Arduino Forum

Using Arduino => General Electronics => Topic started by: npulver on Jul 18, 2013, 07:54 pm

Title: Crystal behaving badly
Post by: npulver on Jul 18, 2013, 07:54 pm
I've pieced together a 24-bit analog to digital conversion circuit and I'm having some trouble making it work.  I've got a crystal oscillator (P/N: ECS-49-18-5PXEN-TR) connected to the MCLKOUT and MCLKIN pins of the AD7730R Analog to Digital Converter.  The problem I'm having is that I'm getting 4.9152 MHz on the MCLKOUT pin, but nothing on the MCLKIN pin.  The crystal is not polarized and I assume that I should be getting excitation on one side of the crystal and a steady MHz output on the other.  I've attached a snapshot of the circuit with the section in question highlighted.  Can anyone see any reason why the crystal is not working as expected?
Title: Re: Crystal behaving badly
Post by: KeithRB on Jul 18, 2013, 08:20 pm
How are you measuring it? The MCLK in might be high impedance and easily disturbed. If you are getting 5 MHZ at the MClock out you are golden.
Title: Re: Crystal behaving badly
Post by: npulver on Jul 18, 2013, 08:29 pm
I'm just sticking an O-scope probe on the pin.  4.9152MHz on MCLKOUT, 0 Hz on MCLKIN.
Title: Re: Crystal behaving badly
Post by: pwillard on Jul 18, 2013, 10:36 pm
Do you have 12pF caps at each of the XTAL pins to GND?  Datasheet - Page 32 says they may be optional...  but in reality, it depends on your specific crystal.  If I read it right, based my interpretation of the datasheet, you will ONLY see the clock frequency on MCLKOUT pin... even with a crystal.

Personally, I would use TCXO and feed ttl level clock to pin MCLKIN

Title: Re: Crystal behaving badly
Post by: mauried on Jul 19, 2013, 02:03 am
The oscillator is a parallel mode type, and the impedance on the input side is very high.
Its likley that the CRO probe is causing the oscillator to stop when you probe the input side.
You should also have the parallel caps installed .
Title: Re: Crystal behaving badly
Post by: npulver on Jul 19, 2013, 01:04 pm
I do have caps attached between each leg and ground.  I spoke with a rep from Mouser and they said the crystal is working the way it is supposed to.  Looks like I have an issue elsewhere.


Thanks for the time, everyone.
Title: Re: Crystal behaving badly
Post by: AmbiLobe on Jul 19, 2013, 01:14 pm
The crystal should be soldered to the circuit. If it is stuck in a solderless breadboard, its wires may be too small to make a solid connection. If it is wire wrapped, the crystal leads may not be connected with a low enough resistance. Solder the crystal to the other components and solder the crystal assembly to the integrated circuit.
Title: Re: Crystal behaving badly
Post by: npulver on Jul 19, 2013, 01:16 pm
The crystal is soldered to a circuit board.
Title: Re: Crystal behaving badly
Post by: MarkT on Jul 19, 2013, 05:31 pm

If it is wire wrapped, the crystal leads may not be connected with a low enough resistance.

Proper wire wrap is _superior_ to soldering since there is a gas-tight weld directly from copper wire
to gold-plated pin.
Title: Re: Crystal behaving badly
Post by: TomGeorge on Jul 20, 2013, 02:44 pm
Your circuit around the crystal is behaving correctly. The pin that you do not see a signal is being supplied by the crystal this a high impedance source, the CRO will load it down. If you are getting the correct frequency on the other (output) pin then all is well, this is a relatively low impedance output and is not loaded down.
To setup and make any A-D converter chip work requires a lot of setting up, I hope you have a suggested circuit from the IC applications data to work from.
Tom