After hanging out here for a couple of years and building a few pedals I decided it was time to give something back to the community. So here's my latest design, the Slackfilter.
It's an envelope filter/phaser/self oscillating noise maker based on Ken Stone's amazing Synthacon VCF clone. Check out http://www.cgs.synth.net/ (http://www.cgs.synth.net/) for more details. Basically it's a VCF with seperate inputs for bandpass, highpass or lowpass rather than most similar filters that have one input and seperate outputs. The cool thing is you can use any combination of the inputs at once. It will also self oscillate if the resonance control is cranked up!
I adapted his design so it runs on 9 volts and added an envelope follower (from Tim Escobedo's phuncgnosis) (http://www.geocities.com/tpe123/folkurban/fuzz/snippets) and an LFO based on another of Ken's designs, the Phycho LFO.
It does autowah type effects using the envelope follower but the LFO adds a whole load of extra effects. The LFO has a triangle output which does synthy type sweeps and phaser type effects. It also has a pseudo random output that creates sample and hold type effects.
Here's the latest schematic (http://www.eskimo.plus.com/fxstuff/newslackfilter.png)
and original version (http://www.eskimo.plus.com/fxstuff/slackfilter.png)
and some soundclips, the clips are guitar >> slackfilter >> condor >> soundcard. The dirty sounds are done using a BigMuff after the slackfilter.
Auto wah (bandpass) (http://www.eskimo.plus.com/fxstuff/envelopebp.mp3)
Auto wah (lowpass) (http://www.eskimo.plus.com/fxstuff/envelopelp.mp3)
Auto wah (highpass) (http://www.eskimo.plus.com/fxstuff/envelopehp.mp3)
LFO driven effects (http://www.eskimo.plus.com/fxstuff/lfo.mp3)
Random LFO effects (http://www.eskimo.plus.com/fxstuff/random.mp3)
self oscillating weirdness (http://www.eskimo.plus.com/fxstuff/selfosc.mp3)
It's still a work in progress and some things need work, the envelope follower has a fairly limited sweep and the LFO can be a bit noisy.
If anyones got any comments or suggestions on how to improve it I'd be happy to hear them.
If anyone fancies building one then I've got some vero layouts that I'll be posting soon along with some build photos.
Hey, some great "noises" there Slacker ... many thanks and congrats !
MM.
Very very nice, and another reason for me to start looking into the Steiner filter in general. I took the liberty of doctoring up the posted schem so that it makes for an easier view in black and white: http://img.photobucket.com/albums/v474/mhammer/SLACKFILTERBW.png
It's your baby so I won't leave this up for long, but I figured that it would be easier to look over and discuss in B&W for some folks. If you're uncomfortable with that, just let me know and it will be deleted promptly.
cheers for the comments :)
Good idea Mark I've redone the schematic in black and white, does make it a bit easier on the eye.
... good stuff !!
:icon_mrgreen:
Thats so cool ::) I love it.
btw, the vcf is not original work of Ken Stone, its originally from this synth (http://www.synthmuseum.com/stp/stpsynthacon01.html)
ch.
Thanks, yeah you're right it's based on the VCF from that synth, orignally designed by Nyle Steiner (http://www.patchmanmusic.com/NyleSteinerHomepage.html).
I should have said based on Ken Stone's amazing Synthacon VCF clone.
sounds great :D
i love the random LFO, i was thinking of a way to make the LFO of a tremulus lune random.
check out this this page from GEO (http://www.geofex.com/Article_Folders/LFOs/psuedorandom.htm) it's got some great stuff about random LFOs. I started from there and then used the Psycho LFO (http://www.cgs.synth.net/modules/pic/schem_psycho_lfo.gif) to get an idea of component values, even though I ended up with different ones because I thought it was too fast.
If you just took the top 3 schmitt triggers from my LFO and hooked them up via a switch to the depth pot of the tremulus lune that would probably work.
I've just updated the schematic, the random switch on the LFO wasn't how I'd done it on the final build. I've also added the power supply connections to the CD40106 used by the LFO.
I've used a DPDT for the LFO mode switch, which gives triangle or random output, but you could use 2 switches and have a square wave output as well.
Oh yeah, heres one more modified "Steiner VCF" (http://www5b.biglobe.ne.jp/~houshu/synth/VcfStn0505b.gif).
