Reducing Furnace Cycling

[jpanhalt]

My home has hot-water heating (aka hydronic) with only one furnace. Domestic hot water is heated with that furnace as well and uses a separate heat exchanger loop. Excluding an outside wood burner that can be used as an alternative heat source, there are 6 zones in the house: domestic hot water, basement, garage, upstairs, downstairs, and an additional space. The furnace is propane fired and is rated at 105,000 BTU/h (approximately 31 kW or 12 hp). House is reasonably energy efficient as it was built in 1993. There is no need for the furnace to run continuously; moreover, frequent cycling of the furnace wastes propane.

Domestic hot water (DHW) has priority over house zones. The house zones are all wired in parallel (OR'd). If only one zone calls for heat, the furnace turns on. It is the latter fact that can lead to frequent cycling of the furnace. Original thermostats were Honeywell 87A's and had mercury switches, and by carefully setting them differently, I could reduce cycling a little. Some have gotten a bit funky and unreliable. Of course, direct replacements are no longer available. Honeywell's digital replacements do not allow adjustment of the spread (i.e., hysteresis between call for heat and satisfied). They do have an adjustment for number of cycles per hour. The values vary from 1 to 9 (i.e., 1,3,5,9). I am not 100% certain what the number represents. It appears from blogs that the duty cycle is set at 50% and that, for example, a value of "3" means the furnace is on for 10 minutes, then off for at least 10 minutes, until the demand for heat is justified. It is not clear whether the furnace will stay on for more than 10 minutes to satisfy the demand for heat. However, if the demand is met before the 10 minutes elapses, the furnace will turn off. (I am currently testing one I installed yesterday.) The alternative is a thermostat that has an adjustable spread. I have tentatively decided on using a Honeywell version, modelRTH6350D.

Here is the wiring for the house zones:

When the "End Switch" is closed, the furnace comes on. Since there is no interconnection between the thermostats, it is possible that one will cycle until satisfied, then another and so forth, which will lead to frequent cycling of the furnace regardless of the cycles/hour setting for each thermostat.

QUESTION:
The purpose of this post is to get opinions on prioritization schema for the in living space . I am considering inserting an MCU between the household zone end switches and the furnace relay that will prevent rapid cycling by sequential calls for heat from each zone. Here is a tentative scheme:

Any comments or suggestions?

John

 

[MikeMl]

Have you tried an Arduino? I am getting to really like these. I wire them right into a project, and they become a permanent part of it...

Timing things, and logic like you show is easy to code... They run on 5V, so can switch things easily. I don't bother with shields; I just build the mini into a protoboard and surround it with other stuff...

 

[jpanhalt]

I have not tried Arduino. My experience so far is just with PIC. I don't see programming it as being difficult. The internal oscillator will be more than adequate for timing the 20 minutes. I will need a relay (magnetic or solid state) as the rest of the system is 24 VAC.

John

 

[nsaspook]

I have not tried Arduino. My experience so far is just with PIC. I don't see programming it as being difficult. The internal oscillator will be more than adequate for timing the 20 minutes. I will need a relay (magnetic or solid state) as the rest of the system is 24 VAC.

John

The first priority is a fail-safe for the system, if it crashes for any reason the default should be for it to be removed from heat control. I might try weight based 'fuzzy' decision logic to allocate the heating resource. A static state based model will require some tinkering to get it just right under changing conditions but a learning type system that uses feedback to optimize the heating resource (by modeling rooms as a heat reservoir needing to be refilled) while more complex to implement should quickly 'learn' how to keep the system in the needed parameter windows if possible or if not display the reasons for it.

 

[tcmtech]

How often does your propane furnace need to run at full burn for extended times?

If its not a lot or if ever it might be worth switching out your burner jet to a smaller lower BTU one so it's putting out less peak heat thus giving it longer run times.

 

[jpanhalt]

How often does your propane furnace need to run at full burn for extended times?

When it runs, it is only at full blast. When working properly, the on periods are relatively short, say 20 minutes during the day. Longer at night if it is cold. I keep most of the house fairly cool at 66°F (62° at night) and where I am working is at 68 to 70°F during the day.

There were problems with the control relays and some of the wiring last Winter. The burner was both cycling frequently and getting "stuck" on. I controlled it many times (at least daily) by going to the basement and physically opening the run relay. Last Winter was pretty bitter for my area, and I didn't feel that was a good time to mess with it. Now I have fixed most of the wiring, I hope, as well as done some repairs/PM on the plumbing. It has been behaving well for the last three days, but the weather has been mild too.

The burner is the right size for the house.

