using an LM2917/LM2907

[liquids]

Anyone have any tips on how to utilize the LM2917 to create a voltage from a frequency?

I understand that to get accurate F/V with such a device in use with a harmonic-laden signal like a guitar note, lots of signal pre-conditioning has to occur...

I've seen some samples of the Analogue Systems rs35 unit which seems to make use of the lm2917 and, even if maybe slow, it seems like it can get 'decent' results.
Likely a whole bunch of other circuitry is involved there to turn hz/v into v/oct, and of course then one needs an oscillator to respond to a control voltage...

But for now, I don't really know how to utilize the LM2907/LM2917 chips.

There is an old schematic here:  http://yusynth.net/Modular/Commun/STEINERVCF/gtr-synth.jpg

But I always find that schematic kind of funny...since it's using the 2907 F/V signal into pin 14 of a PLL which responds to a frequency...not a voltage!  Nevertheless, is that 'schematic' portion for the 2907 and it's output the best/ideal starting point for feeding it a frequency (if done correctly) and getting a voltage corresponding to that frequency, hz/v?

[PRR]

You have read http://www.ti.com/lit/ds/symlink/lm2907-n.pdf ?

[liquids]

Have I read it? Sure I have.   But really, not sure I understand much at all.  I can assume I do, if that helps, but I tend to think that will just lead to me blowing up the chip and right now I'd rather not do trial by error, because I only have two on hand.

In fact, I just read it again and am wondering to myself if this chip can even do what people say it can if fed a 'properly' squared off audio signal, except there are units out there that utilize it as aforementioned.

All the datasheet examples of circuitry use the 8 pin version as their reference.  Most of the time on the datasheet examples, the output is taken off the internal BJT arranged like a buffer.  

However, the schematic I linked takes it's output off of pin 2 of the 14 pin chip (which are what I have), which is internally noted as a pin out of the 'charge pump' circuitry.    

I am lost.  I have two chips (a 2907 and a 2917), and if I blow one, I'll be compelled to place another order.  I'd rather not do that, for one.

The schematic I linked is the only place I've seen a schematic of such a chip utilized for audio, and I suspect it's utilized poorly or improperly - however, I have no way to be certain.

A while back, when I build a portion of the linked schematic, and audio probed at the connection between the 2907 and 4046 as shown in the schematic, it just sounded like a big squared fuzzed up version of the guitar signal. Not necessarily a frequency dependent DC voltage.

So I'm asking for help, if anyone can and/or is willing to offer it.

[R.G.]

A while back, when I build a portion of the linked schematic, and audio probed at the connection between the 2907 and 4046 as shown in the schematic, it just sounded like a big squared fuzzed up version of the guitar signal. Not necessarily a frequency dependent DC voltage.

I'm guessing that they only use the input/squaring up part of the chip and drive the 4046 as a pll. Pin 14 on the 4046 is a signal in, not the voltage input to the VCO.

If so, yes, that's what you should hear there. They're only using the input squaring function of the 2907/2917.

[liquids]

A while back, when I build a portion of the linked schematic, and audio probed at the connection between the 2907 and 4046 as shown in the schematic, it just sounded like a big squared fuzzed up version of the guitar signal. Not necessarily a frequency dependent DC voltage.

I'm guessing that they only use the input/squaring up part of the chip and drive the 4046 as a pll. Pin 14 on the 4046 is a signal in, not the voltage input to the VCO.

If so, yes, that's what you should hear there. They're only using the input squaring function of the 2907/2917.

Okay...so the first datasheet example for a super simple tachometer circuit (on P8 per PRR's datasheet link) shows connections similar to the 'synth' schematic I linked...but it uses pin 5 (via the internal bjt) as an emitter follower output/buffer/feedback point...and turns it into a pot to drive a VCF in part...is it potentially as simple as tapping off of pin 5 to get a voltage that is input-frequency dependent/variable?    Is there likely a better way to utilize this chip for said purposes?

[liquids]

Well Okay, here goes...

