# The Money Mustache Community

## Learning, Sharing, and Teaching => Ask a Mustachian => Topic started by: dragoncar on November 15, 2016, 05:04:25 PM

Title: Manual Thermostat Overshoot
Post by: dragoncar on November 15, 2016, 05:04:25 PM
I searched around, but couldn't find a good answer to this.  Maybe I'm using the wrong terms.

Ignoring the merits of the absolute temperature values (an endless debate), lets say it's 63 in my house and I want it to be 68.  If I just turn on my furnace to 68, it will warm up to 69-70 and shut off.  Then all my thermal mass will bring the temp quickly down to 66, and the furnace will turn on again.  Thus, it will cycle on and off for 1-2 hours before it's finally at a steady state.

I suspect "smart" thermostats learn this behavior and perform an overshoot -- if the starting temp was 63 degrees all night and you want to get the house up to 68, run the temp all the way up to 75 so that you have extra energy in the air to bring the thermal mass (e.g., granite countertops) to 68.

This makes you more comfortable faster and also keeps the furnace in the most efficient operating region.  I understand furnaces are inefficient in their first and last few minutes of operation.

So my question is do I actually save any energy doing this manually?  Lets say I wake on on Sunday and it's been 63 all night (thermal mass is at 63).  Can I turn my thermostat up to 75-80 (exact value determined by experimental trial and error) and then set it back to 68 once it turns off?  This will keep my furnace from short cycling while trying to bring the thermal mass up to temp, but on the other hand the temperature differential with the outdoors will be higher, so I'll lose more energy through my walls/windows/attic.

Do I come out ahead, behind, or negligible?

If this hasn't already been analyzed, I'm willing to do some first approximation calculations myself, but any info on the inputs is appreciated.  For example, what does the efficiency curve of an operating furnace look like?  (e.g., 1st min: 50%, 2nd min 70%, 3rd min+ 80-90%)
Title: Re: Manual Thermostat Overshoot
Post by: Bracken_Joy on November 15, 2016, 05:05:56 PM
Following, because I wonder this too. I've seen both sides argued.
Title: Re: Manual Thermostat Overshoot
Post by: MDM on November 15, 2016, 09:02:59 PM
Unless the furnace can deliver an adjustable fuel flow, you'll always have the "on/off" control you describe, with the room temperature moving back and forth around the nominal temperature setpoint.  The only way to get a steady temperature is to have the combustion heat delivered by the furnace exactly match the heat lost from the house to the outdoors.

Don't know what algorithm smart thermostats use, but Bang–bang control (https://en.wikipedia.org/wiki/Bang%E2%80%93bang_control) is the usual mode for the common thermostat.  If they use "fuzzy logic" or other heuristics, the intent may still be to keep the actual temperature as close to setpoint as possible.  If so, the thermostat would see a large overshoot as a bad thing.

See PID controller (https://en.wikipedia.org/wiki/PID_controller) for what might be done if the fuel gas flow could be throttled, similar to how an automobile cruise control adjusts gasoline flow.

Good question about the efficiency curve - don't know the answer.  I'd guess it's not significant, but real data (if available) beats guesses any time.  Most thermostat/furnace designers probably treat "keeping temperature close to setpoint" as the measure of success.  That would lead to frequent cycling.  E.g., keeping the temperature +/- 0.1 degree would have faster cycling than keeping the temperature +/- 1.0 degree.
Title: Re: Manual Thermostat Overshoot
Post by: Dezrah on November 16, 2016, 07:39:00 AM
I don't know enough to answer your question but I can vouch that thermal mass

We agreed to housesit for my parents for two weeks in winter.  The worst winters in our state never get cold enough to burst pipes, so we just turned off the temperature control completely.  When we got back it was about 50F inside but it was SO COLD.  The couches were cold, the cabinets were cold, the bed was cold.  We literally had to sleep close for warmth that night just to stop shivering.  It took about a full day running the heat and electronics before we felt comfortable again.
Title: Re: Manual Thermostat Overshoot
Post by: Prairie Stash on November 16, 2016, 08:17:47 AM
More important; what's your furnace efficiency? If I have a 95% high efficiency furnace then it means I'm wasting 5% of the natural gas; that 5% is heat that's vented out the stack. Basically with a high efficiency furnace trying to squeeze more out of it is negligible, all the heat that's possible is already going into the house and leaving through the walls/ceilings/windows (low efficiency furnaces can send 40% of the heat through the chimney). The second variable I can think of, I spent about \$600 on my mid range furnaces gas last year. How much did you spend on heating your house; outside of hot water, delivery charges, and basic monthly bills? Assume a few % difference; that's how big/negligible your potential savings could be by decreasing your heat loss.

