Little Angel Chorus Issue

[yaruhas]

Hello,
I am currently building a PT2399 Little Angel Chorus right now for a friend of mine, but I came across an issue as follows:
The effect works, however, if I disconnect the power supply and connect it back again, the effect disappears. It is instead replaced by a clean signal. Then when I make a connection between 9V power supply and ground for a brief second, it will reboot the effect back again. This did not occur on breadboard, and I have also rebuilt the board from scratch, replaced the DC jack etc, but still cannot resolve the issue.

[madbean]

The Little Angel is not designed for rapid power cycling so the issue may be that you are plugging the power back in too soon.

Here's what happens: The PT2399 needs a minimum of at least 1k resistance between Pin6 and ground at power up. Less than that and the chip will lock up. However, the chorus effect works best when the delay time is as short as possible with the chip. To get around this, there is an anti-latch circuit to prevent the lockup and keep the short delay time. This is the 2n3904 and associated circuitry attached to pin6.

When the circuit is first powered, the PT2399 sees the fixed resistor attached to pin6 (usually a 4k7) so it boots up. Then as the anti-latch circuit charges up current flows to the base of the 2n3904 and thus resistance between collector and emitter goes minimal. The transistor essentially acts as a parallel resistance (some versions also have a 47R reisistor attached to the collector) to the 4k7 which drags the total value down to a minimum, thus making the delay time go as short as it can.

When you unplug and plug power back in the 10uF cap takes a while to drain its stored current to ground so the transistor may still be conducting. When that happens, the PT2399 now sees too small of a resistance to boot and it locks up.

How to prevent this: don't rapid cycle the power like you described or put a resistor in parallel with the 10uF to ground to speed up the current drain. Probably a 100k or 220k will create a high enough time constant for charging and discharging but it may take some experimentation to get the right value.

PS All this is assuming you have the anti-latch circuitry and there are no problems with it in your particular build.

[madbean]

Oops I just realized my drain resistor idea is not ideal because it would create a voltage divider with the 68k resistor attached to the 10uF cap. Maybe putting a 100k before the 68k would be better.

You could probably design the LA to incorporate draining that cap as part of bypass switching with the LED attached somehow. But, it's a lot easier to just no unplug and plug power in too quickly.

[yaruhas]

Thanks for your response. My current circuit does not involve the anti-latch circuit. I will try it out though!
Here is the link to the schematic that I am currently using
https://i210.photobucket.com/albums/bb292/frequencycentral/LittleAngelMiracle.png