Omnistat2 Auto Learning and Automation

[mattw22]

I know the Omnistat2’s have an auto learning feature that can figure out how long it takes to reach a desired temperature. My question has to do with setting up rules..

Is there an automation rule that can tell the thermostat to auto reach a temperature by 5am? Or do I have to make a setpoint at let’s say 4am to reach 72 degrees by 5am?


I guess I’m wondering if the Omnistats are smart enough to see that I have a Automation rule/setpoint for 72 degrees at 5am and to auto start at a certain time to reach that goal.

 

[RobertPickral]

Matt,
The feature that you are referring to has to do with cycles in the HVAC system. The omnistat2 will auto adjust the cycles in your system to make it run more efficiently. Bascially, HAI has a patent on an algorithem that monitors what is happening and can adjust the run time of the cylces to keep your actual temperature from fluctuating too far above or below your desired set point. (Other thermostats will overshoot the desired set point. EX: if your desired set point is 72, the thermostat will actually overshoot it and cool to 71 or 70 in anticipation that the system will sit dormant for several minutes. Then your actual temp will overshoot in the other direction by 1 or 2 degrees before the system kicks on again.) The learning feature of the omnistat2 is a comfort feature that will keep your temp within a half of a degree of your desired set point.
Regarding your solution, yes, HAI has the logic to obtain your desired goal.

 

[sic0048]

Matt,
The feature that you are referring to has to do with cycles in the HVAC system. The omnistat2 will auto adjust the cycles in your system to make it run more efficiently. Bascially, HAI has a patent on an algorithem that monitors what is happening and can adjust the run time of the cylces to keep your actual temperature from fluctuating too far above or below your desired set point. (Other thermostats will overshoot the desired set point. EX: if your desired set point is 72, the thermostat will actually overshoot it and cool to 71 or 70 in anticipation that the system will sit dormant for several minutes. Then your actual temp will overshoot in the other direction by 1 or 2 degrees before the system kicks on again.) The learning feature of the omnistat2 is a comfort feature that will keep your temp within a half of a degree of your desired set point.

I assume there is some sort of mechanism to prevent the system from short cycling however?

I actually just replaced my thermostat because it was so sensitive to temp changes and tried to keep it at exactly the setpoint, that it would run a minute or two and then rest only a minute or less before turning on again. From what I understand, this short cycling of the unit is very bad. So I replaced it with a thermostat that allows me to set the min run time as well as the min rest time. I might have bigger temperature fluctuations now, but at least I am not short cycling the system. (What I really need to do is replace my 30 yr old single pane drafty windows, but that's another project ).

So if the algorithem tries too hard to keep the temp to within a 1/2 degree, I could see a problem if it cause the system to short cycle.

 

[RobertPickral]

Sic0048,

My explanation wasn't exactly the greatest. The learning feature has do with the patterns in one's home. The algorithem will learn these patterns and adjust accordingly. Yes, short cycling is not good and that isn't really what the algorithem does. (Although my poor explanation made it seem that way.)

 

[sic0048]

Thats fine. I really don't need an better explaination. It just caught my attention since I was trying to get away from a system that did short cycle. As long as there is something to prevent short cycling (and obviously there is), that is great.

As a side note, I really like the HAI Omnistat2 thermostats. The RC-2000 was my number one choice in that price bracket (competing with the RCS TR-40 and Aprilaire 8870). But I was able to find the TR-16 on a clearance sale at my local supply house. That put it well below the price bracket of the Omnistat2 and too good of a deal to pass up. But I had looked long and hard at the RS-2000 and it is an excellent product.

 

[RobertPickral]

Here's a better explanation:
The Omnistat2 operates each HVAC system uniquely. Each HVAC is unique by the rate it heats and cools a dwelling based mainly on the size and layout of the dwelling. Ambient sources also play a large factor into this rate (such as outside temperature, insulation, the oven, etc.). The OMNISTAT2 tracks how well the HVAC heats and cools in the individual house and compensates using the HVAC system accordingly. The Omnistat2 “learns” how well your HVAC operates in your house with the current conditions.

