LED flickers in sync with input

[slashandburn]

Not the only issue I'm having here, otherwise I'd be tempted to just call it a feature and forget about it. I built (yet  another) lm386 amp. My usual crude design, Tubescreamer style gain circuit before a 386 amp with a Big Muff tone control wedged in between. This time, though,  I had space in the box for a bigger board so for the hell of it I threw in  Muff Fuzz circuit on a DPDT http://4.bp.blogspot.com/-XFz_PHgaZaI/T6tqlUuAbXI/AAAAAAAABfQ/hWi7Wqfgkpc/s1600/EHX+Muff+Fuzzes.png  (after the BMP tone control, before the lm386). It's ridiculous and totally unnecessary, a really could've used the space for something more useful. Anyway, the problem seems to start when i use the switch to cut this Muff Fuzz circuit out of the chain. Muff Fuzz On, the led works fine.  Muff Fuzz Off, it blinks along with the input.



Cool feature, but I'm guessing it means something is definitely wrong. LED is is an ultra bright blue wired in series with a 20k resistor from V+ to ground.

Probably worth mentioning the tone control is also borked. I get signal (not a nice sweep, but signal nonetheless) on the first 50-60%. It crackles a bit like a SHO so I'm guessing I've got unwanted DC on there. Emphasis on the word guessing. I have no idea.

Any idea what's going on with this? I'll update with a schematic and photos soon if needed. Cheers

[GibsonGM]

Re. the LED:

First guess is a totally miswired LED.  Is this an indicator for "muff fuzz in, muff fuzz out"?  How is it wired?   Or is it the main "all on/all off" for the effect, on the bypass switch?

If it's wired correctly, it could be a power supply issue...not enough decoupling (?)   Is the muff fuzz on the DPDT wired up true bypass?  Hard to say without more info.

[slashandburn]

Hey Mike! Ah yeah I should've said. The led is an indicator for the entire circuit, normal operation I'd expect it to come in with the jack getting plugged in and stay on until the jack comes out. It does this, but only with the Muff Fuzz On. With the Muff Off it does what my guitar is doing.   Switch for MF is True bypass, no indicator.

As far as decoupling goes, I've got a big electro cap on the opamp gain stage output (I think 2.2u), between that and the BMP tone. The tone wiper goes to the switch, and is sent either to the MF, or true bypass straight to the lm386 input pin 2.


Edit: here's a quick sketch of how I jammed the Muff Fuzz in there.


And the obligatory pic of the strip board spaghetti. For all the good it'll do. 9v rail is right up top.

[GibsonGM]

Man, I'd really look for some miswiring!   If it was a PS issue, I'd expect it to do it all the time.   

So you LED is grounded to 1 side of your input jack, circuit completes with the plug insertion, right?   Could your signal somehow be getting into the ground line?  Amplified, I am picturing the LED rectifying the couple volts AC and flashing with it.....ODD.    Are you grounding the muff fuzz with the DPDT in 'off' position, and (some of) the previously amplified signal is bleeding to ground thru the LED, or some such very weird and almost impossible thing?

I'd try measuring AC voltage at the MF bypass with it off, just for curiosity's sake.  See if there's a couple volts AC there...fishing in the dark, but theoretically maybe it has the slightest merit...

[slashandburn]

Cheers Mike. Yeah I'm kinda baffled by this one. I've triple checked my switch wiring, all I can think is that maybe I should ground the input of the Muff when it's bypassed, though Im struggling to believe that this alone could cause the weird LED blinky issues and the goosed tone control.

I'm about to have another poke around and see what I can find. I wouldn't mind keeping the blinky led feature though it probably needs to be fixed, and  in any case I need to fix the tone control.

Very strange fault. Outside of these issues its somewhat functional if I keep the tone pot centered and stop worrying about the LED.

Edit: by somewhat functional I mean it sounds as stupidly over the top as you expect from a lot of gain hitting an 386 driving a 50mm speaker.

[GibsonGM]

Not 100% sure, but should the 386 have SOME resistance in series with its input?   MF should have a decoupling cap in its power pins if it doesn't already.

And you mentioned DC symptoms on the tone cap...should measure for that. If a coupling cap is leaky, that might 'do' something here...maybe the crackle is telling you something...not seeing a big issue w/how the MF is wired if there are no errors on the board, altho grounded bypass might be good.    If all else is working ok, sounds good, I'd revisit how the LED is wired, perhaps.

[slashandburn]

Oh those are good points. The 386 has series resistance only when the Muff Fuzz is enabled. There's a jumper I could replace with a resistor to see what kinda difference it makes.

And re the coupling cap for the Muff Fuzz. Doh! It's not there.

As for how its wired, kinda hard to explain my on-the-fly stripboard wiring. There's likely other missed components, missed cuts and shorts galore. I should know better than to jump into these things without proper planning. It's a wonder it works at all.

[GibsonGM]

I'd go 1K, maybe 10K.  Doesn't need poo for current to operate!  I think it may help stability, but then, I don't use 386's much, I don't like their peculiarities.

