Overshoot and underhoot

Thread Starter

Annu007

Joined Jun 12, 2006
11
Hi all,
I am using a 80 MHz crystal oscillator. What I am seeing a 3V negative undershoot and a 2V positive overshoot. I am using 5V Vcc. Also I am not seeing a clean square wave.

Can somebody help me.

Thanks
 

beenthere

Joined Apr 20, 2004
15,819
Hi,

Do you have the crystal bypassed to ground? Usually a capacitor on either side of the crystal helps. They run in the range of 6 - 12 pF.
 

Thread Starter

Annu007

Joined Jun 12, 2006
11
Usually a capacitor on either side of the crystal helps...
What do you mean by this?

Pin 1- Enable/disable
Pin 4- Ground
Pin 5-Output
Pin 8- Vcc
 

beenthere

Joined Apr 20, 2004
15,819
That's a TTL oscillator you're using. Or perhaps CMOS. The junk you see may be an artifact. The bypass capacitors work for crystals.
 

Papabravo

Joined Feb 24, 2006
21,159
Hi all,
I am using a 80 MHz crystal oscillator. What I am seeing a 3V negative undershoot and a 2V positive overshoot. I am using 5V Vcc. Also I am not seeing a clean square wave.

Can somebody help me.

Thanks
At those speeds without very careful layout and impedance control your expectations for a "clean square wave" are quite simply unreasonable.
 

Ron H

Joined Apr 14, 2005
7,063
The ground lead on a scope probe can easily induce transients of this magnitude on signals with fast risetimes. Some probes have optional tips with extremely short grounds.
 

Papabravo

Joined Feb 24, 2006
21,159
Ron makes an excellent point. It is often the case that observation for the purpose of measurement significantly changes the system under observation. As an exercise you might want to compute the reactance of the trace inductance at that frequency. I think you'll be surprised.
 

Ron H

Joined Apr 14, 2005
7,063
Ron makes an excellent point. It is often the case that observation for the purpose of measurement significantly changes the system under observation. As an exercise you might want to compute the reactance of the trace inductance at that frequency. I think you'll be surprised.
Even with a perfect (zero impedance) voltage source, this will happen. The current into the probe's capacitance must return to the probe tip through the ground lead of the probe. If this is a 6 inch wire, the inductance will probably be over 100nH - possibly over 200nH. The ringing is due to the ground wire. It may be exacerbated by the inductance of the trace being probed.
 

Ron H

Joined Apr 14, 2005
7,063
Are you using a "Times 10" probe? If you are, have you tried adjusting the capacitor on the probe?
Good point, except at 80MHz, the probe compensation cap will probably cause the amplitude of the entire waveform to change - not just the edges. It would be worth checking out.
 
Top