next topic »

[Reply #15 on:]

That's interesting Westfw,

I didn't mention it but while I was waiting for the breakout board to arrive I hacked an UNO board with the 328 removed, soldered a wire to the auto reset cap (because I wanted to test the clone circuit as much as possible), connected this and GND, TX, RX to a 6 way leaded socket (pin-out as per FTDI cable, CTS tied to GND and 5V not necessary as clone has own) and was able to program with 100% success. The clone board has a 16MHz crystal with 2 x 22pF caps, auto reset cap 100nF. The supply rails are clean with plenty of decoupling.

When I've got time I will make some measurements with a 'scope and post the findings.

ATB Dan

[Reply #16 on:]

The 115200bps serial rate used by optiboot is not particularly accurate on a 16MHz AVR.  If (say)
neither the AVR nor the FTDI had a crystal-controlled clock ...

I can imagine this problem would get worse at higher baud rates. Why on earth do they
put a 5% tolerance ceramic resonator in the UNO along with using optiboot at 115200?

[Reply #17 on:]

a 5% tolerance ceramic resonator

That's extremely unlikely.  I couldn't even find any 5% resonators.  (I *did* find a bunch of 0.5% resonators (which is more typical) whose "short descriptions" were missing the decimal points, so it LOOKED like they might be 5%.  But they weren't.)

[Reply #18 on:]

Oops, misfire, it is 0.5%, not 5%. Still many times worse than 20-50 PPM for crystals, which
is 0.002-0.005% [if in error again, well, maths isn't strong in the mornings].