John

 

[tcmtech]

How often does it do the ~20 minute on cycles?

What I am suggesting is lowering its BTU output to increase the duration of the run times so that it take maybe half an hour to an hour to warm the house up every time it does a cycle opposed to only running for 20 minutes.

It's just a thought otherwise you either need to get different thermostats that have a wider temperature hysteresis band because there is no other way to make the 105,000 BTU's of heat being produced go someplace beneficial during a longer run cycle.

 

[jpanhalt]

I am not concerned about long cycles. It is the too short cycles that waste propane.

The thermostats I bought are Honeywell RTH6350D, which have an adjustable cycle time and are 5-2 day programmable. None of the digital Honeywell's have adjustable hysteresis that I can find. Lux makes versions with adjustable hysteresis. That Honeywell version is inexpensive ( about $33 at Lowes when you press Lowes to price match Walmart +10%). Plus, I liked the cycle time idea.

My plan, for which I asked comments, was to develop a simple logic that would avoid rapid cycling. A secondary or future purpose is to redesign the whole priority system with an MCU. The system I have now uses White-Rodgers Type 1311 zone valves to set the house vs. DHW priority:

There has to be a simpler way to do that today. So, if the MCU house-zone priority system works, I will likely rip out those old valves and go to simple on/off valves with a switch as shown above.

John

 

[tcmtech]

I am not concerned about long cycles. It is the too short cycles that waste propane.

Yes I follow that. What I am saying is that by reducing the burner output it would take longer to produce the same amount of heat thus lengthening the shorter cycle times. Other than switching out for different thermostats the only other possible solution I can think of is to add some sort storage tank that greatly increases the overall thermal mass of the system thus also extending the burner run times when it has to reheat it.

With that the circulator pumps draw from that heat source and the main burner only cycles on to heat the storage tank up of which that part of the system could have a thermostat with a considerable hysteresis bandwidth.

I used a similar system with my boiler when I used to primarily burn wood. In order to get my refueling times to more than 6 - 8 hours apart I added a old 250 gallon fuel oil tank to the side of my boiler thus nearly tripling the thermal mass. Down side was when burning wood was I had to reload it back to back 2 - 3 times to get the now 350+ gallons of water up to temp but once it was there I had a good 16+hours of normal run time before I needed even start thinking about loading it up again.

As for your system electronically I think you can change your shorter burner run times to some degree but you are still stuck with how much total thermal mass you have and how fast your burner brings it up to temp and I just can't see an electronic control way around those physical attributes without subjecting the whole system to wide temperature swings which in themselves if big enough will start to create their own problems elsewhere.

The way I see it either the burner has to have less output or the system has to have more thermal mass to heat up to get the short cycle run times longer being just increasing the span of the temperature swings will at some point result in system temperatures that go too low to be efficient at heating the house and if low enough will likely start causing condensation problems when the burner starts up due to the heat exchangers around the burner being too cool when it restarts.

That's my thoughts anyway.

 

[alec_t]

I have to agree with those thoughts. The physics of the system dictate that the only solutions (that I can see) are (1) increase the thermal mass, (2) improve the house insulation efficiency, (3) increase the thermostat hysteresis or (4) reduce the burner output.
Some boilers have a pseudo-proportional output: two or three solenoid valves supplying burner jets of different sizes. Any chance you could do something along those lines?

 

[KeepItSimpleStupid]

John, I took a brief look at a zoning control and some hydronic manuals.
A couple of things stood out:

1. Furnace (electric/gas) always runs 4 minutes regardless. It shuts off when all zones are satisfied.
2. Circulator runs all zones 1 min every 24 hours regardless. You don't have to turn on the furnace.
3. Freeze control by zone? Guessing something like: Temp drops below some value turn on hydronics for 4 minutes.

If the zones are programmable, they have an effective priority.

I don't know how hydronics fit into zone controllers, but it's probably similar to damper control although dampers can get pretty complicated.
You have power/spring/feedback/ r they can modulate knowing the time to fully close.

The Vision Pro stats from Honeywell are complicated and expensive. I'd look at some of their manuals. I'd look at a zoning panel.
There may also be zome "communicating stats" that can use zoning.

http://zonefirst.com/products/zone-control-panels/

**broken link removed**


Dual fuel may complicate your setup.

In any event, a number of Hydronoc zoning system are described in this document: http://www.taco-hvac.com/uploads/FileLibrary/100-9.0.pdf

Earlier, I had a different notion of priority. But, it likely, that the other zones are shut off until the priority zone is satisfied (heating water).

