Bi-Filter w/Hammer mods: Progress report

[Dave_B]

I finally breadboarded the Anderton Bi-Filter with the Hammer mods.  I've never heard samples of this effect before so I wasn't sure what to expect.  It sounds really nice with the a clean guitar and chorus in front of it, but unless I've wired it incorrectly, it's definately not an in-your-face effect. 

I don't know if some of this is normal for the BF, but I'd appreciate any feedback: 
I had no luck with the inverted straight sound.  I know what it's supposed to do, but I couldn't get anything out of it.  It definately wasn't an improvement over the filtered-only sound, so either it's a careful balancing act or I should take another look at my wiring. 

When I turn the envelope sensitivity up too high, the audio distorts.  For my humbuckers, this happens around 12 o'clock.  I'm guessing this is an issue but I won't get to it before tomorrow.

The sweep knob has a definate sweet spot.  If I venture too far beyond that spot, the tone changes, but the envelope stops working.  Is this supposed to be a trimmer?  I couldn't make out the knobs on Mark's BF's on ampage. 

I'll try to bring the breadboard upstairs to record some clips for anyone who might benefit.  Lastly, I was trying to figure out a way to put a distortion pedal in front of it.  After some consideration, I'm thinking about splitting the clean signal going into my Obsidian, then sidechaining it into the envelope follower of the BF. 

Thanks all!

[Eb7+9]

...  definately not an in-your-face effect.

... the reason why is because there's only one stage of phase-shift - ok, there's two stages but they're running in parallel, so you still get 90deg shift max at any frequency ... you'll get more phase-filter depth if you wire the BiFilter stages in series with identical caps instead of in parallel with hi/low filter caps as Craig had in his original design (your choice on the cap values) ... if you want an in-your-face Phase-Filter effect you'll have to go with a four-stager like Korg's "Mr. Multi" ... I posted an un-official schem for it ...

~jc

[Mark Hammer]

I finally breadboarded the Anderton Bi-Filter with the Hammer mods.  I've never heard samples of this effect before so I wasn't sure what to expect.  It sounds really nice with the a clean guitar and chorus in front of it, but unless I've wired it incorrectly, it's definately not an in-your-face effect. 

I don't know if some of this is normal for the BF, but I'd appreciate any feedback: 
I had no luck with the inverted straight sound.  I know what it's supposed to do, but I couldn't get anything out of it.  It definately wasn't an improvement over the filtered-only sound, so either it's a careful balancing act or I should take another look at my wiring.

Look at the wiring.  It should make an easily noticeable difference.

When I turn the envelope sensitivity up too high, the audio distorts.  For my humbuckers, this happens around 12 o'clock.  I'm guessing this is an issue but I won't get to it before tomorrow.

This sounds suspicious.  You may get more noiceable envelope ripple on the tail of the decay (assuming you picked a time constant cap that is fairly small/low), but the envelope sensitivity should have no bearing on the audio other than that.  Someting sounds amiss.

The sweep knob has a definate sweet spot.  If I venture too far beyond that spot, the tone changes, but the envelope stops working.  Is this supposed to be a trimmer?  I couldn't make out the knobs on Mark's BF's on ampage.

That 10k unit IS a chassis-mounted pot on the original.  And you're right.  I found it to be of limited utility on every one I've built.  I suspect it ought to be replaced with a 5k pot and better chosen resistors tying it to ground and V+.  Only part of the overall range works.  The other thing to keep in mind is that, unlike the manual shift often found on phasers and flangers, this control simply adds adds a DC offset...blindly.  If the amount of picking-produced sweep and DC offset sum to produce more sweep than the optoisolator is capable of, that's what you get.  In things like the BF-2 flanger, the impact of the manual shift depends on the setting of the depth control (as it does in many other modulated effects).  Here, there is no such tradeoff to "guarantee" the utility of the manual shift over its whole range.[/quote]

I'll try to bring the breadboard upstairs to record some clips for anyone who might benefit.  Lastly, I was trying to figure out a way to put a distortion pedal in front of it.  After some consideration, I'm thinking about splitting the clean signal going into my Obsidian, then sidechaining it into the envelope follower of the BF. 

