Honeywell thermostat turns on compressor 1 minute before schedule off

heatmenow

I recently purchased and had professionally installed a Honeywell "VisionPRO® 8000".

It worked well and I could set a manual schedule as desired, which it followed.

However, almost every day, when the heat is scheduled to turn off at 9:00 a.m., the thermostat will turn on the compressor at 8:59 a.m., only to run briefly before shutting off.

I know I did not purchase a "smart" thermostat, but this is just ludicrous. Why can't the thermostat figure it needs to be off in less than a minute? It's like absolutely zero thought went into the programming of this device.

mattmia2

Set it to setback at 8:50? It evolved from an electromechanical device, there will be some boundary conditions that could be improved with a microcontroller that will get missed.

Jamie Hall

NO time based scheduler can look into the future. None. In fact, most humans can't. All it knows is that you have a certain target between two times, and the measured target is below that value, so turn on the heat.

What happens if it is a colder day, and the heat turns on two minutes early? Five minutes early?

Yes, it is possible to design and build predictive controls (up to a point -- there are limits). They are a good deal more difficult to do, however, and have their own quirks. You wouldn't be happy paying for them.

heatmenow

Jamie Hall said:

NO time based scheduler can look into the future. None. In fact, most humans can't.

Respectfully disagree. The ability to see into the future and objectively evaluate it is what makes humans human. To wit, I made this post explaining the situation based on temporal constraints.

Jamie Hall said:

What happens if it is a colder day, and the heat turns on two minutes early? Five minutes early?

I believe that would be reasonable, however I'm not an expert and do not know the recommended minimum time a compressor should run. Less than a minute seems like it's not optimal, possibly even damaging. I'm seeking clarification on this point.

Jamie Hall said:

Yes, it is possible to design and build predictive controls (up to a point -- there are limits). They are a good deal more difficult to do, however, and have their own quirks. You wouldn't be happy paying for them.

I agree: I don't want any "smart" gadgets in my home, and perhaps I got what I wanted.

It just confuses and vexes me as a programmer that more thought was not put into the design of the scheduler on this thermostat.

Edit: I also want to add that the inability to contact anyone at Honeywell who cares, let alone understands what I'm describing, is beyond appalling and frankly kind of embarassing. Don't think I will be buying Honeywell devices again based on their support reception.

Thank you for reading and commenting.

pecmsg

Sorry but ALL thermostats are basically DUMB. If it calls for heat 30 seconds or 30 minutes before the next schedule it will bring it on. 

LRCCBJ

heatmenow said:

However, almost every day, when the heat is scheduled to turn off at 9:00 a.m., the thermostat will turn on the compressor at 8:59 a.m., only to run briefly before shutting off.

.

I am a bit more curious as to what exactly is going on here. If this is not a smart thermostat, and I believe it is not, then it will engage the compressor whenever the indoor temperature falls below the setpoint (with the associated differential). This COULD happen at 8:59 but it would absolutely be a coincidence because the thermostat does not know that it is scheduled to turn off at 9:00 because it is NOT a smart thermostat.

Therefore, on the following day, it is absolutely impossible for this thermostat to engage at 8:59 am because the statistical odds of that happening are probably 1 in 1 million.

So, it appears that the data you have provided regarding "almost every day" is incorrect.

ethicalpaul

But they don't have to be. This logic could easily be added, but every addition of logic like this increases the chance for failure and adds to QA costs and the chance of unintended consequences (edge cases)

mattmia2

Running the compressor for a minute won't hurt it. It wastes a little energy but it won't hurt anything.
With a gas or oil burner that could be more of a concern where the burner fires enough to start heating the HX but not to get it hot enough to evaporate any condensation or to start the blower.

Jamie Hall

ethicalpaul said:

But they don't have to be. This logic could easily be added, but every addition of logic like this increases the chance for failure and adds to QA costs and the chance of unintended consequences (edge cases)

Sort of. It would take quite a bit of additional computing horsepower -- easy enough to do, perhaps, but why?

Off topic -- I am rather aware of @heatmenow 's comment on predicting the future and how that makes us human. Curiously, just last Sunday the sermon I preached was on exactly that topic -- and how foolish it is in some senses. True, we need to see where we are and where we think the future might be going -- but as to actually predicting it... no. But you all probably don't want a sermon!

ethicalpaul

I don't see additional computing horsepower required. It has to be easy to pick a time in the future and see what the setpoint is going to be at that time...after all, the thermostat displays it during programming.

You pick a possibly adjustable "look ahead time" (60/120/300/whatever seconds) and before you begin a call for heat you ensure that the setpoint isn't going to be less than the current temperature at that time. Cooling works the opposite way, something all heating/cooling thermostats already have to deal with.

Why? Because the current behavior annoys some people (as evidenced by this thread). Whether it annoys enough people for a marketing department to ask the engineering department for the feature, I have no idea.

Jamie Hall

Seems to me, @ethicalpaul , that that would just offset the short time problem...

To really do it right -- and with thermistor based thermostats rather than bimetal -- you need to measure two things: present error (how far you are from the setpoint) and rate and direction of change in that error. Then you really can write a predictive algorithm which would be easy enough to program...

Or you can look up from you iPhone and get up from your chair and walk over to the T87 and change the setting...

ethicalpaul

No, it wouldn’t offset the problem. Changing the thermostat would though. 

But don’t look at me, I’m not bothered by this issue—I have a CycleGard 😂