The CANary project

I would make the legs shorter, perhaps so that the center
just barely touches the small "hump" over the Nav screen.

I would prefer to have a flat bottom that I could paste a piece
of stiff paper over to keep out dust and debris. Then, I could
stick on 4 rubber "feet" in the 4 corners, to make it non-skid
and raise it, if necessary. For placing the feet, it would be nice
to have a horizontal flat area at each corner of the bottom,
perhaps 1/2 or 3/8 inch "square-ish".

But, perhaps you want it higher for "touch-ability" of the screens?

Speaking of screens, mine arrived from Serbia today, via DHL.
So, I could use some coaching on wiring these displays, please.

Very nice screen of the cell voltages. :mrgreen:

A vertical bar graph of the number of cell-pairs at each voltage
would give a quick-glance view of equalization.

Are you missing a "leg" on the left side of the front surface
of your rendering of your v2 enclosure?

Am hoping to find a message on the canbus indicating the condition of the cars light sensor so I can dim the display whenever the LEAFs built-in instrumentation is dimmed. Worse case, I can add a cadmium sulphide cell to one of the spare analog inputs,

Click to expand...

There is a message for the headlight status (0x625 on the can bus) Not quite the same but usable. The light sensor may also be in that message.

garygid said:

Are you missing a "leg" on the left side of the front surface
of your rendering of your v2 enclosure?

Click to expand...

No, it's just odd shaped due to the shape of the enclosure:


I will make them all shorter. I think I will also hollow the feet out down to the optimal distance for each corner so it will sit flat as it arrives, but you can cut off the hollow, thin-walled, section to set the right height for each. They will all have to be different lengths for it to sit with all four feet on the dash. I added rubber feet to the corner of rev1, and when I push three feet flat on the dash, the fourth is ~3/16 above - the surface back there has twist in addition to the curvature.

turbo2ltr said:

Am hoping to find a message on the canbus indicating the condition of the cars light sensor so I can dim the display whenever the LEAFs built-in instrumentation is dimmed. Worse case, I can add a cadmium sulphide cell to one of the spare analog inputs,

Click to expand...

There is a message for the headlight status (0x625 on the can bus) Not quite the same but usable. The light sensor may also be in that message.

Click to expand...

Thanks - I'll check it out. Headlight status should work but I'll see if there are other options. One of the displays I added was one that just displays which canbus messages changed so you can easily see cause and effect for the various sensors.

OK. Got a few blocks of time over the weekend to crank on the Rev2 enclosure and PCB. Gary, I took your suggestion on the snap in fixtures for the LCDs and PCB. In fact it was necessary. There isn't enough room inside to use the peg method for both PCB and LCD (once one is in you can't install the other). So I used the "snap" fixtures on the tight spots and the pegs where I had room. I should be able to push one end into the snap and then slide th other over the pegs.

I also figured out how to make the PCB mounted USB work. If I place it toward the bottom (but about an inch up to clear the dash adequately), I can tilt the PCB out far enough that the USB is flush with the back. Also added a support near the USB aligned to a hole in the PCB for a machine screw (so pushing a USB device in and out won't strain the PCB or push it off it's pegs.

I hollowed out three of the feet to the desired distance to make it rest flat on the dash so the walls can be cut off it one chooses to. Finally, there is a tiny hole up and right of the USB which aligns above a reset switch I added.

Here's the WIP PCB to help illustrate the USB and mount locations. Top layer (red) is toward the front of the enclosure. I would have liked to have put the RTC battery on the top side, but the LCD's are almost touching the PCB on that side. It'll be a little tricky to change the battery but should be possible without disassembling.

commects:
1. wall between the displays, you might want to round the sharp edge.
2. access to the mini/micro USB connection on the Mbed board for
re-programming or logging to the PC - seems dificult.
3. it appears that the only reason to have the circuit board forward so far is
to keep most of technology USB connector inside.
4. I would probably...
make the break-off part on the top of the board, to keep the bottom edge stronger.
Mount the board back almost touching the back wall of the box,
leaving more room for the headers, battery, and access to the mini usb.
I would slide the Mbed to the right edge of the board to make access to
the mini usb a lot easier.
I would move the USB to the right end of the box, on the pigtail, or possibly
use a right-angle connector on the left side. If one used 3 slide-in supports
instead of two, and one short pin support, one might get the USB connector
through the left end.
The support in the center of the board is probably not required.

more in next post

garygid said:

commects:
1. wall between the displays, you might want to round the sharp edge.

Click to expand...

I debated that.. Aesthetically, I like sharp clean lines. Maybe I soften it a bit but with a smaller radius

garygid said:

2. access to the mini/micro USB connection on the Mbed board for
re-programming or logging to the PC - seems difficult.

Click to expand...

A little, but easier then accessing the programming port on the gid-o-meter. I will probably leave the programming USB cable plugged in to the mbed when actively debugging/programming. There is room there but I agree may be tough if you have thick fingers.

garygid said:

3. it appears that the only reason to have the circuit board forward so far is
to keep most of technology USB connector inside.

Click to expand...

Yes, I like having the logger USB connection in the back and the cabled USB port is bulky (at least mine is). The only place I could fit the cabled connector was on the right side wall and then it interferes with the programming USB connection. I could have rotated the mbed orientation and cut a hole for programming on the left side leaving the space on the right for a cabled USB datalogger port, but once I realized I could swing the bottom of the PCB out to make the PCB mount work I decided to go that route (didn't want a hole left over once I was done programming). Even pulled forward, there is still more clearance on the bottom than on the top. I am working on a 3d mechanical model of the board and the LCDs to ensure there are no collisions. You may be right that I will have to adjust to ensure the LCDs or center wall don't block the programming port. It's is tight in there.

garygid said:

4. more coming

Click to expand...