My bench is kind of backed up with other things at the moment (too damn distractable!) but one of my planned initiatives is a BFF that has a simple Dist+ clipper stage between the input buffer stage and the filters.  A bit like the Q-Tron, but with a fuzz hardwired into the "loop" instead of simply being able to insert one.  So, the first op-amp stage (with the 33k feedback resistor) sends its output to a bypassable fuzz and to the envelope follower.  This way you get to filter a harmonically richer signal, courtesy of the clipper, but do so with more dynamic responsiveness available.  If you clip before you get to the rectifier, you have less dynamics to work with because of how clippers restrict dynamics.

[Dave_B]

Day two.  Thanks for all the feedback, guys.

The distortion was an easy fix.  I simply backed off the filter output level.     Somehow I had it in my head that it was unity gain and couldn't possibly be the culprit. 

The out-of-phase straight signal still isn't working correctly, as simple as it should be.  A closer look is obviously needed. 

JC's explanation regarding the mild phase shift makes sense, thanks.  For those who haven't heard it, it's not Bad Stone level swooshing as you could guess, but it's a really cool sound.  Especially when used with a detuner/chorus/etc.  I played with it until 12:30am last night, knowing I would be screwed when I tried to wake up this morning (I was late).

My next step is to put an LED in parallel with the CLM6000's to get an idea of what they're doing.  As I mentioned, when it's not working, I don't know if it's because they're pegged or not getting current.  It'll be better when I have the pots mounted and I know I've wired them up in the right direction.     I'll definately be adding some trim pots to either side of the Sweep control and probably reduce it to 5k as you suggested.

Thanks again!  More later... 

[Eb7+9]

... one thing many people don't realize about Craig's topology is that the input resistor (47k) and the grounded photo-resistance combo at the input of the pass-filters form your typical compressor/limiter divider network ... so this circuit's got an extra mojo-compression quality built into it - kinda cool ... when I first played Mark's unit  in Ottawa I had a nagging feeling about it - it's only when I spent a few hours with mine that things clicked ... thanks again Mark !

[Mark Hammer]

Following up on Jeorge Huge Tripps' suggestion, I spent some time trying out the Tone Core Otto Filter in stereo, rather than in the mono dock it gets sold in.  Plugging it into two amps and switching them each off and on, one at a time, revealed how to do something interesting with the BFF, if you have the flexibility to do so.

The Otto Filter has one filter mode called "talking filter".  Very funky sound, that has some elements of the Schaller Yoy-Yoy, a touch of Baseballs, and some relatively distinctive sounds.  When you separate it in stereo, it becomes clear that the talking filter mode is essentially two counterswept bandpass filters.  One goes up while the other goes down.  I don't think they are absolutely symmetrical in tuning and motion, but they sound pretty close to it. 

The BFF uses two staggered bandpass filters that are swept in the same direction.  If you can arrange for it, the Otto Filter suggests that it would be cool to stick an envelope inversion stage in front of ONE of the optoisolators, such that one of the LDRs starts out high resistance, and the other low resistance (i.e., LED on).  By "inversion" I don't mean inverting the audiop signal that drives the rectifier.  Rather, I mean something like what is seen in the Mu-Tron, where the strength of picking drives a higher voltage lower.

When pooled into mono, they would sound good, but if the Otto Filter is any indication, they would sound terrific if the two filters were fed to separate outputs and amps.

[Dave_B]

Day three:
Figured out that the lower filter was outside the range of my guitar.  It didn't do anything audible so I assumed a wiring error.  Replacing the .01's with .0047's got it going.   With both filters, I'm getting a nice amount of quack now.  I still need to tune them.  I forgot to mention the really nice sustain Eb7+9 pointed out.  To me, this alone makes the effect worth having.  It works really well with light picking. 

There's still some wierdness I can't explain without putting it on the scope.  If I turn the sensitivity pot down to a point that the envelope is too short, it acts like it "locks up" the EF.  I have to nearly turn it all the way up to get it to quack again.  Almost like it there's some feedback in the EF.  Dunno.

