PT2399 Trouble in Delay/Reverb Circuit

[CooperL]

I've built a combination delay/reverb circuit using a PT2399 for delay and Belton brick for reverb.  Everything works fine on the breadboard; however, after assembling it on a perf board I have some problems with the PT2399 that I can't seem to find other people having (point me in the right direction if that's not the case!). 

Here is the schematic I'm using with the delay section zoomed in.  SW1 is a momentary stomp switch that when engaged acts like turning the feedback all the way up.
http://imgur.com/e5gacNe,LzHV6rK#0
http://imgur.com/e5gacNe,LzHV6rK#1

The problem is that at seemingly random times, the input to the PT2399 will stop working.  When it stops working, the input voltage appears at pin 16, which it shouldn't because pin 16 is the inverting input of an op amp whose output goes to pin 15.  And since that op amp is set up in feedback, that voltage should be held at 1/2 the power supply to the PT2399.  I've heard of PT2399s latching up and smoking (this hasn't happened..yet) and I have a fixed 1k resistor on pin 6 to prevent too much current sinking from that pin.  When it stops, I've been able to get it working again by pressing SW1 a few times.  It stops working so randomly though that I haven't figured out how to recreate the problem.

Has anyone experienced a similar problem?  This circuit sounds amazing when it's up and running so any help would be greatly appreciated!

[samhay]

First thoughts - it is a large circuit, so the possibility for errors when translating this to perf is high.
The resolution of your schematic is pretty poor, and I can't make out any fine detail - which switch is SW1?

I don't know what happens if you exceed to the input range of a PT2399, but one thing that can help with overload is a red or green LED from pin 7 to ground - http://www.diystompboxes.com/smfforum/index.php?topic=107226.msg973536#msg973536

[Fender3D]

Are you using single supply aren't you?
Then POT2 shouldn't be tied to Vref?

[CooperL]

First thoughts - it is a large circuit, so the possibility for errors when translating this to perf is high.
The resolution of your schematic is pretty poor, and I can't make out any fine detail - which switch is SW1?

I don't know what happens if you exceed to the input range of a PT2399, but one thing that can help with overload is a red or green LED from pin 7 to ground - http://www.diystompboxes.com/smfforum/index.php?topic=107226.msg973536#msg973536

If you look at the more zoomed in photo, SW1 is the switch connected to POT2.  I'll try using the LED, when the switch is engaged, it drives the input of the PT2399 really hard and gets distorted (which is what I was going for) so maybe that's causing a problem.  Thanks for the help!

[CooperL]

Are you using single supply aren't you?
Then POT2 shouldn't be tied to Vref?

Not sure what you mean, POT2 goes between the output of an op amp and ground with the wiper feeding back into the input of the PT2399.

[Fender3D]

Yep,
lug 1 should go to VR (or whatever you have as voltage reference, bigger pic is too much blurry to check...) if you're using single supply, not to GND

[CooperL]

Something else I've discovered...If I remove the PT2399, the voltage at the input pin sits at Vref (5V), which is what the input from the previous stage is biased at; however, I have DC blocking caps between the stages so I'm not sure why that bias is still there.

[CooperL]

Yep,
lug 1 should go to VR (or whatever you have as voltage reference, bigger pic is too much blurry to check...) if you're using single supply, not to GND

I see what you're saying.  Tying Lug 1 to Vref would keep the signal biased around Vref.  I'm confused because I'm blocking any offset with the caps right after that pot.  Am I still off?

[CooperL]

Thanks for helping, here's a bigger schem

https://drive.google.com/file/d/0BxLJIo7qTDmaX3JJaGdqMTVGcGM/edit?usp=sharing

[CooperL]

^you can download and zoom

[Fender3D]

Something else I've discovered...If I remove the PT2399, the voltage at the input pin sits at Vref (5V), which is what the input from the previous stage is biased at; however, I have DC blocking caps between the stages so I'm not sure why that bias is still there.

Whether you followed thoroughly your schematic, you should have no voltage without the chip inserted...
Check for shorts or bad part...

[samhay]

As others have said, POT2 needs to go to VR.
You are not applying THAT much gain in the feedback loop, so not sure if your problems arise from overloading the PT2399 input.

An aside - is the 47R resistor (R 28) in series with your power supply? If so, this could cause some big problems. I would expect your circuit to draw ~ 100 mA, so you will drop nearly 5V across this resistor. This will cause your 7805 to drop out. However, this is ~ 1/2 W of dissipation across R28, which would burn a 1/4 W resistor pretty quick, so I guess I am missing something?

[samhay]

Whether you followed thoroughly your schematic, you should have no voltage without the chip inserted...
Check for shorts or bad part...

+1

[CooperL]

As others have said, POT2 needs to go to VR.
You are not applying THAT much gain in the feedback loop, so not sure if your problems arise from overloading the PT2399 input.

An aside - is the 47R resistor (R 28) in series with your power supply? If so, this could cause some big problems. I would expect your circuit to draw ~ 100 mA, so you will drop nearly 5V across this resistor. This will cause your 7805 to drop out. However, this is ~ 1/2 W of dissipation across R28, which would burn a 1/4 W resistor pretty quick, so I guess I am missing something?

That 47R was a mistake in the schematic and was not implemented, good catch!

Something else I've discovered...If I remove the PT2399, the voltage at the input pin sits at Vref (5V), which is what the input from the previous stage is biased at; however, I have DC blocking caps between the stages so I'm not sure why that bias is still there.

Whether you followed thoroughly your schematic, you should have no voltage without the chip inserted...
Check for shorts or bad part...

Did a more thorough inspection of the bottom of the board and found a microscopic little short from C15 to 5V, so small that it had significant resistance and thus didn't show up on the continuity test.  This seems to fix the problem but it has always been hard to recreate so I will wait and see. 

Thanks so much guys! 

[samhay]

cool - fingers crossed that you have it working now.
Would be interested to see a gut shot...