By far the best approach is to spend the time spent fiddling with a thermostat on home improvement projects (time better spent if you haven't done so already). After you fix all the drafts and leaks the potential savings gets even more negligible (which is a good thing).
Title: Re: Manual Thermostat Overshoot
Post by: AlanStache on November 16, 2016, 08:23:51 AM
The thermal mass may or may not be a factor.  It depends on the how much heat your heater can put out, if it is a modest heater and the home large everything will be brought up to temperature more evenly.  As for being most efficient this is attached with the furnace off.  Maybe the 'best' path would be to gradually heat the home in the morning as the outside heats up too, but then it would be cold until 10am....

I put in a nest a few weeks ago and it does not seem to overshoot more than a degree or so.  Nor have I noticed it cycling.

I am not sure how much the penalty for cycling on/off is beyond wear and tear on the equipment.  But even that if a furnace works for years and years with little maintenance what is the cost of one cycle?
Title: Re: Manual Thermostat Overshoot
Post by: arebelspy on November 19, 2016, 06:34:03 PM
This is an interesting question.  Following!
Title: Re: Manual Thermostat Overshoot
Post by: Metric Mouse on November 20, 2016, 04:29:23 AM
How much did you spend on heating your house; outside of hot water, delivery charges, and basic monthly bills? Assume a few % difference; that's how big/negligible your potential savings could be by decreasing your heat loss.

By far the best approach is to spend the time spent fiddling with a thermostat on home improvement projects (time better spent if you haven't done so already). After you fix all the drafts and leaks the potential savings gets even more negligible (which is a good thing).

This was my thought. Not only is forced air such a terrible way to heat mass (radiant would be much more efficient), the potential gains seem small for the amount of effort.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 20, 2016, 02:11:46 PM
How much did you spend on heating your house; outside of hot water, delivery charges, and basic monthly bills? Assume a few % difference; that's how big/negligible your potential savings could be by decreasing your heat loss.

By far the best approach is to spend the time spent fiddling with a thermostat on home improvement projects (time better spent if you haven't done so already). After you fix all the drafts and leaks the potential savings gets even more negligible (which is a good thing).

This was my thought. Not only is forced air such a terrible way to heat mass (radiant would be much more efficient), the potential gains seem small for the amount of effort.

I mean, I already have a forced air system.  I'm trying to optimize my use of an existing system, not asking about replacing it with radiant.  Fiddling with my thermostat is in no way comparable to all the alternatives suggested.  Sometimes I can see why we get new posters who just disappear after they ask a question.

I've already fixed drafts and leaks
Title: Re: Manual Thermostat Overshoot
Post by: arebelspy on November 20, 2016, 03:28:12 PM
I've already fixed drafts and leaks

But have you found that smell?
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 20, 2016, 07:56:29 PM
I've already fixed drafts and leaks

But have you found that smell?

Haha, I never found it and it went away again, so I'm still suspecting it's either one of my plants that smells different when flowering or a neighbor smoking the ganja.  I've spent extensive time in the crawlspace and attic since then and never smelled anything (although I did find three mummified dead rats in my attic while doing sealing and insulation, no smell).  I blocked off their point of entry and it was not fun cleaning up their little gifts.
Title: Re: Manual Thermostat Overshoot
Post by: rhadams1988 on November 20, 2016, 08:19:34 PM
What a great question! I used to be an energy engineer so I can give you my guess, but I certainly haven't done any calcs on this. I would bet that raising your inside temp to 75 degrees increases your heat loss to the outside pretty dramatically. I think this would end up costing you more in wasted gas than the fact that furnaces are inefficient starting up.

I guess turning it to 75 could be more efficient if your furnace started up literally every minute and then cut off after 30 seconds, but I'm assuming your furnace doesn't start up more often than every 10 minutes (at a minimum). So my advice is that your should just let the thermostat cycle a little more often in the morning until the thermal mass is up to temperature.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 20, 2016, 08:37:27 PM
What a great question! I used to be an energy engineer so I can give you my guess, but I certainly haven't done any calcs on this. I would bet that raising your inside temp to 75 degrees increases your heat loss to the outside pretty dramatically. I think this would end up costing you more in wasted gas than the fact that furnaces are inefficient starting up.

I guess turning it to 75 could be more efficient if your furnace started up literally every minute and then cut off after 30 seconds, but I'm assuming your furnace doesn't start up more often than every 10 minutes (at a minimum). So my advice is that your should just let the thermostat cycle a little more often in the morning until the thermal mass is up to temperature.

Thanks!  Do you have a rough guess for how long it takes a furnace to come up to efficiency?  Like are we talking 30 second or 3 minutes?

