Marcos-Munky's Auto-Wah Build Report!

[Marcos - Munky]

Is great to know that a lot of people is interested in this schematic. I think it would be cool if someone add the depth pot, and Gez's Swing pot, to make an inproved version.

[darren]

i don't have a 4001 to breadboard, but maybe you could scale all the LFO components?

ie instead of 470k + 4M7 pot + 330nF, have 100k + 1M pot + 1uF.  worth a try  

[gez]

i don't have a 4001 to breadboard, but maybe you could scale all the LFO components?

ie instead of 470k + 4M7 pot + 330nF, have 100k + 1M pot + 1uF.  worth a try  

You don't need a 4001.  When the 1M is pulled to ground, the circuit functions just like a standard CMOS inverter LFO.  This whole circuit can be constructed with a 4049 inverter chip or any other CMOS inverter chip.  

Tying the inputs of a NOR (or NAND) will give you an inverter, so with the exception of the gate with connections 1,2&3, it's a pretty standard circuit.  Imagine the NANDs are inverters, imagine inputs 1&2 are tied together and ditch the 1M tied to the +ve rail and you've got your oscillator.

It's an idea to include a 10M 'protection resistor' (though not essential) from the junction of the 470k and 330n cap to the input of the ist inverter (1&2 if this gate were wired as an inverter), to prevent the protection diodes conducting.

As for values you're bang on the mark.  I was giving values from memory and the 470n cap will give a nice range for a tremolo, but probably not for this circuit.  Then again, why not switch between caps (wobble wah?).  The frequency is determined by 1/(2CR), so the original circuit gives 0.29 - 3.2 Hz, your values give 0.45 - 5Hz and mine 0.97 - 10 Hz.  Unbuffered inverter chips work better in these circuits (can't remember why, but I've got some notes on this somewhere if anyone's interested)

I'll do some number crunching and come up with some values for a swing pot.

[gez]

You usually see this circuit with inputs 1&2 wired together so that this gate functions as a inverter.  However, the designer has been really clever and tied input 1 to the lower end of that 1M resistor.  With the switch open, input 1 is held high and the output is ALWAYS low - no oscillation.  With the switch closed, input one is pulled to earth and the output can be either low or high, depending on the state of input 2 - the gate effectively functions as a inverter so the oscillator kicks into action.

If you set up a envelope follower to drive a Schmitt Trigger, input 1 could be connected to the output of the Schmitt and this would synch the filter with the initial attack of a guitar note.  If you’re canny, you could set the thing up to synch with a down OR up-sweep.

Seeing as a envelope follower has been used, why not connect that up to the filter as a switchable option?  This simple little circuit could be a real über wah!

If I have the time today I’ll sort out some values for a ‘swing’ pot.

[gez]

Before I start on the swing pot, I forgot that pins 8&9 would have to be connected to pin 3 (not 4) via a reverse bias diode for the flashing millennium to be in synch with the filter!

I’ve sorted some values for the ‘swing’ control.  For those whose eye’s are about to glaze over, if you intend to breadboard this circuit in the next few days and, having read through this post, you don’t have a clue what I’m talking about, mail me and I’ll send you a provisional schematic.  Once the values have been verified I’ll send a proper schematic to Marcos for posting.  For those who do understand, post your results back to the Forum.  All my breadboards are full, so unfortunately I can’t have a crack at this circuit myself.

I’ve had to make a few changes and compromises due to the original values I chose being a little hard to come by, so this mod hasn’t been verified!  Change the 4M7 rate pot to a 470k lin pot, wire up a 1M resistor in parallel with it and change the 470k stop resistor to 20k.  Disconnect this resistor chain from input 2 and wire up it’s ‘loose end’ to the wiper of another 470k lin pot.  Now connect a Si diode to each outer lug of this pot so that they’re back-to-back.  Join the ‘free ends’ of the diodes together and connect them to input 2.  Change the 330n rate cap to 470n (.47u) and wire up another cap of the same value in parallel with it.  Make this second cap switchable for a slow/fast switch.  That’s it!

