Distortion Help--

[diemilchmann]

I have just gotten the PCBs for a 2 stage op amp distortion ive designed, and i seem to have run into a bit of a problem. About every 5 minutes, the pedal will slowly crackle out, to almost no sound. However, then all i have to do is even barely tap any pin of the TONE knob, and it comes back to normal. Ive double checked everything, its not a connection problem, tried another op amp, tried a large resistance from the pot to ground, etc. nothing works...

heres the schematic, sorry that its crappily hand drawn and a phone photo...

Thanks,
Julian

[PRR]

Re-read your post and the non-attached/linked image.

Hand-drawn is fine. Crappy diagram and crappy photography doesn't invite attention.

If you want free design consultation, we have to see what the heck you are talking about.

[deadastronaut]

cold joint on your tone pot?....re-solder it...

[jcgss77]

Or just possibly the pot itself?

[diemilchmann]

its not the pot or the connection, because when i breadboarded it before i did any soldering, i switched out pots etc, and i just reasoned that it was the breadboard........ stupid me. Here
s an ACTUAL schematic...sorry. Could it be some kind of weird charge building up on C3? and it discharges when i ground it with myself?

http://imgur.com/5m8zx

Thanks y'all

Julian

[slacker]

You've got nothing to correctly bias the first opamp stage so it's just luck whether it works or not. Have a look at this which  explains what you need to do http://focus.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=sloa058&fileType=pdf or just look at virtually any opamp based stompbox circuit

[diemilchmann]

ding ding ding we havea  winner. I had gotten rid of the biasing in a previous design, oops
Thanks guys!

[Ice-9]

Unrelated to your problem but should you also not have D1,D2,C9 connected to gnd or vbias. It looks like these have no connection.

[diemilchmann]

my mistake on the schematic thanks

[PRR]

> You've got nothing to correctly bias the first opamp stage

Yup.

> it's just luck whether it works or not

Not luck. If an input has nothing to tell it where to be, it is sure to drift to an unworkable state.

In CMOS the input current is exceptionally small and in theory two devices cancel. They never cancel to zero, but in practice CMOS will often assume an in-between operating point for long periods of time.... or not, depending on humidity and moon-phase.

Vacuum tube grids can sometimes hit a "valley" where they can stay stable, if plate current is low and plate voltage drops with plate current. It isn't trustworthy. And it fails utterly at high currents.

In this case the NPN-input LM358 has a for-sure input current around 40nA. If there is NO path for this current the inputs drift up, ideally to infinity and actually going dead around 1.2V below the + rail.

So why did it work at all?

He has 0.057uFd of capacitance on the input. If the input current is a constant 40nA, this will slew at 0.7 Volts per Second. So it takes about 11 seconds from zero to +7.8V, where the LM358's inputs nominally crap-out.

If you recreate the circuit, and meter pin 1 (amp 1 output) you may be able to watch it rise from zero to over 7V in about 10 seconds. And it will work over most of that range.

So why was it observed to run for 5 minutes?

What the "crap-out" really is, is that the LM358's input 6uA stage current loses voltage and starves. Input stage current falls. But transistors will work at quite low current. Performance is degraded but in this application maybe not enough to notice.

The lower input stage current means lower input current. Slower cap charging. So it may linger near +7.8V for a long time.

When does it really quit working? The LM358's second stage needs 4uA at emitter, is a lateral PNP, say 0.4uA at base. So the circuit may not quit until input stage current is well under 1uA. Input current could be near 1nA. Slew is near 0.02V/S, time to slew 8V is 450 seconds or 7 minutes.... uncannily close to the reported 5 minutes.

Side-notes: I'm not sure what R1 R2 C4 bring to the party.

[askwho69]

Weird design

[PRR]

> Weird design

Nah, not much. Booster and diode clipper with frills.

[Johan]

what's the point of the wet/dry?..it doesnt do anything other than what the "gain" does...other than perhaps filling up a pre-drilled hole in the chassis and limmit the range of the gain-pot..?..
J

[diemilchmann]

It's obviously NOT a wet/dry. But in my opinion, it almost be haves as such. Yeah, its basically a secondary gain, but to me it seems like wet/dry is a good name. Its weird.