If 30 seconds, I totally agree with your guess.
Title: Re: Manual Thermostat Overshoot
Post by: Metric Mouse on November 21, 2016, 06:52:12 PM
How much did you spend on heating your house; outside of hot water, delivery charges, and basic monthly bills? Assume a few % difference; that's how big/negligible your potential savings could be by decreasing your heat loss.

By far the best approach is to spend the time spent fiddling with a thermostat on home improvement projects (time better spent if you haven't done so already). After you fix all the drafts and leaks the potential savings gets even more negligible (which is a good thing).

This was my thought. Not only is forced air such a terrible way to heat mass (radiant would be much more efficient), the potential gains seem small for the amount of effort.

I mean, I already have a forced air system.  I'm trying to optimize my use of an existing system, not asking about replacing it with radiant.  Fiddling with my thermostat is in no way comparable to all the alternatives suggested.  Sometimes I can see why we get new posters who just disappear after they ask a question.

I've already fixed drafts and leaks

Sorry for the confusion. I was not suggesting a replacement system.
Title: Re: Manual Thermostat Overshoot
Post by: Sibley on November 22, 2016, 08:02:05 AM
Slightly off topic but seems relevant. If you have steam boiler and radiators - do NOT use any old programmable thermostat. They cycle too quickly and can damage the boiler. In theory they make programmables that cycle slower and are safe for the boilers, but personal experience has shown they're not great. Stick with manual.
Title: Re: Manual Thermostat Overshoot
Post by: nobody123 on November 22, 2016, 10:02:44 AM
Thermostats a simple state machines that measure the ambient temperature and have some threshold of what it considers within tolerance (varies by manufacturer).  It can only energize or deenergize the heat or compressor control lines to signal what it suggests furnace controller should do.  There is a delay built in so a gust of cold air from opening your front door doesn't instantly flip the heat on.  When the thermostat hits the correct point where it decides to turn off heat or AC, it signals the control board in the furnace.  The furnace then has logic to do a graceful shutdown.  The amount of extra run time to shut down the AC compressor, make sure gas has been burned off, etc., varies by furnance controller.

Learning thermostats might "learn" how quickly it takes your house's ambient air to get from 63 -> 68 and can help in a programming sense (it knows exactly when to turn on the heat so when you return home from work at 5:45 PM it will be exactly 68 degrees).  Some also take humidity into account so you can say "make it feel like 68" and the actual temperature it targets may be slightly above or below to adjust for humidity.  Smarter ones will get the outside temperature from the web to factor that into the equation on how early it needs to start heating or cooling.  But they still can only send the same "heat on/off or AC on/off" signal to the furnace control board.

In general, if you turn your thermostat up to 75, then back down to 68 when it turns off, you just heated your air to 75 when you wanted it to be 68.  Yes, if you have some cold things (granite counters) they will suck heat from the air and the furnace will turn on again sooner, but I would argue that you're better off with an extra cycle or two getting the house stable at 68 than potentially wasting the gas of a long run guessing on how much to overheat your house.  Think about it, in furnace mode you're essentially burning gas at a given temperature and running a fan that uses some electricity to push the air around.  The inefficiency is essentially the time it takes to purge the cold air from the duct work, which is what, maybe the first 15 seconds of a run?  I guarantee the one startup cycle you saved was far outweighed by heating your entire house an extra 7 degrees.

Title: Re: Manual Thermostat Overshoot
Post by: nobody123 on November 22, 2016, 10:05:17 AM
And to clarify, most modern furnace control boards have guards against "short cycling" by ignoring thermostat calls for heat or cool if they are too close to the last run and might damage the equipment.  Assuming you don't have a mis-configured system, you should just let the thermostat do its job.
Title: Re: Manual Thermostat Overshoot
Post by: robartsd on November 22, 2016, 10:30:36 AM
Thanks!  Do you have a rough guess for how long it takes a furnace to come up to efficiency?  Like are we talking 30 second or 3 minutes?

If 30 seconds, I totally agree with your guess.
I would imagine that most furnaces reach operational efficiency in a few seconds rather than a few minutes. I could see residual heat remaining in the furnace and being lost. I would want to get a thermostat with a wider temperature range if the furnace was regularly cycling on again after being off for less than ten minutes.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 11:45:27 AM

In general, if you turn your thermostat up to 75, then back down to 68 when it turns off, you just heated your air to 75 when you wanted it to be 68.  Yes, if you have some cold things (granite counters) they will suck heat from the air and the furnace will turn on again sooner, but I would argue that you're better off with an extra cycle or two getting the house stable at 68 than potentially wasting the gas of a long run guessing on how much to overheat your house.  Think about it, in furnace mode you're essentially burning gas at a given temperature and running a fan that uses some electricity to push the air around.  The inefficiency is essentially the time it takes to purge the cold air from the duct work, which is what, maybe the first 15 seconds of a run?  I guarantee the one startup cycle you saved was far outweighed by heating your entire house an extra 7 degrees.