The diodes create a differential charge/discharge ratio for the cap which sets the rate.  Moving the pot from it’s middle to one of it’s extremes changes the mark/space ratio from 1:1 to 1:2 OR 2:1. which will set up a swing feel; 2:1 accents the beat on a up-sweep, 1:2 on a down-sweep.  With the pot in the middle the wah is normal (no swing).  The combined resistance of the stop resistor and the max resistance of the rate pot MUST be the same value (or less) as the swing pot’s resistance to guarantee a 1:2/2:1 ratio for the slowest setting of the rate pot.  The higher the speed the less need to go to the extremes of the swing pot’s travel - you’ll get the idea when you try it.

The frequency range is limited but you can tweak cap values to suit (smaller = faster, larger = slower).  The rate of the oscillator ‘should’ vary from 0.85Hz to 3.8Hz , giving a tempo range of about 51 - 228 BPM, which should cover most situations.  Only problem is, these buggers are temperamental; if the protective diodes come into conduction it alters the rate slightly, as does the output resistance of each ‘inverter’ (which is dependant upon supply voltage and varies with each chip) - even including the 10M ‘protective resistor’ can increase gate capacitance and change timing.  However, it should be pretty consistent with the right values, I just hope the ones I’ve given are.

Any takers on this one?

[Arn C.]

Thanks Gez!   Swing pot-great idea,  Now if we could get M.Hammer     to give us a hand with a depth pot, we would have one fantastic auto-wah!!!
Arn C.

[idlefaction]

especially if you cascade filter sections!!  this would be an easy one to build a 4P out of.  what does the filter actually sound like, i've never built one of these one-transistor T filters?

[gez]

I bought a mini breadboard and wired everything up - it's fine.  I only had high-gain trannies so I had to do some re-jigging to get it to work.  I had to make the 15k smaller to get the sweep right; a trimpot would be a good idea but I ran out of jumper leads so I had to use a resistor which meant I couldn't get the filter sweep perfect.  I got a rough idea though and it sounds good enough for me to buy some lower gain trannies tomorrow!

The swing control works a treat.  If you give this a whirl, try it at a slow speed first to get the hang of it.  With the rate on max, the speed switch on fast and the swing pot just slightly off-centre you get a pseudo Leslie sound which I really liked!  

I also rigged up the flashing millennium and it works fine too.  I've sent the schematic to Marcos so it should be up soon.

[gez]

Hey Arn, I’ll do a bit more work on this over the next week.  If I come up with anything more I’ll let you know (keep the perfboard on hold!   ).

The RC low-pass does a reasonable job of converting the square into a slightly distorted triangle, but the swing control, when slightly off centre, seems to correct this by altering the mark/space ratio and you get a pretty good triangle (as far as I can tell anyway - the image is a bit fuzzy).  Mine ticks, but then I haven’t paid ANY attention to layout so we’ll see how things go.

[Arn C.]

Thanks Gez!   I will leave the original because I like it as is.  But I definately want to make a fully modded one once it is all together!
Thanks again!
Arn C.

[gez]

For anyone thinking of having a go at this circuit, hold your horses!  There are a couple of things that need fixing.  I’m working on it and hopefully I’ll have something within a week.  

Normally  I use the swing control with triangle oscillators, but it seems a little problematic with this circuit as the range of sweep seems to shift a bit as the control is turned.  I think it’s because of the square to ‘triangle’ filter coming off the buffer.  Anyway, a quick fix is to replace the 15k resistor with a 50k log pot which acts as a ‘range’ control (set it up so that it’s resistance increases as you turn it).  It’s probably the nearest thing to a depth control you’ll get with this circuit Arn, and I’d recommend doing it even if you don’t have a go at the swing control as it compensates for different trannies.

The second problem is clicking.  I know Arn’s is quiet, but mine clicks like buggery (does buggery click?).  I know I haven’t paid any attention to layout, but even when I have in the past I can never get it to go away totally, and I think that many who build this circuit will come across the same problem.

Recently I’ve been following the advice given in ‘The Art of Electronics’ and started running CMOS oscillators off 3V and it works like a charm!  I tried it on this circuit today (used a micro-power 5V reg and MOSFET source follower with gate tied to a trimpot) and I had to put my ear to the speaker to hear anything! (not bad considering there’s a ton of spaghetti crossing all over the breadboard).  

