Anderton Bi-Filter with Mark Hammer Mods

[Chico]

Mark posted a schematic of some mods that he recommends for the Anderton Bi-filter a while back.  My copy of the schematic labels the mods as "untested", so I thought I'd test them.

This is a really cool build for anyone looking for the envelope controlled filter effect.  With Mark's mods, the device provides an extremely flexible and versatile array of sounds.  Unlike the Quack/Q variety that use a transistor as a variable resistor, the Bi-Filter uses a pair of LDRs.  Each LDR affects an independently tuned filter section.  

This design gives you not only an envelope sensitivity control, but also a bias control that allows you to change the way the LDRs are driven by the envelope.  Mark adds a control that allows you to mix in dry signal to the filtered signal.  Not only that, you can chose between inverted and noninverted dry signal content in your output.  Finally, Mark adds a switch that allows you to alter the characteristics of the high frequency filter.

I built this pretty much straight off the schematic, and used NSL-32SR3 optocouplers for the LDRs.

The design is a fair degree more complex than the Quack/Q variety, but sounds great and is extremely versatile.  It is also readily adaptable to different guitar/pickup combinations due in part, to the more versatile envelope section.

If you are looking into a filter, check this one out.

Thanks for the cool mods, Mark

[Mark Hammer]

Well thank YOU, Chico.  Mine is currently sitting 85% done in a chassis on my over-cluttered bench, so for my part the suggested mods for adding in un/inverted dry signal are theory only.  You have my gratitude for demonstrating they work in reality too.  

I added a very simple MXR Dist+ style clipping section (preset clipping/output level) between the input stage and filters that could be switched in and out for texture as well as for uncompressed envelope detection (sticking a fuzz ahead of the filter reduces dynamics), but it isn't critical.

The wet/dry mix is a very nice feature that bass-players will love especially, since it keeps the animation in there at the same time as it preserves bottom.

I know that when the two filter sections on the stock unit are staggered, it sounds very much like an envelope controlled 4-stage phaser, especially when the decay is stretched out.  When the filters and dry signal are mixed in so that notches are produced, it must sound even more like an envelope controlled phaser.

This is a VERY moddable design that Craig provided, and what I've suggested barely scratches the surface.  It is remarkable, though, how just a couple of tweaks 'n toggles here and there can expand the palette so much for so little money.

Thanks again, to you and Craig Anderton, and congrats.

[bwanasonic]

Does this circuit resemble the EH Bass Balls at all, at least in effect? I always really liked the BB and would consider building something like it. Got the *plain* env. follower covered with an EMMA Discumbobulator.

Kerry M

[Eb7+9]

... and used NSL-32SR3 optocouplers for the LDRs.

I used the same cells in Mark's take on the circuit ... btw, did you notice the compression in this Envelope follower ?  

... the scaling resistor at the filter inputs forms a voltage divider with the grounded photocell resistors, definitely a compressin gain block - I think it adds to the unit a great deal ... I just wish I could figure out how to avoid having it peak so much at times, like maybe having some kind of non-linear Q limiting in there ...

[Mike Nichting]

Hey Mark,
I would like to thank you for all of yopur help and very interesting mods on all of the different effects~!!!

Much appreciation bro~!!

Mike N.

[Mark Hammer]

JC,

Hi from Ottawa!  Rumour has it you got our weather.

Mike,

Thanks, and you're most welcome.  Our recent change of government and legislation gives me a little too much free time at work while we wait for direction, which accounts for my output these days.  Glad you enjoy it.

Kerry,

The BB and BFF are cousins.  Both use staggered bandpass filters for a thicker sound (although Anderton used them first), although the filter topography is a little different in each case, so the filter response and tone is a little different.

Though they both also use a simple half-wave rectifier to extract the envelope, the BFF suffers less of the envelope ripple issues that the BB has.  There are two reasons why.  First, the BB aims for a much faster attack time (which is why the stock design *sounds* like it only sweeps down, even though it sweeps up THEN down).  Second, the BFF uses optoisolators, rather than transistors for control elements, and these tend to smooth out ripple during the decay cycle.  You don't notice it so much when playing single-note lead lines, but if you happen to strum a chord and let the filter gently fall back as the chord dies out, there is much LESS "gentleness" in the way the BB does it.  As a consequence, I far prefer the BFF for rhythm work which is, quite frankly, a lot of what envelope-controlled filters do best.