I utilized the synth schematic's use of the LM2907, component for component, buttapped off of pin 5 at the junction of the 10k resistor (was a 10k potentiometer in the schematic) as aforementioned.

Since I recognize d in advance that the 'tachometer' is not going to create a 1v/oct (exponential scale, as are frequencies) kind of relationship, but a V/hz relationship (linear), I utilized this:



The Paia Fatman uses linear control voltages...and scott gravenhorst kindly merged the linear aspect of the fatman's response to CV into a simple vco by Rene Schmitz that uses CMOS 4069 stages for the oscilator.

Well, when feeding the tachometer a compressed/filtered/CMOS-type squared signal, it worked.  I did not grab exact parts per the schematic shown above in all cases, and tuning the oscillator is a moving target as expected, but I personally think it sounds really good, is really fun to use (that's what this is all about for me, personally), and though I'd like to learn to refine it and get it to work properly of course, I don't see why this approach shoudln't be further pursued.

Drawbacks - the closest I got to tuning caused the VCO to oscillate at an audible frequency when no notes were being played; likewise, when I dialed in the oscillator to silence when I wasn't playing, I couldn't tune it up properly.

Another drawback - voltages....my bad math lead me to believe that I couldn't get a corresponding voltage out of a 1k frequency with said chip unless I ran it at 18v....may or may not be true.  But Trying to figure out how to arrange a simplish power scheme that doesn't utilize a regulator or two for every chip in the circuit is somewhat difficult with the particular circuit stream I was attempting.

Anyhow, all of these issues may be more due to lazy parts substitutions and my speaking too soon more than genuine limitations or real vco tuning difficulties.  I'm really glad to say that I didn't find monophonic guitar into CV into oscillator a fruitless endeavor (as many many many posts have lead me to assume), and tracking latency is not something I was thinking about at ALL while I was playing with it either - it's there but I don't think it's as noticable as I originally envisioned it would be.   It's circuitry worth trying to refine, IMO.   

At the same time, I stumbled across the sonuus monophonic guitar-to-midi tracking box that seems to put any guitar in reach of any midi synth rather simply and affordably (and the fatman is analog and controlled by midi at that). That is quite tempting avenue to pursue as well.

Anyhow, if anyone else wants to attempt this concept (for the first time or again), and/or help and offer suggestions here, I ain't too proud to beg...but if nothing else, if this interests you, don't rule out the possibility of using your guitar to CV an oscillator.

[liquids]

I guess this is a bump...

at the utter surprise that not one person has commented on the last post...not even steven giles to chime in just to say the EHX synthesizer did this all much better years ago and the schematic is readily available 

[liquids]

A semi-related question...
Rene's CMOS VCO runs off of since sided supply...right now I am running it +12v/ground.
If I take the output of that - which I presume is ramping between 0v and 12v - and send it to a non-inverting op amp buffer that does not have rail-to rail capability, your run of the mill TL072, or 4558 say
...said op amp is running off of bipolar powered with +/- 12 rails...does the op amp output/input it as if it is +/6v signal with headroom to spare, or will I clip the positive rail of the op amp? 

If clipping is inevitable with/without a decoupling cap and reference to ground at the op amp's input, to get +/-6v do I do something simple like decouple the VCO and reference the op amp's input to -6v?

[PRR]

> a bump... ... to chime in just to say

The ARP did better LFOs 40 years ago, and schematics can be had.

But maybe you have pulled in SO many different ideas, you are ahead of us all.

You asked about an anti-lock brake chip used for F/V conversion. Now you seem to be on to a CMOS oscillator.

> I presume is ramping between 0v and 12v

I could write a thousand words about charge-pump. Or about astable CMOS. Except it's been a looong time since I danced with CMOS. (And I was ditch-dig supervisor today so my fingers and brain are worn down.)

I think if you use a cap a thousand times bigger, then you can follow the action on your volt-meter, out of the oscillator, through Rene's 100K pot, and through whatever buffer. SEEing it is often better than reading it.

[liquids]

> a bump... ... to chime in just to say

The ARP did better LFOs 40 years ago, and schematics can be had.

But maybe you have pulled in SO many different ideas, you are ahead of us all.

