Minimum furnace run time?

[Daox]

I've always read that your furnace shouldn't short cycle. But, I was hoping that we can help further define short cycling. Is it 10 minutes, 5 minutes?

I just checked my TED usage and it seems that my furnace kicks in for a little over 30 minutes at a time to maintain temperature (ambient is fairly cold right now, say ~15F). I'd say 30 minutes is plenty long, but I'm not 100% sure. It then doesn't run for about an hour, and then turns on again for a little over 30 minutes.

So, what would you guys say is short cycling? What should be the minimum run time we should look for? What are other variables that we should be looking at to answer these questions to our best?

[Piwoslaw]

Originally Posted by Daox

I've always read that your furnace shouldn't short cycle. But, I was hoping that we can help further define short cycling. Is it 10 minutes, 5 minutes?

My answer would be that the ON cycle is not long enough for the all of the air (or water, in the case of a boiler) in the house to get warmed up. In other words, if the return air (water) temperature starts to rise, then it is no longer short cycling.

[MN Renovator]

I've always wondered this too. My brother has a house that used to have a Honeywell digital setback/timer thermostat that would anticipate heat but it had a fixed anticipation and didn't 'learn' the anticipation. This was very bad because the furnace it is connected to takes 3.5 minutes from the call to heat until it actually kicks the blower on. It usually only had a call for heat for 5 minutes so the burner would fire at about 2.5 minutes and after about 5 minutes the thermostat figured the job was done and shut off leaving a 2.5 minute burner cycle with only 1.5 minutes of blower time.

That was a very short cycle. Oddly enough my house has the same furnace and I used to have a thermostat that would basically have the gas burning for 5 minutes with the blower going for about 3 minutes and 45 seconds of that. On a 0f degree day it would need to run 5 times in an hour to keep 70 degrees. So it was firing up every 7 minutes after it had turned off. Pretty much firing up after it shut off over and over. I had replaced it before it got any colder because I considered it ridiculous. This was all before I started any insulation or sealing though too.

I replaced the thermostat with one that has a wider span and doesn't anticipate the heat, now it has the call for heat between 12-14 minutes so on average it runs the blower for about 10 minutes with the gas firing a little longer than that. It takes about 10 minutes for the output temperature at the farthest supply registers to stabilize at their maximum temperature due to the cold ductwork warming up so I think I'd prefer a 20+ minute call for heat so it could better make use of the warmer temperature. I wouldn't notice the change in temperature over that period. In reality going to a 40k 90+% efficient condensing furnace would bring the runtimes to that level versus my 75k(57k output) current furnace. Newer furnaces also fire up quicker, usually in about a minute instead of 3.5 minutes.

What the ideal time is, I'm not certain but I personally would like longer cycles than I have, especially considering it would make ignition components last longer and have less heat cycles for the heat exchanger which should make the whole furnace last longer and require less break-fix maintenance. Granted I haven't had a single problem with my current furnace, if I paid big money for a new one, I'd tie it with a thermostat that I could make sure it cycled on less than once per hour and ran long even on the coldest days. I think the longer the better as long as the temperatures between on and off are still comfortable.

[stevehull]

There is no set amount of short cycling. It all depends on the outside temperature and the difference between that and the conditioned space (delta T).

For example, consider a 24,000 BTU single stage heat "machine" (resistance, gas, geothermal, propane, whatever). Assume the neutral point for heating is 65 degrees F in a well insulated 2400 sqft home.

At a moderate heating load (outside temp of 40 F), the duty on time might be 10 minutes on 20 minutes off. But at zero F (large heating load) this might be 30 minutes on and 1 minute off. Assume that zero is the design low temperature.

Yes, this heating unit would "short cycle" when it was 55 degrees F out, but for a single stage, non-proportional unit, this unit is perfectly sized as at the maximum load it is on virtually 100% of the time.

However, if it is zero out and the unit is not on at 100%, then waste heat is generated and/or you have paid for a unit too large for your application. If it is purely resistance heat, there is virtually no standby loss, but if there is any vent, then there is standby loss with decreased efficiency.

Many well intentioned HVAC installers put in too large a heating system as they think that it will "heat up the house faster". It may do that by 1-2 minutes, but at a large waste of efficiency.

Newer gas furnaces (and some electric resistance heaters) use modulating gas values (variable ac amps for resistance heat) to keep the furnace on for as long as possible. This minimizes the heat blast and then cool down so despised by many on forced air heat.

