Lpc2000

Does anyone have an example program yet for using the CAN interfaces 
on the LPC2129? FullCAN mode is what I am trying to use. I have a 
simple example going, but reception seems to stops as soon as I 
increase bandwidth :-(

Peter

On Thursday 08 April 2004 02:50, you wrote:
> Does anyone have an example program yet for using the CAN interfaces
> on the LPC2129? FullCAN mode is what I am trying to use. I have a
> simple example going, but reception seems to stops as soon as I
> increase bandwidth :-(
>
> Peter

What do you mean by "stop":
* Do you get a CAN-level failure (eg. bus off)
* Do you just stop getting interrupts?

-- CHarles

I didn't want to bore you with details, but now that you ask for it...

--- In lpc2000@yahoogroups.com, Charles Manning <manningc2@a...> 
> > Does anyone have an example program yet for using the CAN 
interfaces
> > on the LPC2129? FullCAN mode is what I am trying to use. I have a
> > simple example going, but reception seems to stops as soon as I
> > increase bandwidth :-(
> >
> > Peter
> 
> What do you mean by "stop":
> * Do you get a CAN-level failure (eg. bus off)
> * Do you just stop getting interrupts?

I only expect low-priority service messages on CAN (doing some 
background data logging) so I plan to use the FullCAN mode with no 
interrupts.

I set the FullCAN filters (4 of them) and once in a while (with the 
current LPC2000 performance this means about every 250 microseconds) 
check the semaphore bits of the FullCAN filters if any filter 
received something.

I use 125kbps and several CAN monitoring and analyzer tools on the 
network. This works just fine at a 1% bus load. But if I increase 
busload (and note that on the CAN side there are no errors, monitors 
and analyzers catch up and can work with every message) the 
semaphores at some point seem to "lock" and say they are always busy 
receiving something - but all 4 IDs simultanously busy at all times 
seems weird so I guess there is still some setup I am missing...

Peter

Peter,
     Unfortunately full_CAN mode is not functional in the current 
devices.  There is a problem with interrupt contention, when this is 
fully characterized it will be detailed in an errata.  In the mean 
time please do not use full-CAN mode.


Philips apps.

--- In lpc2000@yahoogroups.com, "embeddednut" <embeddednut@y...> 
wrote:
> Does anyone have an example program yet for using the CAN 
interfaces 

> on the LPC2129? FullCAN mode is what I am trying to use. I have a 
> simple example going, but reception seems to stops as soon as I 
> increase bandwidth :-(
> 
> Peter

--- In lpc2000@yahoogroups.com, "philips_apps" <philips_apps@y...> 
wrote:
> Peter,
>      Unfortunately full_CAN mode is not functional in the current 
> devices.  There is a problem with interrupt contention, when this is 
> fully characterized it will be detailed in an errata.  In the mean 
> time please do not use full-CAN mode.

Please confirm that the CAN peripheral is fully functional (apart from 
the FullCAN issue mentioned here).

Personally, I've never used FullCAN or equivalent hardware message Id 
matching. I have always done software message filtering, mainly 
because I've always used J1939 protocol which is easiest to filter in 
software.

Thanx

-- Charles

--- In lpc2000@yahoogroups.com, "philips_apps" <philips_apps@y...> 
wrote:
> Peter,
>      Unfortunately full_CAN mode is not functional in the current 
> devices.  There is a problem with interrupt contention, when this 
is 
> fully characterized it will be detailed in an errata.  In the mean 
> time please do not use full-CAN mode.
> 
> Philips apps.

Its always good to know that it is somebody else's fault...

Please inform us as soon as the errata is available and alternative 
usages outlined - some code examples would really be nice...

Peter

Dear Philips,

If the "full CAN" capabilities don't work on the LPC2129 devices, then - as
far as I am concerned - the CAN features are of no use, and I may as well
use an LPC2106 and and external CAN transceiver (e.g. MCP2510).

Had I known this, I wouldn't have bought 20 MCB2100 (LPC2129) boards from
Keil earlier this year.   (20 boards may not mean much to you, but it is a
large chunk of my teaching budget).