The output stage is modified to make up the gain change with resonance. You might wanna try that :)
This inspired me to figure out that you can use digital noise generator as a sample-and-hold generator by itself !
ch.
I cant wait until there will be a pcb layout for this thing! :icon_biggrin:
Quote from: christian on May 14, 2006, 06:04:36 AM
Oh yeah, heres one more modified "Steiner VCF" (http://www5b.biglobe.ne.jp/~houshu/synth/VcfStn0505b.gif).
The output stage is modified to make up the gain change with resonance. You might wanna try that :)
Cheers I'll give that a try, those diodes look interesting as well :)
I've changed the design slightly to address a couple of problems with original.
Here's the revised schematic (http://www.eskimo.plus.com/fxstuff/newslackfilter.png), the changes are shown in red.
On the original the frequency and depth pots were highly interactive. This was due the fact that adjusting the depth pots changed the loading on the base of the bias control transistor. I added diodes between the depth pots and the base of the transistor which has solved this problem. Now adjusting the depth pots doesn't affect the frequency, it also means that the LFO and envelope follower can be mixed with some nice results (sound clips to follow). Removing the 1Meg resistor from the envelope depth pot has also improved the envelope follower sweep.
I also added a capacitor to ground to the inputs of the LFO output buffers which has reduced LFO bleedthrough.
I got the idea for the diodes from this site (http://yusynth.net/Modular/index.html) which has a some alternative Synthacon filter designs. Including one using an opamp instead of transistors and an interesting looking guitar synth. It's also got a copy of Nyle Steiner's original article and design and loads of other good synth stuff.
If anyones interested here's a couple of photos
This is just the filter section of the pedal, the resistors connected with sockets are to adjust the sweep of the frequency pot.
(http://www.eskimo.plus.com/fxstuff/slackfilterphoto.jpg)
Here's a guts shot, showing my messy wiring, like a said it's still a work in progress :)
(http://www.eskimo.plus.com/fxstuff/guts.jpg)
Brilliant!! I'm obviously gonna have to take a couple of weeks this summer and do nothing but filter projects, this one and Stephen Giles' multi- spacefilter idea in particular! Nice implementation on this one, slacker, especially getting that kind of feature set without letting the design become totally out of control. OK, who's workin' on the vero layout? :icon_wink:
Thanks, I was pleasantly surprised with how easily it all came together, that's probably more a testament to the original filter design rather than my skills. There's probably still a few things that could do with tweaking but I'm going to live with it for a while to get an idea of what it can do and what might need improving.
I've got vero layouts for how I built it, I'm tidying them up and adding the tweaks I've made then I'll post them. As you can see in the photo it's 3 separate boards, one for the filter, one for the IO and envelope follower and one for the LFO. I did it like this to make it easier to experiment with different ideas and because I already had the LFO board from another project that never quite worked out.
I'm hoping someone will take up the challenge of doing a PCB layout, because I don't have any experience in that area.
This looks wicked cool, slacker! 8)
I'm thinking that it might also be cool to mod the front-end so that you could actually mix three different input signals, one each to the LP, BP and HP inputs (ie: bass, drums, synth). I guess you'd repeat your input buffer two more times? Then you might also want three more switches to be able to select which input(s) will drive the envelope response.
I've only got about a dozen different projects in various states of completion. But, I'd love to tinker with this one. 8)
~ Charlie
Very Cool! 8) 8) 8)
Nice work!
Quote from: moosapotamus on May 22, 2006, 12:18:31 PM
I'm thinking that it might also be cool to mod the front-end so that you could actually mix three different input signals, one each to the LP, BP and HP inputs (ie: bass, drums, synth).
Thats a cool idea, I'd thought of having an FX loop but separate inputs would give even more possibilities. I quite like the idea of say running the wet signal from a delay into one input with dry going into another.
