OBD MPGuino gauge

[Yoshi]

Originally Posted by ishiyakazuo

SparkFun has another (maybe even better?) board for AT90CAN128: SparkFun Electronics - Header Board for AT90CAN128
They also sell the bare board very cheaply:
SparkFun Electronics - ATMega128/103 Bare PCB

Please note that the AVR-CAN includes MCP2551 (CAN transceiver) and ST232 (RS232C transceiver), then assembled and tested.

CANについての情報がありましたら、教えていただけますと大変助かります。
よろしくお願い申し上げます。

こちらこそ、よろしくお願いします。

Yoshi

[dcb]

I think the elm327 is starting to look like even more of a bargain You don't *really* want the rs232 bit anyway. Mouser has the mcp2551 for $1.50 (the 327 reference schematic calls for it also for CAN), and it is pdip:
http://www.mouser.com/Search/Product...xOd%2fTw%3d%3d

[ishiyakazuo]

Originally Posted by dcb

I think the elm327 is starting to look like even more of a bargain You don't *really* want the rs232 bit anyway. Mouser has the mcp2551 for less than $2 (the 327 reference schematic calls for it also for CAN)

Seriously, I think so too. After looking into it, it seems like for the two more difficult protocols (PWM and CAN), there are chips out there to do the work for us (for PWM, there are others aside from the ELM parts, that assume you know the basics of J1850, which we would have to for VPW anyway).
And hey, $60 is still substantially cheaper than the $150 for SG2...

[dcb]

So we have pretty universal coverage with the elm 327 and reduced cost obd options for specific protocols (i.e. the 323) that don't require any additional coding, and you can always install a mpguino if you can't afford that. The tough question is, does the standalone iso or vpw approach still make sense (arduino plus a couple transistors)?

[ishiyakazuo]

I think so. It still means less tinkering with the car itself, which means more people would be willing to use it.
I know I have plenty of friends who would have no trouble hooking up an OBD-II connector, but if I showed them hooking up to the ECU's pins directly, they'd casually back away.

Honestly, ISO should be fairly trivial if you use the USART rather than bit-banging. And when you look at $1.50 for the Freescale vs. the $16 ELM (actually, does the ELM also need the Freescale? I haven't looked), there's a significant cost difference in the solutions that should appeal to many. (Keep in mind that the top-selling car over the past decade is Toyota Camry, followed by Honda Civic... both ISO through at least 2005.)
I don't know about the J1850 standards. I still think VPW is probably within reach, but I don't really have a car to do testing with.

We can think about how best to fit into the 14k without configurations when it becomes an issue.

[Magister]

Maybe the standalone ISO with just the arduino and the MC33290 still makes sense, it covers a lot of cars and is very cheap to build. The ELM version does not need the freescale, take a look at the PDF page 56
http://www.elmelectronics.com/DSheets/ELM327DS.pdf

I got the MCP2551 free from Microchip as a sample, and I will order the ELM327. What's nice with the ELM327 is its cover of all the protocol. Also the reference design is available and if you just want CAN for instance you just have to buy/solder the CAN part and put some pins to ground/5V.
see page 58 of the PDF for an example for VPW.

See ¶ "Where can I buy a circuit board for my ELM IC?" of this url, almost at the end, for schematic, gerber, etc
ELM Electronics - Chip Help

Also OBDuino and MPGuino share a lot of code. LCD and buttons are the same (I optimized some LCD functions). In OBDuino there is a level of abstraction for the protocol layer. There is a generic get_pid() function that return the result, inside there is a call to iso_send_cmd() that can be changed by a parameter and call vpw_send_cmd() or elm_send_cmd() for instance.

Also if we have an hybrid setup that can work with the MC33290 and/or the ELM327 and the code is less than 14k, why not?

[dcb]

Originally Posted by ishiyakazuo

Freescale

?

I thought you were doing all through hole and a couple transistors? Do you need to add a smd to the mix?!? The elm 323 (iso) is just using a couple transistors.

http://elmelectronics.com/DSheets/ELM323DS.pdf

[ishiyakazuo]

@dcb: Just relax To the code, it doesn't matter which you're using, so long as the signaling is correct. The transistor method will likely work fine too. But I don't want to risk having to debug both HW and SW at the same time, so I'm gonna use the Freescale on my end, so I know the levels are correct in my particular application. (If we were using the same PIC, there'd be absolutely zero risk, but who knows how the pullup/pulldown values match up with the PIC's?)

EDIT: Nevermind again. I didn't notice that on perquin it even says it's for interfacing with ATmega8, which should be fully compatible on the UART pins. Actually, I like it a lot because it's all 10k resistors, so a resistor network (http://search.digikey.com/scripts/Dk...10X-2-103LF-ND) and two BC547s (http://search.digikey.com/scripts/Dk...=BC547BTACT-ND) would do it much cheaper than the Freescale. I'm sold on it!

@Magister: Sorry, looks like I glossed over the get_pid() function earlier Nevermind about a lot of what I said then.
I'll take a peek at the OBDuino code again tonight to get more familiar with it

[dcb]

Originally Posted by ishiyakazuo

(If we were using the same PIC, there'd be absolutely zero risk)

I wouldn't even bother reporting on a non atmega168 results, yah it *should* work, bla bla, don't care , wait for your arduino before you draw any conclusions please.

[Magister]

Originally Posted by ishiyakazuo

would do it much cheaper than the Freescale. I'm sold on it!

The freescale is free as a sample, even 2 samples are free
It is a chip used in professionnal tools for years, it includes about 75 transistors.