[sdiy] TB-303 humming noise problem

[cheater cheater]

Hi guys,
the TB-303 which I have here has a fairly annoying issue. I was
wondering if someone could help me figure out what to do about it.

Short description: there's buzzing associated with the pattern
selection LEDs blinking. Below I give a long description of the
symptoms, then what my measurements showed, and then what I tried.


The setup is, I have a nearly completely vanilla 303 (changed the
output capacitors for larger, and changed the cutoff adjust resistor
for a trimpot), and it is operating from batteries.

When I  turn it on, the pattern selection LED is blinking. As it is
blinking, when it's on there's a faint buzzing noise, around 1 kHz,
audible through the headphone output. It's not audible if the
headphone is connected to the line out; it can be heard if I use a
headphone amplifier. It is actually an alternating sound: when the LED
is on, it's a more buzzing, lower tone, and when the LED is on, it's a
much fainter, higher tone (maybe around 8 kHz?). The faint high tone
can be barely heard via the headphone output.

If you e.g. press the pitch mode button and hold it, the selected
pattern's LED stays on. The buzzing is constant at that point. Bear in
mind that all the while the normal mode LED is on (constantly, not
blinking).

It does not change when I change any of the 6 small synthesis
parameter knobs. Naturally when I change the tempo knob the LED starts
blinking faster, alternating the sound more rapidly. The volume of the
buzzing does not change at all when the volume knob is turned. The
buzzing does, however, change in sound when I select another track,
and when I select the pattern section (A/B). It does so in groups:

section A:
patterns 1, 2, 4: alter between high-pitched faint tone (led is off)
and a sort of 5th tone, low (led is on)
patterns 5, 6, 8: alter between same high-pitched faint tone (led is
off) and the base tone of the 5th interval from above (led is on)
patterns 3, 7: only the high-pitched tone (led is off and on)

section B:
patterns 1, 2, 4: alter between the 5th low tone (this time when the
led is OFF) and a tone maybe half an octave higher, just as loud (led
is on)
patterns 5, 6, 8: there is a nearly constant pitch somewhere between
the tones of patterns 1, 2, 4, however there's still some modulation
of the tone when the led changes state
patterns 3, 7: only the 5th low tone (both when the led is off and on)

Notice that in both sections, the LEDs for patterns 3 and 7 are
darker. Is this normal? Also the section A led is darker.

This is what I've measured:
1. the BU terminal is completely clean, as is the +12V rail.
2. The 4.5 rail and the 6V rail and the 9V rail all have this buzzing
easily visible on the scope. It turns on and off with the LED, and so
on. According to the gain setting, the ripple should be about 5mV each
for the "5th interval low tone" buzz, and only barely visible
otherwise.
3. Testing across C65, the buzz is clearly visible. It is also clearly
visible testing across C1.
4. The buzzing cannot be seen on either the line or headphone output,
even on the highest magnification setting.


Here are things I have tried:

1. new batteries, of course.
2. using a power supply. I tried several.
3. checking if the gnd cable from the switch board is attached
correctly. I also tried crocodile leads either going between the same
junctions, or going to the battery pack's negative terminal.
4. testing capacitors in the power supply. They all tested low ESR -
between just under 1 Ohm up to 2 Ohm. Similar new elco's measured in
the same range.
5. checking for cold solder joints. Tried redoing all I could
identify, and can't find any more. I'm left with simply redoing every
single solder joint.
6. placing a 10nF bypass capacitor across C65 (it's the one that the
arrow points to on the PSU schematic, the "5" is unreadable). I
believe it was an MKT.

I would be thankful for any suggestions. This buzzing is ruining an
otherwise perfect synthesizer, and I hope I can improve its operation
somehow.

Thanks,
D.