using an LM2917/LM2907

[liquids]

Thanks and sorry, PRR.  

Guess that kind of thing happens when I only reference the one page I printed out from a (Fairchild) datasheet for a pinout reference while breadboarding....

[liquids]

Resurrecting this old thread/idea...because I finally resurrected what was on my breadboard for all this time...and tweaked...

I got it working again.

I have to say, the Deluxe Octave Multiplexer (DOM) self-tracking filter circuit is better than I previously remembered, for unknown reasons, but it sure does work!  
I added it in via some tweaking...and it really was some notable improvement.  
Some tweaking for desired results didn't hurt.  It's not perfect.  Nothing in guitar is perfect.  If I wanted perfect tracking...I'd have a dedicated guitar with as large a string gauge as my finger bends could handle, and they'd be flat or tape wound, and hex pickups...but I'm not there.  At least, not yet.

For standard stuff, the DOM circuit is fantastic, and it's easily duplicated with LM13X00 OTAs instead of the CA3094 OTAs called for in the original circuit.

My previous use of the circuit snippet of the DOM tracking filter - as is similarly found in the Electro Harmonix (EHX) guitar synth - was not impressive to be in comparison to some fairly precisely chosen static filtering.

I did proceed the DOM tracking filter with a 3-pole 'static' Low Pass Filter (LPF) like you find in most tracking circuits (EHX microsynth, boss OC-2 and almost every octave-down pedal scheduled I can find uses this approach).  
Most of those tout a "700hz" frequency break in print for harmonic filtering.

Whatever their frequency cut off is in reality, I simulate things with LTspice, and I tweak values so that the frequency cutoff plot allows the highest notes on the guitar - above the 12th fret on the high strings - not to be as filtered out or gain-reduced as they do with the "stock" cutoff static LPF.  That makes it more liable for glitching out, though, certainly.

I didn't a/b between only using the DOM filter in comparison to the static filter feeding the DOM filter.  And "both" is a lot of filtering, but it's worth the chips/space/time to me IF the results are there.
That said, I sharply cut off everything that is notably above 1khz with a static filter, and then fed the DOM self-tracking filter with that.  I would imagine that helped, but I have to experiment on if the static filter actually does help the way I imagine it may.
I suspect that cutting off harmonics above 1k before the DOM tries to "find" the frequency means the DOM does less "searching/scanning" of frequencies and harmonics and more quickly "lock in" to the note being played.

The speed issues is important, because if one audio probes, not playing a note, or skipping from a low note to a high note in particular causes the DOM to spend some time "searching" for the high note that is now being played.  The cut off of the filter is sharp, so when one first hits the high note, it's quiet...then it audibly swells in volume once the DOM filter locks into that frequency.

That makes me wonder...

1) Why or what benefit is it to have a compressor BEFORE the DOM tracking filter, rather than after...if not both (the DOM circuit uses a compressor, though one may use any compressor).
- compression both before and after the filtering may be overkill, because the assumption here is, something following the tracking filter is going to amplify/square/distort and hence "compress" it at least to some degree.
2) If the DOM is outputting what is, effectively, a sine wave or similar to it...ignoring VCO's....what might be a high-quality wave-shaping circuit that takes a steady sine wave and turn it into a sawtooth "directly" without Pitch to Voltage (P-2-V) and Voltage Controlled Oscillator (VCO) workings, which is what I currently have going?
- Taking the DOM's "sine wave" out and yielding a saw-wave via direct waveshaping may involve compression given the aforementioned "swell" that happens with some notes, as well as some serious amplification for a big swing - I am aware.  I am running +/-12V rails right now and may eventually be using 15v, so amplification to yield a sine with whatever amplitude AC voltage swing is needed should not be an issue at all.

(Note that my earlier sound recordings got moved into a sub-folder in the gallery, but if one looks for my folder and such, they should remain somewhere in there)

Thanks!

[PRR]

Guitar signal varies >100:1. Your tracker may not be able to handle that. It may be easier if it knows any 'good' signal will come to it already inside a, say, 2:1 range.

Sine to sawtooth: square it off. Use the (say) positive rise to trigger a ramp generator. Sawtooth! Problem: the amplitude falls as frequency rises. There's no way to make the sawtooth hit the right voltage at the right time unless you know the frequency, or post-compress the sawtooth (which takes a few cycles).

[bms19]

I know it's quite late but I have a question: for a project I wanted to use a LM2907 to drive a led: basically intensity been proportional at guitar frequency. BUT I have a hard time trying to find a way to use the guitar signal to drive the LM2907 like a tachometer would do (ideally a square signal centered on 0V?). Any oidea?
Regards
Ben