I've read that it can take 10 minutes to reach peak efficiency, which may be true but it's still possible it reaches 80% peak within one minute.  I'm sure this varies by furnace too!

I did a little test last night but unfortunately had to throw out some good data because I got distracted

Just setting it to 68, leaving it alone, it ran for 15 minutes, off for 5, then on for 10, off for 6, then on for ???... I think over time the off time kept increasing but I'll need more good data.

So id like to compare this to running for 25 straight then off for 11.  Same total run time but will the temperature be 68 at the end of the off period or slightly higher?  If it's 69-70 at then end of the off time, then I got an efficiency boost.  On the other hand, if it cools down to 66-67 by the end, it's either a wash or less efficient
Title: Re: Manual Thermostat Overshoot
Post by: tljohnsn on November 22, 2016, 11:54:35 AM
Slightly off topic but seems relevant. If you have steam boiler and radiators - do NOT use any old programmable thermostat. They cycle too quickly and can damage the boiler. In theory they make programmables that cycle slower and are safe for the boilers, but personal experience has shown they're not great. Stick with manual.

Thanks for this - I just put one in this year and now found the setting to change the cycle time.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 11:58:37 AM
Slightly off topic but seems relevant. If you have steam boiler and radiators - do NOT use any old programmable thermostat. They cycle too quickly and can damage the boiler. In theory they make programmables that cycle slower and are safe for the boilers, but personal experience has shown they're not great. Stick with manual.

Thanks for this - I just put one in this year and now found the setting to change the cycle time.

its also relevant because, if I had a thermostat with an adjustable range (they exist), I would increase run times.  But I doubt I'd ever get back the cost of even a cheap new thermostat, much less a smart one like a nest.

Wife wants a nest mostly because it looks cool and she can turn on the heat remotely.  May have to rig up a cheaper alternative since we are already good at keeping the heat off when we are out of the house or asleepk
Title: Re: Manual Thermostat Overshoot
Post by: nobody123 on November 22, 2016, 12:58:44 PM
Just setting it to 68, leaving it alone, it ran for 15 minutes, off for 5, then on for 10, off for 6, then on for ???... I think over time the off time kept increasing but I'll need more good data.

So id like to compare this to running for 25 straight then off for 11.  Same total run time but will the temperature be 68 at the end of the off period or slightly higher?  If it's 69-70 at then end of the off time, then I got an efficiency boost.  On the other hand, if it cools down to 66-67 by the end, it's either a wash or less efficient

You need to think about comfort.  After a 25 minute run time, your house won't be 68 or 69 at it's peak temperature, it'll be something north of that like 73.  Now when it settles back to 68 you will feel cold.  The run times even at a steady setting of 68 are going to vary because of outside temperature & humidity, the amount of sun/shade on your house, appliances and people giving off heat, etc.  There's too many variables and the potential savings isn't worth it.  Using a programmable thermostat to compensate for when you're away is a much bigger bang for the buck than trying to micromanage out one heating cycle during the day.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 01:35:41 PM
Just setting it to 68, leaving it alone, it ran for 15 minutes, off for 5, then on for 10, off for 6, then on for ???... I think over time the off time kept increasing but I'll need more good data.

So id like to compare this to running for 25 straight then off for 11.  Same total run time but will the temperature be 68 at the end of the off period or slightly higher?  If it's 69-70 at then end of the off time, then I got an efficiency boost.  On the other hand, if it cools down to 66-67 by the end, it's either a wash or less efficient

You need to think about comfort.  After a 25 minute run time, your house won't be 68 or 69 at it's peak temperature, it'll be something north of that like 73.  Now when it settles back to 68 you will feel cold.  The run times even at a steady setting of 68 are going to vary because of outside temperature & humidity, the amount of sun/shade on your house, appliances and people giving off heat, etc.  There's too many variables and the potential savings isn't worth it.  Using a programmable thermostat to compensate for when you're away is a much bigger bang for the buck than trying to micromanage out one heating cycle during the day.

I am thinking about comfort.  I'll definitely be more comfortable going up to 75 and back to 68 than 70-66-70-66

Potential savings isn't worth what exactly?  A programmable thermostat will achieve nothing
Title: Re: Manual Thermostat Overshoot
Post by: Sibley on November 22, 2016, 01:36:42 PM
Slightly off topic but seems relevant. If you have steam boiler and radiators - do NOT use any old programmable thermostat. They cycle too quickly and can damage the boiler. In theory they make programmables that cycle slower and are safe for the boilers, but personal experience has shown they're not great. Stick with manual.