For AC, the same is true. Too large a unit does not run for long times thetreby removing latent heat (condensation). The thermodynamiocs of AC merans you remove water andf that phase change allows the temperature to decrease. But if too large an AC unit it too will short cycle and cool dowen too quickly without removal of latent moisture. Mold is the common result (where I get called in).

The ideally sized heater/AC will run 100% at the design loads for your climate.

The new variable speed (10-120%) Copeland compressors in some new geothermal heat pumps allow tremendous SEER and COP (50+/6+) by maximizing run times.

So, check the duty cycle on times on your coldest night (about now in January) and I can discuss later how to modify natural gas/propane units to increase the duty cycle on time and thereby increase your heater efficiency.

[AC_Hacker]

Originally Posted by Daox

...what would you guys say is short cycling? What should be the minimum run time we should look for? What are other variables that we should be looking at to answer these questions to our best?...

I can't answer this question for all situations, but I did do some testing on my little homemade heat pump, which was drawing heat from my ground loop. I was set up to measure and calculate quite a large number of variables, and "instantaneous" COP was one of them. By instantaneous, I mean over a period of a minute.

I noticed a repeating pattern of COP ramping up to a 'shoulder' then increasing more gradually to a max, then declining a bit and finally settling into a steady level.

It seemed to me that a Calculus approach would be needed to determine the maximum best time interval. A good approximation could be done by imagining the curve as being composed of a series of thin strips and counting the area in successive strips. This incremental area, divided by time would tell the story.

I did notice that the initial ramp-up and then settling curves were pronounced enough that it was clear that a series of optimum bursts would be more efficient that a single long run.

But the behavior I saw was of my particular setup and I would dis-advise trying to generalize to other systems.

I am quite sure that electric resistance heating would not at all have characteristics similar to my heat pump. In fact the PID controllers count on very short pulses of current to achieve and maintain a temperature.

And, regarding heat pumps in general, in the days of single speed heat pumps, the optimum size of the heat pump was slightly less than the size needed to maintain comfort on the very coldest days, and to plan for axillary heat (electric resistance, wood, gas, etc) to fill the heating gap on those days.

Now with multi-speed compressors, the logic is not so clear, but bigger is not always better.

-AC

[MN Renovator]

"I noticed a repeating pattern of COP ramping up to a 'shoulder' then increasing more gradually to a max, then declining a bit and finally settling into a steady level."

Question: Did the output temperature rise change with the duration of the run where COP declined?

I've noticed at cold indoor ambient temperatures that my refrigerator will pull wattage at a high level and slowly drop but at the coil warms up the wattage increases about 10 watts to about 130. During standard room temperature ambients it settles to 150 watts after about 3 minutes and stays there for the whole run cycle without any climb after the drop. Not that this means anything, just an observation. If your COP is dropping, it could be because the output temperature reached its max and the geothermal field/coil temperature dropped slightly after the water made a few rounds. I'm speculating, of course. Was the COP drop significant to where it would mean that going through the low startup COP again would be beneficial to operation?

[stevehull]

AC Hacker is correct about the instantaneous COP responding to infinitely small loads. This Neutonian approach is indeed the basis for the calculus integral!

In the end, however, a load is a load and we can look at it as an average load over some time period. I have used a data logger to examine short cycling and the thermal mass that the load is coupled to. An example:

Compare/contrast a radient large thermal heat load to an air load with very small thermal load. The thermal capacity of the radient load is FAR greater than the air load. Thus a heating system that uses a radient thermal floor (PEX in concrete for example) is going to be far more stable and use much less short cycling compared to a purely air load which has very little thermal capacity.

The reality of course is that an average house has a large thermal mass - even without a radient floor concrete mass. You can see this by taking a well heated home and simply turning off the heat on a cold day. The slow change in temperature vs time (heat decay curve) reflects the massive amount of stored heat in the home. This assumes that the home does not have a lot of air infiltration that "washes out" the decay curve.

On the flip side, go into an unheated home (non radient thermal floor heat) in the winter After turning on the air heat, it only takes a few hours to get the air temp up to 70 F ("normal"), but the house "feels cold" for days. The reason is that the thermal loads of the walls, cabinets, and other home thermal infrastructure have not yet warmed up. The longest to warm up is the floor. The carpet actually insulates the subfloor from being heated up and that cold on the feet is hard to take. The home "feels" cold and many will turn "up" the thermostat in order to get the perception of warmth.