Regarding the stereo mod, now that you mention it, Mark, I do have a (not so) short list of things I planned to do to this circuit, including that.  Most of the suggestions have been covered, but for the sake of completeness (in no order):

1 - Stereo out
2 - Hi-Gain/Q switch (I've got them both hardwired at 740k at the moment)
3 - Separate envelope circuits for each filter
4 - "Invert" circuit for each envelope follower
5 - LED display for each envelope follower (probably a single LED, though I've got a couple of LM39xx's in the box)
6 - Attack-Release circuit (switchable between A/R and EF)
7 - Add trimmers to either side of the sweep pot
8 - Hammer Mod: 1Meg pot in parallel with LDR
9 - Two additional filters in series with the two existing.  (This is a big maybe, inspired by Eb7+9's comment)

I'm forgetting a couple, but that's really too much already.   

The thing is, the EF phasing sound is what attracts me to this circuit.  Since R.G.'s 180+ is next on my build list and I plan to include an EF on it, I'm not sure if I'm wasting some effort once I get that and the Ross Compressor finished.   

Sorry about the lack of samples.  I've been working on this stuff late at night and the PC is very near the bedrooms.  Even unplugged, the guitars wake the kids.   Hopefully by the weekend I can post something. 

Thanks as always!

[Mark Hammer]

Your long list of potential mods validates my regular assertion that this is a VERY good project for learning about envelope control, about "feel" in such devices, about filters and about simple things/changes that can make a world of difference.

For the time being, I'd say that staggering the filters farther apart is an excellent way to mimic what an envelope-swept phaser would sound like.  Anderton has the filter caps at .01uf and 3300pf in the stock design.  That puts the higher filter about an octave and a half higher up.  If you drop those 3300pf caps to 2200 or 1500pf, the overlap between passbands is reduced enough that it feels like a couple of resonant peaks seeping, rather than one wider passband (as was Anderton's original intent).  From your comments, the stock values didn't work out very well so you'll have to use something other than what I've suggested, but the stagger/spread should be attempted with those other values.  For example, 6800pf for the lower filter and 1000pf for the upper one wouldn't be completely out of whack.

Combined with the clean output of the front-end buffer (remember, the filter sections are inverting) that should pretty much nail an envelope-swept phaser, because it will produce two swept notches.

As for having a different sweep feel for the two optoisolators/filters, you don't need to make separate envelope followers.  Really all you need to do is have different RC combinations (R11/C10) for each LED.  The stock arrangement is a 470R series resistor and a 100uf cap to ground.  This gives a slightly delayed attack, and a longish decay.  You could easily leave that in place for the upper filter, but run, say, a 100R resistor and 10uf cap to ground directly to whichever of R1 or R2 is going to the LED driving the lower filter.  So, to recap, instead of 470R going to two parallel 1k resistors, 470R goes to one 1k resistor and 100R goes to the other.  A 100uf (or 22uf-100uf) cap goes to ground from the 470/1k junction, and a 10uf cap goes to ground from the 100R1k junction.

What that gives you is a lower filter that sweeps faster than the higher one.  The lower one is the "meat" of the tone anyways, so it sort of needs to reset faster to be ready for the next note/strum/pick.  The other filter is the articulation of the note, and can add some nice ambience by taking a bit longer to settle down.  The other thing I might add is that if you decide to split the filter outputs to stereo, having one settle faster than the other produces some nice stereo movement effects....not unlike the Otto Filter.

Sounds like you are well on your way to doing all the right experiments, though.

[Dave_B]

This got me thinking.  If I set my Moog to sweep from a rumble to a tweet with a ramp LFO, I could actually see the notches on the scope, right?  If that's true, a VCO modulated by a ramp would be a nice DIY project for our benches.  It would allow us to 'see' the frequency response of whatever effect we were testing.  In the case of the Bi-Filter, it would allow me to visually tune the filters and see how they 'move' with changing amplitude.

Or I could be way wrong.

[Mark Hammer]

Nope, you're right.  Sweep generators are a fundamental part of testing in the audio industry.  Just remember to set the sweep of the scope right or else you won't see the entire spectrum distributed across a single screen.

[Dave_B]

I've only done a little DIY since my last post on this thread, but I did manage to patch in a distortion circuit tonight.  Since I have an Obsidian (still!) on a breadboard, it was easy to patch in.  As Mark suggested, I put it here:

(hope that makes sense).  For anyone contemplating this circuit, adding the option of distortion is probably a no-brainer.  It sounds great to my ears. 

[Mark Hammer]

Perfectly clear.  I'm working on it right now.