Long pulse accurrancy

[ewaldharmsen]

I am on 147 RC8.
I use a sonoff 4CH device to control my sunscreen.

I use longpulse for it.

However it seems to be not very accurate.
When I send /control?cmd=LongPulse,15,1,1 it runs for 2.3 seconds
When i send a longpulse of 5, it runs for 5.5 seconds.

Is there a way to make a long pulse of 1 second really 1 second?

[grovkillen]

I don't know the behind the scenes code that is handling this but you could try to use rules instead.

Code: Select all

On 1secondPulse do 
  gpio,12,1
  timerSet,1,1 //timer 1 set for 1 second
endon
 
On 5secondPulse do 
  gpio,12,1
  timerSet,1,5 //timer 1 set to 5 seconds
endon
 
On Rules#Timer=1 do  
   gpio,12,0
endOn

Trigger through:
http://<espeasyip>/control?cmd=event,1secondPulse
http://<espeasyip>/control?cmd=event,5secondPulse

[Martinus]

I am on 147 RC8.
I use a sonoff 4CH device to control my sunscreen.

I use longpulse for it.

However it seems to be not very accurate.
When I send /control?cmd=LongPulse,15,1,1 it runs for 2.3 seconds
When i send a longpulse of 5, it runs for 5.5 seconds.

Is there a way to make a long pulse of 1 second really 1 second?

1 second should be considered short here. So better use the normal pulse command if you need this level of accuracy.
(Longpulses are timed in background and with single threaded processing this can be inaccurate, unless it's build with hardware timers and interrupts)

[ewaldharmsen]

Thanks all!

I am using normal pulse now, that seems to work better. Problem solved

[toffel969]

Thanks all!

I am using normal pulse now, that seems to work better. Problem solved

Yes normal pulse is much better for accurate timing. The reason for that accuracy is also the biggest downside of normal pulse: it blocks the ESP --> rules and webinterface do not respond in that time. AFAIK even the Timers in Rules get delayed. So its your choice: accurate but blocking, or less accurate and non-blocking

[ewaldharmsen]

clear explanation, thanks!