You asked about an anti-lock brake chip used for F/V conversion. Now you seem to be on to a CMOS oscillator.

The anti-lock brake chip used for F/V conversion from guitar gives me a voltage...I can take that that I can run through circuitry similar to the image I posted, a VCO.  The VCO is Rene Schmitz CMOS oscillator.  It works as pictured off of linear F/V...I'm by no means "ahead of [you] all," or I wouldn't be asking for help here.

If ARP schematics are worth looking at for linearly controlled VCO's I'm all ears...

> I presume is ramping between 0v and 12v

I could write a thousand words about charge-pump. Or about astable CMOS. Except it's been a looong time since I danced with CMOS. (And I was ditch-dig supervisor today so my fingers and brain are worn down.)

I think if you use a cap a thousand times bigger, then you can follow the action on your volt-meter, out of the oscillator, through Rene's 100K pot, and through whatever buffer. SEEing it is often better than reading it.

It did occur to me sometime after posting, that I could use my scope and use the (never before used) second channel to compare and/or more or less estimate jsut how how the VCO sawtooth was ramping.  It was at peak about 8.5 volts, which explains why it was not clipping the positive rail of an op amp when DC coupled...

[liquids]

Realized that I never came back to this, but between the incomprable help of Harry Bissell and the inspiring efforts of 'soffa' here on the forum, I wanted to post some samples I recorded a little while back, that might whet the appetite of those curious about using your guitar to P->V.  This is with the LM29x7 chip doing that for me, eventually controlling Rene Schmitz simple CMOS VCO, a sawtooth.  These samples were recorded for demo purpose only, they also are mid-process for design and tweaking, do not implment multiple simultanous VCO (or at least VCO octave) voices, nor the standard VCA and VCF acompaniments.  Just a guitar controlling a VCO.  

Little of this circuitry is creative or new - and without Harry Bissell's willingness to answer questions and help me think things through, the pace of this effort and project would have dead ended, as it had once already, before I started asking for his input.  Thanks Harry!!!

Hopefully these snippes inspire a few others who like the sounds of true VCO synths, and as a result decide to give it a go and using their guitars as a controler rather than the traditional keyboard method.  The results are almost as fun as all one learns along the way...and I'm hardly approaching finalization of even the core VCO circuit(s).

Of note, there are no CD4046 chips in this circuit or PLL of any kind, though those chips and methodology are also a fine approach at guitar--->VCO.

http://www.aronnelson.com/gallery/main.php/v/liquids/Short+Demo.mp3.html
http://www.aronnelson.com/gallery/main.php/v/liquids/Sawtooth_Clean_amp_Compressed.mp3.html

[soffa]

Awesome, Liquids!

I was hoping that you'd post some more of your stuff. You likewise have inspired me to keep movin' forward!

[liquids]

Thanks Soffa...can't tell if I am surprised, or not at all surprised that no one else has commented on the sound of a VCO more or less glitchlessly tracking single-note guitar playing (minus the overcompression that was dialed in with the first clip as noted), now that there are links to soundclips...

but away and onward we go, eh?    

[Jordan A.]

It sounds wonderful, the tracking is very impressive.  I have stayed away from guitar -> VCO stuff so far because I imagine it will consume my free time and breadboard for a solid month once I start, and I have assumed a high probability of the end results being underwhelming.  The tracking you are getting so far has me wanting to dive in.

Any complete or semi complete schematics of what you have so far?  I'm still not sure exactly what you have going on, and pitch  -> voltage is a pretty new world to me.

Nice stuff!

jordan

[igor12]

Godamn Liquids!!!  I was all over your synthbox and now this?  It sounds freaking incredible to me!

[~arph]

I am watching you... Closely :-p your posts don't go by unnoticed

[slacker]

Sounds really good, be interested to see where this goes.

[liquids]

A month could be short amount of time spent on this kind of circuit if you're brand new to this realm...or plenty of time if you're at all informed...I was the former.