Hmmmm. I may just rotate the mbed 90 with the programmer pointing down... then I can better place it behind the more open areas. Ah well.

First, please see the additional comments in my previous post.

If the mbed is pointing down, with the mini-usb plugged in, it
cannot be set on its base, so, not the best idea.

However, slide the mbed down and right, and the battery left,
and you will have good access for both. Or, just slide the mbed right.

I would probably keep the regulator and zener (leave air space around the body)
closer to the bottom of the box, to give them more cooling air circulation.

Consider adding a fuse, or a self-resetting fuse.

Since plugging and unplugging is best done when powered off,
I prefer having a switch, even if it was a right angle switch inside,
perhaps hanging off the right side of the board. Also, I prefer to
be able to turn it off, rather than have to unplug the OBD connector
when leaving the car for a significant time.

I do not see the connections for the two displays.

Did you want to allow for your beeper?

Tipping the board back, you could have the reset button on the
bottom edge of the board, not need a reset hole, and keep
all the components, including the battery, on the front side of the board.

I would probably spread a lot of ground area around the board,
that rarely ever hurts.

For the two CAN transceiver chips, add some pin-1 or orientation
markings to the board... I do not see any. Also, they could be
moved to the top of the board, to keep all the surface mount on
one side of the board.

Consider adding holes for optional test points, to easier connect a scope
or a meter to help debug defective components or boards.
They cost nothing now, but can save a lot of time later.

Were you planning to be able to disable writing to the CAN buses
with a jumper or two? Might be good for testing, at least.

Oh, I see the two jumpers, but it looks like they short the mbed output
to (power or ground?), unless there are two resistors that I cannot see.

Do you want a small filter capacitor on the analog inputs,
especially the one from the switched 12v divider?

Probably better to make the metal (heat sink) around the
regulator (and zener) pads a lot bigger?

Probably the connector for the ribbon cable can move up a little,
and the coin battery can be under the ribbon?
Then move to OBD 2x6 connector down near the ribbon connector
to give room (maybe) to use a header there?

If possible, move the "Canary rev2" down near the bottom edge
so that it can be read without taking the board out?

Just my thoughts, and enough for now.

Did the rev1 board arrive yet?
Cheers, Gary

Thanks for the suggestions Gary. One question. What's the motivation to move the SMDs to one side? Using both sides gives me a lot more real estate and it's easier to route the wires. They are very short so don't really create any additional clearance requirement above the through solder joints.

A few quick answers to some of the questions:
1) there are two jumpers already there to disable/enable TX - all they do is disconnect the mbed pin from the VP230 transciever. Hi Z is sufficent (actuall ideal since the RS pin is a current reference pin) to disable the TX
2) there will be a lot more heat sink once I add the ground and power planes (will do that last just before ordering the board since it makes it a pain to modify after that)
3) all the lcd, buzzer, touchscreen, etc connections that directly connect to the mbed are done using the second row of through holes inside the mbed socket. I brought them all out so other devices cold be added later if desired. The wires can be fed from the back to the LCD through to hole created when the flex breakout section is cut out.

With all SMD on one side, the manufacturing or fabrication
has only one surface to do, which might? be easier than 2 sides?
But, maybe you are manually soldering the SM components?
Also, probing two-sided components is not as easy as 1.
However, your point is well taken.
I do not have all the best answers, but I just want you to
consider the questions, in case you have not already.

Leaving the RS floating might be OK for your chip,
but it did not seem to be that way for the chip that I have.

When something does not work, it is a real pain to have to
un-solder wires or chips to try another component.

I would leave room to stick a 1x20 row of male header pins through
from the back side of the board, and be able to plug a female pin
holder on from the front side.

I would use a 2x3 and a 2x7 on each display and 1xN as needed
to connect to the Mbed. But, you have 8 (4 xy, 1 re, 3 isp) pins to
connect to two displays, so the easiest way to do that is to make
TWO rows of holes along the top side of the Mbed Board, outside
of the socket for the Mbed of course, to be able to connect to the pins.

Even if you are soldering wires, the extra row of holes will make the
job a lot easier.

To get more space above the front side of the PC board,
(for example, for headers)
I would consider sloping the back of the enclosure out a bit
at its top so that the PC board has sufficient (perhaps more
or less uniform) clearances from the sloped displays.

With the increased front sunshade on the enclosure,
you do not want it to tip over on fast acceleration.

To save some significant $ on the enclosure, you might thin all
the large surfaces, when more than a half-inch from structural
corners, edges, and support points. If needed, put in (leave in)
an occasional full-thickness stiffener, perhaps 1/4"wide.
Going down to 1/16" (from 1/8") over 60% of the box could
save significant plastic, thus 3D printer cost?

Did the v1 $88 price include shipping (I expect not)?

For ease of debugging, I tend to use DIP chips and sockets, rather
than surface mount. For some quantity of production, surface mount
can be better, especially if one is using automated fabrication,
and is confident in having boards without manufacturing defects.

Thanks for all the suggestions and feedback - lots to think about.

Thanks for letting me be involved.

I can hardly wait to get a CANary breadborded.
I read the docs on the display last night,
and I am closer to understanding its connections.

Please keep up the good work, and thanks for
keeping me in the loop as you progress.

Investigating EAGLE layout and schematic software.

Generally expensive, except for the Free version,
which is limited to boards a maximum of 80 x 100 mm
(3. 2 x 4 inches).

So, is that one reason why the CANary board is not longer?