What do you do that fries the PLL out of a Propeller?!?!?

Microcontrolled

I've been building a large remote for the Stingray I won in the Schmartboard contest (they won't ship it to me until March, so this is something to pass the time) when, after installing a power grid, the device that previously worked perfectly, stopped. It shows up in a scan (e.g. "A propeller has been detected on COMX") but no serial communication takes place. After a day of troubleshooting I finally tested it with a serial com program, to find it didn't work. However, when another propeller was put in it's place, it did. So I guess the PLL was fried, but my question is, how? I've done this to more then one and I would like to avoid doing it in the future.

Mike Green

The Propeller has several power supply pins (Vdd / Vss). They all have to be connected and you should have a bypass capacitor across the pairs of pins (usually 0.1uF). You also have to be careful with the layout of the supply pins and lines under the chip to avoid induced noise on the supply lines. For examples of good layout, look at the Demo Board and the several board designed for > 80MHz use. The Propeller Backpack is a good example too.

What is believed to happen is that there's a spike on either Vdd or Vss that forces the net supply voltage over 4.0V which is the absolute maximum voltage between Vdd and Vss. The weakest part of the chip appears to be the PLL select circuitry, so that blows most of the time. If all the Vdd / Vss line are not connected together or the PCB lines connecting them are too long or run alongside other power lines, you can get different potentials on Vdd and/or Vss points on the chip and that can add to the existing Vdd-Vss voltage at different points on the chip.