Re: phase locked loop SSTC

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "by way of Terry Fritz <twftesla-at-qwest-dot-net>" <free0076-at-flinders.edu.au>

Quoting Tesla list <tesla-at-pupman-dot-com>:

> Original poster: "Jan Florian Wagner by way of Terry Fritz
> <twftesla-at-qwest-dot-net>" <jwagner-at-cc.hut.fi>
> 
> Hi,
> 
> I'm trying to do a simple phase locked loop to keep the SSTC driver &
> sec always in tune, schematic is 
>  http://www.hut.fi/~jwagner/tesla/SSTC/PLL-schem.gif
> for the PLL and the mains side is the half bridge as in
>  http://www.hut.fi/~jwagner/tesla/SSTC/half-bridge.htm 
> and half-wave rectifier mains as input.
> 
> But, well, something seems to be wrong with it... Has anyone else been
> doing something like this? PLL coil?

I'm not at home so I don't have much time to chat, but yes I have been doing 
something like that =)
 
> If I got it right then in this circuit the frequency lock range ==
> frequency tracking range? At least, testing with a square wave input
> suggested this is the case... I was careful to make the upper freq vs
> lowest freq (R1 and R2) ratio 1.8 so it shouldn't lock on to TC sec
> harmonics (with a ratio >2 this could happen??).
> 
> The drive signal reacts just nicely to moving a hand towards the coil.
> But, there's a changing phase shift in the signal, from about 40
> towards 60 deg, so while the frequency is correct, the phase just

That's my problem too - you need to keep the phase shift just right.

> isn't. => mosfet switching losses go up and streamer output goes down.
> In addition to this, slowly increasing the bridge input voltage keeps
> everything in tune, but then most of the time if I try to draw an arc
> to a screwdriver it just makes *zip*, one short spark, and then it is
> totally out of tune and mains must be powered down and then up again.

Mine didn't need power cycling, but rather stayed locked onto the fundamental. 
It was just slightly off though, since the phase shift was something like 10 
degrees out.

> Just to get a second small and unimpressive *zip*.
> 
> On occaison it gives out some wimpy 5cm streamers, though. But that's
> about 1/6th of the output possible with manual tune... :o(
> 
> Any ideas whats going on here?
> 
> every help would be welcome!
> thanks, 
>  - Jan
> 
> PS: A small current sense transformer senses the current into/out of
> the primary coil (half-bridge). Specs: 1 pri turn, 100 sec turns, 820
> ohm resistor, coupled via 10nF to CD4046 input incl two back to back

Although I haven't looked at the schematic yet, that's the same chip I
used. It 
should lock fine from the signal provided by two back-to-back signal diodes, 
since it has some kind of amplifier built in. My problem was that my phase-
shift network was just too lossy to provide the required signal into the chip, 
since I was shifting by almost 90 degrees to get what I wanted. With more
turns 
on the current transformer, I'm sure it would have worked since the phase
shift 
was rock-solid with changing resonance frequency.

> 6.2V Zeners. Then I'm using phase comparator II in the CD4046, which
> should keep a 0deg phase shift between the sensed current and the
> output drive signal, over the entire frequency range. Loop filter: 1nF
> and 1.2kOhm

I forget which comparator is II, but if you didn't know, one is 0deg and the 
other is 90deg.

Have fun,
Darren Freeman