AW: more frequency tracking..

[Magnus Danielson]

>>>>> "HJ" == Haible Juergen <Juergen.Haible at nbgm.siemens.de> writes:

Hi all!

 >> The idea of using PLL to have a VCO track along isn't new, the Moog
 HJ> 921
 >> oscillators have a PLL mode.

 HJ> Great surprise for me. Guess I have to look at these schemos again ...

There is this cursuit sitting close to an input and a mode switch
which takes one as the generator waveforms as input (this is from
memory, I am 630 km from my schematics). You will find it in the
middle left on the 921A schematic and it appears on the 921 schematic
as well. I haven't bothered to actually look the chip up but I think
it is a 4-quadrant multiplier or something, thus a classic phase detector.
The 921A clones that I have clearly says PLL and I even recall that an
add says PLL as well. It is a simple curcuit to add to any VCO, so
there is no reason to leave it out.

 >> A first degree is pretty useless for most applications, it has no
 HJ> loop
 >> filter and will make abrupt jumps and has no memory what so
 >> ever. Whenever there is a phase difference it will cause a jump on
 HJ> the
 >> output, not very neat.

 HJ> I have to disagree with "pretty useless", because I worked with
 HJ> the first degree PLL in my diploma thesis (I hope this is the right
 HJ> word), and this PLL made a hell of a good demodulator for
 HJ> FM radio.

I know that PLLs are used as FM demodulators, but I don't strictly recall
their grade. But if you look at what PLLs mostly seems to be used for,
namely frequency and phase locking of a regenerated signal as compared
to a reference signal I think that a first degree is pretty useless.
Saying this will of course not mask out the possibility of them being
very usefull in other fields, such as FM demodulation. I was a bit
wreckless with my language, so sorry...

 HJ> BTW, it can be shown that the schmitt trigger / integrator
 HJ> triangle / square oscillator that is widely used in LFOs will
 HJ> act as 1st order zero crossing PLL when you add one more resistor
 HJ> to the integrator input summing node and use this to feed in your
 HJ> signal. The ratio between the two resistors will then set the
 HJ> hold range of the PLL. I mention this, because I remember seeing
 HJ> a similar circuit in Buchla's famous noise and random module,
 HJ> but I may be wrong here. (Have to look again.)

This is interesting. I will try to meditate a bit over it, but I think
I can see how the schmitt-trigger can become a phase comparator.

The PLL world really consists of many many tricks... in my GPS book
there is things like FLL and DLL as well. Same but diffrent.

 HJ> As for no memory: If there is a pole, there must be some memory,
 HJ> and while there is no capacitor in the (nonexisting) loop filter, there
 HJ> is surely one in the VCO.

Sorry folks, but Juergen and I have started to throw poles at each
other, so keep you toes out!

The pole in the VCO will remember the phase position, not the
advancement of the phase position. What I also mean is that when the
phase detector does detect a difference it will make a direct jump
instead of a smoth transitition from the memory position.

When I say memory I really refer to a memory of the frequency control
voltage. Again I am not being explicit enougth in my lingo, so sorry.

 HJ> So I couldn't resist to comment here. Great post, nevertheless!

Thanks!

Cheers,
Magnus