Hysteresis - above and below the setpoint

[ufergus]

I would like to add my two cents to the discussion. Before switching my setup to BrewBit, I was using an Auber controller to control my heating and cooling outputs. I was able to configure the Auber with two set points, one for heating and one for cooling each with their own hysteresis values. For me, these options would be ideal for the BrewBit. For now, the 1.3.0 update has fixed the heating/cooling fight condition due to overshoot which was a show stopper. Thank you.

[hebnern]

So it sounds like you want the setpoint associated with the output rather than the controller. We discussed this a lot before releasing, and ended up associating it with the controller since it makes for a simpler setup. Basically the user just says "keep my beer at X degrees" and we try to make that happen. It was a conscious to go with simplicity at the expense of flexibility. I just feel that it would really complicate the setup process especially when you consider temp profiles. To support this, you would have to input profiles for both heating and cooling, which would add more complexity to an already somewhat busy UI. I am not saying I am against this, but I want to consider all of the implications before I commit to making any changes.

[ufergus]

I can achieve the equivalent behavior with the current code base, I just wanted to provide some additional feedback. Believe me, I understand the keep it simple game plan. The other option that would give a similar result would be to have the same temp sensor be an input to either controller. It seems like the controller is tied to a temp sensor, but if you could tie both controllers to the same sensor, you could achieve the result of having a set point per output.

[SLOBrewer]

Hmm, thats interesting. Yeah a hysteresis of 0.5 should trigger it at 60.5. I'll take a look.

Any chance to take a look? The whole fermentation went that way: https://brewbit.com/dashboard/devices/588/sessions/2294?token=db87c63217bf12a6e49df119cf541ef3

It looks like the cooling is turning on 1 degree above my setpoint (2 x the 0.5 degree hysteresis setting) and cutting out 0.5 degrees below my setpoint (1 x the 0.5 degree hysteresis setting).

Regardless of the other discussions on this thread, that seems to be totally different than any intended behavior.

[jamin]

Could the discrepancy be the temperature absorption of the liquid affecting the reading? Either way, that is a TIGHT temperature range.

What kind of variance is the goal? At what point does the inefficiency of repeatedly turning the compressor on/off outweigh a tighter variance?

[SLOBrewer]

Could the discrepancy be the temperature absorption of the liquid affecting the reading?

I don't believe so. I'm eagerly awaiting the new functionality to show exactly when the heating and cooling cycles are on but if you look closely to the data you can see that at exactly 1 °F above the setpoint (2x the hysteresis setting) the fridge is kicking on and the temperature drops very quickly. The fridge kicks off and you get a small amount of overshoot at 0.5 °F below the setpoint (1x the hysteresis setting.) If the fridge really were kicking in as it was supposed to at 0.5 °F above the setpoint then you wouldn't see that slow rise followed by a sharp drop that's delayed and exactly coincides with the 2x hysteresis point.

What kind of variance is the goal?

The goal should be that the fermentation temperature be kept within 1x the hysteresis value of the setpoint. In this case with a setpoint of 60 °F I would expect the fermentation to stay within a band of 59.5 °F and 60.5 °F. Currently fermentation is staying within a band of 59.5 °F and 61 °F.

At what point does the inefficiency of repeatedly turning the compressor on/off outweigh a tighter variance?

That's the point of letting me chose the hysteresis value. If I wanted to have the compressor turn on and off less frequently I would chose a greater hysteresis value. I could chose 1 °F as the hysteresis value but based on what I've seen here that would result in the fermentation band being 3 °F wide ranging from 59 °F up to 61 °F. That's just weird.

In summary, I can make this work with the current behavior but it doesn't match what any other temperature control system does. Furthermore it doesn't match how the BrewBit team said they intended for the hysteresis value to work.

[SLOBrewer]

BrewBit Team,
Here's a guess on what might be going wrong. Since I have both a heating and cooling control on this fermentation, is it possible that this is related to something you've done to avoid the tight loop of the two battling it out? Obviously I wouldn't want to have my M-T cool down so much that the heater kicks in, which raises it to the point that the fridge kicks in to cool, etc.

[hebnern]

I'm not seeing anything obvious. I'll dig a bit deeper this weekend.

[SLOBrewer]

Okay, interesting. While I've set the hysteresis for both devices to 0.5 °F on the website it's still showing 1.0 °F on both devices. That explains the fermentation behavior but it doesn't explain why the setting didn't get sent down to the devices. Clicking update device again seems to have pushed the correct setting down.

Sounds like there isn't a problem with the temp control code that interprets the values but that there was some case where either the hysteresis wasn't sent down to the device or it got reset at some point.

[hebnern]

I'll bet it got reset after the 1.3.0 update since that will clear all settings. We probably need to add settings syncing when the device first connects to the server to ensure that they are always in sync.

[SLOBrewer]

I think you've nailed it. I didn't realize that reset the value and when I checked the website it still showed the old value. If the website is going to show values it seems like they should be in sync with what's on the device.

[grantanddenise1]

I am currently using s/w v1.3.1. I have a fermentation chest freezer with a heat pad and I am using a single thermowell probe to control heating and cooling. I have a set point of 10.0C and hysteresis of 0.3C with a 10 minute delay on cooling and 3 minute delay on heating. The session data indicates that the temperature is cycling between 9.6C and 10.2C, which seems pretty good. However why does the cooling continue after dropping below the set point and why does the heating continue after rising through the set point. I am constantly cycling between hearing and cooling. Furthermore why is the cooling being switched on when the temperature hasn't even hit 10.2C when the S+H is 10.3C? It appears to be behaving like the trigger points are still SP +/- H/2. I thought this had been fixed in 1.3.0. Please clarify.

[hebnern]

Can you post a link to your session so I can take a look?

[grantanddenise1]

Sorry I have had a few issues so have lost my session data. I think may be I am a little confused. My comments above were when I was using the brew bit in PID mode. So my question is hysteresis applied in PID mode or only in on/off mode?

[hebnern]

It is applied in both PID and ON/OFF

[grantanddenise1]

So recently I am noticing that at times the output does not switch on when it should i.e.. when the temp reaches the set-point plus hysteresis value. For example I currently have a set-point of 19.0C with a hysteresis of 0.1C and therefore would expect the cooling output to be switched on when the temp goes above 19.1C and most of the time it does just that. But occasionally it does not go on. As you can see in the attached profile, the temperature rose to over 19.6C before the cooling output switched on. Clearly this is undesirable and needs to be fixed. Why does it only happen occasionally?

[hebnern]

Hmm, that is odd. I am not sure what that is about. Is this in PID mode? We are planning to add more logging data to give us better insight into what the controller is doing internally that might help track this sort of thing down a bit better.

[grantanddenise1]

No this is on/off mode.

[grantanddenise1]

Any thoughts or fixes?