This will be my first attempt at something this complicated and I'm just learning Fritzing and basic board layout so it may be a bit funky.
The bottom is power in going to a 7805 which powers the 328 and L298. A couple of places for decoupling caps and a power LED. The next chunk under the header holds the diodes, 4 LED's and output sockets for the motors.
Above that is a header for a little L298 breakout board 1-15 numbered from the left. I was going to build the L298 onto the board but I figured I could make some other plugin driver boards if I wanted to fool with other drivers.
Above that is the 328. I still need to route the reset button. I put a place for another decoupling cap close to the chip.
The VR spot in the upper left is for a DC/DC converter for driving servo motors from the pins to the right of it. The DC converter will power them and the inputs are wired to the chip.
The 2x6 is a programming plug. I just pull the chip on my UNO to program standalone stuff...
The 8 pin header on the right are just I/O pins but the bottom 4 pins have 5V and ground for plugging in an ultrasonic sensor.
I still need to review my notes to check all the 328 "life support" stuff.
I built one driver board plus another driver board with a 328 and both worked well. They were on protoboard but I want to do something with PCB's. It's pretty cool messing with it
In general, make power traces wider, use a ground plane on both sides, avoid right angles (can etch funny, creating thin-ness and altered resistance in the corner).
You need 100nF caps from Vcc, AVcc, Aref to Gnd as you noted.
I'm not familiar with ground planes although I've heard reference to them. Interesting about the 90 deg bends! I always wondered why boards seem to have a lot of 45's.
You haven't seen the articles where high speed signals fly off the corners and cause electromagnetic interference among other things Or funny reflections back to the source. Or ...Some interesting reading.
I would say at these speeds, its more a matter of trace quality.
Then you should also straighten/shorten some traces (in the upper right side, there is one awkward sticking out).
Pull-up resistor on the reset pin?
Power-indicator LED with no resistor?
A schematic would be also helpful.
I added room for another cap, plus the pull up resistor for reset (To 5v right?) and the reset button grounds the reset pin. I'm not having much luck generating a schematic from fritzing. Mainly because I used 2 pin headers to make 3 pins and the 15 pin...
I just used the PCB view to sort of point-to-point
I'm going to carefully go over everything before attempting to etch. I added some vias at the corners to aid with registration.
The power led resistor is in the cap spot beneath it. Fritzing doesn't seem to have an "up and down" resistor footprint so I substituted a cap. But thanks for scoping it out florinc!
I appreciate all the input and I'm working on straightening the schematic out...
A few comments, as once again, no schematic to go from. Some reaffirmation of
comments by others:
use very wide traces on power busses and grounds.
you need a large reservoir cap, in addition to bypass cap, from Vdd to Gnd on
the h-bridge. S/B in the range of 330-470 uF at 50V, or larger. Note - this cap
will be nontrivial in size.
I always place small series-Rs [330 ohms or so] in the lines between the cpu and
the h-bridge input/enable/PWM pins. To prevent any possible motor spikes from
blowing the cpu.
make sure the gnd and power traces from the cpu to the h-bridge are separate
from the ones from the h-bridge to the motors. IE, no motor currents in the
cpu gnds. No ground loops.
L298 definitely needs external clamping diodes on the motor pins, but I
believe you have those on the pcb, however, best to have those as close to the
h-bridge as possible.
