Close

Serial I/O kicking my ass

A project log for GPS Disciplined xCXO

A DIY GPS disciplined 10 MHz reference clock

nick-sayerNick Sayer 05/29/2016 at 13:360 Comments

I've wasted 4 revisions of boards trying to get the serial I/O interface to the oscillator right.

I wanted to add the option of changing the oscillator interface to use TTL level instead of RS-232 level, since that (in combination with allowing the oscillator input to come from an external connection instead of pin 7) would allow supporting additional FEI products. I think now I've got a correct set of solder jumpers. The two-way jumpers are both on the output side of the MAX-232 chip, plus there's an additional jumper to disconnect the RS-232 input side (because there's a pull-up there to ~10V).

But the latest surprise was that the dual NAND gate chip I was using was not pinned as I expected. I expected gate A to have inputs on pins 1 and 2 and an output on pin 3, then gate B to have inputs on pins 5 and 6 and output on pin 7 (pin 4 is ground and pin 8 is Vcc).. It turns out that the two outputs are swapped from that. Gate A's output is on pin 7 and gate B's output is on pin 3. What this meant was that the latest prototype was not blocking programming signals from the oscillator - it was merely inverting the serial-out pin, turning everything into garbage.

So... v1.6.1 will have that error fixed. And hopefully that will be the last one.

With the errors worked-around, though, the performance seems quite good.

Discussions