Thanks for this - I just put one in this year and now found the setting to change the cycle time.

I'd pay attention to it to be sure it's ok with the longer cycle time. It takes time for the boiler to kick in, distribute steam throughout the system, and then ramp back down. If you interrupt that cycle too often that's when you get problems. My parent's have an older boiler, so theirs may take longer than newer boilers, but even with the long cycle time it wasn't letting the boiler go through the whole process. Different boiler/thermostat might be fine though.
Title: Re: Manual Thermostat Overshoot
Post by: mskyle on November 22, 2016, 03:08:50 PM
Slightly off topic but seems relevant. If you have steam boiler and radiators - do NOT use any old programmable thermostat. They cycle too quickly and can damage the boiler. In theory they make programmables that cycle slower and are safe for the boilers, but personal experience has shown they're not great. Stick with manual.

Thanks for this - I just put one in this year and now found the setting to change the cycle time.

its also relevant because, if I had a thermostat with an adjustable range (they exist), I would increase run times.  But I doubt I'd ever get back the cost of even a cheap new thermostat, much less a smart one like a nest.

Wife wants a nest mostly because it looks cool and she can turn on the heat remotely.  May have to rig up a cheaper alternative since we are already good at keeping the heat off when we are out of the house or asleepk

I got my wifi thermostat (not a nest, a Honeywell that looks like a regular old programmable thermostat) free after a rebate from my natural gas company - \$99 on Amazon, \$100 rebate. Installed it myself, which I think cost \$1.50 for some wire cover thingies because one of the wires in the wall wasn't connected to the wire that went into the thermostat. It's very handy, although I would say the two most common things I use it for are 1) turning the heat on from bed on weekend mornings and 2) turning the heat off when we go away for the weekend and realize we forgot to turn off the program before we left.

The thermostat also has some kind of learning aspect, the kind of thing people are talking about here, where it anticipates how long it will take to warm up to the desired temperature and starts ahead of time - so if I say I want it to be 65F when I wake up at 7, the heat will actually come on at 6:30 or so. You can turn that off, though - my dad has the same thermostat and he doesn't use that feature because his heat is slower to heat up and noisier and he'd rather wake up on time to a colder house than an hour early to loud clanking.
Title: Re: Manual Thermostat Overshoot
Post by: James on November 22, 2016, 03:22:21 PM
I think it's more complex than you have figured. You can think of the high thermal mass items, like granite, but you also have to consider low thermal mass items, like windows. If you heat the house to 75 degrees, things like walls don't matter much because they are higher thermal mass. But your window glass will heat up pretty quickly (and things like metal doors, etc), and that heat will spread outside faster than if it came up to only 68 degrees. Also, there is always air leak in the house, no way to make it perfectly tight, and you wouldn't want it perfectly tight. So whatever small air leak you have is leaking 75 degree air for a short time. I doubt it's a big factor, but I also doubt there is any big factor in the inefficiency from the furnace cycling a few times over half an hour or so.  I would bet setting at 68 and letting it come up slowly is most efficient, but by how much I have no idea, and I'm not very certain I am right... :)
Title: Re: Manual Thermostat Overshoot
Post by: nobody123 on November 22, 2016, 03:25:23 PM
Just setting it to 68, leaving it alone, it ran for 15 minutes, off for 5, then on for 10, off for 6, then on for ???... I think over time the off time kept increasing but I'll need more good data.

So id like to compare this to running for 25 straight then off for 11.  Same total run time but will the temperature be 68 at the end of the off period or slightly higher?  If it's 69-70 at then end of the off time, then I got an efficiency boost.  On the other hand, if it cools down to 66-67 by the end, it's either a wash or less efficient

You need to think about comfort.  After a 25 minute run time, your house won't be 68 or 69 at it's peak temperature, it'll be something north of that like 73.  Now when it settles back to 68 you will feel cold.  The run times even at a steady setting of 68 are going to vary because of outside temperature & humidity, the amount of sun/shade on your house, appliances and people giving off heat, etc.  There's too many variables and the potential savings isn't worth it.  Using a programmable thermostat to compensate for when you're away is a much bigger bang for the buck than trying to micromanage out one heating cycle during the day.

I am thinking about comfort.  I'll definitely be more comfortable going up to 75 and back to 68 than 70-66-70-66

Potential savings isn't worth what exactly?  A programmable thermostat will achieve nothing

I thought you were trying to save money by having several longer runs of your HVAC by manually touching the thermostat multiple times per day, as opposed to letting it just operate, because there is a slightly inefficient portion of the heating cycle at startup.  Your hypothosis is that if you need to heat your house to 68 from 63, it is better to do one long run to overheat the air to 73ish and potentially skip the inefficient parts of a few shorter heating cycles.  Am I misunderstanding?

