Guitar LED circuit sucking tone

[artifus]

never trust a maplin goon. (no offence to any maplin employees who may be reading) they just pick parts from a labelled draw. have you checked your resistors with your multimeter? how many bands to the colour code? check under good strong white light to confirm value with online calc if no multimeter.

*edit* only just spotted your photo on page two - red, red, black, yellow, brown - 2.2MΩ 1% Warning: Non-standard 1% (E96) value. Standard E24 value. Reverse: 14kΩ(2%)  reverse as in brown, yellow, black, red, red. from: http://www.hobby-hour.com/electronics/resistorcalculator.php

?

[deadastronaut]

grant is  colour blind IIRC...
sometimes that red can look brown in certain light to me too..

[MrStab]

i despise Maplin with a passion! lol. they're the biggest joke of a company in existence. i just happen to have a bunch of 1M's from there, but the 2.2M resistors i have are from a fairly-popular UK retailer for pedal parts and i wouldn't like to discredit them as they've been great with everything so far. i can't imagine i would've gotten unlucky twice from different sources, when i've used parts from the latter to successfully build much more complex boards.

the schematic as it stands in my desperation is just 2 resistors soldered together with a 9V battery clip hooked up to either end. can't get more basic than that!

thanks for tolerating me here lol

[MrStab]

please excuse my not-so-great camera yet again

1M resistors (from Craplin):


2.2M resistors (from online):


Battery test:

[Gurner]

Since you've got the most basic voltage divider in play....

1. connect 9V across these two identical resistors - note down the voltage at the junction.

2. Also, note down the battery voltage whilst the battery is physically connected to the same two resistors (you've already measured it @9.73V but that was with no strain on it)

3. reverse the battery wires to the same resistors (i.e. put the battery black wire where the red wire was & the red wire where the black wire was) - again, note down the voltage at the resistor junction

The above test should tell us something about the relationship of the value of the resistors

(I'm beginning to think your multimeter is loading this most basic of circuits down.....indeed if you get the same voltage reading at the junction for the tests above, that's the likeliest problem)


FWIW ....IMHO you should only be using Alakaline batteries....not zinc chloride from poundstretcher!

[MrStab]

had a wee laugh out loud moment at the poundstretchers comment - however did you guess? i'm kinda broke. in my defence, i favour less-cheapskate batteries in stage applications. that's not really a defence at all. lol
tried a Duracell out of my wah at some point, just to rule out the battery causing wrong voltage.

right! so i understand direction #3 and i think #2 (basically test from the battery terminals whilst resistors connected?), but i don't get what you mean by #1? haven't i already done that in the pics, or do you mean something else?

i can only think of it being the multimeter, too - yet i used it to confirm PT2399 voltages just the other day. maybe there's some nuance at play which i don't understand.

it's quite inconvenient that "Voltage Divider" makes for a potentially-dangerous acronym.

[Gurner]

right! so i understand direction #3 and i think #2 (basically test from the battery terminals whilst resistors connected?), but i don't get what you mean by #1? haven't i already done that in the pics, or do you mean something else?

Yes you've done it (but I didn't know which resistors you'd choose!), so again...

1. connect 9V across these two identical resistors - note down the voltage at the junction (yes, you've done this once, butr we need to be scientific here!)

2. Also, note down the battery voltage whilst the battery is physically connected to the same two resistors (you've already measured it @9.73V but that was with no strain on it...so you need to also test the battery voltage when 'loaded' with the two resistors connected)

3. NOW reverse the battery wires to the same resistors (i.e. put the battery black wire where the red wire was & the red wire where the black wire was) - again, note down the voltage at the resistor junction

Don't over think this...just carry out the above tests explicitly & report back!

[MrStab]

2.31-2.32 V (slight fluctuation) with positive & ground at EITHER side.

9.72 @ battery with load.


overthinking tangent: if it is the multimeter, and the 4.5 reading at the at the output on pin 1 (& inverted input via. jumper) did suggest correct biasing, why might the "raw" circuit be loaded down, but not post-opamp?

[Gurner]

2.31-2.32 V (slight fluctuation) with positive & ground at EITHER side.