This is why I REALLY like thermal mass floors and geothermal heat in it. This is a match made in thermodynamic heaven and also for human comfort. Bursts of heat are absorbed by the thermal mass and are evenly distributed in it. The infrared (IR) component mans that you "feel" warm, despite an air temperature lower than what you would otherwise be comfortable in. Many home I work on have a radient floor set at 72 with an air temp of 67 sometimes 66.

The key is to make sure the thermal mass heat is constrained in a thermal "bathtub" and that the undersides, walls and especially corners of the thermal mass have large amounts of foam insulation.

And most importantly, it is far better to actually undersize the GT heat pump as the load developed in the floor during the day can be used at night as supplimental heating. This is because the time constant of the thermal mass (decay curve) is longer than a 24 hour period. So in the daytime when outside temps are high, the thermal mass gets a full load of heat and at night it then releases it.

This concept is VERY hard for HVAC technicians to understand, yet every ancient ancient Roman engineer knew of this and used it to keep Roman stone house warm at night. A small fire, heating a large thermal mass, develops a lot of accumulated heat (stored BTUs) and then at night releases that heat when the temperature in the home drops. In fact, the larger the delta T (temp difference between heated stone and house temperature) the larger the amount of heat released. Romans engineered and built built small fireplaces under the house stone floors and had a serpentine flue distribute the heat below the floors.

This is why I often suggest a GT heat pump installed at about 2/3'rd of actual peak load as there are options for those few times in the year when you need supplimental heat.

The kitchen oven, for example, is a great emergency 20 kW space heater for use on those few really cold days for a couple hours in the AM. It is already there, meets code and I really like muffins on a cold morning! Yes, the meter spins when you use it, but this is for a very small "slice" of time allowing the moderately "undersized" heating system to meet the average heat need (as coupled through a large thermal mass).

Again, we do NOT need to put in huger capacity heating/cooling units. Far smaller units, coupled to a large thermal mass, allow far better comfort, reduce "short cycling" and cost far less.

Hope this helps . . .

[MN Renovator]

This is very true about thermal mass, as a test I've had my house at 40 degrees, raised the temperature to 70 and after three hours it will be at 65 degrees but getting it to 70 from there battling all of the thermal inertia behind the thermal mass of everything in the house really slows things down. Did this last January heating the house up and allowing it to be at the set temperature for 12 hours (after the thermostat shut off the first time) and my heat load on the house went from 285BTUhr/f measuring it when it was at a solid temp before the start of a very cold period to about 400BTU/hr. Even after 12 hours the thermal mass still had a significantly large part of the load. I learned not to measure heat load that way. This morning it was 260BTUhr/F and it was -3.5f average outside from midnight to 7:30am. -13 is the design temp for my area 260BTUhr/f * 83 = 21580. 10.28BTUhr/sq ft, I'll take it. Actually I won't, still have work to do. 40k is the smallest condensing furnace I can buy but I'm not sure I'm going to go with a 2-stage since I keep reading that the lower stage is still less efficient than running high stage. As long as I can get a good thermostat that will keep the cycles efficiently long I think I should be fine though. Either that or using moderately deep setbacks like I do when I'm not heat load testing isn't such a bad idea. With this colder temperature the gas runtime is now at 17 mins instead of the 10-12 mins I get with warmer outdoor temperatures. I'd probably have 30 min cycles with a 40k condensing furnace, maybe a little more if I can set the temperature swing a little higher.

[Daox]

So, I reprogrammed my thermostat tonight. Instead of allowing a +/- 1.5 degree F difference before kicking on, it now is back to its default setting of +/- 1 F.

The really weird thing is in my manual it says:

Your system should cycle on about 3 to 6 times per hour. A smaller swing number
increases the number of cycles, so room temperature is more constant. A larger swing
number decreases the number of cycles, to save energy in most cases

3-6 times per hour!!! I don't think so. We'll see how this new setting works.

[AC_Hacker]

Originally Posted by Daox

...3-6 times per hour!!! I don't think so. We'll see how this new setting works...

As I recall, the latest thing in furnaces is modulating (the furnace can turn it's flame size down to suit the heat load) condensing (they are able to recover the 15% of energy in the water vapor that would otherwise go up the chimney) or ModCon furnaces for short.

There are also water heaters & tankless water heaters that do the same thing.

But, with all the EcoWork (new word?) you have done, your heat loss should be much more gradual.

-AC