Opinions on LIN

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
I'm looking at short-haul comm protocols. LIN seems to fit
my needs. One-wire is especially attractive. Anyone have any
suggestions in this regard?

Thanx,
Doug



Re: Opinions on LIN

Quoted text here. Click to load it

Saving one wire against RS232 means the protocol knows when
to switch the direction or the message length is fixed.
Is saving one wire really that attractive ?

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net

Re: Opinions on LIN
This is not point-to-point communications. I thought about using RS-485
but that takes 4 wires for full-duplex. 2-wire is half-duplex so I might
as well go 1-wire with LIN. Another limitation I am facing is the I/O
capabilities of the uP I am using. With only 5 I/O pins, an extra pin
for full-duplex comms does become a potential issue.

Doug

Quoted text here. Click to load it



Re: Opinions on LIN
On Mon, 9 Aug 2004 23:25:25 -0400, "Doug Dotson"

Quoted text here. Click to load it
What  length is short-haul ?  Is it between PCBs, between adjascent
boxes ?
Anyway I have been looking at LIN myself, and for what it is designed
it looks quite nice. At the moment there are very few proper LIN slave
devices which is not a MCU + LIN Driver. So ATM if you are going to
connect two or more MCUs together then the more traditional UART, SPI,
I2C or CAN might be a better option.

Regards
   Anton Erasmus

Re: Opinions on LIN

Quoted text here. Click to load it

Probably not more than 100 feet. If it were between PCBs than I would
probably use I2C or SPI.

Quoted text here. Click to load it

There is code for the ATMEL chips that implement LIN driectly on
a single pin. Using an external driver such as the MC33399 requires
multiple pins and with only 5 I/O pins on the chip I plan to use, that
becomes an issue.


Quoted text here. Click to load it



Re: Opinions on LIN
On Tue, 10 Aug 2004 09:22:53 -0400, "Doug Dotson"

Quoted text here. Click to load it

LIN bus has been designed for in-car use, hence I do not think you are
going to get 100 feet.  Having said that it, is basically a
half-duplex protocol that can be implimented using a normal UART +
suitable driver. One can probably impliment the LIN bus protocol using
a RS-485 driver. This will easily do the 100 feet, although you will
need 2 wires (plus GND). With this you still only need one MCU pin for
the half-duplex protocol.
The one-wire bus from DALLAS/Maxim can do long distances and can
also be done in software, although the timing restrictions are more
severe than for the LIN Bus protocol.

Regards
   Anton Erasmus


Re: Opinions on LIN
Comments below.

Doug

Quoted text here. Click to load it


Ya think! Obviously I'm not planning on using it for a car.

Quoted text here. Click to load it

I suppose so.

Quoted text here. Click to load it

Perhaps, but the LIN spec describes the line characteristics. RS-485
doesn;t seem to comply.

Quoted text here. Click to load it

But this isn't LIN.

Quoted text here. Click to load it

But this isn;t LIN. Perhaps I wasn;t clear. Tell me about the pros and
cons of LIN. I know of numerous ways of communicating. I'm asking
about LIN.


Quoted text here. Click to load it



Re: Opinions on LIN
On Tue, 10 Aug 2004 17:36:33 -0400, "Doug Dotson"

Quoted text here. Click to load it

OK. AS I understand it LIN Bus was designed to be a sub bus to very
cheaply connect sensors to a local CAN Bus node. I.e. one can use a
US$15 MCU for the CAN bus node, and to connect sensors and outputs
such as switches and lights to this node one uses LIN bus. The LIN
slave can be a cheap 1-5US$ MCU or an IC designed as a LIN slave.
Currently there are only a few dedicated LIN slave ICs I have found.
There are a few to drive LEDs (Including programmable PWM for
brightness). Once there are many dedicated LIN slave ICs, it would
be quite a nice and cheap solution in many cases. ATM for most
things one need to use an MCU for the LIN Bus master as well as a
MCU for the slave - hence currently there is not much advantage
in using LIN bus in stead of the more traditional approaches.
For the high volumes used in the motor car industry the difference
in pricing between the CAN MCUs and non CAN MCUs are significant.
For lower volumes the price difference are insignificant. For many
of the projects I work on - using a US$50 MCU will still make sense if
I can save a week or two in development time.
The full LIN Bus specification provides a way for the software to
configure itself to handle similar types of devices from different
manufacturers in a transparent way. So if you have a module that
senses a door switch provided by Bosch - one can replace it with a
switch sensor provided by BMW without having to change any of
the software. I think the idea is to start with MCU based slaves, and
then gradually replace these devices with dedicated slaves. This can
then be retrofitted to older models without problems.

My 2c worth.

Regards
   Anton Erasmus

Re: Opinions on LIN
Thanks Anton,

This helps alot. I have been looking at MicroMessaging and noticed
the relation between LIN and CAN.

Doug

Quoted text here. Click to load it



Re: Opinions on LIN
Actually the LIN spec is 40 meters max.

Quoted text here. Click to load it



Site Timeline