I think that even if you could successfully do what you're proposing and it works, it's not worth your time to find out if it's true.  Do the math, let's say your natural gas bill is like \$200/month during the winter.  That's less than \$7/day.  Let's say your heat cycles on 4 times per hour, based on the numbers you provided above.  So 96 cycles per day means \$0.073 per cycle.  Let's assume the first 20% of a cycle is 100% inefficient and is a waste of 1.46 cents.  If you're only talking about doing the 63 -> 73 -> 68 change once per day when you get home, you're going to go through that work for less than 2 cents.

Think about it, if there was any monetary benefit in what you're trying to accomplish, a thermostat manufacturer would already be doing it.  They would be advertising it as providing huge cost savings for the consumer, and demonstrating the efficiency it provided.  It would be a 5 minute coding effort to implement on any electronic (non mercury switch) thermostat to temporarily go to 5 degrees over setpoint if starting temperature is more than 5 degrees less than the displayed set point.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 04:14:48 PM
I think it's more complex than you have figured. You can think of the high thermal mass items, like granite, but you also have to consider low thermal mass items, like windows. If you heat the house to 75 degrees, things like walls don't matter much because they are higher thermal mass. But your window glass will heat up pretty quickly (and things like metal doors, etc), and that heat will spread outside faster than if it came up to only 68 degrees. Also, there is always air leak in the house, no way to make it perfectly tight, and you wouldn't want it perfectly tight. So whatever small air leak you have is leaking 75 degree air for a short time. I doubt it's a big factor, but I also doubt there is any big factor in the inefficiency from the furnace cycling a few times over half an hour or so.  I would bet setting at 68 and letting it come up slowly is most efficient, but by how much I have no idea, and I'm not very certain I am right... :)

Yep, that's the trade off I'm considering.... basically does the higher heat loss from higher air temperature offset the losses from furnace startup inefficiency.  Data will be required.

First, I'll need to determine my steady state heating loss.  For example, how many therms does it take to maintain a temp of 68 inside when it's 60 outside.  I can potentially determine this over 1-2 hours by watching my meter and for the most part extrapolate assuming that heat loss scales linearly with temperature difference (this is my understanding... I know we are exacting nitpickers here, but let me know if this is wrong even at a generalized level).

Then, I could somehow determine what kind of efficiency losses I get from starting/stopping the furnace.

I'm nowhere near even #1.  I've been running a little test all day:

Entire house started at 61.  It's 61 outside, and I set the thermostat to 68.

Over two hours, the air temp has been 66-70, but the granite has only come up to 63.6 and interior walls to 65.8.  That's a lotta thermal mass.

Run times are around 9 minutes on / 6 minutes off.  The burner is only actually on for around 7 of those "on" minutes.

At this point, I'm not even sure I there's a reasonable way to quickly bring up the thermal mass temperature (I'd be willing to overshoot to 75 because that's a really comfortable air temperature for me, but It's ridiculous to overshoot to say 85)
Title: Re: Manual Thermostat Overshoot
Post by: FerrumB5 on November 22, 2016, 04:24:52 PM
What's your average (over 1 year at least) monthly gas bill and average outside temp? Mine is 57 terms (average cost is \$37, out of which \$15 is unavoidable delivery cost + some fees. I had a zero usage month, and bill was still \$15), average temp is 54F, 14 months of data. Even right now it's 38F outside, set to 69F and it runs may be 2 times in 1 hr?
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 04:26:15 PM

I thought you were trying to save money by having several longer runs of your HVAC by manually touching the thermostat multiple times per day, as opposed to letting it just operate, because there is a slightly inefficient portion of the heating cycle at startup.  Your hypothosis is that if you need to heat your house to 68 from 63, it is better to do one long run to overheat the air to 73ish and potentially skip the inefficient parts of a few shorter heating cycles.  Am I misunderstanding?

You understand, but you seem to think this means more work.  I already manually touch the thermostat multiple times a day.  I usually leave it off.  Sometimes I turn it on or off at some point during the day based on waking up, going to bed, leaving the house, expecting guests, expecting wife home, and so on.  None of this follows a set schedule and I've never found it to be an issue to manually adjust the thermostat.

Quote
I think that even if you could successfully do what you're proposing and it works, it's not worth your time to find out if it's true.