I’m going to have to re-jig a few things, maybe even convert the oscillator into a true triangle job, but I’m confident it will be problem (read ‘click’) free…stay tuned!

[gez]

P.S. Now I've got the 'range' control wired in, I'm in a better position to judge the filter.  It's not bad!  Pretty subtle, but that's good - I've rigged up 'better' filters to LFOs in the past and they can sound a tad ridiculous IMO.

[Arn C.]

GEZ,
    Thanks for all your work on this circuit.  I think it is definately worthy of your time.

[If any of you are having problems with finding a 4M7 pot, then try a 1M pot, reduce the 470k to 100k and increase the 330n cap to 470n]
-Can I still do this without changing anything else in the circuit?

[a quick fix is to replace the 15k resistor with a 50k log pot which acts as a ‘range’ control (set it up so that it’s resistance increases as you turn it). It’s probably the nearest thing to a depth control you’ll get with this circuit Arn, and I’d recommend doing it even if you don’t have a go at the swing control as it compensates for different trannies. ]
-This is cool, thanks!

[The second problem is clicking. I know Arn’s is quiet, but mine clicks like buggery ]
-I used shielded wire from both jacks!

Peace!
Arn    
[YOU GET OUT OF LIFE WHAT YOU PUT INTO IT!]

[gez]

Hi Arn

"[If any of you are having problems with finding a 4M7 pot, then try a 1M pot, reduce the 470k to 100k and increase the 330n cap to 470n]
-Can I still do this without changing anything else in the circuit?"

I've only breadboarded the values I gave for the 'swing' control version.  The 1M and 100k will work fine, it's just a matter of finding the right cap value to suit.  470n will get you into tremolo territory and will probably be too fast for this circuit.  I think Darren was closer when he said 1u.  If this is slightly fast you can always wire up a smaller value cap in parallel (.22u?).


"I used shielded wire from both jacks!"

Good advice, and I'm sure this is one of the reasons I've had problems in the past.  However, I like the idea of designing something that doesn't need all the attention to detail usually associated with click free circuits (also, the last time I bought shielded wire the thinnest they had was almost finger thick!).  Just indulge me!

"Thanks for all your work on this circuit. I think it is definately worthy of your time."

I'm enjoying it!  The thing I like is its subtlety, which really suits an auto-wah in my opinion.

[Arn C.]

Thanks for the responses!   I bought a little spool of shielded wire and it is quite thin, sort of the wire for headphones but only a single wire not two for stereo.  Can't remember where I bought it, so long ago.  But, it was aound $5.   I have also used rca patch cable, the cheapest ones I can find.
I will try the pot change on the next autowah I build.   Soon.   The other one is all boxed up, painted and in use!
Peace!
Arn C.

[gez]

I've sent a revised schematic to Marcos, so hopefully it'll be up soon.  Rather than totally redesign this thing I kept faithful to the original and have tried to leave alone as much as possible.

The LFO runs at 3V.  The only changes I've made from my first version are to replace the 15k with a 50k log pot ('Range control'), reduce the 100k to 20k and I've made the .47u caps smaller (.22u).  This was necessary as the LFO runs slower at lower voltages.  I also found that using 22n caps, instead of 10n, for the filter gave a better wah sound.

I wired up a few trannies and they all seem to work.  The only reason I had problems with high gain ones was a wrong connection - duh!

Anyway, that's me done on this one...now time to mutate it for purposes of my own! (insert sinister cackle here)

[Marcos - Munky]

Is uploaded it now. Thanks again, Gez.

[Arn C.]

Thanks again Gez!  I will try making another one of these with your mods!

Now if we can get someone to work on the "DEPTH MOD" , we could have a real happening pedal!

Peace!
Arn C.

[gez]

The Range control acts partly like a Depth control.  With max resistance on the pot your normal guitar tone is heard.  Decreasing the pot’s resistance gradually brings the wah effect in.  Unfortunately, at really low resistance the range of the sweep isn’t enough to break past the mud barrier, but that’s the down side of this design.

The only way I can see of getting a true depth control would be a mixer at the output, blending in the clean signal with the filtered.  However, seeing as the wah effect is pretty subtle (not a criticism), I think this would be a waste of time - full wah ahead Scotty!