Leon, Bigglez, thanks very much for sparing your time to look at my board!
leon_heller:
You should have a 100 nF decoupling capacitor on the PIC.
Good point, thanks.
leon_heller:
One or more TVSs for transient protection would be a good idea, car supplies can have some large voltage transients. The zener diode might not be very effective.
bigglez:
(1) Not sure what D1/D2 are for?
I must confess, I cribbed the power supply schematic from another design, a community effort fuel injection computer called MegaSquirt. Apparently these diodes are commonly used by commerical EFI computers. Since I have built a MegaSquirt and it's been totally reliable for the last 5(?) years...
But I’m also mystified as to the D1/D2 purpose!
bigglez:
(2) R8 is ten megohms? Seems very high
Yes, this is more of a place-holder at the moment. I did a project years ago that used a photodiode and 10M resistor feeding into a Schmitt trigger gate to provide light-level sensing. I lifted the old circuit for that, but despite the PIC datasheet claiming similar specs for the RA4 Schmitt input (thresholds, leakage), it seems not to work so well on this prototype.
I will probably move the light sensor off-board into a separate module (would cut down on noise as well). However I need to crack on and get the wipers installed so I can get the vehicle on the road, clever light sensing comes later
bigglez:
(3) R7 limits Q2 base to under 1mA, what load is on Q2?
It's switching a standard car relay. 1k ohm is a placeholder, I may change this if necessary.
bigglez:
(4) Never seen a ‘profet’ is it flyback protected?
Yes, in theory it is fully protected. My relays also have flyback diodes built in, and this circuit is only going into one specific vehicle, so I have been a little lazy with the idiot-proofing.
I think the profets are overkill but I wanted a high-side switch, to limit the amount of rewiring that will be needed in the vehicle (the existing relays are all common-ground coil).
bigglez:
(5) I too get a DRC error on the polygons, but can see why?
I found Eagle to be a bit inconsistent on the polygon DRCs - sometimes it complains about clearances, sometimes not (even with no board edits in between).
bigglez:
(6) LED1 needs a ballast resistor (already noted by others)
Thanks. I'm using a LED with integral resister as I has a surplus of those.
bigglez:
(7) Does “+12V” refer to a vehicle battery? If so is it fuse protected?
Yep, all fused :-)
bigglez:
(8) RC6, RC7 Unprotected off board uC pins? Caution needed to
protect these from abuse and/or noise entry to the uC
These are an insurance in case of needing to add / change things later.
I learned on m first jobs years ago that prototypes are a bad place to start optimising board size!
I prefer to waste some space but have the option to use spare pins later…
bigglez:
(9) Took a deeper look at the planes (poly fills), why not lasso the
entire board and have one continuous fill? (looks like the poly fill
was done piecemeal.
This is because of point 5, the "random" DRCs on the polygons.
I have a +12v polygon on the solder side, in a half-hearted effort to reduce the resistance of the +12v connection to the profets. If I then also draw a polygon round the whole board and connect it to GND, I get lots of DRC clearance errors between the +12v and GND polygons.
How do other people solve this issue? Should I move the +12v polygon out of the solder side?
Thanks for the comments, it’s giving me a bit more confidence in what I’ve designed.
Steve.