It would be great to see a moosapotamused version, I dread to think how many knobs and switches it would have :D
thanx for the link to the original Steiner notes Slacker !!
http://yusynth.net/archives/ElectronicDesign/N-Steiner-VCF-1974.pdf (http://yusynth.net/archives/ElectronicDesign/N-Steiner-VCF-1974.pdf)
I notice in the conversion to Voltage-Control the op-amp in the first circuit is replaced by a two stage bipolar amplifier ... you can probably squeeze more Q out of your topology by changing back the two-device gain stage to a variable (higher gain) op-amp stage ...
or, for now ... in your improved schematic there's a non-inverting op-amp with a x5 gain at the output, to test this theory out you could take the resonnance feed after the op-amp instead of after that transistor stage and see how much Q increase you get ... you can rig a second pot at the output of the op-amp for the variability ... if that works then try replacing the 220k by 2meg pot in the feedback of that op-amp - at some point your filter should morph into an oscillator and give you a nice sine wave ... from your nice clips it sounds like you might still have some gain to go before you reach that point ...
also, I looked at the 47k input impedance thing - it's probably not crucial ... you could always buffer the input and stick a 100k pot in series to see what changing that value does to the response but it's likely minor ...
all in all, the VC conversion is a very neat trick - it's somewhat analogous to the way varistors work in parallel DC circuits in the Magnetone 480 pitch shifters ...
best regards ...
~JC
Quote from: Eb7+9 on May 23, 2006, 02:21:31 AM
I notice in the conversion to Voltage-Control the op-amp in the first circuit is replaced by a two stage bipolar amplifier ... you can probably squeeze more Q out of your topology by changing back the two-device gain stage to a variable (higher gain) op-amp stage ...
Yeah I've thought of trying an opamp version just to see if the sound or performance is any different.
Quote
or, for now ... in your improved schematic there's a non-inverting op-amp with a x5 gain at the output, to test this theory out you could take the resonnance feed after the op-amp instead of after that transistor stage and see how much Q increase you get ... you can rig a second pot at the output of the op-amp for the variability ... if that works then try replacing the 220k by 2meg pot in the feedback of that op-amp - at some point your filter should morph into an oscillator and give you a nice sine wave ... from your nice clips it sounds like you might still have some gain to go before you reach that point ...
The filter will already self oscillate in lowpass or highpass mode with the resonance control above about 80%. Bandpass won't quite self oscillate. I tried your idea and it does increase the Q, needs some playing about with to find the right pot value because it gets completely out of control very quickly. I've also tried increasing the value of the resonance control like in the design christian posted and that works nicely.
those 2 BC109's are acting as variable resistance to ground right? why are there 2? can you just use one of both? or use difefrent modulations on both?
Quote from: birt on June 03, 2006, 09:23:36 AM
those 2 BC109's are acting as variable resistance to ground right? why are there 2? can you just use one of both? or use difefrent modulations on both?
They basically give you mirrored image of the voltage, so that once they go thru the diodes (which also respond mirrored to the mirrored voltages), they cancel each other out. This is mainly a concern if youd had high-frequency oscillator modulating the VCF or some "thumping".
This is a standard thing to do in a "synth-world" to keep the control voltages out of the audio path.
For guitar fx with controlled circumstances like having just envelope follower as a CV for the filter you wouldnt need this, but to make this work with a single transistor, youd probably had to change a lot more stuff here.
ch.
good to know. i was maybe going to put this one in my big filterbox as the 4th filter but i'll stick to three i think
Quote from: birt on June 03, 2006, 11:10:33 AM
i was maybe going to put this one in my big filterbox as the 4th filter but i'll stick to three i think
:D:D:D
A little question: can someone tell me how work the two buffers in the lfo section? are both necessary?, im confused about it cause pysho lfo looks different.
The way my LFO works is a bit different to the psycho LFO. To get the triangle waveform I took the signal from the input of the inverter, so the voltage goes smoothly up and down as the capacitor to ground charges and discharges. This signal needs to be buffered before it can drive anything, otherwise the signal gets loaded down and the LFO stops working.
You only really need the buffer that's attached to the depth pot. The other buffer just drives an LED that flashes in time with LFO.
If you just wanted the square wave or random mode you could probably leave the buffers off completely as the 40106 can easily drive the transistor or even LEDs.
Thanks for replie :icon_mrgreen:
What are you using for a power supply?
It runs off 9volt DC. I use a wallwart but a battery should work, don't know how long it would last though. I'll have to measure the current draw.
Cool design and cool sounds, Slacker.
