| ??? 09/14/07 03:49 Modified: 09/14/07 04:21 Read: times |
#144603 - No, no... Responding to: ???'s previous message |
Rubem said:
I've already replaced D1 with a jumper (so C4 provides decoupling)... It "looks" like decoupling, but is not. The connections are way too long! Using a solid ground plane would decrease the associated inductance by a factor of 10...100! Also, along the same ground wires not only the decoupling is connected to ground pin of micro, but also the currents of burden caps. This is a very very bad "decoupling". No, it's no longer a decoupling, but even a positive feedback, which could effectively block the oscillator, indeed. You must provide different paths for the currents of burden caps and supply decoupling caps, unless you use a solid ground plane. And the connections must be as short as ever possible, if you use wires. Again, every millimeter counts when taking wires! Rubem, your setup is not suited to make your micro work! Decades of experience tell me, that your micro would run properly if you would only use a proper layout, optimally containing a solid ground plane! Besides, your burden caps carry yellow marks, telling me, that these are no NP0 ceramics! Are you sure, that you have proper caps soldered?? Kai |
