Making a generic envelope follower

[soggybag]

I have been trying this design and can't get it to work. There is a < at the end of R1 I assumed this should go to ground? But I'm not sure. Any suggestions? I used an LM324 as the chip and 1K for both R3 and 4.

[Paul Perry (Frostwave)]

Soggybag, where did that circuit come from? I don't like the look of it!

[soggybag]

This came from Stompboxology 13, 4 Envelope effects. The Author Nicolas Boscorelli has a fair amount of recognition, I know he's written at least one book on the subject of stompbox circuitry and apparently 13 issues of the Stompboxology publication. I think Stompbology came out quarterly, so thats potentially 52 issuses. The issues are pretty packed with stuff and it all looks pretty well thought out.

This is from the second page of Vol 13, issue 4. I posted a link to the whole page in an earlier post to this thread. The caption for this example says:

Integration smooths the control envelope; choice of R3, R4 & C2 determine relative attack and decay times.

I tried to get this to work with no luck. The < off of R1 was a little perplexing. I couldn't decide if this should go to ground or to V+ or Vb or ? I used an LM 324 and 1n34 diodes. I used 1k for both R3 and 4.

[Vsat]

Try putting a 100K or 220K resistor across D1, the op amp (+) input needs a path to ground for the input bias current - as shown, this depends on the diodes' reverse leakage current (rather tenuous) - also the resistor provides a discharge path for the capacitor and  sets the cutoff freq for the RC combination. The 10K going to the op amp (-) input probably goes to ground if this is a single-supply env follower, rather than V/2.
Cheers, Mike

[puretube]

another (old) approach, here: ripple free by filament-lag...

[Paul Perry (Frostwave)]

Thanks Vsat & soggybag, I'm happier now!

[Vsat]

LM358 LM324 should be fine in the circuit (with added resistor across the diode). If you assume a typical input bias current of 100 nA for the  op amp, the 0.1 uF cap will pull up towards  (+) 9V at a rate of one volt/sec, when the other side of the cap is connected to the output of another op amp, or buffer, etc, in the absence of any input signal.... making the env follower put out full voltage. The resistor should cure this. TLC272 would be another good one to use. None of these single-supply op amps like to see input signals go more than a few hundred mV below B(-) or ground, particularly if the signal comes from a low-impedance source. Schottky diodes would work well here throughout the circuit.
Cheers, Mike

[Vsat]

Also good idea to make R3 bigger than R4, so the  capacitor C2 is able to charge up properly, without completely discharging between the (+) half-cycles of the input signal.
Mike

[soggybag]

Thanks for the help vsat. Can I run a guitar signal directly into this or should I put a buffer in front of it?

[Vsat]

The circuit as shown has a gain of two for the positive half-cycle (the two 10K resistors on the first op amp define this)... minus one diode forward drop. To minimize the envelope "error" from the diode drop make sure the first op amp is capable of putting out lots of voltage, and that either Schottky or germanium diodes are used.  It would like to see a large input signal (say 8V peak-peak max or 4V peak max ) with a 9V supply, alternatively you can boost the gain of the first op amp by reducing R1 to around 470 ohms to 1000 ohms, or by increasing R2. There has to be enough signal coming out of the first op amp else D2 won't even conduct. With a big signal coming out of A2 now, put a potentiometer to ground on the output of A2 and you will have a variable envelope amount control for setting the level to a VCF or flanger  etc. And don't forget to put a 100K resistor across D1.

Can't really hook it directly to a guitar because D1 will short out the pickup on the negative half-cycles. No reason not to try it though. A buffer in front would be better. It would also be a good idea to put a resistor in series with C1 (say around 10K) to limit the current flow through D1 when you use another op amp or other low-impedance source to drive this follower.
Mike

[soggybag]

Thanks Mike that was very helpful and informative. I will try all of these ideas out this afternoon and report back with the results...

[Joe Kramer]

Stompboxology was a really nifty publication with a lot of good stompbox ideas. Too bad it isn't around any more. Does anyway know how many issues are out there? I have a few, but I'm guess there were a lot more.

Soggy,

Thanks for posting that STMPBXLGY page.  I agree it was a great publication, sorry to see it go.  I have a couple of issues and would like to trade with you or anybody else that has them.  I wonder what the ethics are concerning that?  Seems like it would be safe as long as it didn't involve posting the pages or selling them.  Anyway, maybe "Trading Issues of Stompboxoglogy" should be a separate thread.  What do you think?