Do take into consideration that when I emailed these clips to Harry B., the master of guitar synth design and one of the most helpful persons I've ever met in the DIY scene, he was highly comlimentary, but given him knowing the circuit(s) and circuitry I've considered/evaluated/used along the way exactly (its changed many a time), he noted that the quality of the tracking he heard in the clips may have as much to do with a cleaness of guitar playing as it does the circuitry itself.  Going back to the mention of the synthbox work - I think that may apply and should be taken into consideration here...my friends who haven't put a lot of time into learning how to play single note melodies cleanly might make this same circuit sound like a bleep labs thingamapoop (http://bleeplabs.com/thingamagoop2/), not sure. Nothing wrong with that either if you're not hoping for perfect guitar p->v!  I haven't gotten perfect guitar p-v either though!

Otherwise, if you can play notes legato and stacatto cleanly (also note that there is no plastic pick in my hand during this recording, or nearly ever for that matter), than you shouldn't struggle too much to get similar results than you hear here.

As for sharing schematics, I rarely get down to values on paper or pixel these days publically - synthbox taught me that posting a schematic is not worth the pain it causes me and the creativity of other's to take it in different directions is cooler and more valid.  No schematic, not because the circuit you hear is special - but mostly because I am hoping to offer detailed typology info, to inspire others to start from and tweak intelligently for their own goals and needs that will likely outdo mine and challenge me back to designing.  

However there is nothing so far in my circuit that is unique at all IMO - lots of typical circuits you'd expect, strung together. All combined, it's still a great number of places for error if attempted to be duplicated pin for pin....involves a number of chips, uses bipolar power rails in many places, and overall lots of components.  That comes with the territory.  One has to be familiar, willing to struggle, fail a lot, learn a lot, etc.

At the same time, guitar->VCO is by no means something impossible, nor is it one in a million attempt for the initiated who understand circuits, not just merely solder veros....that's what I've learned from my work here.

I didnt dare attempt VCO stuff for so long since it seemed that everyone but the old korg roland etc engineers and their units, including Harry B., indicated it was generally futile, unsucessful, at least highly imperfect and unreliable in it's results...in general, the thought is 'don't waste your time.'

I'm here primarily to suggest it's not completely futile, in fact, IMO far from it.  
I provided these clips to further demonstrate that and encourage others to give it a try in the DIY sprit.

I will say that the CD4046 PLL method of VCO (if you work with a 4046 you are working with a VCO, absolutely) is fun, but I actually found that in the end, using a 'discrete' VCO rather than a PLL was more to my liking and all the effort I put into trying to get a 4046 track well translated directly to a 'discrete' VCO attempt, so if you havent tried a 4046, you should.  If you've tried a 4046, you may as well try an all out VCO!

As for typology of the circuitry you hear in these clips, certainly I can share much of that: first is compression.
Here in the clips is an old dod milkbox (THAT chip based comp), but it has been modified to get rid of the high expansion.  That is a great feature of that compression, but not for this purpose.  So post-mod, its essential a boss cs3 with no tone control (or tone pot centered).  THAT chips are very good for compression....however, good sustain has been achieved by simple opto comps like the flatline, tweaking values and using the most responsive optocoupler i can find.  OVerall you want compression to help you get more apparent sustain from the VCO and cleaner 'tracking.'
But in regard to compression and every area of theguitar-->VCO concept, there is 'more than one way to skin this cat' and oftentimes it's simply a lot of balancing things out the way you (I) find least disapointing; compromises are the key to overall satisfactory results IMO, which means patience and tweaking.

From the compression, you are then hearing attempted 'fundamental' filtering from the typical static 3-pole filter, like ehx microsynth/oc-2, etc, with a bit of tweaked values for my tastes via spice simulation and breadboarding work.

Then, I suggest some gain control if/as needed, followed by comparatoring of the signal (i like real comparator chips here).

Then, the subject line LM29x7 chip:  The circuit out there from Terry Tsutsumi (sp?) that uses a 4046 and a 2907: the LM2907 in that schem is not used not as an p-v, but it could be, and those values worked well from the go IMO; from there one could also figure out what the components there are doomg and tweak them as you see fit, if you see fit.  However I think a JFET is a little less than what's needed to properly drive the LM29x7 chip for p-v as shown in Terry's schem.
 
