I will say that it is very likely that if you switch to a PREEMPT RT
kernel, that determinism down to 1ms should be easily achievable. So long
as you do not have too much in the way of background processes running at
the same time.
On Thu, Dec 3, 2015 at 7:10 AM, William Hermans <yyrk...@gmail.com> wrote:
> *The GPIO is high for about 1.5 Milliseconds. The usleep()-function is not
>> linear, so it did not work with any other value e.g. usleep(500). Sometimes
>> I decrease the value by one (1 microsecond) but the time decrease by 100
>> microseconds.*
>> *Has anybody an idea what could be wrong? *
>>
>
> First of all, usleep() is not your problem. The problem most likely is
> related to kernel latency. You could improve this by using a PREEMPT RT
> kernel, but that will only at best give you 95% determinism.
>
> Second, since you have not made your code public, we can not tell if
> servo.setValue() is well written or not. In the context of determinism. I
> will say however if you're not using mmap() with /dev/mem/ on the GPIO
> registers directly. Your code will be "slow", and not very deterministic.
>
> Lastly, if you need absolute determinism, you had better start using the
> PRUs . . .
>
>
> On Thu, Dec 3, 2015 at 4:30 AM, <andreas.bu...@hs-owl.de> wrote:
>
>> Hi,
>>
>> I am trying to control a servo with my BeagleBoneBlack using Angstrom. I
>> must have a high-level on a GPIO for 1 millisecond. I use the following
>> simple code:
>>
>> while(1){
>> servo.setValue(BlackLib::low);
>> usleep(19000);
>> servo.setValue(BlackLib::high);
>> usleep(1000);
>> }
>>
>> The GPIO is high for about 1.5 Milliseconds. The usleep()-function is not
>> linear, so it did not work with any other value e.g. usleep(500). Sometimes
>> I decrease the value by one (1 microsecond) but the time decrease by 100
>> microseconds.
>> Has anybody an idea what could be wrong?
>>
>> --
>> For more options, visit http://beagleboard.org/discuss
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "BeagleBoard" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to beagleboard+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to beagleboard+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