9.72 @ battery with load.

Well that proves the resistors are of equal value (and you haven't a bad 'un in there)......I reckon you've not only bought nasty poundstretcher batteries, but also a nasty poundstretcher multimeter! (or equivalent)....false economy - look how much time you've wasted for the sake of an extra fiver!

overthinking tangent: if it is the multimeter, and the 4.5 reading at the at the output on pin 1 (& inverted input via. jumper) did suggest correct biasing, why might the "raw" circuit be loaded down, but not post-opamp?

because the junction of two high value resistors is very high impedance - easily loaded down by a poundstretcher multimeter, whereas the output of an opamp is very low impedance....even a sh1t Chinese sources poundstretcher multimeter won't makes it's voltage output waiver!  You are right though...knowing what we know now, the opamp biasing was fine, but your pin measurements were very askew (due to your multimeter), which made it look like there was something seriously amiss with the opamp.

remember this...

If you've the circuit as outlined in my last post yet still are seeing just 1V at the junction, it can only mean three things...

1. You resistors aren't the values you think they are (but they look fine in the photo - so I doubt it)
2. There's still something else connected to the track that those two resistors join.
3. The device (or person!) doing the measuring isn't working properly!!

To give this mini peculiar diversion some closure, it was reason number 3!

[MrStab]

hey, i'll have you know this multimeter cost me £6! lol
good to know that it's not ideal, especially given that i bought it to bias an amp in the first place... why do i smell burning...
i've only been at this pedal-building lark since mid-february and this is the first issue i've run into with it (or so i think), so i didn't think i'd get far enough to warrant a better one.

right, so after this unmitigated disaster, next time i run into dough i need A. a better multimeter and B. a bulk supply of decent batteries. it wasn't so much a case of saving an extra fiver as not having an extra fiver!

i'm gonna get back to work on this project, hoping i don't burn out my last LM386 (no output when i audio probed one of the 2 burnt ones), and to get this thread back on track i'll let you guys know how it goes.

thanks a lot for helping me figure it out! makes a lot of sense that it's the multimeter.

[Gurner]

FWIW a decent multimeter needn't cost the earth...I have a reasonably priced & spec'ed Uni-T UT61E  multimeter (Unit-T are a respected brand), which cost about £35 on Ebay, but their range starts much cheaper, for example...

http://www.ebay.co.uk/itm/UNI-T-UT-33B-LCD-Palm-Multimeter-Ohm-Volt-meter-UT33B-/330744556929?pt=UK_Mobile_Phones_Communication_Radio_Meters&hash=item4d01e96d81

[MrStab]

bookmarked, cheers! i was a bit of a fish out of water when hunting for one, always been more of a computer nerd until recently. all i knew was i needed a 200mV resolution at the time. the irony was i rejected a few of the ones i looked at because they seemed a bit too flimsy. i can probably do without a transistor testing function for the time being.

[slacker]

Perfectly decent multimeters can produce  similar results in that situation. It's one of those things you learn to watch out for, Gurner's little test is a useful check. You can do the same thing it a circuit, where you wouldn't want to reverse the battery by measuring from the junction to ground and then from the junction to the supply voltage. Courtesy of Gus Smalley.

[Gurner]

Perfectly decent multimeters can produce  similar results in that situation.

I've just tried this same test with 2 x 1M resistors with my multimeter.

Battery voltage ....8.5V (loaded), junction voltage 4.182V (ideally, this should be measuring 4.25V, but it's close enough that in a troubleshooting scenario I wouldn't be knocked off course with a seemingly out of whack reading) ....but yes, Ian makes a good point .....even with a reasonably spec'ed meter there'll inevitably be a degree of loading/impact wrt high resistances  ...just one of those things (but can be negated a fair bit by sticking with a reasonably trusted multimeter brand)

[MrStab]

i'd much rather have 4.something than half what it should be!
trying to absorb as much as i can, this'll make for handy knowledge in future. will try to score a multimeter by the end of next week.

