EA tremolo works great... for 2 minutes and then quits.

[Bishop Vogue]

Ever have a client actually return a pedal for a refund?  This was my first and hopefully my last.  When I built this pedal, everything worked great and it passed my QC.  But maybe my QC isn't rigorous enough because the buyer returned it saying the tremolo just quit working after about 2 or 3 minutes.  Turns out, he was right... but I have no idea what might cause this (I read somewhere that the LFO can 'catch up' but I'm not really sure what that means) or how to correct it.  Anyone here have any ideas?  Thanks a million!

[Rob Strand]

It sounds like a leakage problem.  Maybe a leaky part or even a cap in reversed.

BTW, the best production technique for catching these types of things is to do a burn-in.
The simplest case is to check the pedal (undisturbed, not powered cycled or reset) at the
end of burn-in.  Better is to monitor it throughout burn-in.

[Bishop Vogue]

Thanks Rob, I'll definitely check the caps.  I had to google burn-in... at the least I learned something. 

[Rob Strand]

I'll definitely check the caps. 

I was thinking a solder joint could do it, weird it would be so repeatable though

I had to google burn-in... at the least I learned something

I've had a lot of success with it.  It's really good at preventing the dead on arrival stuff
and early fails which make you look bad in the eyes of the customer.

Sometimes it won't pickup bad connections. I've seen devices working for a week
then someone touches it and it fails - which turned out to be poorly crimped cables
(which is another painful story).

[PRR]

> google burn-in...

Bathtub Curve

When you put stuff together, sometimes it is wrong from the start, sometimes it goes wrong a little later. "Usually" you get a clump of early failures. The survivors may have low failure for a long time. In late life you have wear-out failures. "Bathtub".

I don't call a clothes-line "done" until I have seen it hold clothes on a couple windy days. Then it may be OK for some years until the rope or tree rots.

The classic example is incandescent light bulbs. Hammering tungsten into filaments is not an exact thing. Lamp makers try very hard to make a filament 100% perfect and sized to give exactly 1,000 hours life. But thin spots happen. These run much hotter and may burn-up in a minute, an hour, a day. Lamps get at least a little burn-in to find the ones with thin-spots.

Westinghouse(?) "discovered" that they could double the in-service reliability of aircraft radio tubes with a 52-hour burn-in before final testing. (That's actually a very old practice: 2 hours for $1 home-radio tubes, 100 hours for costly transmitter tubes; Westinghouse just needed a theme for that month's magazine ad.)

[Bishop Vogue]

SOLVED.   I just wanted to add this note here for posterity.  Maybe someone else will have the same problem.  My problem as that whenever I set the rate to min or max, the led stopped pulsing and the tremolo vanished.  Here are the changes I made that have solved the issue.  The 'pulsing' LED should be white. The RATE pot should be changed to a 25K.  Add a 680 ohm resistor between RATE 3 and board.  The result seems solid, though the rate can't achieve motorboat speeds now. Thanks for all the help, folks.  It was another learning experience.

[Rob Strand]

Here are the changes I made that have solved the issue.

The root cause could be something to do with the transistor gain and the oscillator not starting.
I remember when that LED version came-up on the forum(15 years ago?) it was pretty
tricky getting it to hold together.

[duck_arse]

I've found that running the flashing led and its clr in parallel to the pso collector resistor [and using a largish clr and superbrite led], isolates the oscillator some from the led. so you get the oscillator running right and steady without the led, and then adding the led doesn't affect osc operation adversely.