Regarding a fast huff puff frequency stabilizer
I'm trying to get my head around how a fast huff & puff frequency stabilizer works. I'm familiar with huff & puff designs or at least I thought I was until I started studying them more closely. Reference: http://www.hanssummers.com/huffpuff/library.html
The fast huff puff uses a shift register delay while the original huff puff design did not.
But while trying to design a fast huff puff for VHF frequncies it seems like some things don't add up.
One article says:
"The formula for step size is:
Step = 10^6 x VFO^2 / ( z x M x xtal)
where VFO is the VFO frequency in MHz,
z is the number of stages of delay,
xtal is the crystal reference frequency in MHz, and
M = 2^n where n is the number of divide-by-2 stages in the VFO divider."
However, other articles say that to maintain a constant step size as the VFO frequency is tuned that a person needs to swap the D and the Clock inputs into the D flip flop used as a digital mixer. That would be swapping the xtal based input with the VFO based input.
That is what is hurting my head because the above formula will never allow the step size to remain constant regardless of which feeds which input since both frequency sources are part of the formula and only one of them changes. The xtal base frequency never changes and the VFO frequency always changes so the step size must change regardless of swapping the 2 variables in the formula.
It seems like if the xtal based frequency clocks the D-flip flop and the shift register instead of the VFO based frequency then the D flip flop and shift register output values would change at the constant xtal based rate and it seems like maybe that would make the step size constant but the formula doesn't allow that to happen.
Is this formula correct? Is there a better formula some place?
Maybe the persons saying the step size would be constant were just wrong in their thinking?
Cannot mention if the formula is fine or not (should be), but what I can say is the huff & puff VFO stabilizers have poor phase noise performance and a lot of spurs (hard to remove) at the output.
In VHF these spurs are even higher, and compared to a standard PLL (where the spurs are attenuated by the loop filter) huff & puff needs more filtering.
At low frequencies is better to spend some time stabilizing a standard VFO, which definitely gives cleaner output.
I'm confused now because others were saying a FLL circuit similar to this has less phase noise than a PLL circuit and that is why I became interested in it. But these HUFF and Puff circuits only correct for low frequency drift and other techniques are needed to reduce higher frequency phase noise. I have a very low phase noise VFO design already and only need this low frequency drift correction.
http://www.mail-archive.com/soft_rad.../msg01121.html
http://www.sps.ele.tue.nl/members/m....pc/ayranci.pdf
huff puff...man that worked great back with vacuum tubes and cat whisker diodes. But, since the earth cooled, most engineers only use PLL's to control frequency. In very rare cases maybe a FLL with very high q frequency discriminators.
The PLL can do most anything u want it to. If you want it to control a long term drift and not effect phase noise, you would lock your oscillator to a very stable crystal oscillator, but use a very narrow control loop bandwidth--perhaps 1 Hz or 10 Hz open loop bandwidth. That way by 100 hz offset from the carrier, the PLL would have almost no effect on phase noise.
If you need to clean up an oscillator's phase noise as well as control frequency drift, you use a very low phase noise crystal or saw reference oscillator, and lock in a bigger loop bandwidth, typically 50 KHz or so.
In some cases, they do still use counter based PLLs, like locking a crystal to a gps beat, but they are still PLL's, just very low loop bandwidth and all digital. Sometimes these are tricky to implement due to the transport lag...makes the loop tend to be unstable.