Crystal Capacitor Formula (Parallel Mode)

[Costas Vlachos]

Hi all,

I'm looking for the formula to calculate the value of the two capacitors C1
and C2 used in crystal oscillators for microcontrollers (you know, the ones
you put between the Xtal terminals and GND). I've found this on the net:

Crystal Capacitor Formula (Parallel Mode)
-----------------------------------------
Let: CL = Crystal's load capacitance
CP = Circuit's parasitic capacitance
CI = Chip's input capacitance
C1 = Crystal oscillator capacitor 1
C2 = Crystal oscillator capacitor 2

Then: C1 = C2 = 2 * CL - (CP + CI)

Example: For a crystal with 12pF load capacitance (CL)
and assuming CP + CI = 5pF (typical), we have
C1 = C2 = 2 * 12pF - 5pF = 19pF. We can then
use two 18pF capacitors (closest standard
value).

Is this correct? Can someone please confirm?
Many thanks.

Costas

 

[Spehro Pefhany]

Hi all,

I'm looking for the formula to calculate the value of the two capacitors C1
and C2 used in crystal oscillators for microcontrollers (you know, the ones
you put between the Xtal terminals and GND). I've found this on the net:

Crystal Capacitor Formula (Parallel Mode)
-----------------------------------------
Let: CL = Crystal's load capacitance
CP = Circuit's parasitic capacitance
CI = Chip's input capacitance
C1 = Crystal oscillator capacitor 1
C2 = Crystal oscillator capacitor 2

Then: C1 = C2 = 2 * CL - (CP + CI)

Example: For a crystal with 12pF load capacitance (CL)
and assuming CP + CI = 5pF (typical), we have
C1 = C2 = 2 * 12pF - 5pF = 19pF. We can then
use two 18pF capacitors (closest standard
value).

Is this correct? Can someone please confirm?

Yes.

Best regards,
Spehro Pefhany

 

[Larry Hatch]

Hi all,

I'm looking for the formula to calculate the value of the two capacitors C1
and C2 used in crystal oscillators for microcontrollers (you know, the ones
you put between the Xtal terminals and GND). I've found this on the net:

Crystal Capacitor Formula (Parallel Mode)
-----------------------------------------
Let: CL = Crystal's load capacitance
CP = Circuit's parasitic capacitance
CI = Chip's input capacitance
C1 = Crystal oscillator capacitor 1
C2 = Crystal oscillator capacitor 2

Then: C1 = C2 = 2 * CL - (CP + CI)

Example: For a crystal with 12pF load capacitance (CL)
and assuming CP + CI = 5pF (typical), we have
C1 = C2 = 2 * 12pF - 5pF = 19pF. We can then
use two 18pF capacitors (closest standard
value).

Is this correct? Can someone please confirm?
Many thanks.

Costas

On a side note. There is a fellow on the internet (Ricardo Richi, or
something like that). He shows how to make pf caps with wirewrap. I
have done about 10 PIC and AVR projects, with a xtal and two little
twisted wires to run it. And no errors in the uP running ever. One is
on a motorcycle that does 150MPH wheelies and viberates a lot. Only
a cell phone too close have cause a problem, and I do not believe it is
in the wire caps, but the lack of caps and the long wire from the back
on the bike to the front.

Anyway, I though the wire twisting was interesting, now to have a good
meter to tune them. Any ideas on that? Good way to measure 33pf, etc caps.

 

[Michael A. Terrell]

Interesting! Never done anything like this before, although I've thought
about it. I have a Fluke 189 DMM which does capacitance metering. I'll try
it and see what I get. BTW, my DMM has a "relative" function which zeroes
the reading by subtracting the capacitance from the leads, etc. I think most
modern DMMs have this.

The trick to measuring low capacitance is to build a small insulated
test jig that plugs directly into your meter to minimize stray
capacitance, and to keep the value stable. Pomona, Keystone, and other
companies make banana plugs with either a threaded stud, or threaded
hole for a 6-32 screw. A small piece of PC board with a v shaped
section of the copper clad scored and peeled off gives you a good
surface to check surface mount capacitors, or leaded parts. You can
tack solder gimmicks or leaded caps to the board for critical
measurements, or use a cheap, self wiping IC socket cut in half to let
you plug leaded parts in for test. Just remember that the sockets wear
quickly, and you need to make sure you can replace them easily.

About the success of the wire caps, I wonder if this is due to the fact that
it would normally work even with NO caps in the circuit. I mean, in the
early days of learning electronics, I used to simply connect the Xtal on its
own (no caps) and it worked fine (still does after many years). So I'd say
check them out on a more critical circuit to see their effect/performance.
But it's a good idea and I will try it. I particularly like the fact that
you can theoretically tune the capacitance to exactly the value you want.

Many years ago while I was in technical high school we visited Hameg (a
company that makes oscilloscopes and other instruments). One of the
engineers showed us the inside of an analogue oscilloscope and there were
many little wires coming out of the PCB and just standing on one end, like
little antennae. He explained that during calibration they move them around
and change their shape to adjust certain things! Could those be similar to a
wire capacitor? Interesting stuff...

Costas

Yes, but that is a pretty cheap method for test equipment.