Pretty Years clone only works when briefly shorting regulator to ground

[slowpogo]

I built the newly-released Pretty Years clone from DEFX (Petty Tears). Initially it wasn't working, but as I was troubleshooting and measuring voltages at various points, I accidentally shorted pin 1 and 2 of the LM317 regulator with my probe. Then the pedal suddenly came to life and worked perfectly in all other respects. Shorting the regulator seems to "jump start" it somehow, I have to do this anytime power is turned on again. I reflowed the solder on the regulator and nearby parts but no change.

Linked the build doc/schematic below. Any idea what might be the issue? Bad regulator?

Seems like possibly a grounding issue so I'll mention this as well... the build doc suggests using shielded coax cable and only connecting the shield on the input ground lug, and output ground pad. However it didn't pass a signal at all when I did this (active or in bypass). I connected the shield at both ends of input/output and then it passed the signal. I have not housed the pedal in an enclosure yet, if that matters.

https://drive.google.com/file/d/1_b0BPyUz5bh0LZNeIGyKICXwt4M49_M_/view

[slowpogo]

After trying a million other things I popped in a different LT1054 I had laying around and, voila, it works great now. Unsure if I damaged it somehow or it was just faulty.

WICKED pedal, by the way. I knew it did all kinds of crazy stuff with the resonance and filter matrix and whatnot...I didn't realize it would also be fantastic as a "normal" low to medium gain pedal. Really nice saturation from the tubes, and it has a unique dusty sound to it, like an aural polaroid. Fitting since the tubes are about 65 years old. Among the larger more involved builds I've done but really glad I did

[antonis]

It should be a good idea to make R2 120R..
LM317 needs 10mA current (and NOT 5mA as is erroneously refered in datasheets for over than 40 years..) between OUT & ADJ pins (internal circuitry operating current returns through the load) so it might cause start-up issues when feeding light load..
(about 240R load needed to be set in parallel with R2 here - dunno if tubes' parallel cathode resistance fulfill it..)

[digi2t]

When I built the validation unit, I initially tried a TI LT1054 that I bought from Tayda. None of them worked. It would not fire up at all. I went into my stash and found some Analog Devices branded 1054's (from Mouser iirc), and they work fine.

Maybe Tayda is selling rebranded MAX1044's? Who knows. Not buying those from them again.

Nice to hear that someone has a successful build going. Are the instructions regarding room for the jacks adequate? Please send me some pics of your build.

[MikeA]

This is from the TI LT1054 data sheet, and the Analog Devices datasheet says the same thing.  I experienced a similar issue to the OP with LT1054 startup in a dual-supply circuit, and adding the Q and R to the negative bus fixed it.





 

[digi2t]

Seems like possibly a grounding issue so I'll mention this as well... the build doc suggests using shielded coax cable and only connecting the shield on the input ground lug, and output ground pad. However it didn't pass a signal at all when I did this (active or in bypass). I connected the shield at both ends of input/output and then it passed the signal. I have not housed the pedal in an enclosure yet, if that matters.

The way the grounding scheme is illustrated in the build document requires that the audio jacks be contact with the enclosure for continuity, with ground being provided by the sleeve of the CV jack through the enclosure, otherwise yes, you won't have ground at the audio jacks with the works out of the enclosure. The single sided coax shielding for the audio jacks maximizes shielding, and prevents forming a loop.