You do need the coupling cap if one is not there (you showed .1u in the schematic)....bet that would put DC on the tone pot! 

I build on perf board, don't do MUCH for layout, just enough to know things will fit, to avoid jumping signal wires all over each other etc.   Just have it make some sense, then start building, looking 3 steps ahead to I don't solder into a corner!

By your description, it seems like something is modulating your LED.  Power supply fluctuations by not enough decoupling capacitance COULD do that, possibly.  Drawing on the power supply at your audio rate...I can't tell where your LED is powered from - maybe making sure it is coming right from 9V at the source is best?  It's just a "plugged in" indicator.   Don't power from a chip's pin!  It's actually "high current" as far as these ICs see it.    And if not there, put a 1 to10u cap across the power pin and gnd of each chip (some also add a low value like 100n too).   

What are the 2 BFC's doing (big freakin' caps)?

[slashandburn]

Problem solved. I muntzed the LED and CLR out of the circuit and normal service resumed.  Even the tone control. Perhaps the CLR and led wires placement on the strip board was causing the problem? It was very close to the output capacitor from the 386. I'll have to upload a sketch later to see if someone can pinpoint what I did wrong. I'm assuming taking the led and CLR off the strip board and wiring it directly to power jack and input will give me a normal functioning LED and tone control. I guess I'll find out later when I've finished my chores.

Yeah those BFCs are the power filter and output caps for the lm386. They've been lying around for ages taking up space in my parts drawer, I figured I should use them where I had space. Originally planned an all axial layout but that went out the window very quickly along with most other forms of planning.

Also, my bad, the coupling cap on the Muff Fuzz input was there after all, I did however ditch the 47uF power filter cap since I already had some filtering  (470u and 100n in parallel, then the lm386 has its own 100u right at the pin.) and the MF shared the same V+. Probably no harm throwing one in there.

Still not sure how I managed that though. Weird feature. Sorta liked it. Probably should have measured things like you suggested. I'd probably be less baffled.

Still confused, but happy it's working now.

Cheers Mike.

[GibsonGM]

Glad it worked out!  Sometimes we never do figure out what caused the weird thing, I wouldn't sweat it, as long as the tone control is OK.

  What's a "CLR"?  Guess I haven't had enough coffee yet! 

[duck_arse]

awww, am I too late? looking at your circuit dia shows the 386 input has a DC path to ground thru the tone pot when the muff fuzz is bypassed. it probably should have a DC blocker at the 386 input, which will also cover the muff fuzz output if it is lacking an output cap.

voltage readings on the 386 input with the MF in/out of circuit? - or, conversely, voltage readings on the tone pot wiper as you work the MF bypass?

[slashandburn]

Cheers Duck (and Mike, again!) yeah it seems fixed and works as good as any of my other 386 amps. I'll definitely take you up on that extra coupling cap though, might solve a few intermittent peculiarities that I put down to the 386 being a special case and/or the crappy little speakers struggling with the output.

If we forget about the Muff Fuzz for a second (most of these little Ruby amps I've thrown together already are just preamp > passive tone > 386) does that mean I've been getting my coupling cap placement wrong all this time?  Opamp (preamp) output > coupling cap > tone > 386.

Should I add another coupling cap between Tone and Lm386 or simply move the existing cap? I see no issues with adding a second cap but if I can cull some components to keep the parts count down I'd be all for that.

Oh yeah, Sorry Mike. CLR = Current Limiting Resistor. Not sure I'm using the right term, I just meant the resistor in series with the LED indicator!

[GibsonGM]

Sure, CLR is ok, just didn't know the abbreviation, ha ha!  It IS a current limiting resistor.  That term should be more popular, wonder why it's not.  You'll still probably have to define it all the time tho!

386 data sheet says you can get DC offset problems which are avoided if you capacitively couple the thing, so I'd add the cap after the tone stack at 386 input, myself.   It's a good idea, there are a few mV there.   Yes, a new cap...the 'existing cap' is needed where it is!   

You need one wherever DC can flow into some place you don't want it...into the tone stack is bad, so you have one at preamp output to keep DC bias in there where it belongs.   I never consider the 386 as a DC source, but it is, so one goes at its input to keep it in there!     The MF has one at its output to keep ITS DC where it belongs, so that is all set.     If you just moved the preamp out coupling cap, you'd still get DC in the tone stack, and the pot would crackle.   Better to just add one...value isn't that important; maybe 1u or better?  A place to experiment to see where it rolls off....

[PRR]

The LM386 inputs can be DC-coupled to ground and it will work fine.

It may not quite meet the Power Output (clean) spec in the specifications. In our world this hardly matters.

Do not bias the inputs half-way, as we do on general purpose opamps. The LM386 does not work that way.

[stallik]

Playing with led's on my grandsons breadboard, switching between 1 and multiple led's, we noticed this effect due to the resistance in the switch. Just wondering if your led was flashing for the same or similar reasons