The resistor is needed because the crystal doesn't conduct DC. If a DC
path isn't present, then the inverter will not self-bias.
The inverter is weak, but this is all kind of relative. The channel
length is made long to minimize the current of the inverter since it
sits in the linear part of the transfer function. But the inverter is
still plenty strong.
If the resistor is internal, you can still jam the signal. It just means
power will be lost in the resistor.
On 3/5/2011 4:58 PM, Stan Searing wrote:
The on chip inverter is usually pretty weak and often designed to
have a high impedance output. When intended as a transconductance
amp, the output mostly looks like a current source (hence the
high impedance). Some oscillators use a resistor outside the IC in
series with the inverter output to help control maximum crystal
drive level, reduce chances of overtone oscillation, or in hopes
of improving the overall Q.
If the series resistor is present, there are three places you can
inject your signal: at the inverter output, between the resistor
and crystal and at the inverter input.
With a weak on-chip inverter, you can overdrive the inverter output
over a large frequency range. Phase noise could be an issue,
but was not a problem with my application when I've used this method.
Be careful not to exceed the voltage range of the chip.
If you stay close to the crystal resonate frequency, injecting
the pulling signal in between the series resistor and crystal
can be similar to injecting at the inverter input. But it's not
a good place to inject the signal as you get further from the
frequency the loaded crystal wants to sing at.
Injecting at the input of the inverter has worked OK for me.
The further you go off frequency, the larger signal you need.
If you want to drive a ways off, the load capacitor and the
crystal shunt capacitance in series with the other load capacitor
just become a capacitive load.
If the resistor is not present outside the IC, then the first
two cases become the same. You may be able to overdrive the
inverter output, but have to also consider the load of the
crystal and crystal load caps.
If you stay close enough in frequency that the crystal is singing
along with your source, than I'd guess the phase noise should be on the
order of the un-pulled oscillator or the pulling source (whichever is
worse).
(And in the cases where I pulled way off [20 %], the crystal went
pretty quiet and I didn't notice any phase noise [although I wasn't
really looking for phase noise].)
Good luck,
Stan
-----Original Message-----
From: time-nuts-boun...@febo.com [mailto:time-nuts-boun...@febo.com] On
Behalf Of Don Latham
Sent: Saturday, March 05, 2011 11:52 AM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] pulling oscillators
The crystal oscillators are reference oscillators in 2 meter ham radios. I
wanted to make the least perturbation in the radio, hence pulling instead of
simply introducing the correct reference. I will be using a reference of the
nominal frequency the radio should have, so really should not need to pull
more than a few ppm if that? also not woried about the phase noise, either.
Will derive the driving signal from my gps 10 mhz or from one of my FEI Rb
devices. I need the accurate frequencies, or at least I think I do, because
the 2 meter radio will drive a transverter to 2.4 GHz moonbounce.
Sorry about not knowing how far I need to pull, or the specs of what's
there. I suspect there has been some drift since the radio was made.
Putting in a vcxo or varactor simply puts off the problem, as then I have to
monitor<that> frequency and control it. I could cobble in a little tuning
cap, but still would be left with a pretty temp sensitive reference.
Hope the question is clearer, and thanks to all who replied!
Don
Adrian
Don,
if you have a reference oscillator of the right frequency, what's the
the purpose of trying to pull a CMOS oscillator to the same frequency,
rather than just using the reference frequency?
Please be more specific on whant your requirements are.
Adrian
Don Latham schrieb:
Hello all:
I've developed a need for pulling crystal oscillators built in to pll
circuits. These are cmos, and have the common style oscillator
circuit built in. The crystal is across an inverter in the chip, and
there is a small cap between each end of the crystal and ground.
The chips are pll's in radio transceivers, early at that.
I could carefully remove the crystals and caps, simply driving the
non-inverting input on the chip with the reference, but I would
rather simply tack on a very small cap and "pull" the crystal
oscillator with an external reference signal of the right frequency.
Anyone out there tried this?
Thanks
Don
_______________________________________________
time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