Did you say vero? For this thing? That'd be awesome, but I imagine it'd be huge.
yeah I built it on vero, I did it on 3 separate boards so they're not too big. It would be fairly big on one board but seen as how you need a box big enough for 6 pots, 4 switches and a stomp switch I don't think size matters too much.
I really must get the vero layouts finished and posted here, I'll try and make time next week.
Wow. That'd be great. Thanks in advance!
I am on the edge of my seat.
Quotethose 2 BC109's are acting as variable resistance to ground right? why are there 2? can you just use one of both? or use difefrent modulations on both?
Not exactly. The BC109's are acting as a differential current source. What makes this thing tick is that if you replace the diodes with resistors and ignore the BC109 modulation scheme, it's a fairly common active filter with three modes. The diodes act as variable resistors. The input to the gains stage comes from the center of the diodes, the input and output are mixed at the junction of the two 2.2nF caps. The BC109s only supply a variable current to change the diode resistances.
QuoteThey basically give you mirrored image of the voltage, so that once they go thru the diodes (which also respond mirrored to the mirrored voltages), they cancel each other out. This is mainly a concern if youd had high-frequency oscillator modulating the VCF or some "thumping". This is a standard thing to do in a "synth-world" to keep the control voltages out of the audio path.
Correct. The diodes are set up so that one side (the right hand BC109) is fixed, and as the input control voltage increases, it causes more current to shift to the left hand BC109. This forces more current through the diodes. With no control voltage, most of the current flows in the right hand BC109 and bypasses the diodes.
The voltages on the diodes are symmetrical about the sensitive input to the gain stage, and cancel out as noted. Same idea as was used in the Magnatone vibrato amps.
Did you ever finish the veros? ::) or somehow continue developing the circuit? Might be very interesting ...
bump:)
Sorry I never finished the vero layouts, I don't think I've even got them any more. They were a complete mess anyway, because I built different sections of it at different times.
awww :(
i was really looking forward to this one, i dont imagine im the only one either
Every filter design can have separate low band and highpass inputs, just like it can have separate low and highpass outputs.
Its just adding and subtracting the inputs and outputs.
(http://www.xs4all.nl/~rhordijk/G2Pages/Resources/Art3_4.jpg)
Cool thing is to use it as a spectral crossfader. And you can use different filter outputs to change the filter curves.
(http://www.xs4all.nl/~rhordijk/G2Pages/Resources/Art3_2.jpg)
Lots of experimental fun to have with those filters. And its a shame not more synths have this.
I've just built an Uglyface with LFO and incorporates switched diodes to give ramp up and ramp down as well as regular triangle modulation. This has proved to be very useful and I wonder if you could incorporate this feature into the design.
Quote from: dellamorte on April 20, 2008, 10:41:11 AM
awww :(
i was really looking forward to this one, i dont imagine im the only one either
Sorry, if it's any consolation I'm working on something similar and hopefully better at the minute. It's based on the Oscar synth filters (http://www.diystompboxes.com/smfforum/index.php?topic=65918.0) that Paul Perry posted recently. It's probably going to be a month or two before it's ready but I'll definitely post a layout this time.
Quote from: Mick Bailey on April 21, 2008, 08:13:55 AM
I've just built an Uglyface with LFO and incorporates switched diodes to give ramp up and ramp down as well as regular triangle modulation. This has proved to be very useful and I wonder if you could incorporate this feature into the design.
I'm not sure if the existing LFO could be modded to incorporate that, but you can splice basically any LFO you want into the design.
Quote from: slacker on April 21, 2008, 01:20:56 PM
Sorry, if it's any consolation I'm working on something similar and hopefully better at the minute. It's based on the Oscar synth filters (http://www.diystompboxes.com/smfforum/index.php?topic=65918.0) that Paul Perry posted recently. It's probably going to be a month or two before it's ready but I'll definitely post a layout this time.
I'm really looking forward to this, your Echo Base was a great design, one of my favourite pedals!
/LARS
I'm ordering the parts for this today. I'm excited. I'm planning on building it with a send/return so that clean guitar can control the envelope generator and fuzzed out guitar gets sent through the filter. I'll be making my own layouts for this and if I can get it to work, I'll post em for the rest of you.
I had one question, though. After each of the switches to select what type of filtering is a 47k resistor. The other sides of all of these switches meet at a common point. Would it be a bad idea to put a single 47K on the common side of these switches instead?