Example:

It is hot outside in the afternoon (90+). The HVAC would keep the AC on longer because it knows it will not take long for temperature to rise after it is turned off. This will gradually change as the day goes on and the house does not heat as fast.

 

[Ryetree1]

Robert is correct in his latest brief summary of the Omnistat2 learning. It does not effect short cycling. Another example: You decide to turn on the oven (in a small home) introducing an outside heat source. The Omnistat2 recognizes that heat is rising faster than it is used to for that time of day, so it may turn off the heat sooner so there is not as much overshoot.

The Omnistat2 has configurable settings for the minimum on and off times for both heat and cool so the user/installer can set how long the cycle they want. A shorter cycle (not to be confused with a really short cycle that can damage a HVAC) typically maintains the temperature closer to the set point, where a longer cycles let the temperature drift from the set point a bit farther. In most cases a longer cycle is usually more efficient and will lengthen the HVAC lifetime. The Omnistat2 takes precautions to not let the HVAC continuously short cycle in a way that will damage a HVAC.

Matt,
To answer your original question, the Omnistat2 does not look at the schedule to pre-cool/heat to reach a desired set point. You normally would have the set point change earlier than u want the temperature to be that set point. Its not that the Omnistat2 is not smart enough (you hurt its feelings), but rather chooses not to. If the thermostat would do this pre-cool/heating, it would make a drastic difference if the thermostat was programmed by its internal schedule (where it would know what was coming up), or an automation controller (where it does not know what is coming up).

Thanks,
Ryan
Embedded Systems Engineer
Home Automation Inc.

 

[mattw22]

Ryan-

So your saying that if I use the internal scheduler over the controller rules. And have a temperature change set for 5am for let’s say 72 degrees. The Omnistat2 will ramp up before 5am to meet that temperature?

I would be putting probably 15 of these in a warehouse. I could use the internal scheduler if it will give me better efficiency overall. My other question is if I use the internal scheduler and at some point needed to make a rule that lets say changed the temperature for 7-8 hours during the day, would the omnistat make the setpoint change and then default back to its internal program at the next interval?

Thanks

 

[PaulD]

It is hot outside in the afternoon (90+). The HVAC would keep the AC on longer because it knows it will not take long for temperature to rise after it is turned off. This will gradually change as the day goes on and the house does not heat as fast.

How does the thermostat know the outside temperature unless you have an outside temperature tied in to the thermostat?

 

[Ryetree1]

Paul,
You can of course connect an outdoor sensor for accuracy, but the Omnistat2 does not direct use that for its learning. It instead uses the heating patterns as a factor for its algorithms. If the heat is rising a lot faster than normal, the Omnistat2 may not know it is very hot outside, or perhaps you are having a party with 100's of people in the house. It does not matter because it aims to keep the temperature correct and does not differentiate with what the actual additional heat source is.

Thanks,
Ryan
Embedded Systems Engineer
Home Automation Inc.

 

[aard3]

Matt,

No, the Omnistat2 by itself will not do that. It will not turn on the heat in anticipation of a setpoint change that is going to happen at a certain time in it's scheduled programming.

As to your second question, If you command an OmniStat to say, 72 degrees at 1:00pm, and at 3:00pm the Omnistat is programmed to goto 75, it will go to 75.

Ryan-

So your saying that if I use the internal scheduler over the controller rules. And have a temperature change set for 5am for let’s say 72 degrees. The Omnistat2 will ramp up before 5am to meet that temperature?

I would be putting probably 15 of these in a warehouse. I could use the internal scheduler if it will give me better efficiency overall. My other question is if I use the internal scheduler and at some point needed to make a rule that lets say changed the temperature for 7-8 hours during the day, would the omnistat make the setpoint change and then default back to its internal program at the next interval?