Do the math, let's say your natural gas bill is like \$200/month during the winter.  That's less than \$7/day.  Let's say your heat cycles on 4 times per hour, based on the numbers you provided above.  So 96 cycles per day means \$0.073 per cycle.  Let's assume the first 20% of a cycle is 100% inefficient and is a waste of 1.46 cents.  If you're only talking about doing the 63 -> 73 -> 68 change once per day when you get home, you're going to go through that work for less than 2 cents.

Maybe, but look at it this was: Is it worth your time to reply to me?  How much are you getting paid for this conversation?  Some people have an intellectual curiosity (as I suspect you do) and don't do everything for ROI.

I'm not sure I buy your free market argument that this won't work.  As mentioned by others, there are a lot of variables, so it could be a wash for the general population but work out well for a few people.  There are a lot of niche applications that don't get their own products because they don't pencil out in manufacturing.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 04:29:31 PM
What's your average (over 1 year at least) monthly gas bill and average outside temp? Mine is 57 terms (average cost is \$37, out of which \$15 is unavoidable delivery cost + some fees. I had a zero usage month, and bill was still \$15), average temp is 54F, 14 months of data. Even right now it's 38F outside, set to 69F and it runs may be 2 times in 1 hr?

I dunno, I just added some insulation so who knows?  Some years I'm out of town for December, others I have house guests who like it toasty.  Last year it was \$400/year or \$33/mo?
Title: Re: Manual Thermostat Overshoot
Post by: FerrumB5 on November 22, 2016, 04:33:18 PM
So, if you estimate it at \$33/mo, and if you have comparable delivery fees, consumption is roughly \$16/mo. That is divided between water heater, furnace, stove (if gas), drier (if gas). That's not a lo of savings if you optimize the cycles. Besides, there are times of day when you sleep and simply not home. Just saying, it could work potentially
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 04:45:24 PM
So, if you estimate it at \$33/mo, and if you have comparable delivery fees, consumption is roughly \$16/mo. That is divided between water heater, furnace, stove (if gas), drier (if gas). That's not a lo of savings if you optimize the cycles. Besides, there are times of day when you sleep and simply not home. Just saying, it could work potentially

I don't have a base charge, and I left out the summer months (no heat).   So \$33/mo is probably pretty close.  Maybe \$30 if you subtract hot water/cooking.

The question really isn't specific to my usage, percentage savings, etc., it's about whether doing this is a net savings at all.
Title: Re: Manual Thermostat Overshoot
Post by: AlanStache on November 22, 2016, 08:10:11 PM
If we assume the furnace uses a constant amount of power/gas when on but produces less output while starting up during its inefficiency period all this really comes down to is measuring the total time the furnace is on given the two control sachems.  I think you mentioned up thread trying to record times but getting distracted?

As systems transfer heat faster when the temperature difference is larger going to a higher temperature will lose more heat to the outside but this could be offset by increased furnace efficiency with fewer cycles.  This sounds like we are getting into some small numbers that could be hard to measure given other uncertainties.
Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 09:46:48 PM
If we assume the furnace uses a constant amount of power/gas when on but produces less output while starting up during its inefficiency period all this really comes down to is measuring the total time the furnace is on given the two control sachems.  I think you mentioned up thread trying to record times but getting distracted?

As systems transfer heat faster when the temperature difference is larger going to a higher temperature will lose more heat to the outside but this could be offset by increased furnace efficiency with fewer cycles.  This sounds like we are getting into some small numbers that could be hard to measure given other uncertainties.

Yeah, based on my observations I'm thinking there's not much to do here.

It took about 5 hours to raise the temperature of my countertops from 61 to 67 degrees.  That's not something I can speed up with a manual overshoot approach.  If it was more like 1 hour, maybe.

The only time things were really inefficient was the first shutoff - System ran for 30 min, shut off for 3, then ran for 10 more, shut off for 7.  An efficient system could easily skip that 3 min shutoff period.

But then for the next 4 hours, it ran approximately 9 min on, 9 min off.  Maybe 18 min on 18 min off would be preferable, but I'm not doing that manually.  A better thermostat could help but it's not likely worth spending any money.

Finally, within the last hour (AFTER sunset) it's started running longer off times -- now 9 min on / 20 min off.  In other words, the off time got successively longer as my thermal mass came up to temperature.  This makes sense, that it's mostly going to take the same amount of run time to heat the air 2-3 degrees, but the air will cool down slower without the thermal mass effect.

If I have similar temperature baselines tomorrow, I'll try a long burn / long off and see how it compares to my readings from today.  But most likely, I need a thermostat with a wider trigger range (i.e. 4-5 degrees instead of 2-3 degrees) for a longer run time.  I don't personally find a 4-5 degree swing uncomfortable.  On the other hand, my wife is always cold so I don't think she finds either swing comfortable.

