Parasit Studio 8-Bitar clone motorboating problem

[antonis]

As for the 470uF cap, I only have 100uFs (and 1000uFs...) at the moment so I put that in, but do you think a 470uF would be crucial?

It depends on what "crucial" means..

Motorboating occurs due to power supply dips at frequencies below double the mains frequency (see about AC rectification)
(not enough "recovery" time for a complete cycle..)

10R & 470μF form a low-pass filter of corner frequency at about 34Hz (well below 100 or 120 Hz for rectified mains -  see about 1st order RC filters)

A cap of value bigger than 470μF lowers down particular corner frequency, hence forms a more effective power supply filter, in an analogy of 34Hz X new cap/470μF..
(e.g. a cap of 1000μF forms a LPF of 16Hz corner frequency with 10R resistor..)

P.S.
In case of space availiability, you can place 2 or more caps in parallel resulting into a total capacitance equal to the sum of individual capacitors..

edit: I tend to believe that Rob never sleeps..

[Rob Strand]

edit: I tend to believe that Rob never sleeps..

The covid period hasn't helped  .

As a bit of a update. 

I had a look at how the circuit responds to power supply noise.   It's quite bad.    The way the power supply noise gets into the circuit is inherent in the way those CMOS amplifiers work.    To make matters worse it looks like reducing the amplifier gain does not help that mechanism much.   I still recommend trying the added gain resistor, except I'm less confident of the result.

The big cap idea looks like a possible fix but it might need to be quite large.   I'm thinking feeding this circuit with a regulator would be a better idea.   The reason is regulators maintain a low impedance down to DC whereas with caps the supply impedance starts to rise at some low frequency.   The motor-boating is at a low frequency.

I'm pondering over sneaky ways to fix it.

[Rob Strand]

Some ideas to try:


No guarantees they will work or not change the sound.  You might have to play around a bit.


Normally the DC bias point is set by a balance of the two MOSFETs inside of the CD4069UBE.
That setup naturally moves the bias point up and down with the supply voltage. 
Mod 1 tries to make the bias point depend more on the MOSFET gate threshold VTO, which is constant,
and less on the power supply.  That way the power supply variations and glitches are rejected more.
It seems you need to shift the DC bias voltage on the output of the first gate by about 2V before
significant supply rejection is obtained.
For this mod best to keep the resistor in series with the PSU rail that was mentioned in earlier posts as small as possible to give more scope of adjustment (say <= 100R).

[mrmet5]

Work keep getting in the way but I should be able to get back to this in a few days.. I'll keep you updated, thanks guys!

[mrmet5]

Hi guys!

So I finally finished with a few deadlines and went to grab some 470uFs, and voila, it worked! And the effect being not working was the input wire was not quite connected to the breadboard... my fault! I can't thank you guys enough for helping/teaching me!!

And if you don't mind teaching me a bit further... I feel like the output is a bit low. I know Rob mentioned actually taming down the gain after pin6 in order to reduce motorboating, but is there anything I can try to increase the output? (Is the initial gain stage(?) between pins5 and 6?)

[iainpunk]

the volume is decided by the last inverter, (pins 11 and 10), which outputs directly to the volume control. as far as i remember, when i had it on my breadboard, it was really loud! maybe a bad connection between the output of the board and the volume control?

cheers

[mrmet5]

Hey Iain!

You know what, it turns out it was a faulty jumper cable... I need to stop being cheap and buy some good ones! haha. Thanks for your advice though!