No internal oscillator in xcore-200 chips

Technical discussions around xCORE processors (e.g. General Purpose (L/G), xCORE-USB, xCORE-Analog, xCORE-XA).
babazaroni
Experienced Member
Posts: 94
Joined: Sun Feb 10, 2013 4:47 am

No internal oscillator in xcore-200 chips

Post by babazaroni »

Too bad there is no internal oscillator in the new chips. We use an xcore-u series chip and could boot up on the internal oscillator, then program an external clock chip (5351a) to our correct cpu frequency, then switch to the external clock source.

The 5351A can be programmed at the factory for extra cost, but one also risks a board manufacturer accidentally building a lot of units with unprogrammed parts.

Anyone know of a programmable clock chip with at least 2 outputs that also has a default output and about the same low cost as the 5351A ($1 in quantity 2000)?


User avatar
mon2
XCore Legend
Posts: 1910
Joined: Thu Jun 10, 2010 11:43 am

Post by mon2 »

Hi babazaroni. Some comments raised from our pending XMOS design work to date which will feature an external PLL:

a) believe there is a concern about the clock switch over to be seamless from the one oscillator to another and should be glitch-free otherwise you could halt the code execution (?) - just an educated thought.

b) due to the above concern, we are considering to use the following device although the cost is higher but the device contains an internal reference crystal value of your choice to ensure that there is a default factory clock "out of the box" when soldered. Then you can use the I2C interface to tweak to your desired frequency.

P/N SiLabs Si570

http://www.silabs.com/Support%20Documen ... /si570.pdf

c) another possible choice that may work and at a lower cost:

http://datasheets.maximintegrated.com/en/ds/DS1085.pdf

Not sure if the precision of this clock is suitable for your needs so please review in detail.

Again, we are concerned about glitch-free switch over to prevent code lock ups. Not sure if that is a real issue with these devices. One idea is to write a watchdog routine and use an external reset supervisor. However, there is still the requirement for an external working clock to recover from a possible lock up state. We have been bashing our heads on these topics for a while and even have provisions to use alternate methods for in the field programming / recovery of the flash in case the flash is corrupted - not likely to happen but nice to have considering we do not wish to pay for shipping the product back to our side.
babazaroni
Experienced Member
Posts: 94
Joined: Sun Feb 10, 2013 4:47 am

Post by babazaroni »

Thanks for the chip suggestions.