PT2399 tempo indicator needs demystified!

[R O Tiree]

The oscillation frequency from the PT2399 is well over 1MHz even at the longest usable delay time and ramps up smartly as you decrease the delay time. The highest frequency that the human ear can discern is widely quoted as 20kHz - 50 times lower! Ergo, it is the logic circuitry that is causing the ticking that you can hear, not the PT2399.

If it's space on the board that's a worry, try point to point wiring the electrolytic caps and the protection diode to the 7805 and solder the 1nF caps that I suggested direct to the power and GND pins of the 4040s. It's not ideal, but it will prove one way or the other what's going on. There might also, given that this is a vero layout, be some interference being picked up between traces - not short-circuits, but inductive/capacitative between tracks. When I made a similar circuit for an Echo Base +++ (tap tempo, rate LED, better filtering, pre- and post-emphasis) it was my first attempt at isolation routing. Having a GND plane around the digital chips (and a separate one for the analogue side as well) clearly helped, as well as keeping digital GND separate from analogue GND and properly smoothing the supply to the digital side of the circuit and each and every digital chip meant that this aspect of the circuit was absolutely silent as the grave.

Actually... have you routed the GND from this daughter board back to the DC jack (or Input jack Ring terminal) or have you taken a feeder wire from the main board?

[MrStab]

so as per your experience, "geography" plays an important role here - first time i've ever run into that beyond keeping power/signal wires perpendicular to each other. so, from what i understand, keep analogue/digital grounds separate, never meeting?  all i know in this regard is that when it comes to the PT2399 (even though we're talking about the 4040's), the analogue & digital ground are connected by an internal resistor. at the moment the ground connects straight to lug 3 of the Small Time's buffered on/off SPDT, which basically goes to a star ground along with everything else. i'm yet to implement a 2.1mm power jack on this build, as i never made the push into making it "stageworthy", but the 4040 board is indeed grounded to the battery negative via. the input ring jack. up until now i've just grounded everything together in a paranoid fashion..

not to make a habit of saying this and not honouring it, but i'll get round to rewiring to the regulator and trying out those caps over the weekend or next week, once i'm done with this build for a friend (stupid mislabelled non-clearcoat...). shame i didn't leave some arbitrary connection sticking out of the PT2399 clock pin, so i could give him a tempo indicator as a "Super Awesome Upgrade", as was popular in 1980's computing lol.

[R O Tiree]

Ultimately, all the GNDs have to come together at some point - ideally the DC jack - otherwise it just won't work. You can see, I hope, that 20mA from an LED switching on and off that is running along the same conductor as your tiny little signal GND is going to affect things somewhat? On the other hand, running that same 20mA along a different wire to your "star GND" point will not have the same effect.

When I built my Echo Base +++, I didn't even bother with a battery - just not enuff room and not enuff mAh in the battery to power the damn thing for any useful period Hence, I was able to send everything back to the DC jack for the GND connections. I connected the 2 GNDs from the PT2399 to their own GND planes - DGND (digital GND) and AGND (analogue GND) and, even though they are connected internally by a resistor, as you said, this helped. It's good that you kept the GND connections from the 2 boards separate - good instinct, there.

What's actually happening here is that the 4040s are busily counting away and the diodes are a poor-man's AND gate... once they're all "active" the counters reset back to zero, the transistor fires up, lights the LED for a short time and round we go again until the next time everything lines up. It's the pretty big current spike inherent in the cumulative effect of the 4040s going back to zero, the transistor switching between cut-off and saturated and the LED current that are all adding together to make this ticking noise. It might be that this current spike is causing V+ to sag just a little. Caps act as a reservoir to smooth this out. Quite why you are finding that it doesn't happen all the time... I have no idea.

[MrStab]

update (finally): tried a 47uf cap right after the 78L05, & a 10nf connecting each 4040 Vcc to ground, and a 1n4001 "wrapped" around the regulator. no change, it still ticks. sounds like a cricket or something.

i've tried both on mains and battery - MerlinB's notes suggest running this circuit off a battery would be futile so earlier i just wired up a 2.1mm DC jack and got rid of the battery clip i had in there. it seems to tick all the time now. it's not obnoxiously loud but it's annoying and would probably be audible in a practise or live situation. should i try a bigger electrolytic?

i'm aware all grounds have to come together, maybe the way i type makes me seem 0.01% more dumb than i actually am or something lol.

[R O Tiree]

Try putting a 100µF cap between Vin and GND on the 7805 as well?

Try a 1k resistor for R5 - that will reduce the current draw through the LED. The LED will be a bit dimmer, but the current through it will be a lot less, which might help.

I didn't ask before, but are all the diodes the correct way around? Check, and re-check, go and have a cuppa, noodle some choons on your guitar, whatever, and then go back and have another look... while we're at it, check values of resistors as well.