as i'm the kind of impatient moron who's gonna keep trying stuff until then (not too "mandatory" a project, i have fun just putting this stuff together):

probing reveals that audio passes all the way through both IC's as intended, but the LEDs stay on permanently. assuming the obvious, i tried a 1k resistor after the 10uF gain cap on the LM386 to lessen the gain a bit, but still no. no overheating (yet), though! my limited knowledge tells me that maybe something's getting into the 386 that shouldn't. or would the buffer stage change the characteristics enough that i should maybe consider taking out the gain cap altogether? as Rob's original design (and its rip-offs across the web) seem to assume a pure guitar input. (i would just go ahead & try this but unable to right now & it'd be interesting to know why the buffer may or may not affect the power-amp/LED stage)

[Gurner]

So back to the original circuit...this is a very inelegant circuit!

For a start, the voltage across the leds is gonna change over the fade life of the LEDs...so even if you biased them perfectly on a fresh battery, as soon as the battery voltage drops, the LEDs are gonna be incorrectly biased.

If you measure the voltage coming out of your LM386...you'll see it's about half your supply voltage....this is with no signal.

So even without a signal you've got about 4.5V (assuming a fresh 9V battery) DC present ...if you place an led & series limiting resistor across it - the LED always be on.

What colour LEDs are you using? (also measure the voltage across one of your LEDs when its on)

[MrStab]

i'm currently only testing with 2 blue LEDs in series, as per your advice i'm not gonna try running 4 in series in the final thing - will no doubt need to experiment with parallel and/or red LEDs. i've since taken the electrolytic cap labelled C2 out of the circuit (what would've been the buffer output cap), in case that matters.

voltage at LM386 output reads 7.5 (although my DMM could be lying!), 5.6 after the current-limiting resistor (1k), 2.75 before the 2nd LED. i take it something's wrong there if it's not 4.5, or are you assuming the power to the 386 comes after the voltage divider? 9v+ feeds into the LM386 directly, i figured i should keep both "halves" separate in that regard.

is there any way to elegant-ise it?

[Gurner]

i'm currently only testing with 2 blue LEDs in series, as per your advice i'm not gonna try running 4 in series in the final thing - will no doubt need to experiment with parallel and/or red LEDs. i've since taken the electrolytic cap labelled C2 out of the circuit (what would've been the buffer output cap), in case that matters.

voltage at LM386 output reads 7.5 (although my DMM could be lying!), 5.6 after the current-limiting resistor (1k), 2.75 before the 2nd LED. i take it something's wrong there if it's not 4.5, or are you assuming the power to the 386 comes after the voltage divider? 9v+ feeds into the LM386 directly, i figured i should keep both "halves" separate in that regard.

is there any way to elegant-ise it?

Last question first ....IMHO ...no.

Can it be made to sort of work? Yes......but well? That depends on your expectations.

Ok, first problem, your LM386 output voltage is wrong (this is why your two blue leds are on all the time)...the LM386 output pin should be sitting at about 4.5V (half its supply voltage) - remove the LEDs and remeasure (btw the voltage here is low impedance so you can trust your multimeter!)

with all the haze of that earlier problem, I've sort of lost the plot a little & perhaps it's worth you posting voltage for all you LM386 pins (it's not a chip I use much so it's probably better to let other chime in here)

[MrStab]

LM386 WITH 2x blue LEDs connected:

3.54
0
4.59
0
7.48
8.17
9.14
2.01

WITHOUT LEDs:

3.58
0
4.62
0
7.76
8.38
9.22
2.01

i've wasted more than enough of your time, G - no worries if you wanna take some time out!

i'm just surprised quite a lot of people seem to have tried the 386-LED thing with no regard for tone suck. is it an inherently problematic concept, or only so with the buffer in front?

[Gurner]

pin 3 should not be at 4.59V...you need a cap between your tl072 output & the LM386 input (I've just gone cross eyed trying to follow your board layout & mapping it into a schem in my head!)

you'll also need a pull down resistor to ground on the right hand side of the cap (LM386 side)...something like 10k ought to be ok.

Try those things & see if your LM386 output returns to a more normal 4.5V(ish)

Re tone suck...I guess those who make the circuit are so won over by lights that they overlook the impact on their tone!