• 27.04.2024, 22:06
  • Register
  • Login
  • You are not logged in.

 

Dear visitor, welcome to Aqua Computer Forum. If this is your first visit here, please read the Help. It explains how this page works. You must be registered before you can use all the page's features. Please use the registration form, to register here or read more information about the registration process. If you are already registered, please login here.

cptninc

Full Member

Is it possible to make a liveliness watchdog in Playground?

Thursday, August 11th 2022, 4:29am

Had a weird failure today: My D5 Next stopped updating the temperature that it was reporting to aquasuite but was working normally aside from that. It responded to inputs from aquasuite, and it showed the correct temperature on its display, but the temperature it was reporting back to the computer was stuck. When I noticed, I tried restarting the service and that had no effect. Rebooting the PC fixed it.

I have all of the fans keyed to this temperature. That means when this temperature is incorrect, the loop itself could overheat. This failure mode also means that the aquasuite service does not change the sensor to its fallback value because aquasuite thinks everything is operating normally.

I was thinking that one way to make this a safer failure in the future would be to set a watchdog timer which would change the temp to a fallback value if the sensor value has not changed over a specified time interval, like 30 seconds. When things are operating correctly, there will always be some level of variation in temperature. What happened today was it was stuck at exactly 38.08C for several hours.

Is this watchdog something that could be performed via the Playground?

RE: Is it possible to make a liveliness watchdog in Playground?

Thursday, August 11th 2022, 8:16am

Had a weird failure today: My D5 Next stopped updating the temperature that it was reporting to aquasuite but was working normally aside from that. It responded to inputs from aquasuite, and it showed the correct temperature on its display, but the temperature it was reporting back to the computer was stuck. When I noticed, I tried restarting the service and that had no effect. Rebooting the PC fixed it.

I have all of the fans keyed to this temperature. That means when this temperature is incorrect, the loop itself could overheat. This failure mode also means that the aquasuite service does not change the sensor to its fallback value because aquasuite thinks everything is operating normally.

I was thinking that one way to make this a safer failure in the future would be to set a watchdog timer which would change the temp to a fallback value if the sensor value has not changed over a specified time interval, like 30 seconds. When things are operating correctly, there will always be some level of variation in temperature. What happened today was it was stuck at exactly 38.08C for several hours.

Is this watchdog something that could be performed via the Playground?


I love getting my little grey cells working in the morning:)

So this is a quick stab at a watchdog sensor that I think will do what you have asked for.

https://www.youtube.com/watch?v=mxoOKzBpcnA

cptninc

Full Member

Thursday, August 11th 2022, 2:13pm

Wow - awesome!

This is perfect. Thank you!

Thursday, August 11th 2022, 3:44pm

Wow - awesome!

This is perfect. Thank you!


I was still on my first coffee of the day lol

cptninc

Full Member

Thursday, August 11th 2022, 5:55pm

lol. I'll save the really tough ones for when you've had three cups.

After watching it run for a bit, I made one small change. I multiplied the D5N temp by 100 to try and increase the effective resolution of the trend function. Without the multiplier, small variations of .01C didn't seem to be registering on the trend. Adding the multiplier seems to have cleared that up.

Thursday, August 11th 2022, 6:55pm

lol. I'll save the really tough ones for when you've had three cups.

After watching it run for a bit, I made one small change. I multiplied the D5N temp by 100 to try and increase the effective resolution of the trend function. Without the multiplier, small variations of .01C didn't seem to be registering on the trend. Adding the multiplier seems to have cleared that up.


Cool.

Hadn't even thought about that :)

cptninc

Full Member

Thursday, August 11th 2022, 8:13pm

I had initially assumed that the calculation was handled to more sig figs than on display, but then I noticed that the switch was triggering sometimes even when I saw the raw temp value moving .01C. Adding the multiplier doesn't add any precision to the temp value, but it effectively does add them to the trend calculation. The .01C changes that had been shown on the Trend function as 0 are now getting processed by the Equals function as non-zero.

Thursday, August 11th 2022, 8:18pm

I had initially assumed that the calculation was handled to more sig figs than on display, but then I noticed that the switch was triggering sometimes even when I saw the raw temp value moving .01C. Adding the multiplier doesn't add any precision to the temp value, but it effectively does add them to the trend calculation. The .01C changes that had been shown on the Trend function as 0 are now getting processed by the Equals function as non-zero.

[attach]9593[/attach]


I think another way that might achieve the same result is to add a delay step between the Trend and the A=B steps.

And this is what I like about the Playground :)