From the LM29x7 chip, I'm using circuitry that tweaks off scott gravenhorsts LINEAR CV VCO which more or less is a simple tweak of rene schmitz simple CMOS VCO.

Why linear VCO?
I've learned that my guitar's pitch is already expo....so driving a linear p/v chip yields expo voltage response due to expo'd pitch at the input of the linear p-v...which makes this chip that much more usable.

So following, linear V->I control of a vco is what you want here. And since the linear V->I and most guitars are hence less sensitive to temp than an exponential V-I converter, tuning seems a lot easier IMO as compared to a linear resistor string synth into a expo VCO circuit....

Overall, nothing I'm using here is uncommon, unavailable, or unique, it just would need someone willing and interested in investing their time to do some googling and breadboarding and/or posting and/or e-mailing, trying, failing, trying, suceeding, refining values, tuning, etc.  That's 'all' I did. I'm no one special.  Just happened to be encouraged and determined enough to keep trying and working on it...and still am.

That said, these clips are weeks or months old, snapshots of the circuit in process...improvements are desired and needed IMO, even from what you hear. For example, with this arrangement, the vco i use 'ticks' with no input signal (proably putting out a 0.2hz signal or something at 'zero' volts from the p->v), but I have by now worked that out, more or less, with a lot of conversational help and leads to other circuit snippets for problem solving, offered to me from other people. I am also working on using the cv, some form of S&H, triggering, and cmos switching to ADD an element of pitch/cv based filtering, sort of post initial note detection which is yielded originally, again, from the static filter etc. This would add an OTA tracking filtering based on initial pitch detected, etc. Similar in concept to what thye ehx DOM does essentially. Thats in hope of squelching the octave hop that invariably occurs from monophonic 6 string guitars being filtered by static filtering. Who knows how much it will help vs hurt.  

Overall, the skys the limit and theres lots of 'refinements' one can add, too. Suboctave sawtooth can be had by mixing some perfectly tweaked ratio of a sawtooth with an octave down squarewave (see arp little brother, picsynth project, etc). Waveshaping is easy with a big swinging sawtooth wave, such as this yields (and what you are hearing).
But if you check out all the synth diy stuff, you can't miss the commons ways to get a sawtooth to triangle, square, pulse, sine and beyond once you have a sawtooth or triangle wave VCO.

Again, it just takes some major interest, persistence, and time! As the first clip demonstrates in the 'smearing' between notes and the manner in which the final note 'decays' (if you can call it that), too much compression, for example, will yield poor results...balance is key, and is where I've spend the most amount of my time with this circuit...finding balance...easy to over filter, over compress, over gain, over "design."  

Hopefully more people will dare try it - i didnt dare since it seemed everyone but the old korg roland etc units indicated it was futile. My own initial inspiration was the videos and pictures and info about of the analogue systems rs30/rs35 unit...never did find too much of how that circuit approaches thing even with good photos and some unanswered interrogation of the company!  

I'm here primarily to suggest it's not impossible, and to provide clips to further encourage others to try guitar control of VCO.  I'm CERTAIN someone else here is far more capable of creating something quick after reading this post that equals or betters my results in half the time...I'm just an amateur hack, I promse!

[liquids]

Question as it pertains to this circuit.

I have a bipolar supply.

My logic runs one sided off V+ and ground.

I use bipolar powered op amps into an actual comparator (LM311).

If I run the comparator off bipolar power, it would be easier for me...rather than adding a Vref...

With an LM311, if Pin 4 goes to V- and Pin 1 goes to ground, will the comparator's output still swing 0V to V+?

I guess what confuses me is that while I think that is true, the EHX DOM (http://www.freeinfosociety.com/media/images/4113.jpg) uses a LM311 to drive a 4013, but the LM311 is powered as single sided.  Don't want to blow any more chips up than I already do. 

[PRR]

http://www.ti.com/lit/ds/symlink/lm311.pdf pg 8