Great pity - and not the best of adverts for your company.

Michael.

Dear Michael,

although it does not help you too much now, we, the apps_team share 
your frustration about the full-CAN mode. The fact that it does not 
work as expected is fairly new and we are still investigating the 
best work around. 

Members of this group have had problems with the Full CAN mode and we 
did not want them to waste additional time. Usually we try to post 
problems together with a work-around, unfortunately this time, we are 
still working to ge tthe work-around ready to go. It is going to be 
an Application Note how to use the CAN in a mode I would call 
extended Basic-CAN. In close cooperation with CAN experts we hope to 
have something that will enable you to use the boards in a way you 
intended for teaching by the end of next week. 

In the unlikely event that this is not going to work for you, we will 
find another solution.

Sometimes it is a difficult decision to publish problems that are 
still under investigation in order to prevent the members of the 
group waisting time but at the same time we frustrate others like you 
at the moment :-(( if we publish without a resolution.  

Best regards, Robert


--- In lpc2000@yahoogroups.com, "Michael J. Pont" <Michael.Pont@t...> 
wrote:
> Dear Philips,
> 
> If the "full CAN" capabilities don't work on the LPC2129 devices, 
then - as
> far as I am concerned - the CAN features are of no use, and I may 
as well
> use an LPC2106 and and external CAN transceiver (e.g. MCP2510).
> 
> Had I known this, I wouldn't have bought 20 MCB2100 (LPC2129) 
boards from
> Keil earlier this year.   (20 boards may not mean much to you, but 
it is a

> large chunk of my teaching budget).
> 
> Great pity - and not the best of adverts for your company.
> 
> Michael.

Hi Michael, Peter and the others...

--- "Michael J. Pont" <Michael.Pont@...> wrote:
> If the "full CAN" capabilities don't work on the
> LPC2129 devices, then - as
> far as I am concerned - the CAN features are of no
> use, and I may as well
> use an LPC2106 and and external CAN transceiver
> (e.g. MCP2510).

Even without the FullCAN mode these devices have some very powerfull 
filter mechanisms that you'll hardly find anywhere else - and the 
three transmit buffer queue for each CAN channel is a very nice 
feature, too.

I am confident that "someone" will publish code alternatives here 
soon that help overcome the non functioning FullCAN mode ;-)

Olaf
Tutor at ESAcademy
www.ESAcademy.com
www.MicroCANopen.com

While Michael is entitled to his opinions, I'd like to disagree with 
him taking the stance that this is a bad advertisement for Philips.

I have yet to find any manufacturer that does not have bugs in their 
silicon. Get over it. What makes the difference is how the 
manufacturers handle the errors. The standard response seems to be 
denial for as long as possible.

I personally think Philips' approach of cooperating with a community 
such as this is very positive. I'd rather be told about problems 
immediately even if the information is incomplete at the time. This 
saves time (otherwise wasted chasing unknown bugs) and builds 
confidence.

-- Charles

Hi Robert

How is the CAN appnote coming along?

-- CHarles

To Philips Apps,

So is CAN on the LPC2000 devices a 'go' or 'no go'?

Which conclusions do you expect us to draw now that 17 days passed 
since your last posting on this issue? Is it that difficult to get 
the CAN interface working? Please know that three other members of 
the group contacted me privately to ask me if I have any updates on 
this issue.

I appreciate your openess of telling us that CAN does not work as 
documented, but my question now is: does it work at all? And please 
do not try to get away with a simple 'yes' (assuming 'no' is not an 
option), because the next question will be 'how'?

