The CANary project

[nananaLeafman]

I just measured mine - I am reading about 330mA. Rather than lift the diode, I pulled the 10A fuse powering the port ( top, middle 10A fuse in the fuse compartment to the left of the steering wheel ) and inserted my meter there with and without the CANary connected. You can probably just insert a #20 or #18 wire, but I soldered two leads to an old blown fuse to ensure good contact and no damage to the fuse slot.

So I did the same thing, got a reading of 48ma without canary, and 261ma with canary, this is with the screens dimmed with headlights on. so seems reasonable.

I'll try pulling the canbus choke now and see what happens.

edit: after less than 30 seconds the diode reaches 65 degrees celcius, at which point i pulled the plug as it was still warming.

I removed the canbus choke, and teh diode no longer gets extremely hot but the car still locked down. I'm not removing the vp230 on carcan.

edit 2: removed teh carbus vp230 and still same problem, moving on to evbus...

 

[nananaLeafman]

SUCCESS!

I replaced both vp230 chips and both chokes and its now working as far as I can tell.

I havnt gone for a long drive yet as i need to charge before i can go anywhere, but it looks promising...

 

[nananaLeafman]

Heres a shot of the canary in action:
http://imgur.com/lVzKBWE" onclick="window.open(this.href);return false;

Of course I was too excited and put it in the case before I put the battery on, or shorted the write enable header... so I have to take it apart but its working for today

Now I need metric, but I need the wife off my back first :lol:

 

[TickTock]

Hurray! Now we know what a good and bad canbus looks like and how it impacts the leaf. Are you going to paint it?

 

[nananaLeafman]

maybe eventually, I'll paint it but the white isnt that bad, I want to get a couple things working before I'm too worried about that, mostly a good charge timer and metric

 

[nananaLeafman]

So over the weekend, the canary has worked pretty well.

I've noticed twice now it gets into some strange loop where it dims,turns off, then starts up again. May be correlated to evening time as the sun is going down but I have been unable to isolate it. At first I thought it might just be temperature related as the car had been parked in the sun and my cheapo thermometer reached 37C, but it did it again last night and it was much cooler (10-15c)

I'll see if i can narrow it down if you haven't seen the issue before, seems like software though if I unplug it for a couple minutes then plug it back in its usually fine.

Also, how are you managing the code? I'll hopefully get my metric stuff tested this week, and I may have a way of doing firmware upgrades off usb. I'm not very familiar with how mbed manages changes, I'm assuming I can pull a diff somehow and send it to you?

 

[TickTock]

This is also a new behavior I have not observed. Mine has been working [almost] flawlessly for months and I'll wager my dash gets a lot hotter than yours. The only problem I have been encountering is sometimes the USB drive fails to write to the log file (and disables logging until I manually reenable it). You said it goes into a loop. Are you saying that it keeps repeating the dim,off,on pattern? Does it happen while operating the car?

 

[nananaLeafman]

Yes, this is when the car is on, although i first noticed it as i had remotely turned on the A/c to cool the car down (im canadian, 30+C and I'm ready to die), and as i walked up to it I noticed it was flashing.

Bright->dim->off about 0.5s per step.

The second time i had turned the car on and started backing out of the driveway when it started looping.

If it keeps happening I'll throw in some logging and try and figure it out. It could even be hardware related, not sure if applying 14V to things through that bad zener could have killed anything, like that 5v reg or the mbed?

 

[garygid]

Overheated regulator shutting down, perhaps?
Maybe because it is getting too high of an input voltage?

 

[TickTock]

Good thought. You can touch the 3.3V and 5V supplies easily without removing from the enclosure (far right two pins on the mbed looking into the bottom). Measuring Zener drop is probably not possible without at least some disassembly, but that should be 6V if you are able to test.

 

[nananaLeafman]

I was just looking over the datasheet for the lm1117-5, and the maximum dropout is around 1.3v, if the leaf isnt charging the 12V battery, wouldn't it be possible that after the 6v lost to zener, that we don't have enough voltage to regulate properly, say the 12V battery is 12V:

12V-6V=6V -1.3V=4.7V or below the 5v threshold for stable regulation...

I'll do some probing if i can get this to occur again, it did happen this morning and was at a much slower frequency than last time.

So for items to check:
3.3V rail
5V rail
6v Zener
gate on the screen led transistor
input to lm1117
anything else?

 

[TickTock]

I have seen my battery drop to as low as 11.1V without any failure. The only thing that uses the 5V is the USB port and the LCD backlight so unless you are logging, I wouldn't expect a dropout issue to affect functionality (thanks for the thought though, since this could be an explanation for the intermittent USB logging issue I have been seeing ). As long as the 3.3V is within tolerance the mbed and VP230s should work.

When you measure your battery voltage (top of the Zener), take note of what CANary is reporting as the battery voltage. There is a scalar in the config file you can use to adjust the ratio of the resistor divider to battery voltage. Tweak this to make CANary match what your DMM reports.

 

[nananaLeafman]

So I noticed an interesting effect today, the rate of the flashing is proportional to the battery voltage.

