SI570 Temperature Stability

11 posts / 0 new
Last post
Last seen: 2 months 1 week ago
Joined: 2011/01/11 - 12:48
SI570 Temperature Stability
Plain text icon ui_soft_tcxo.txt1.55 KB
Image icon lo.jpg49.12 KB


Like many of you, i have been using Softrock type of direct conversion transceivers for Wspr. They have in common the fact that the mixers are clocked from SI570. It is excellent chip, but the problem that it can not be easily locked to external gps or rubidium source. So i have decided to try to code a simple software tcxo routine and use external temperature sensor, physically mounted on the SI chip and then use the measurements and a look up table to issue small frequency changes on small intervals.

Right now i am testing it on 20m:

- will get decodes as long as i don't do very fast changes to LO temperature(i use a hairdryer for that)
- even breathing close to the PCB causes changes in the .1C range
- lookup table precision is 1C, don't have good way to measure with better resolution
- the spectrum analyzer is not locked to external stable clock

Here a video of the process of cooling down from 60 to 30 degrees(do not watch with loud volume!)

The schematics and the lookup table are in the post attachments.

Maybe somebody have seen precise temperature chart of SI570, i couldn't see this in the pdf. Is there
a built in temperature sensor in this chip, that 'drags' the frequency while i try to stabilize via
software tcxo?

Any suggestions are welcome.

73, Chris