LM567 as a VCO?

[anotherjim]

I've a feeling this may not be practical with this chip.

I might expect pin 2 is the only access to the VCO control. It's where the loop filter cap normally goes, so if it were a 4046, would connect to the VCO control voltage input (pin9).
Problem is, I can't get a CV to its pin2 to change the VCO frequency by much. Some, but not much. Possibly, the chips internals are fighting my external voltage? I'm using a 50k pot for external CV & it may be I need a lower impedance source to overcome the internals, but I'm wary of damaging it.

Now, the chip I have has been in my stash for years and years. May even have been an Archer one from Tandy.
So, I could have a bad chip, however, I can get the basic oscillator functions to run. I'm feeding it 8.3v because the absolute max is 9v for this chip.

Can anyone say if it can be done? Before I order some others to try.

[diffeq]

If I undrstand internals of LM567 correctly, there isn't much you can do to increase pin 2 sensitivity. It taps into load of Q34-Q35 diff. pair and input of Q61. Maybe using a potentiometer with less than 10k of resistance would increase the swing (but maybe not).

There is another way: using LDR as timing resistor between the capacitor and pin 5, leaving pin 2 voltage trimmer for fine tuning.

[anotherjim]

It taps into load of Q34-Q35 diff. pair and input of Q61.

Thanks Tony.
I looked at the internal diagram, but it isn't easy to work out the loading - but it must be possible for the loop cap to  control the VCO without getting dragged off. Then I can't find any design out there using it as a VCO.
As you say, an LDR timing resistor is probably the best thing to do, although the data sheet says only 10:1 adjustment that way.

I may yet try brute force on pin2.

[anotherjim]

The CMOS version may be more promising. Also a much more informative data sheet for that one.
You can see how the phase detector output is connected. I think it might be easier to overcome than the BJT version. The data says 1/8th capacitance of the BJT loop filter which suggests a higher source impedance from the CMOS detector. Also, it might help to tie the signal input pin3 one way or the other

I will have to get some CMOS ones.

[Kipper4]

Is it  a hard to find part Jim?
Is it an expensive chip?

Can't wait to see how it goes. Following.

[R.G.]

The 560 series were good for their intended uses, but suffer from limited range on the VCO. They're fine if your application can live with about a 10:1 VCO range. Audio is about 1000:1.

[anotherjim]

If we start by admitting that pitch tracking audio is very difficult, then we might say to hell with it. 
Make something that tracks so badly it readily jumps around harmonics instead
Restricted VCO range may be helpful in this case.

10:1, if that can mean (say) 80Hz to 800Hz, can cover most of the notes we play?

The LM566 is better for this, but obsolete. You can find them. Banzai have stock. I have a few, but I may not spend time working out stuff for obsolete - better saved for existing designs that need it. Penfold "Metal" effect is a good start with room for expansion if you want to play with the 566.

XOR phase comparators can readily jump harmonics, but you will need a VCA/Mute output, because the VCO will not stop with the XOR but fall back to the same "centre" frequency in a pll application.

LMC567 isn't obsolete, but not stocked by many hobby suppliers. RS Components have them.
LM567 is commonly stocked. Neither 567 type is expensive.

Avoiding the 4046 because I suspect the signal input path in the chip is jittery with slow input signals. When it's tracking glitches, it can be very harsh. I did some experiments with a 556 as a tracking synth many years ago and I seem to remember it being a lot more musical despite been a good deal less accurate than the 4046.

[R.G.]

In PLLs, what really matters are:
1) how clean the input signal is
2) what phase detector you use
3) what the characteristics of the loop filter are.

The PD on the 560 series is something simple, like and analog multiplier IIRC. That's great for low-swing range PLLs. The CD4046 has only digital phase detectors, and I suspect that this forces the input cleanup to really, really matter so you have real edges, not slopes.

If you wanted to use the 4046's wide range VCO, you could probably pair it up with the PD from the 560 series to get an analog multiplier PD and wider range, using only the 4046 VCO. There is probably some tinkering of voltage levels and such to be done, but a VCO is a VCO once you have satisfied its input voltage and filtering requirements.

[anotherjim]

After further tinkering, I conclude the LM567 cannot work usefully as a VCO. Actually, it is not a PLL as such, since it's meant to sit on its centre frequency and switch an output when an input signal comes close, so perhaps not surprising.

I was reading about a modular synth patch using the VCO with a ring modulator (phase detector) and slew (loop filter) to simulate a PLL. Not something commonly attempted. Doepfer's PLL module is a 74HC4046!

A 555 may make a good VCO (small & simple) for a home grown PLL system. Only a few more parts than the 567 would have needed.