IE higher voltage (14+) and it flashes extremely fast, drop it down to 13 and its very slow, below 13 its rarely occurs.

Also, when i plug the embed into usb it no longer powers up, and i get 3v on vin to gnd.

Going to try and borrow a variable power supply from work to take more measurements on the bench...

 

[nananaLeafman]

interesting, my zener diode is toast again.

I ordered a couple spares, but thats really bizarre.

edit: Wonder if I got a bad batch, or the brand is inferior to the ones your using which arent causing you issues. I'm getting tempted to hack in a switching buck regulator in place of the zener to do the 6v drop, I probably have parts that would work.
Still, I'm not sure why I cant run the mbed anymore off usb, I wonder if the lm1117 died from the high voltage, although the absolute max is 20v so it should have survived and thermal regulated itself.

edit2: according to: http://mbed.org/forum/mbed/topic/1802/" onclick="window.open(this.href);return false; it sounds like i also fried the FPF2123 on the mbed, which isnt a big deal. The main issue is why do my zener diodes keep melting...

edit3: ordered http://cgi.ebay.ca/ws/eBayISAPI.dll?ViewItem&item=350753515699" onclick="window.open(this.href);return false; to replace zeners, if its any good.

edit4: replaced zener, noted ~230 ma draw, and a 6.14 v drop accross the diode...

Testing with the bench supply told me it is definitely looping because of thermal regulation of a regulator, they get 80+ degrees very fast at 14V on the mbed, nothing seems damaged other than the FPF2123, which I can replace with solder supposedly, although I'll probably order a new one eventually as I cannot upload firmware now.

at anywhere below 6.2 v things start degrading (after zener) so if your battery is sagging, this will cause issues, I am hopeful that replacing the zener with that SMPS will resolve 99% of these issues, provided the noise it introduces doesnt cause other problems... thoughts? at 5.9V the USB 5V will begin to rapidly drop so thats likely your issue, 6.1V has noticeable screen dimming...

 

[TickTock]

Thanks for the updates. A DC-DC converter is definitely a slicker solution which will be easier on your battery. You can always add caps if the noise proves to be problematic. Turbo3 went with the DC-DC converter, too, on his version. However, the Zener should work. At 230mA @ 6.14V you are at 1.4W which is well below the 5W rating. Did you buy your Zener (1N5340BG) from the same eBay seller (using the link I provided on the Wiki)? I thought I tossed in a couple extra Diodes in with the boards I sent (but that might have been someone else).

I've built three of these now and have never had an mbed die. I have locked up mbed's with bad firmware, though, and was unable to upload new code (wouldn't even mount the drive when connected to a PC). However, the trick of pressing-and-holding the reset button as you plug it in worked for me (you have to keep holding it as you drag the new firmware over to the mbed drive). Did you try that? Does the mbed run when powered by the USB and you just cannot upload firmware or does it not even turn on? If the latter, then I agree the 2123 probably got toasted somehow.

 

[nananaLeafman]

I got the first batch of Zener's from mouser, the second set is from digikey (which is in it now), I couldnt find a decent deal on them on ebay as that auction was no longer active.

I'll try testing the reset but I'm fairly certain I killed that FPF chip, which is not a big deal, the mbed processor is still fine.

 

[TickTock]

I went ahead and ordered a DC/DC converter (http://www.ebay.com/itm/151037086416?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649) since it was cheap enough and seemed like a handy thing to have on-hand. The one I picked actually has a 3.3V output, too, so it could potentially replace both regulators.
I also ordered some 5.1V Diodes (http://www.ebay.com/itm/2-pcs-1N533...470?pt=LH_DefaultDomain_0&hash=item53ee4d04c6), too, to see if an extra 0.9V headroom will make my USB logging more reliable.

When you add the metric option, please make it configurable. I can help with adding the entry to the config file. Also, I think you have a 2013 and as such only have 3 temp sensors so the temperature readout will have an error (I display average of all four on the main screen). So maybe we should add a model year entry into the config file as well to accommodate such differences.

The best way to manage multiple authors is to create a branch and make edits there. Then, once you think you have a rev ready to go, give me the link and I can use the merge feature to bring in your enhancements.

 

[TonyWilliams]

So maybe we should add a model year entry into the config file as well to accommodate such differences.

Why not just read the highest two temp probes? That way, you don't need to configure for a particular year car, and if a single probe becomes inop, you will still display the average of the two highest temps (instead of an error and no data).

 

[nananaLeafman]

My metric stuff is a config option, I just have to test before I'm ready to send you a diff. I also added firmware update from a config button that reads a new bin file off a usb key, so no more trips to the computer.

My car is actually a 2012, as Canada is still waiting for the 2013's to be released, if ever. Odd, considering we would benefit more than most with that new heat pump!

 

[TickTock]

So maybe we should add a model year entry into the config file as well to accommodate such differences.

Why not just read the highest two temp probes? That way, you don't need to configure for a particular year car, and if a single probe becomes inop, you will still display the average of the two highest temps (instead of an error and no data).

Good thought. Maybe I should just display the max on the main display. Anyone really concerned about the min? I do display all four in the cellpair screen if further information is desired.