MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Harry Bissell's Fast Response Low Ripple Envelope Follower
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2, 3, 4, 5, 6, 7, 8, 9  Next [all]
Author Harry Bissell's Fast Response Low Ripple Envelope Follower
roglok
Harry Bissell's Envelope Follower has been mentioned quite a few times here and over at electro-music.com, but it seems like no one has ever built it. As posted in the Show Us Your DIY Builds Thread here and here I decided to take the plunge and make a stripboard version of the circuit.



My version of the schematics





Input Signal: TR-606 Snare



Top: Bissell Envelope Follower
Bottom: Doepfer A-119 modded for faster response


Quote:
I am stoked to make a layout for it, but it would be good to understand perhaps some optimisation Roglok did.
thumbs up


I haven't really optimised anything. Quite the opposite - There's a neglected opamp and 5 grounded Schmitt triggers that could've been put to good use... If you have any specific questions, fire away!
It's peanut butter jelly time!
infradead
oh wow.
craque
ooooh very cool
dropmotif
Great! I've been thinking of building this for a while now. After seeing the scope shots this will be next in line.
emdot_ambient
I thought I had the schematic for this...I know I've looked at it several times... hmmm.....
roglok
emdot_ambient wrote:
I thought I had the schematic for this...I know I've looked at it several times... hmmm.....


schematics are linked above...
widdly
Based on the scope shots..Wow!

So you built it exactly as per the magazine article?

for people like me who took a while to finid it ---> http://m.eet.com/media/1141271/18042-122602di.pdf


The left over opamp and schmitt triggers could be used to add a gate output. Maybe even a trigger.
diablojoy
that is impressive
Quote:
The left over opamp and schmitt triggers could be used to add a gate output. Maybe even a trigger.

to have both would be useful.
A couple of the schmidt triggers and a couple of 2n3904's would do for gate and trigger outs and provide for LED's as well
the op amp could be used for something else maybe a preamp on the input with adjustable gain .
guest
that is a very interesting circuit, and a nice build. what is the fourth transistor in your build doing? i only saw 3 in the schematic.

the extra opamp might be best used as a precision fullwave rectifier before the whole thing. the circuit only follows positive going attacks at the moment, which could cause an audible delay on low frequency signals that strike negative first. that would also allow you to clock it a bit faster, and get a faster response time for the same lowest frequency of interest.

are the 2 diode drops at the output a problem? do small signals not get detected?

i wonder what the tradeoffs are for the number of peak followers. it seems it could work equally well with only 2.
amdagan
Nice one! Been ogling this circuit for a while.

Another idea for optimisation (completely untested): drop the 4017 and use 3 of the 40106 gates as a 3-phase oscillator. Potential disadvantage: if you want to change the oscillator frequency, you now need to change 3 resistors instead of just one in the existing circuit. But it seems like that frequency doesn't need to be adjustable anyway.

That still leaves half of the 40106 and the spare opamp for deriving gate / trigger, as others have already noted. Maybe use the opamp to make a Schmitt trigger (so threshold & hysteresis can be tuned using resistor values, not possible with the 40106), use two gates from the 40106 to turn gate into trigger, and maybe 2x3904 to buffer the gate and trigger outs? Something along those lines I guess.
roglok
thanks, all!

widdly wrote:
Based on the scope shots..Wow!
So you built it exactly as per the magazine article?


well, i swapped the dual op amps for quad ones. i've also added a transistor LED driver, which i've removed later on (the two resistors and tranny below the 4017). that's probably where i should've used the opamp, but the LED was a bit of an afterthought and there was some space left in that corner smile...
apart from that, no changes or mods were done.

Quote:
The left over opamp and schmitt triggers could be used to add a gate output. Maybe even a trigger.


yeah, that's what i thought, too. but all the standalone comparator circuits i've come across use an inverting configuration and thus require two op amps (extra stability?). i think the threshold would definitely have to be adjustable to be useful. i'm not good at whipping up circuits myself (yet?), so any pointers on how to do this properly with a single opamp are appreciated...

oh, and here's the full schematic, including TL074 pinout - i found it easier to build after sketching out the two hidden processors...

fonik
roglok wrote:
yeah, that's what i thought, too. but all the standalone comparator circuits i've come across use an inverting configuration and thus require two op amps (extra stability?). i think the threshold would definitely have to be adjustable to be useful. i'm not good at whipping up circuits myself (yet?), so any pointers on how to do this properly with a single opamp are appreciated...

to make the comparator working the other way around just swap the inputs (theshold and signal input).

the threshold is set by a voltage divider, so just replace it with a potentiometer (maybe with limited range, achieved by wiring the potentiometer between the rails via additional resistors).
and
you could even put a normalled CV input jack in front of the potentiometer.
amdagan
Quote:
any pointers on how to do this properly with a single opamp are appreciated...


