Modernizing the EHX Knockout Attack EQ

[Bill Mountain]

I bought one of the reissues on whim with some birthday cash months ago but never really jammed on it until last night.  I used it with a TS clone and it added a super tight and thick bottom end with my P-bass.  I felt like I had a pair of subs behind my amp.  So as any DIYer is bound to do I started to think about how I could make one for myself.  The current idea is to combine it with a fuzz of sorts in a single box.

One look at the schematic and you can see the design of the filters are not ideal even though they sound just fine.

Schematic:

http://www.freeinfosociety.com/media.php?id=4117

Even the original designer said he'd design it differently today:

http://electroharmonix.vintageusaguitars.com/index.php?option=com_content&view=article&id=15:the-attack-eq-fact-and-fiction&catid=13:electro-harmonix-articles&Itemid=45

So now I am wondering if it behooves us to update the design for future generations.  My first though was to do a series of Sallen-Key filters or a state variable filter but I worry It wouldn't have the same feel?  Could a Bax EQ be designed with steeper slopes I wonder?  Is it fine how it is and maybe I should just leave it alone?

This is not an immediate project for myself but It will be nice to think about it for a bit while I finish up other projects.

I also wonder what fuzz would pair really well with it.

Decisions...decisions...

[slacker]

We need someone smarter than me to tell us how many poles the filters are then we can figure out what sort and how many active filters we need to replace them with. My gut feeling is there's not much to be gained redesigning it unless you want to make the filter frequencies variable.

[Mark Hammer]

When you think of it, the Knockout is essentially a BMP tone control gone absolutely berserk.  Six poles (and maybe 7 if one considers the components setting the gain of the op-amps) on a lowpass and highpass path.  Instead of the BMP control where one pans between the lowpass and highpass sides, this allows you to dial in how much of each path - not unlike a BMP tone-control mod Jack Orman suggested some years back.

Perhaps the most sensible "modernization" of the circuit might be to use a rotary switch to vary the number of stages and slope of the filtering, (either overall, or separately for each path) by shunting/bridging select stages (R/C pairs).

Alternatively, how about making each path 2-pole, but incorporating a dual-ganged pot to tune each path.  So adjust where the highpass cut begins, and then dial in how much of the content from that passband you want to keep.  Same for the lowpass path.

Now here is where it could get really interesting....  The Knockout fixes the filtering so that the high path tends only to the bite, and the low path fixates on the fundamentals.  If the lowpass and highpass filters were adjustable, then the lowpass could be used to filter out high stuff, and the highpass could be used to filter out deep bottom.  In other words, they could be combined to produce a bandpass function.  Creating a specific passband, and dialing in how much of it you want to blend with full bandwidth, can mimic a variable resonant boost.

[Bill Mountain]

First pass of a schematic for a DOD 250 with the Fundamental control of the Attack EQ.  No modernization yet.  Just wanted to see what it would look like.

[lion]

I've been looking at the AttackEQ circuit a couple of times. Did a bit of research and found some info on one of the forums- quoting:

".....it is just a (clever?) eq. The signal gets split in three components: dry, low and high. Low is everything below 85 Hz, high is everything above 6.5 kHz and dry the input signal.

Every knob controls how much of every component goes in the output signal. In other words, if all knobs are set to 0, there is no output. If low is set to 12:00, dry to 0 and high to 0 only everything below 85Hz pass through and so forth.

According to the manual the filters are:
85Hz, 7-Pole Low-Pass Filter
6.5kHz, 6-Pole High-Pass Filter"

For the life of me I can't get the filter component values from the schem (especially the HP section) to fit with the pedal specs mentioned.
Am I the only one wondering?

Edit. I better add that I understand multi pole RC's "add's up" the effect.
With the AttackEQ's component values for the LP filters, the roll of fc is 268Hz for each, so putting 7 in series will shift down the roll of point - which might bring the total LP down to 85Hz.
However for the HP section the Fc is 1540Hz for each pole - by cascading 2x3 sections would that bring the resultant roll off point UP OR DOWN - first I would think down, but thinking again this might the part I'm missing

Erik

[samhay]

>According to the manual the filters are:
85Hz, 7-Pole Low-Pass Filter
6.5kHz, 6-Pole High-Pass Filter"

Simulation says that this is about right:


(click for larger version)

As for modernisation. I guess you can save on large value resistors by using 3-pole Sallen-Key filters. Making the frequency variable would still be tricky.
You could reduce the number of poles as Mark suggested - can you hear the difference between a filter with 4 vs. 6/7 poles?
Another option would be to use a switched capacitor filter ICs like the MAX7401, which is an 8-pole LPF. I don't know of a high-pass equivalent, but I'm not sure the high pass filter needs such a steep roll-off.

[lion]

>According to the manual the filters are:
85Hz, 7-Pole Low-Pass Filter
6.5kHz, 6-Pole High-Pass Filter"

Simulation says that this is about right:

Thanks for posting the plot Sam. Multi pole filters keeps being a "mystery" to me 

Erik

[Gus]

Do a search for "Don Lancaster Active Filter Cookbook" a very good book