Title: Re: Manual Thermostat Overshoot
Post by: dragoncar on November 22, 2016, 09:51:32 PM
The only issue with comparing "total time on" is that I have to really pay attention to hear:

1) the thermostat click
2) 20 seconds later, the gas comes on
3) 60 seconds later, the blower comes on

then I think the blower runs a while after the gas turns off, but I don't hear the gas turn off.  I'd have to hang out in the garage to see the timing on that.  In other words, my readings are off by 1-2 minutes on each side, which will likely be too much error in this comparison.

My gas meter is a bit hard to read... it's got a .5cuft dial, a 2 cuft dial, and a 100 cuft dial.  But I can't watch it while I listen for the changes unless I set up a webcam outside.

Crazy people problems.
Title: Re: Manual Thermostat Overshoot
Post by: arebelspy on November 23, 2016, 02:22:32 AM
Obsessive Compulsive Dragoncar?
Title: Re: Manual Thermostat Overshoot
Post by: nobody123 on November 23, 2016, 10:13:36 AM
Fair enough, if you're doing it for the sake of knowledge, more power to you.  In the interest of science, here are some other things to consider:

Thermostats of at least one major manufacturer I am familiar with (wink, wink) use half a byte to store the partial degree.  So, your thermostat may display 63 degrees but the actual ambient air temperature is somewhere between 62 8/16 and 63 7/16 degrees.  So for your experiment, use a more accurate thermometer and ignore what the thermostat says.  If your temperature differential of your thermostat is +/- one degree, a thermostat set at 63 will issue a call for heat at 62 7/16 degrees (or lower) and will end the call for heat once it observes a temperature of 63 7/16.  Again, all it can do is suggest to the furnace that it turn the heat on or off, so the furnace logic board will also determine if the furnace starts up instantly and how much longer it runs after the call for heat is terminated based on whatever logic the furnace manufacturer coded into it.

As I mentioned before, household thermostats also have some delays built in as "sanity checks" so sudden temorary gusts of extreme temperature (open the front door in winter, open the oven door, running a hot vacuum under the thermostat, etc.) do not affect the display temperature or operation.  You can get a sense of the delay built into your thermostat by getting a hair dryer and blowing hot air on your thermostat.  If the hairdryer is blowing 130 degree air, it will take several minutes for the display temperature to climb, and the display will most likely only climb a few degrees at a time because thermostats assume that a normal house only adds or drops a degree every 10 minutes or so.  Since you don't get to view the partial degrees the thermostat is using to decide what to do or know what the sample rates and tolerances are, it will be difficult to time your manual adjustments with any sort of precision.

And lastly, electronic thermostats have a delay to deal with with the user input.  If you have a setpoint at 63 and you hit the temperature up arrow 10 times to get it to 73, what is actually happening is something like: <user interrupt timer start><setpoint +1><reset interrupt timer><setpoint +1><reset interrupt timer><setpoint +1>... <setpoint +1><wait for user interrupt timer to expire><return control to tstat logic board>.  The user interrupt depends on model but is usually somewhere between 5 and 15 seconds.  So your point of having to start your measurements based on the audible click of the call for heat is correct, but you will also need to account for the extra delay when turning the heat setpoint back down.

You might also not be aware that gas furnaces do not always light immediately ON PURPOSE.  In order to ensure that the ignitor is using the lowest possible current and the safety systems are working, the furnace logic board every X ignitions or so (frequency depends on manufacturer) will "reset" the ignitor to the lowest possible current, then attempt a series of ignitions, gradually increasing the current to the minimum necessary until the fire starts.  This saves the wear and tear on the system in the long run.  So you actually do need to observe the gas fire while listening for the click from the call to heat to determine when the true heating cycle starts to get a proper timing.
Title: Re: Manual Thermostat Overshoot
Post by: AlanStache on November 23, 2016, 10:29:38 AM
The only issue with comparing "total time on" is that I have to really pay attention to hear:

1) the thermostat click
2) 20 seconds later, the gas comes on
3) 60 seconds later, the blower comes on

then I think the blower runs a while after the gas turns off, but I don't hear the gas turn off.  I'd have to hang out in the garage to see the timing on that.  In other words, my readings are off by 1-2 minutes on each side, which will likely be too much error in this comparison.

My gas meter is a bit hard to read... it's got a .5cuft dial, a 2 cuft dial, and a 100 cuft dial.  But I can't watch it while I listen for the changes unless I set up a webcam outside.

Crazy people problems.

Could you make a smart phone app that triggers off the sound to start a timer?  Then leave the phone plugged in and the app running by the furnace.  But this is way deeper down the rabbit hole than Alice ever saw.