View Single Post
Old 09-26-2010, 03:58 AM
Aaron's Avatar
Aaron Aaron is offline
Co-Founder & Moderator
Join Date: Feb 2007
Location: Washington State
Posts: 10,845
crystal oscillator accuracy

Thanks SeaMonkey.


Now the 120MHz and 720MHz frequencies seem to be the two
input frequencies that may do the trick and from what Mike
indicates in his posts other input/companion frequencies can work too.

The 120 and 720 seem to just be a result of starting with a 20MHz
exact crystal.

There are definitely 120MHz crystals available and seem to be towards
the upper end of the easily available and lower cost ones.

Q1) If the 120MHz has more deviation from a lower one like a 20MHZ, lets
say a couple thousand cycles per second for example
and I've seen those +/- tolerances referenced for them, wouldn't the
secondary frequency still be proportionate to it's own doubling and tripling
compared to the 120MHz frequency?

Q2) And therefore, the other 2 frequencies (difference of and sum of) would
also still automatically be in the right proportion no matter what the
input crystal is?

Q3) So basically, if the 120MHz crystal deviates, the secondary frequency
through doubling and tripling and the other 2 "heterodyned" frequencies
would all still follow proportionately to the instability of the 120MHz crystal,
would they?

Q4) If so, then it seems it wouldn't matter the stability of a 120MHz crystal
and we could cut out one entire doubling circuit and one entire tripling
circuit and just start with 120MHz.

At least for now, I don't see a downfall of that because I'm pretty sure
the water molecules have no appreciation for "round numbers" as nature
doesn't work on them so they should split according to Mike's diagram
even if the 120MHz crystal does have more deviation than a 20MHz

IF, there is a benefit to having more stability such as some kind of
entrainment or something in the water bath or water mist cloud that
is more effective at a very, very specific stable frequency - 4 mixed
frequencies from the heterodyning - then I can see a benefit of using
a 20MHz with a doubler and tripler to get 120MHz.

I personally don't want to build any extra circuitry I don't have to
and am looking for any way that can simplify the entire circuit. Not in
concept, just in electronics and using a 120MHz crystal at the front
would eliminate a doubler and tripler on the input path.

Mike is an incredible engineer so I can see why he may want to just
make the circuit have a very specific or predictable frequency to work
with - it is good design protocol or whatever you want to call it.

But if a certain amount of deviation is permissible and won't effect
the results, then I'm all for the easier method.

I'd love to hear any feed back on the above concepts from anyone that
has direct hands on experience with crystal oscillators.

My only experience with crystals was when I was a teen playing with
remote control cars and had different crystals for different channels on
the remote control and the receiver in case someone I was racing had
the same frequency. But of course I didn't have to have any idea how
they worked to use them.

I have some questions about the oscillator circuits themselves but I'll
wait for details on the above. If it is crucial
to have a precision frequency on the crystal, then no big deal... just
go with the 20MHz and put it through a doubler and tripler. But I'd at
least like to know.
Aaron Murakami

Reply With Quote