envelope/edge detection

[Gurner]

still error on the schem - the led cathode goes to ground *not* vref.

Just about any small signal schottky will be better than what you're using (you want the lowest fwd voltage drop you can find for diode D1)

Re a rail to rail opamp....like I say I mainly work with 5V circuits so have none I' familiar with & can recommend (hopefully someone else can chime in)

[deadastronaut]

edit:  morning..

i upped the R2 TO 56k..much more sensitive...

i'll try a pot on the r8 too in a bit...

i'll have to nip out and get a schottky.. ....hmmm

http://www.maplin.co.uk/search?criteria=schottky

[Gurner]

looking good. i would drop R5 to 470R or even 330R.......at 7v DC out of the last opamp, a 330R resistor for R5 would be about 20mA through the Red led (briefly)...so it's within saftey zone.


R6 likely needs to be increased (say 4.7k) & R2 something like 20k-33k

[deadastronaut]

hmm not sure on which schottky to get though...i don't want to get the wrong one as maplins is several miles away from me now....(they closed my local one the b......ds )

http://www.maplin.co.uk/1a-schottky-barrier-rectifiers-46408

?

[Gurner]

There's not a huge selection at Maplin, but the 1N5817 should yield a lowish forward voltage (look at figure 2 here  http://www.fairchildsemi.com/ds/1N/1N5818.pdf  ....the fwd voltage increases with current, but you should still be able to attain somthing las low as 0.3V with it)

[slacker]

This is just off the top of my head so it might not work but couldn't you put the diode in the feedback loop of the opamp, there by eliminating the loss due to the diode drop. Output to diode, diode to feedback resistor, with the output to the next stage from that junction. That would save the hassle of going to Maplin.

[Gurner]

This is just off the top of my head so it might not work but couldn't you put the diode in the feedback loop of the opamp, there by eliminating the loss due to the diode drop. Output to diode, diode to feedback resistor, with the output to the next stage from that junction. That would save the hassle of going to Maplin.

Yes, that's often done to eliminate the fwd drop of the diode ....but that said, I've only ever seen it implemented in a unity gain opamp configuration  ...I ain't too sure of the impact of putting it in an opamp setup that has 30x gain - it might work...dunno! (also, all the associated circuits show bipolar supplies...so it'd likely need a bit of thought to implement in the single supply setup in use here)

[deadastronaut]

just tried it...no difference,  if i did it right that is...


@ian: can you recommend a rail to rail opamp?....while i'm at maplins, i know there stock is limited, but hey in for a penny etc...

[slacker]

Yeah, the fact it has gain is why I couldn't decide if it would work or not.

Not sure Rob, had a quick squiz at their website and couldn't see any.

[deadastronaut]

ok, cheers man...no probs.

i was having a nose around...and found this...hmmmm...could i get this to force an led to respond to attack/release...smoothly..just wondering.

http://members.shaw.ca/roma/attack.html

[deadastronaut]

morning, well i got my 1n5817's....amongst other bits while i was there..

but it still needs a good whack to get it to light....hmmmmm....

[deadastronaut]

hmmm just found this ...

http://www.diystompboxes.com/smfforum/index.php?topic=95116.msg822997#msg822997

[Gurner]

morning, well i got my 1n5817's....amongst other bits while i was there..

but it still needs a good whack to get it to light....hmmmmm....

Well don't fret....always good to have schottkys in your component selection box :-) (& they do help in this particular circuit) Are you sure your LED is *just* biased off (reduce the value of R7 until the LED is very faintly discernibly lit, & then increase R7 a tad until the LED looks off - a preset would be best for this)  ...also did you reduce R5 to say 330R?

I'd say you're now getting to the point were you need a little more 'controlled' conditions (whacking the guitar - whilst immediate & fun - ain't that scientific, lol)  - so maybe consider inputing a sig gen (use your PC sound card and siggen softwtare), slowly increasing the input signal amplitude, seeing what happens with the led & taking DC voltage measurements at the output of the final opamp.

[deadastronaut]

lol...whacking a guitar into it is very scientific for me...

right i put a trim in place of  R7, trimmed, led is only just on now...

led resistor is at 100r..


man i need a sig generator..

[deadastronaut]

ok i just got a tone generator...installed...i'll take an output into the breadboard...what ideal freq shall i use?.

http://www.ringbell.co.uk/software/audio.htm

[Gurner]

ok i just got a tone generator...installed...i'll take an output into the breadboard...what ideal freq shall i use?.

http://www.ringbell.co.uk/software/audio.htm

Anything between 100hz & 800Hz (a guitar goes lower....down to 82Hz & it also goes higher...upto about 1400hz)

[deadastronaut]

cool, i chose 400hz...i have control from my headphone level..

its working ok...


edit:  took the signal out of my mixer main out, at 0.3db..on the vu meter.  

[deadastronaut]

right i gave up on that....cheers though..

i came up with this...it responds great, i tweaked a few values and it lights up to the slightest touch...nice and sensitive...and best of all uses a crappy old 072 yay!!!!whoohooo........(led on output) (both red leds)

it has quite a smooth fade too..

also works with superbrights..

[Gurner]

right i gave up on that....cheers though..

Hey, no worries ! As I said when you started...

(the original cct I furnished was to either switch an led on or off, but what you probably seek is a nice smooth fade from max brightness to extinguished)...now that I can see what your ultimate goal is, it's probably not the way to approach.

That said, much of your latest circuit is the same (you're just using an inverting opamp vs. the non-inverting approach of the circuit your were using until late yesterday)   ......the main difference is that since you are using the LED itself to do the rectification, in this case the LED is being switched on & off (lit) by the AC signal (vs a smoothed DC interpretation of the AC signal via the approach taken yesterday)....to my eyes, that C3 looks like it could be hampering things rather then helping. (because any resulting DC charge on the top plate of C3 will start turning the LED off as the incoming signal level increases) ....also, IMHO you'll not be able to get a nice long 'tail' release with such a circuit (because you are using AC to light the LED - once your signal stops, the LED will instantly extinguish). Also, with a gain of 70, I would imagine there'll be little  'middle ground' region for the LED because instantly your incoming signal will clip (even small signals will clip), driving the led hard ....to phrase another way, it's virtually a black or white scenario with no shades of grey in between.

[deadastronaut]

morning...yeah gotcha,  i'm using it with a bright green at the mo, as cadmium apparently likes that colour best according to what ive read....

yep, the led does just cut on/off though....upon further testing..

d'ya reckon i should just lose the 4.7uf/220 on the end then, and reduce the 3.3m....it would be nice to get that smooth fade back..

that bit of the circuit was lifted from this..i built one a while back but it weren't up scratch imo....

http://home-wrecker.com/nurse-quacky.html


edit: the first clip is with 1M gain resistor, second clip is 4.7M gain resistor...

https://dl.dropbox.com/u/7464107/ef%20shit.mp3