AVR to Arduino - VPW

[Tazzi]

Ahhh, 7volts rings a bell. Cheers!

[antus]

Technically for vpw your meant to smooth the wave out also - its in the elm schematic where they switch also a 5v line to provide a step but all the elm clones omit that part and seem to get away with running a square wave. Not idel for signal quality of the bus though.

[Tazzi]

Technically for vpw your meant to smooth the wave out also - its in the elm schematic where they switch also a 5v line to provide a step but all the elm clones omit that part and seem to get away with running a square wave. Not idel for signal quality of the bus though.

Haha.. of course they would skip out on things related to "quality".

[VL400]

That's for edge rate control to help with emc compliance. It reduces the high frequency components. Its of course not high on the clones list of requirements.

[antus]

Heres 2 documents that might help. Firstly the elm schematic, if you wanted to implement the shaped corners you could build all the hardware on pin 3,4,11,13.

Or if you wanted off the shelf hardware that fits that schematic, but omits pins 3 and 13, heres a reference. You'd need a pickit 2 or 3 and the pic tools to build firmware though.

[VL400]

Oh that circuit is for selecting j1850 pwm or j1860 vpw, nothing to do with edge control. Pwm is 5v while vpw is 7v.

[Tazzi]

Heres 2 documents that might help. Firstly the elm schematic, if you wanted to implement the shaped corners you could build all the hardware on pin 3,4,11,13.

Or if you wanted off the shelf hardware that fits that schematic, but omits pins 3 and 13, heres a reference. You'd need a pickit 2 or 3 and the pic tools to build firmware though.

The ELM devices are looking better and better. So a PICkit 2 Microcontroller Programmer would do the job I guess. And can use the MPLAB X IDE that comes with it I guess.

[antus]

Oh that circuit is for selecting j1850 pwm or j1860 vpw, nothing to do with edge control. Pwm is 5v while vpw is 7v.

ahh ok.

The ELM devices are looking better and better. So a PICkit 2 Microcontroller Programmer would do the job I guess. And can use the MPLAB X IDE that comes with it I guess.

yep thats right. Different pins and clock speed to the original code, but might not be so hard to adapt to this hardware. Not all elm clones are the same, but I think they mostly would have the programming port and the rest of it you know will be compatible even if it has a different pcb layout and usb front end chip. You should check to see whats inside one of yours, if you have the programming header and a square pad down one end you should be able to just hook up and use it.

[Tazzi]

Im a bit lost on how to edit the prescaler in the coding.. and even then, will have to modify the timing values for each parameter (eg SOF). Sooo.. might just attempt swapping out the 16mhz crystal for the 4mhz one.

[antus]

Just reading back over your code

// convert microseconds to counter values - Prescaler calculation here
#define us2cnt(us) ((unsigned int)((unsigned long)(us) / (1000000L / (float)((unsigned long)MCU_XTAL / 1L)))) //So, converting microseconds.. into.. ??
// So it is: (int)(long)microsecond/ (1000000L/ (float)(long 4000000 / 1L)) - Im not sure what the L stands for? hmm...

the timer is counting ticks, which are related to cpu speed. so this is defining a macro us2cnt(us) or microsecongs to count (number of microseconds). So...

#define WAIT_100us us2cnt(100) // 100us, used to count 100ms - so all that us2count is converting 100microseconds to 100milli seconds? Is that right?

its converting a measure of 100 microseconds to number of counts at the current clock speed to load the timer with.

When MCU_XTAL becomes faster the resulting number of counts will also enlarge automatically, but at some point the number of counts overflows the size of the counter. You can either calculate it and see if the number will fit or make a test app that just flips a pin every 100us using this function and then watch on a scope to test for accuracy.