Thanks,

Peter


--- In lpc2000@yahoogroups.com, "philips_apps" <philips_apps@y...> 
wrote:
> Peter,
>      Unfortunately full_CAN mode is not functional in the current 
> devices.  There is a problem with interrupt contention, when this 
is 

> fully characterized it will be detailed in an errata.  In the mean 
> time please do not use full-CAN mode.
> 
> 
> Philips apps.
> 
> --- In lpc2000@yahoogroups.com, "embeddednut" <embeddednut@y...> 
> wrote:
> > Does anyone have an example program yet for using the CAN 
> interfaces 
> > on the LPC2129? FullCAN mode is what I am trying to use. I have a 
> > simple example going, but reception seems to stops as soon as I 
> > increase bandwidth :-(
> > 
> > Peter

Peter,   
    Two CAN implementations (For the MCB2100 board) are now available 
in the files section.  These were recently finished by Embedded 
Systems Academy.  The "blinky" example provides a minimal 
implementation of the CAN interface while the "Full" CAN example 
gives a close approximation to the Full CAN mode.

Since these examples have just been completed the documentation will 
lag somewhat behind, I will post it as well when it becomes available.


Richard

--- In lpc2000@yahoogroups.com, "embeddednut" <embeddednut@y...> 
wrote:
> To Philips Apps,
> 
> So is CAN on the LPC2000 devices a 'go' or 'no go'?
> 
> Which conclusions do you expect us to draw now that 17 days passed 
> since your last posting on this issue? Is it that difficult to get 
> the CAN interface working? Please know that three other members of 
> the group contacted me privately to ask me if I have any updates on 
> this issue.
> 
> I appreciate your openess of telling us that CAN does not work as 
> documented, but my question now is: does it work at all? And please 
> do not try to get away with a simple 'yes' (assuming 'no' is not an 
> option), because the next question will be 'how'?
> 
> Thanks,
> 
> Peter
> 
> 
> --- In lpc2000@yahoogroups.com, "philips_apps" <philips_apps@y...> 
> wrote:
> > Peter,
> >      Unfortunately full_CAN mode is not functional in the current 
> > devices.  There is a problem with interrupt contention, when this 
> is 
> > fully characterized it will be detailed in an errata.  In the 
mean 
> > time please do not use full-CAN mode.
> > 
> > 
> > Philips apps.
> > 
> > --- In lpc2000@yahoogroups.com, "embeddednut" <embeddednut@y...> 
> > wrote:
> > > Does anyone have an example program yet for using the CAN 
> > interfaces 
> > > on the LPC2129? FullCAN mode is what I am trying to use. I have 
a 

> > > simple example going, but reception seems to stops as soon as I 
> > > increase bandwidth :-(
> > > 
> > > Peter

Hi Richard, Robert, ESAcademy,

--- In lpc2000@yahoogroups.com, "philips_apps" <philips_apps@y...> 
>     Two CAN implementations (For the MCB2100 board) are now 
available 
> in the files section.  These were recently finished by Embedded 
> Systems Academy.  The "blinky" example provides a minimal 
> implementation of the CAN interface while the "Full" CAN example 
> gives a close approximation to the Full CAN mode.
> 
> Since these examples have just been completed the documentation 
will 
> lag somewhat behind, I will post it as well when it becomes 
available.

Thank you! Just a little feedback: both programs seem to work fine 
for me. I placed the FullCAN demo program under some stress (1MBit, 
up to 100% busload) and it looks like everything is working fine. I 
did not have a chance to determine the correct number, but it looks 
like CPU usage for this case is still below 50% (at 60MHz).

And our current plan is to run at 125kbit anyway...

Recommendations for CANopen solutions for the LPC2129 anyone?

Peter

--- In lpc2000@yahoogroups.com, "embeddednut" <embeddednut@y...> 
> Recommendations for CANopen solutions for the LPC2129 anyone?

Hi Peter,

What kind of functionality do you require? Single CAN interface 
supported or multiple? Any dynamic configuration (where PSO message 
configuration can change during operation) or static?

We are currently working on a LPC2000 driver for
www.MicroCANopen.com

That will only support one CAN interface and static configurations.

We also know that both CMX and Vector are working on LPC2000 drivers 
for their "full-grown" CANopen stacks.

http://www.cmx.com/CMXCANopen.htm
http://www.vector-cantech.com/index.html?../products?canopen.html

Olaf
Tutor at ESAcademy
www.esacademy.com
www.canopen.us
www.canopenbook.com