calculating load capacitance of crystal

Thread Starter

jgershonw1

Joined Jan 31, 2008
2
I need help determining the load capacitance for a crystal used in the circuit shown below. There was no parts list for the circuit, so I am not sure what I should be using. I found this page that explains how to calculate the crystal capacitor value.

The equation is C=2(CL)-(CP+CI)

C = crystal capacitor value
CL = load capacitance
CP = parasitic capacitance (wires, socket, traces)
CI = input capacitance (mcu itself)

So rearranging the equation, to calculate CL:
CL = (C + (CP+CI)) / 2

I am estimating that (CP+CI) is roughly 5 pf.

CL = (33pf + (5pf)) / 2 = 19pf

or is it:

CL = (66pf + (5pf)) / 2 = 36pf

Did I do this correctly or did I miss something completely?


 

scubasteve_911

Joined Dec 27, 2007
1,203
Sorry, this isn't exactly a technical answer, but, I believe that you usually choose the capacitor values the same as the load capacitance of the crystal minus the input capacitance of the microcontroller and the parasitics.

Steve
 

SgtWookie

Joined Jul 17, 2007
22,230
Something that isn't obvious nor apparent is that quartz crystals were once "living" rock, and that crystals can vary significantly in performance; even within a given lot.

One of your caps should be adjustable. I suggest a range of approximately 25-75pF to be able to "tweak" things in. You can always use a few caps in parallel to get the same capacitance as how you've tuned.

Even though you may have a huge pile of crystals that are rated for the same frequency, it's highly unlikely that any two of them will have identical responses.
 

r00zbeh

Joined Nov 6, 2008
1
You've probably solved this by now, but I think what you're supposed to do is get CL from the data sheet of the crystal, then find the capacitance you need. For example if CL was 20pF in the datasheet, and your parasitic+input was 5pF (per pin) as you mentioned, then the capacitance you need is

20pF = (C + 5pF)/2, or C = 35pF per pin.

The point is CL should be known from the datasheet of the crystal you're using, it's what the crystal "wants" to see, and C is what you're trying to find...
 

KL7AJ

Joined Nov 4, 2008
2,229
Something that isn't obvious nor apparent is that quartz crystals were once "living" rock, and that crystals can vary significantly in performance; even within a given lot.

One of your caps should be adjustable. I suggest a range of approximately 25-75pF to be able to "tweak" things in. You can always use a few caps in parallel to get the same capacitance as how you've tuned.

Even though you may have a huge pile of crystals that are rated for the same frequency, it's highly unlikely that any two of them will have identical responses.
HE HE! How right you are! At the risk of revealing how old I am....I used to re-grind my own amateur radio crystals....in the old FT-243 holders. Definitely a lost art.

I was just recently doing some SPICE modeling of a crystal....and the source driving impedance has a HUGE effect on whether the crystal is going to operate in the series resonant or parallel resonant mode. This is a major overlooked factor in crystal circuit design....perhaps of even greater importance than crystal loading...at least as far as overall Q is concerned.

eric
 

mik3

Joined Feb 4, 2008
4,843
Some crystals require a resistor to be connected between one end of the the corresponding microcontroller pin. When you will choose one to buy check its datasheet and it will tell you what if you need a resistor and what values of capacitors to use for the best performance.
 
Top