Wednesday, March 7, 2012

A New Master Clock - Cont'd 3

To establish that the vendor supplied Crystal and Propeller Internal Oscillator is being effected by the temperature of the processor and it's work load activity. I have built a Homebrew external oscillator using the vendor supplied 5MHz Crystal.
5MHz Homebrew Oscillator

Because my intentions is to use it as an external oscillator if it proves to be effective at eliminating the Beacon drift, I have made it as small as my abilities (and keeping with my overall goal of building very small projects).

Oscillator Schematic via CircuitLab
at http://goo.gl/9ly8P

The selected circuit is much like the 9volt battery oscillator that I use for Marker Oscillators (see many of my previous posts), except even smaller.

Oscillator on Six Pin Header
The new oscillator was built "Ugly Style" without a circuit board on a six pin header. The header provides a place to plug the crystal, and mount the few 0805 components and the SOT 2n3904. Wires attached to the pins will be connected to the Prop, it will hover above the circuit board, mounted a little distance from the processor (or at least that the plan).

Initial tests indicate my HB oscillator has little or no "hash", which is much less than that of the attempted use of the Master Clock (see previous post).

Mast Clock signal on left (with it's Hash)
vs
My Homebrew Oscillator on the Right
The signal displayed on the left was from the Master Clock Oscillator, the signal on the right was the same data, but sent with my HB oscillator. Note: there is a little start up drift with both, and both stabilize after about 4 minutes.

Currently my oscillator is "hanging out" in the air on wire legs, and subjected to air currents and local temperature change that effect frequency (as with most crystals).

I inserted a 60 second delay between test transmissions, similar to a normal Beacon schedules.

Stability After an Hour
After an hour or so, the HB oscillator is stable and does not show signs of short term (2 minutes) stability between Beacon transmission, as seen before.

For this test the room temperature is 72.6F, the HB oscillator is 78.2F, and the Propeller Processor is 81.6F.

Future tests with WSPR will verify short term stability.



Conclusion

I think this proves that the short term stability of the Propeller Beacon during transmit; is due to heating of the internal oscillator from within the Propeller chip.

A low noise external oscillator should provide a stable RF Frequency for use with low band-width Beacons.

I need to find an commercially available Temperature Compensated Oscillator (TCXO). Or, a cheap Oven-Controlled Oscillator (OCXO). Or, package my Homebrew Oscillator for my permanent use with the Propeller.

--

No comments:

Post a Comment