RE: Re: BOUNCE: Non-member submission to canlist

"Nyffenegger, Dave" <[email protected]> Mon, 8 Jun 2015 04:50:41 +0000
Newsgroups gmane.comp.hardware.bus.can
Message-ID <BN3PR0801MB09792F529A476577B20759ED85BF0@BN3PR0801MB0979.namprd08.prod.outlook.com>
Re: 2. Use 120 Ohm terminating resistors, 1 per node, connected across CAN H and CAN L.
Only two terminating resistors, one on each end of the  bus.
-Dave

From: [email protected] [mailto:[email protected]] On Behalf Of Madhusudhan Chellappa
Sent: Sunday, June 07, 2015 1:40 PM
To: [email protected]
Cc: [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [CANLIST] Re: BOUNCE: Non-member submission to canlist

Hello Ayushi,
Nice to hear that you are looking into driverless cars in Univ, best of luck with that.
However, there is a layer of info missing to be able to help you - your application is not clear without which it doesn't make sense to randomly propose a specific CAN module etc.

However, let me first try to directly answer

1. Which CAN module is best for cars?
Considering you are in DIY mode, probably with a low budget, you should go for a Freescale development kit (since you said it is automotive).
Alternately, you should build a good project plan and go to a company that supplies production intent ECUs(I can recommend NewEagle and Pi Shurlock) for sponsorship. These will be plug and play, but you lose that learning component...

2. How do I implement CAN physically?
Assuming you will get the full details online, the really brief answer (experts on this mailing list can help you more when you have specifics):
1. Use twisted cables, 18 AWG. One for CAN H, and one for CAN L.
2. Use 120 Ohm terminating resistors, 1 per node, connected across CAN H and CAN L.
3. Connect the nodes you need to.
It really is easy, and a very good learning experience.
Download Busmaster, and purchase a cheap USB-CAN interface hardware for your electronics lab at Univ to use in your project. I highly recommend KVaser (as Abhijeet) although Peak is cheaper. Do not bother with Vector for this project unless they are giving it away!!

3. How do I program it? Using which OS and which software?
Related to 1 above. If Freescale, there will be some low level coding needed. But it will be cheap. Let me know and we (Geometric) can look into providing some code for you to get you kickstarted.

4. Is prototyping CAN module with arduino/ raspberry pi a good idea for
cars? How else I can make CAN work?
Nope, bad idea according to me. Automotive ECUs from Freescale and others come with the needed chips already built in, and sample code to start with. Rasberry Pi will mean a lot more additional work to just get you started working on your application. And what is the price difference? some 75 USD? You should weigh that against a couple of months to get it working.

And please don't spend time working on Autosar, communciation stacks and the such. These are really meant for production intent OEM applications, and you do not care about that!

Best of luck!
Madhu

------------------------------------
Madhusudhan Chellappa
Practice Head - Embedded Systems
Geometric Ltd.

On Sat, Jun 6, 2015 at 12:10 PM, Aayushi Sahu <[email protected]<mailto:[email protected]>> wrote:
Sir,

I'm a student of Manipal Institute of Technology, India and My team is
working on a driver-less car project where we need to use CAN for
automotive applications. I would like you to help me with the queries
regarding CAN bus.
I have questions regarding the implementation of CAN Bus. I've listed them
as follows.

1. Which CAN module is best for cars?
2. How do I implement CAN physically?
3. How do I program it? Using which OS and which software?
4. Is prototyping CAN module with arduino/ raspberry pi a good idea for
cars? How else I can make CAN work?
Kindly help me with these questions by providing me some references too if
possible. I would really appreciate your help Sir. Awaiting your response.

Thank you.
Regards,
Aayushi Sahu
[email protected]<mailto:[email protected]>