-
Categories
-
Platforms
-
Content
There are two factors, rise/fall times and propagation delay.
CMOS is basically a current source output, so the output risetime is dependent on the load capacitance. HC has an asymmetrical output current capability, so that affects the max freq. At low frequencies the output is a trapezoid. You could say the highest freq is when either the top or bottom of the trapezoid collapses to a point.
Another factor is propagation delay. All CMOS oscillators are Schmitt Trigger designs, so the 1 or 2 gate delays create a hard limit on the max freq.
I've done 4.096 MHz, but that was with AC parts. The AC product line does not have nearly as many parts as HC, and they run on 3.3 V to 5 V only, but they are outstanding performers. Love the 20 mA output stage.
ak
CMOS is basically a current source output, so the output risetime is dependent on the load capacitance. HC has an asymmetrical output current capability, so that affects the max freq. At low frequencies the output is a trapezoid. You could say the highest freq is when either the top or bottom of the trapezoid collapses to a point.
Another factor is propagation delay. All CMOS oscillators are Schmitt Trigger designs, so the 1 or 2 gate delays create a hard limit on the max freq.
I've done 4.096 MHz, but that was with AC parts. The AC product line does not have nearly as many parts as HC, and they run on 3.3 V to 5 V only, but they are outstanding performers. Love the 20 mA output stage.
ak
Here's one for you AK:
Sorry for the poor focus.
A simple RC oscillator using a single Schmitt trigger inverter.
R is the output impedance of a gate and C is the input capacitance of 2 gates.
I shorted pins 1, 2, and 3 of a 74HC14 and measured the output on pin 4.
As my probe has a capacitance of 2pF, it may be better to eliminate the second gate.
The answer to my question is currently 84.5MHz and my guess is the propagation delay is no more than 6ns.
Sorry for the poor focus.
A simple RC oscillator using a single Schmitt trigger inverter.
R is the output impedance of a gate and C is the input capacitance of 2 gates.
I shorted pins 1, 2, and 3 of a 74HC14 and measured the output on pin 4.
As my probe has a capacitance of 2pF, it may be better to eliminate the second gate.
The answer to my question is currently 84.5MHz and my guess is the propagation delay is no more than 6ns.
I thought I might be able to get a slightly higher frequency using a loop of 3 inverters.
And I can!
But it's pretty nasty.
Actually, you wouldn't believe how nasty it is. The spectrum analyser shows a really nasty smear up to 1GHz or so whereas the 74HC14 has the sort of spectrum you'd expect for a square wave, with nice clean distinct peaks.
Oh and I'm using an 800MHz active probe for all of this, with a DC blocking adapter for the spectrum analyser.
And I can!
But it's pretty nasty.
Actually, you wouldn't believe how nasty it is. The spectrum analyser shows a really nasty smear up to 1GHz or so whereas the 74HC14 has the sort of spectrum you'd expect for a square wave, with nice clean distinct peaks.
Oh and I'm using an 800MHz active probe for all of this, with a DC blocking adapter for the spectrum analyser.
Reliably, maybe 3 to 5 MHz. Selecting parts you might get 50 MHz.
At what voltage is another question.
.
Yeah, but that oscillator was producing a very weird signal. Look at the one preceding it. That's 1.66V p-p.
They're not practical oscillators, but they're fun.