A simple way: http://en.wikipedia.org/wiki/Schmitt_trigger#Non-inverting_Schmitt_tri gger
In the circuit shown there, make R2 adjustable to vary the hysteresis width; and apply a positive offset to the opamp's inverting input to shift the whole hysteresis curve to the right. That offset should probably be adjustable, maybe take it from a pot wired between +V supply and GND.
Navs
Roglok, your strip-board builds are absolutely gorgeous.

Regarding add-ons, I always found the gate out on the A-119 useful. An inverted gate and trigger would also be good. The latter could be abused for clock doubling.

On the response, ironically I would want to smooth and slow the output. I find CVs derived from pressure or mics need slewing. They're too abrupt for my taste or use. seriously, i just don't get it
roglok
guest wrote:

the extra opamp might be best used as a precision fullwave rectifier before the whole thing. the circuit only follows positive going attacks at the moment, which could cause an audible delay on low frequency signals that strike negative first. that would also allow you to clock it a bit faster, and get a faster response time for the same lowest frequency of interest.


interesting. the funny thing is that bissell himself mentions full wave rectification and filtering as the "usual" way of doing envelope followers. i wonder if he omitted the rectifcation for a reason or just to keep it simple...

Quote:

are the 2 diode drops at the output a problem? do small signals not get detected?


hmm good point. haven't played enough with it to find it a problem, but i guess you could just add a bit of gain to the final buffer to compensate?

Quote:
i wonder what the tradeoffs are for the number of peak followers. it seems it could work equally well with only 2.


another good question. i can easily try that, though. would simplify the whole thing quite a bit...
Moog$FooL$
wow!!

thanx for this...... really helpful you guys.

thumbs up thumbs up
roglok
@Matthias
@amdagan

thanks for the tips!

@Navs

i see your point. but it's easier to smooth out a peak than the other way around 8_)
Jarno
Nice, seems to be quite a bit of interest in this. I'll see if I can plonk this into Eagle tonight. So, did you do any component substitutions?
What did you do with the rest of the 40106, tie to ground?
roglok
Jarno wrote:
Nice, seems to be quite a bit of interest in this. I'll see if I can plonk this into Eagle tonight. So, did you do any component substitutions?
What did you do with the rest of the 40106, tie to ground?


no substitutions except for changing the TL072 to TL074. the unused CMOS inputs i've tied to ground.
megaohm
Jarno wrote:
Nice, seems to be quite a bit of interest in this. I'll see if I can plonk this into Eagle tonight. So, did you do any component substitutions?
What did you do with the rest of the 40106, tie to ground?


Wouldn't it be a good idea to experiment with the circuit more?
To understand it and embellish?
Harry put this out to share and it seems the least we could do is put some work in and build on his foundation.

Some ideas:

have the clock be variable.
have an external input for the clock.
add more stages (the 4017 does up to 10 steps)
glide on the output, switchable to select up/down/both

When I built this years ago I forgot to connect the pulse gen to the peak circuits.
hihi
It resulted in a more typical EF response, which could be useful and implemented with a simple switch.

Just a few ideas, but we'll never know if these are useful if we just take a person's design and slap it onto a pcb.
roglok
fonik wrote:
roglok wrote:
yeah, that's what i thought, too. but all the standalone comparator circuits i've come across use an inverting configuration and thus require two op amps (extra stability?). i think the threshold would definitely have to be adjustable to be useful. i'm not good at whipping up circuits myself (yet?), so any pointers on how to do this properly with a single opamp are appreciated...

to make the comparator working the other way around just swap the inputs (theshold and signal input).

the threshold is set by a voltage divider, so just replace it with a potentiometer (maybe with limited range, achieved by wiring the potentiometer between the rails via additional resistors).
and you could even put a normalled CV input jack in front of the potentiometer.


does this make any sense?

amdagan
Quote:
does this make any sense?

Does the opamp run from a bipolar supply? If so, you'll need to protect the 40106 input from negative voltages. A series diode followed a resistor to ground should do it.
roglok
amdagan wrote:
Quote:
does this make any sense?

Does the opamp run from a bipolar supply? If so, you'll need to protect the 40106 input from negative voltages. A series diode followed a resistor to ground should do it.


yep, it runs from a bipolar supply. but the threshold is wired between ground and +V only and the input already has a diode?
hmmm.....
Jarno
A fair point, but I'm actually also interested in the circuit as is. Could make some provisions for experimentation.

Oops, forgot to quote, will add later.
roglok
@amdagan, of course you are right about the diode. not sure what I was thinking... meh

having played with it a bit more tonight, i have to agree with Navs - what looks good on paper/scope isn't necessarily the most musical. for my taste the follower is a bit too precise. while I love the sharp attack, it's definitely not ripple-free. this becomes especially obvious, when using a comparator. depending on the source material (eg. a 606 bassdrum) i'm getting a shaky envelope while the comparator spits out multiple triggers.

i'll probably try to lower the clock rate and/or adjust the filter to see if i can find a sweet spot...
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2, 3, 4, 5, 6, 7, 8, 9  Next [all]
Page 1 of 9
Powered by phpBB © phpBB Group