Quote from: Boogdish on September 21, 2008, 05:12:33 PM
I'm ordering the parts for this today. I'm excited. I'm planning on building it with a send/return so that clean guitar can control the envelope generator and fuzzed out guitar gets sent through the filter. I'll be making my own layouts for this and if I can get it to work, I'll post em for the rest of you.
I had one question, though. After each of the switches to select what type of filtering is a 47k resistor. The other sides of all of these switches meet at a common point. Would it be a bad idea to put a single 47K on the common side of these switches instead?
I am TOTALLY looking forward to your results!
Quote from: Boogdish on September 21, 2008, 05:12:33 PM
I'm ordering the parts for this today. I'm excited. I'm planning on building it with a send/return so that clean guitar can control the envelope generator and fuzzed out guitar gets sent through the filter.
That's a really good idea, nice to see someone still playing about with this :)
Quote
I had one question, though. After each of the switches to select what type of filtering is a 47k resistor. The other sides of all of these switches meet at a common point. Would it be a bad idea to put a single 47K on the common side of these switches instead?
I think you need the 3 resistors, because they split the signal to the 3 inputs and stop them interacting with each other. I think if you just used one it would work fine if you just selected one input, but selecting more than one wouldn't work properly.
Hi Ian,
sorry to resurrect such an ancient thread. However, I recently tried something very similar but was not as happy with the results as you seem to have been a decade ago. Unfortunately all the links seem to be broken. Do you mind uploading the schematics again? I'd really like to compare them with my fruitless experiments and see if it is worth giving the Steiner another shot.
Thanks,
Andy
Hi,
please excuse the nagging. Does anyone have the schematics? I'd be really interested in an improved Steiner Filter design. Show of hands: who else?
Thanks,
Andy
The Wayback Machine has it here (https://web.archive.org/web/20120517135948/http://www.eskimo.plus.com/fxstuff/newslackfilter.png).
Thanks Marc! Damn, I always forget to check the Wayback Machine. Sorry bout that, I'll make a note.
I'm looking forward to applying this principle to a simplified envelope filter. I've been meaning to cultivate the synth ripple sound a bit but then got sucked into vactol circuits, which do not ripple all that well. With diodes as control elements this should work much better.
Does anyone know why the design would use 2 diodes replacing one resistor (if you fold back the symmetrical expansion) but 4 replacing the other? That reduces the Q of the filter, which is later made up for by the gain of the BC109-BC179 pair. Seems more complicated than necessary. May it have to do with balancing the characteristics of the three filter types?
Once we are at it: Does anyone know which diodes are best suited for this sort of thing? We want them to have a consistent and (ideally) linear R-vs-I behavior over a fairly wide range. Not exactly a spec that many diodes are made for, is it? I'd be guessing regular old 1N4148's would be a good start? I would expect that rectifier diodes have a steeper (in this case worse) R-vs-I curve, right? LED's might have a too high breakdown voltage for a 9V circuit in this configuration. Ge-diodes are sometimes nice and flat in R-vs-I but probably too inconsistent. Can someone enlighten me concerning Schottkys? I would expect them to be steep but I don't really know.
Andy
Plain old Ge/SI diodes have a moderately predictable V/I over many decades of current, and pretty much all the same.
Actually transistors obey The Law very darn good; diodes are doped for more conductance and tend to have a bend in the plot (though often beyond the currents you will likely use).
Yes '914/4148 or 1N400x. This is a chance to use-up the '4001s you bought by mistake (before you realized that '4007 are the same price as '4001 in DIY lots).
Transistors instead of diodes, eh? I had briefly thought about the possibility of doing just that. The working principle of the diodes in a Steiner Filter isn't all that different from the ones in a Diode Ladder Filter after all, and that works a lot better when implemented with transistors, as Bob intended. I put that on the back burner when I realized that the complexity and parts count would be almost identical to a 2-Pole Transistor Ladder Filter, which somehow appeals more to me, mojo-wise.
But how about using the intrinsic diodes of BJT's instead of "diode diodes"? Any benefit there like better linearity, better matching or anything of the kind? Does your assessment that transistors obey The Law apply also when they are used as diodes or only when used "properly"?
Thanks,
Andy