tick

This was a new (to me) and unexpected sound when I turned the gain up on the mixer and really listened. What is it? I don't know, but it sounds like a kind of click or tick sound every quarter of second or so. 

Possible sources - PSU switching, an oscillator (such as the orchestra LFO), other

PSU switching

If the PSU was a switch-mode type you might expect some switching noise (a ripple voltage), but that would be occurring at the switching frequency, which is much much higher. 
The reason I'm thinking of this is that when faulty they can exhibit similar ticking sound. It can't be that because the SS-30 supply is a linear supply.

But in a linear supply, there's no switching at all. It simply rectifying the 50Hz AC and filtering that. 

So, if it's coming from the PSU, I need a better explanation. 

LFO

 This occurred to me after switching off last night, but could this regular tick be coming from something the circuit which is already regular, like an LFO.

Orchestra LFO

The  Orchestra LFO is a pretty good candidate because a) it's operating in the right frequency range and b) there's a loose wire hanging off it at the moment. The wire is for the LED circuit, which I have yet to implement fully. You will recall that I had a breadboard set-up to demonstrate this. I remember when I got this design wrong, to begin with, drew too much current and caused ticking sound at the frequency of the LFO. 
So, I really need to check that wire and play with the Orchestra rate knob. If the ticking changes with the rate, then I have my culprit. If. 

Vibrato

The Vibrato LFO is a fixed frequency and is much higher than the Orchestra one. At around 6.4Hz, it does not fit what I'm hearing




That's all I have on the ticking, for now...

Orchestra Noise

 http://www.stefanv.com/electronics/oh-the-noise.html


That link is a nice article about reducing noise in a similar situation. The SS-30 has a high-cut filter on the output but is it enough? And if I change it, what then? A noise gate would be one way to go, but I can do that externally. 
The main issue that even when there is no signal going in there is noise coming out. Using the separate voice outputs should solve that to some extent, but it's not ideal. Preferably I'd be able to switch off the orchestra output altogether. I say preferably but it would lead to a set of control settings which produces no sound, which is something I'd prefer to avoid. The only way around that would be to have another balance control to replace the wet with a dry signal, but that's getting silly. A switch could be used instead which would bypass the orchestra and the orchestra depth controls. All more wiring, but possible.

The other thing about this noise is that it's masking the other noise. If I decide to leave it on, then there's not much point working on the others. 

 Swarm of Bees

 The so-called swarm of bees effect is the sound of all the notes buzzing away. As we know, the string synth design depends on keeping all the notes running continuously and raising their normally negative DC level towards 0V, so that they start to pass through diodes. If the negative DC voltage is too high you get all of the notes bleeding through the diodes even when no key is pressed. 

A possible source of the problem: supply voltage too high/positive, switching circuits not switching 'off' properly; sustain circuit not decaying fully, switching voltage from MIDI interface not 'off' enough leading to some bleed from the optoisolators, oscillator signal to large (peak-peak). 

So, lots to got at, there! This will be hard to isolate. As they are apparently all going at once, fixing one switching circuit will be hard to hear. However, with careful use of the scope, it should be possible to trace if they really are all bleeding through and if not all which ones. It will take some work though. 

Hiss-trionics

 General noise (hiss) is the bain of all analogue designs. Starting with the power supply and infecting every part of the design, noise is a fact or life. But some things are there to help and must be proven to still be working - i.e. de-coupling capacitance - and somethings can be upgraded - capacitance again, but also newer ICs. Not to mention the noise from electromagnetic interference. 
Good grounding, shielding, decoupling and so on will all help, but I'm not too hopeful. There's a risk of doing a lot of work to get a small benefit. I can't replace some components easily (or even at all) and working on each board is a major task, unless I desolder the wires and test each in isolation. Big jobs! Like, 'I need to quit the day-job' big. And finally, everything is now squashed into a small space. I need to be realistic and decide how bad it really is. 

I suspect that compared to the Orchestra noise the background noise-floor is quite small.