An new twist on a LDR based envelope filter

[rring]

I wanted to duplicate the mutron sound with a simpler circuit, more control of the bandpass peaking and single supply. I came up with this and I am really pleased with the sound.


The only drawbacks are: It needs three LDR's (I just made mine in about ten minutes with some heat shrink and super-glue) and I couldn't get a TL072 to work well. I had to use a higher bandwidth op amp. I offer a few suggestions on the schematic.

Here is a board for a 1590b enclosure...not much to it. It took about thirty minutes to build.


The design has adjustment for resonance peaking, sensititvity and attack. The attack control is extremely useful.

here are some sound samples:
http://circuitsaladdotcom.files.wordpress.com/2014/05/very-slow-attack_high-resonance_demo.mp3
http://circuitsaladdotcom.files.wordpress.com/2014/05/fast-attack-_high-resonance_demo.mp3
http://circuitsaladdotcom.files.wordpress.com/2014/05/low-resonance_-attack_slower_demo.mp3

[Jdansti]

Sounds absolutely great! 

[deadastronaut]

Oh yes, love it, nice work man, the filter sounds excellent,  8 )

[rring]

I have avoided LDR's in the past just because they are hard to find but if you suck it up and just make yur own - its worth it for envelope effects like this. The smoothing of the slow response time and wide range of R is hard to beat.

[samhay]

Sounds good.
Did you consider contolling LDR1 and 2 with a single LED?

[rring]

yes I did that on the breadboard and of course it works well - it saves a little current, but the circuit only draws 5mA when operating and 1mA of that is the bypass/effect LED pumping. I just figured it was easier to make or get commercial single ones. You would probably want to adjust R5 up some if you did that.

[GibsonGM]

Very nice, RRing!!  I'm impressed. You managed to give it 'what it needs', but kept the parts count down quite a bit.

[Kipper4]

Sweet. Good work.
I like the low counts part too.
Sounds great. Time for me to do some data sheet comparaisons on those oppys against the Tl072.
What is the purpose of D2 please.
Is R1 a boot strapping stabaliser?
Thanks

[GibsonGM]

PS - hey, was there any reason for choosing an LM358 for IC 1?    Looks like just about anything would be OK there, huh?

[rring]

How D2 functions is hard to visualize but it allows the negative cycle of the audio swing to charge the cap thereby making a voltage doubler output of the diode integrator. R1 is just bias because there is no feedback....the collector is just tied to V+ - just think of R1 tied to V+ and not the collector. Bipolar Transistors bias up very consistently, unlike JFETS. I tried 3 or 4 and they all biased up to 4.2 to 4.5 volts at the emitter node.

Another good possibility for an op amp choice is: the TL972 it is available in DIP or SOIC from digikey for less than a buck
there is more info on this circuit at my blog: www.circuitsalad.com

[samhay]

^PS - hey, was there any reason for choosing an LM358 for IC 1?
Looks like it needs to be single supply as the input of 1B is tied to ground.

Ray - I like the way Q1 is set up to supply a DC bias, but was wondering how variable it would be. How/why did you pick the value of R2?

[rring]

Yes you are correct - good point.. I need to address that. Numerous op amps will work. The only problem is the actual led driver portion needs to work near 0 volts, so again a TL072 is not a good choice. To your point, any of the filter op amps I suggest will work for the envelope detection portion. I just used the LM358 because I had a bunch. On the breadboard I used two OPA1652's, then tried two TS922..etc, same goes for the TL972 I just suggested - it will work for both amps.

[rring]

I think I just answered that above(mostly). It won't be that variable. Most bipolars have high DC gain and the base/emitter diode behavior is very uniform - so the feedback behavior dominates operating point. The basic design procedure was: decide my operating point(1/2 supply), decide the highest value of emitter resistor I could use(achieve close to unity gain but still have fairly low drive impedance) and then select the bias resistor to turn on the transistor such that the operating point is obtained. I did this in SPICE and then verified with 3 or 4 junk box 2N2222A's and 2N3904's.

[Seljer]

Alrighty, I threw together a through-hole layout for this thing if anyone wants to give it a shot

I haven't built it myself (I only have 1 LDR in my parts drawer ) but at a glance it look OK, maybe someone else can spot something before starting with the acid. I added space for an optional 1Meg anti-switch-pop resistor at the input. I kind of eyeballed the size for the LDR+LED combos but I think it should work....

<look on the next page because this layout had errors!>


edit: wooooo, post #2000, time for celebration  

[rring]

I think you have the - and + inputs of the LM358 reversed on the layout because pin 5(+input) should not be connected to current feedback resistor(3.3k)?

[Seljer]

I think you have the - and + inputs of the LM358 reversed on the layout because pin 5(+input) should not be connected to current feedback resistor(3.3k)?

Hah, good spot! your schematic appears to to have the same issue and I blindly copied it over.

Probably won't do much without the negative feedback  

edit: so on your original schematic both + and - and IC1 are flipped around, I assume IC2A is good which leaves IC2B. Is everything in order there? I'm searching through the Lancaster filter book and google to find such a topology to verify it.

[rring]

ouch- thank need to fix that!

[samhay]

^...I did this in SPICE and then verified with 3 or 4 junk box 2N2222A's and 2N3904's.
Thanks Ray. The resistance is a little higher than I was expecting, but that makes sense.

[Mark Hammer]

Sounds really nice.

I was pondering whipping up a simple 2-pole lowpass with a pair of LDRs.  I think it could be done with a quad op-amp: one op-amp for the input stage, one for the filter, one for the envelope extractor, and one to spare for whatever.

[rring]

Yes that extra op amp could be used for up/down reversal of the envelope following maybe.