The stock BB permits one to retune the filter sections individually via the trimpots (I've mounted them on the chassis on my own), though it does not permit any sort of global-yoked retuning of both.  In contrast, although the BFF can be modded to do so, the stock design does not permit individual filter tuning, though it does provide global retuning.  The taper of the 10k sweep/calibrate pot and additional series resistor, can profitably be played around with.  I'm not so sure that any single set of component values will satisfy every user.  The existing arrangement is useful for wide range filter retuning if you wish to use the BFF as a static filter (i.e., sensitivity turned "off" to produce no sweep).  On the other hand, if you only want to use it for adjusting sweep start-point, a great deal of the pot's rotation is pretty useless.  My adoption of a range switch for the upper filter is a bone-headed way of staggering the filter sections in the BFF; something which individual tuning pots makes a lot easier on the BB.  If you decide to go with the BB, either first or as your ultimate choice, chassis mount the 10k trimpots as pot controls.  The variable spacing of the two filter sections yields many different tonal possibilities.  These same pots also have an effect on how much sweep is generated for the filter section, so there is an added element to the range of possible sounds.

Incidentally, the stock BFF has equal-value (1k) current limiting resistors ahead of the LEDs in the two optoisolators.  This means that they both glow as brightly for the same envelope signal.  No reason why they HAVE to, though.  A 1k pot in series with one or both LEDs would result in less brightness from the LED in question and less sweep.  That could actually produce some interesting effects, and be of particular use to bass players who may want more sweep of the mids and treble for animating the sound, but subtler sweep for the low end to retain bottom.  Conversely, having less sweep of the upper filter than for the lower one might yield some more voice-like sounds.  I haven't tinkered with it yet, but it seems promising.  

Practically, it may be better to use a single 1-2k linear pot with the wiper tied to R11/C10, and the outside lugs going to the existing 1k series resistors.  You should probably replace these with 680-750R resistors.  When the pot is centred, you'll get equivalent sweep width for each filter section.  Move it to one side or the other, and you'll get unequal sweep amounts, with sweep "favouring" one filter over the other.

The stock BB makes use of the fact that the audio signal has to be heavily boosted to derive a useable envelope signal, and this boosted-but-unrectified signal is tapped to yield the "fuzz" tone.  As "fuzz" tones go, it is pretty awful since it is simply a misbehaving 1458 (much closer to a dual 741 than is a 4558).  Perhaps more importantly, the volume level is a function of where you set the sensitivity control.  In the absence of an output volume pot for the BB, this is a nuisance.  E-H tries to "tame" the fuzz signal with a resistive attenuator network, but this is still inconsistent since it only reduces the signal level proportionately, and not to some absolute level.  I opted to replace the resistor to ground for the fuzz signal with a diode clipper.  You can play with diode type and number to get an approximate equal volume level from non-fuzz and fuzz sound.  The advantage here is that the diode-clipper produces better fuzz than a 1458 alone, and that the diodes set an absolute maximum volume level, such that turning up the sensitivity gets you more fuzz but not more volume.  In my view this improves the playability and usability of the unit immensely.

Note that you can NOT automatically do a similar thing (tap the output of the envelope follower *before* the diode for a "fuzz" sound) with the BFF.   Since the sweep/calibrate pot adds a DC offset to the envelope signal (which provides a basic level of LED illumination), you'd need to DC-filter that signal.  The use of something a little better than a 1458 with a bipolar supply also means the output would be cleaner, so what you'd be getting is a purely diode-based clipping effect.  That may be more of what some folks want and less of what others want.

Like the Dr Q, the BB relies a little too much on the properties of specific op-amps to provide its envelope-detection capacity with a single-ended 9v supply.  The BFF can use whatever the heck you want, with op-amp quality limited only by your pocketbook or VISA limit.  Between that and the use of optoisolators, the audio quality is predictably cleaner for the BFF, althouth the BB is "funkier" in some ways.  I have both, and find a use for both.  The BB can nail some quasi "talkbox" sounds in a way the BFF can't,. but the BFF can nail some quasi-envelope-controlled phaser sounds that the BB can't.  Pick your weapon....and enjoy.