Joe

[soggybag]

Here's an envelope follower from a book called "Electronic Projects for Guitar" by RA Penfold. I extracted this from a project title Dynamic Tremolo. This one seems to work pretty well with a minimum of parts. I tested with an LED hooked to the CV out only so far.
The range on R7 could use some improvement. Seems to be useful only in a certain part of the range, and off for the first 25% of the travel.

[soggybag]

I'm happy with this last design, but! I keep thinking that the op-amp could be replaced by a transistor. Does anyone have any suggestions on this? It looks like I would heading in the direction of Snippet posted by Moosapoamus earlier in the thread...

[Vsat]

There is likely a very bad problem with that Penfold circuit, driving the LED directly from the wiper of the 100K pot not a good idea. The wiper should be buffered with an op amp or transistor before doing anything with the LED. The LED could be operated with current drive by placing it in the feedback path of the op amp, or could be simply connected to ground and driven by the op amp output through a resistor. And if the LED is used to illuminate an LDR, you could use the LED for the rectifier diode and simplify the circuit.

BTW tried the Boscorelli circuit on my breadboard, does indeed work when resistor is added across input clamp diode. I used 1N6263 Schottky diodes, it is possible that 1N34A would be leaky enough that resistor would not be required - but Schottky's are cheaper and readily available nowadays. Keep in mind that both of these circuits use half-wave rectifiers, full-wave would give less ripple.
Mike

[soggybag]

Thanks for the comments Mike. I have to take credit for the arrangement of the pot and LED/LDR arrangement since I just tacked that on to a snippet lifted from the work of Mr Penfold. The R7 pot did not seem to work that well for adjusting the CV range. The circuit seemed to work better with out the pot.

Since a single op-amp is the same size as a double maybe I'll add an op-amp buffer on the end.

I'll give another try with the Boscorelli thing. I breadboarded it twice with no luck so far. Maybe third times the charm? I had connected the output to an LED to ground as a visual indicator to see if it was working. But couldn't get it to light up.

I have been using the (Penfold) envelope to control the speed of a tremolo. It seems to work well. I can hear the change in speed. I feel it would sound better if the decay time were shorter. The Boscorelli envelope might be a better choice since the attack and decay times can be set with R3 and R4.

What did you use for R3 and R4? I used 1K for both and then tried 10K for both.

Can you point me to an example of a full wave rectifier in a 9v stompbox? I've seen this in power supplies but I don't think I remember seeing one made with op-amps.

[Paul Perry (Frostwave)]

http://sound.westhost.com/appnotes/an001.htm
may make things clearer.
The keyword is "precision full-wave rectfier" for anyone googling.
Personally, i havn't found full wave (as opposed to half wave) making
all that much difference, in a lot of applications.
There is a single transistor full wave restifier circuit too, but it's VERY tricky to
tweak into operation!

[Vsat]

Soggy,
I used 5k6 for both R3 and R4 during the testing, and looked at the outputs with an oscilloscope. R1 was reduced to 1K. Remember that an LED needs at least 1.6V to turn on, so you need at least 1.6V at the output.

Paul - there is also that "non-precision full-wave rectifier" circuit, using two op amps and two diodes, one op amp inverts the output of the first op amp, anode of each diode goes to respective op amp output, cathodes go to series resistor which feeds capacitor. As in ARP Quadra compander. Making a single-supply version is a bit trickier.
Mike

[StephenGiles]

2 things - firstly the name Penfold always tended to appear in the same sentence as the word "simple" which I think speaks volumes, and secondly - in many cases a dual power supply (2 batteries providing + and -9v) would save a great deal of the design problems involved with a single supply ............. probably to enable the use of a tiny enclosure...........am I right or am I right???

I am having just this problem in playing about with the EH Space Drum and Crashpad envelope generators, (not the small enclosure as I don't care what size I use) both of which use an LM 324, so I anticipate fewer problems using a 4558 or similar powered by +&-9v.
Stephen

[puretube]

after I just had a phenomenal idea for the perfect env-fol, (whose patentworthiness has yet to be checked...),
maybe it would be interesting to see what Mr. Bissell thinks of :this (older)  quadrature approach.

Btw: you`ve been invited to the party already a long time ago...