1
\$\begingroup\$

I can't figure out the purpose of the 0Ohm resistor in bellow schematic. It's crystal oscillator configuration and why the designer have included a resistor there with 0ohms ? enter image description here

Any idea ?

\$\endgroup\$
  • 13
    \$\begingroup\$ So that it can be replaced with something that isn't 0ohm if needed. \$\endgroup\$ – Ignacio Vazquez-Abrams Dec 19 '16 at 7:44
  • 1
    \$\begingroup\$ Or left out (but probably not in this case, because leaving the XTal and the C out would achieve the same) \$\endgroup\$ – Wouter van Ooijen Dec 19 '16 at 7:51
  • \$\begingroup\$ Not in this case maybe, but sometimes PCB designers cannot route a trace from one point to another even jumping to the other layer with vias (because of trace density), so they place a 0R resistor as a "jumper". Also I agree with @IgnacioVazquez-Abrams, but I'm not sure about a non-zero resistor's use in there. For trimming an XTAL's frequency maybe? We perform it via capacitors in production, though. \$\endgroup\$ – Rohat Kılıç Dec 19 '16 at 8:02
  • 1
    \$\begingroup\$ Perhaps the resistor would be omitted if you use an oscillator module rather than a bare crystal. \$\endgroup\$ – Peter Bennett Dec 19 '16 at 8:05
  • \$\begingroup\$ I also think that it could be fine tuning the XTAL frequency. May be for laser etching later, may be when they receive the network card for warranty. \$\endgroup\$ – Standard Sandun Dec 19 '16 at 8:09
8
\$\begingroup\$

So that a resistor can be added to prevent overdriving the crystal. You will probably see it included in the data sheet as an option.

\$\endgroup\$
  • \$\begingroup\$ what you mean by overdriving ? \$\endgroup\$ – Standard Sandun Dec 19 '16 at 8:12
  • 1
    \$\begingroup\$ Too high a signal. Crystal can be damaged, in some cases. \$\endgroup\$ – Leon Heller Dec 19 '16 at 11:08
  • 3
    \$\begingroup\$ This is the correct answer. It's why the resistor is on the XTAL_OUT (drive) side. Some crystals can handle mW others uW. \$\endgroup\$ – Spehro Pefhany Dec 19 '16 at 12:24
  • 1
    \$\begingroup\$ so it simply works as an Attunator/friction? And in brand new cards they put 0Ohm resistor and when it comes back as an warranty claim, they will add an resistor right ? So the card work longer. \$\endgroup\$ – Standard Sandun Dec 19 '16 at 12:35
  • 2
    \$\begingroup\$ It offers flexibility at the design stage. \$\endgroup\$ – Leon Heller Dec 19 '16 at 12:41
12
\$\begingroup\$

CMOS crystal oscillators may require such a resistor, because the amplifier that drives the crystal (a biased CMOS inverter) has low output impedance when it is either HIGH or LOW. That low output impedance is a short circuit across (effectively) the load capacitor connected to the crystal at that output terminal. That has the effect of damping the oscillation.

Mainly, modern processes allow weakly conducting CMOS inverters for this function, and most application sheets do not call for a resistor at the driven node of the CMOS gate. If an unusual impedance of crystal or a different frequency range is used, either the 33 pF capacitors or the zero-ohm resistor may prove unsuitable, so it may be prudent to allow for changing those values.

It would also allow disconnecting the onboard crystal for diagnosis (temporary) or customization (permanent) purposes.

\$\endgroup\$
10
\$\begingroup\$

The chip has internally an unbuffered inverter therefore, for the circuit to oscillate, the external circuits (including the output impedance of the inverter working with the capacitor connected to that terminal), have to produce another 180 degrees. If this extra 180 degress cannot be produced, the circuit will not oscillate.

The crystal (together with the capacitor at XTAL_IN) doesn't quite produce 180 degrees phase shift so, a little bit of help is needed to actually get to 180 degrees and that's where the internal output impedance of the inverter AND the capacitor on XTAL_OUT come in.

If that internal output impedance isn't high enough, a little bit extra on the outside (10 ohms or so) usually tips the balance and it oscillates. This resistor also governs the peak current that can pass through the crystal so if you are using a particularly sensitive crystal, damage won't occur.

But, the primary role is to provide a few degrees more phase shift. Below is a simulation of a crystal oscillator that doesn't use the capacitor at XTAL_OUT (same effect as not having sufficient output impedance): -

enter image description here

You should be able to see that Vout never quite attains a phase shift of 180 degrees and, unfortunately, this circuit won't oscillate.

Picture taken from my answer here.

And finally, if at all in doubt, do what the data sheet recommends: -

enter image description here

\$\endgroup\$
  • \$\begingroup\$ If a 10 ohms or so resistor is needed, why not specify it as such? \$\endgroup\$ – Dmitry Grigoryev Dec 19 '16 at 14:34
  • 1
    \$\begingroup\$ @DmitryGrigoryev what does it say about it in data sheet that might be relevant on this question? \$\endgroup\$ – Andy aka Dec 19 '16 at 16:11
0
\$\begingroup\$

It's much easier to change the value of a component than to add a component that there was no provision for in the design.

IF the problem is caught between fabrication and assembly then changing a component value is a trivial matter of changing the BOM. If it's caught after assembly then some rework is needed but it's fairly easy rework.

OTOH adding a component where there was no provision before can mean much more difficult rework or even scrapping the boards and starting over.

So if you think you don't need a resistor but you are not 100% sure on that you put the 0 ohm resistor in the design.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.