Thanks

 

[dutchyn]

I'm putting in Omnistat thermostats also, tied to an M1. But, I'm hoping that I can leverage the thermostat intelligence by having the Elk change the programmed points, rather than the immediate setpoints. That way, the thermostats can figure out how to warm a bedroom up from 17© to 21© by 8AM, rather than have to calculate and tune it myself. Based on my reading of the Omnistat RS232 protocol, it seems doable - just set the correct internal registers in the thermostat. So Elk vacation mode tells the thermostat to change each schedules' daytime temperature from 20 to 15 (say).

Yes?

Chris D.

 

[Ryetree1]

Matt & Chris,
The Omnistat2 will NOT start early for an future setpoint change. There is no difference in the internal scheduler and having an automation control do it. The automation controller has more flexibilty to do things such as: If 2 windows are open, change setpoint by x degrees, or if security is armed, leave in away setpoints.

Matt,
You are welcome to set rules in the automation system as much as you like to achieve a ramp by 1 degree steps. If doing so, you may want to completely disable the internal schedule or be careful mixing the two. The Omnistat2 will only change the setpoints due to its internal schedule only when the events occurs. If you have the internal schedule set to change the cool to 74 degrees at 1:15pm, it will only do this at 1:15pm and not any time after. If an outside source (manual turning know, or automation controller) changes the setpoint, that setpoint will stick until the next scheduled event or another manual setpoint change.

Thanks,
Ryan
Embedded Systems Engineer
Home Automation Inc.

 

[mstarks01]

Will the omnistat ever turn the heat on before the setpoint is reached as a function of the anticipation mode? I noticed last night that the heat kicked on above the setpoint rate and stayed on awhile.

 

[cornutt]

Will the omnistat ever turn the heat on before the setpoint is reached as a function of the anticipation mode? I noticed last night that the heat kicked on above the setpoint rate and stayed on awhile.

Interesting. I haven't seen mine do that. But I supposed it's possible; if the thermostat noticed the room was losing heat rapidly (maybe because of a door open), it might do that. Someone from HAI would of course know more than I would.

One thing I'd be interested in is learning a bit more about what the EEC parameters do. I take it from the manual description that they provide a "hint" of sorts to the thermostat of how slowly or quickly the stat should expect the room to heat or cool. I'm wondering because I've noticed that my RC1000s do in fact overshoot a bit; if the setpoint is 70, they don't usually stop calling for heat until the display shows 71 or sometimes 72. I take it that they are trying to optimize the cycle -- heat pumps are, of course, fairly slow to heat compared to gas furnaces.

As far as setpoint anticipating, I wrote some rules on my OPII to do that. I have an ER temp sensor connected to the system, and I have written some rules like this:

17.	TIMED Arrive Time 60 minutes before MTWTFSS
		AND IF Outdoor Temp Sensor CURRENT READING IS LESS THAN 25 
		OR
		AND IF Outdoor Temp Sensor CURRENT READING IS GREATER THAN 95 
			THEN RUN Evening Thermostat Settings

18.	TIMED Arrive Time 30 minutes before MTWTFSS
		AND IF Outdoor Temp Sensor CURRENT READING IS LESS THAN 38 
		OR
		AND IF Outdoor Temp Sensor CURRENT READING IS GREATER THAN 85 
			THEN RUN Evening Thermostat Settings

19.	TIMED Arrive Time 0 minutes before MTWTFSS
			THEN RUN Evening Thermostat Settings

The "Arrive Time 60 minutes before" etc. are user settings. I have a bunch of buttons defined that set them according to what time I expect to be home:

37.	WHEN Home at 2:00
			THEN SET Arrive Time 60 minutes before TO 1:00 PM
			THEN SET Arrive Time 30 minutes before TO 1:30 PM
			THEN SET Arrive Time 0 minutes before TO 2:00 PM

38.	WHEN Home at 3:00
			THEN SET Arrive Time 60 minutes before TO 2:00 PM
			THEN SET Arrive Time 30 minutes before TO 2:30 PM
			THEN SET Arrive Time 0 minutes before TO 3:00 PM

and so on. I run the one I want before I leave in the morning, or I call the system during the day and run a button if my plan changes.