==

FWIW (Warning - hijack):
I've been wanting to put a thermostat in the laundry room, but not as traditional zoning. The laundry does not have a "register", but I'd like to add one, that could "cool" or "heat" depending on the stat in the room. It's a Carrier Infinity system, so I also have to have the damper open for about 10 minutes at 1:00 PM so the time to change the furnace filter system continues to work. The system incorporates an Aprilaire 5000 pleated filter + electrostatic filter. Not the greatest, but not bad either. The filters are changed 2x per year and it costs about $35.00 each. I cut an additional comb at an angle to help install the new pleated filter which comes collapsed.\.

UV might help me because I'm allergic to some molds. I think I have the pollens nailed with allergy therapy.

New furnaces look like this http://www.perfect-home-hvac-design.com/forced-air-furnaces.html Instead of the air cleaner on the floor, mine is mounted to the ceiling rotated 90 degrees to save room.

I made two electrical changes: 1) a bidirectional TVS on the 24 V side (The thermostat went wierd on me) and 2) An X-10 filter at the furnace - too much power line pollution.

Sorry to hyjack your thread.

 

[KeepItSimpleStupid]

In your diagram (4), prevents short cycling by making the stat think it's hotter than it is and (6) causes the the water to shut off a little bit before it's supposed to by heating the sensing element while the valve is open to use the latent heat.

I thin the sensible thing is zone controller unless the dual-fuel thing causes you issues.

 

[KeepItSimpleStupid]

here's the cheap Honeywell programmable stat that supports hydronic systems: https://customer.honeywell.com/resources/Techlit/TechLitDocuments/69-0000s/69-1928EFS.pdf

The Honeywell Vision Pro 8xxx stats have all sorts of options including dual fuel and WIFI access. I think some of them have hysteresis.

Honeywell's site is typically difficult to navigate.

 

[KeepItSimpleStupid]

From an 8xxx manual, cycle rate is defined as:

Cycle rate limits the maximum number of
times the system can cycle in a 1 hour period
measured at a 50% load. For example, when
set to 3 CPH, at a 50% load, the most the
system will cycle is
3 times per hour (10
minutes on, 10 minutes off). The system
cycles less often when load conditions are
less than or greater than a 50% load

 

[jpanhalt]

Thank you fellas for the suggestions. Perhaps my question was not clear. My question for this thread has almost nothing to do with thermal mass of the entire house nor sizing of the furnace. The furnace is properly sized for the house, which has 3 baths with showers, and Winter weather in Northern Ohio. For example, a single shower using 3 gpm with a heat rise of 65° (i.e., 5o° to 115°) requires 97,461 BTU/h (Source: Weil-McLain: BTU/h = gallon/h * heat rise (°F) * 8.33 (factor for water density)). Heating needs for the house will depend on outside air temperature (OAT), wind, humidity, heat-loss factors, etc. A local contractor has agreed that the furnace is not too large. It may even be a little on the small size, if the house were fully occupied, which it isn't.
1) The furnace is not adjustable.
2) The new thermostats (X3) do not have adjustable hysteresis. They only have adjustable cycle time (Post#1) per Honeywell's current philosophy.
3) I cannot find anywhere in Weil-McLain literature a recommended minimum "on" time and "off" time, i.e., minimum cycle time, nor can I find that information on the Internet. All of the comments I read just punted on the specifics. In my opinion, and on/off cycle of 3 min on, 3 min off is too short for a hot-water system like mine. A cycle of 10 min on, 10 min off is OK. During cold weather (OAT <35° or no sun and temp <45°), I would like to see a longer on time. Aside from wasting fuel gas, a short cycle time does not allow the flue to heat sufficiently to minimize condensation.

Let me rephrase my question. If I had just a single thermostat in the house system, using cycle time for the house and a prioritization system for the domestic hot water (i.e., the W-R Type 1311 zone valves, Post#8) is pretty straight forward. But, I have 5 thermostats for the house. Each of those thermostats controls an on/off zone valve that is not a Type 1311 valve, but is simply an on/off valve with a set of contacts that control the circulators and furnace. If each simple valve were set as described by KISS, there would the possibility that each thermostat would follow its own 20-min cycle time with on-times shorter than 10 minutes. Sequentially, however, there could be short off times as each thermostat (no hysteresis) called for heat.

My proposed algorithm is intended to reduce that possibility by introducing prioritization and AND'ing lower priority zones. My question was a request for comment on such a scheme. I guess one could compare the AND'ing to increasing the thermal mass for those areas.

Historical FWIW:

There are several causes for rapid cycling in addition to proper furnace sizing. Here are a few that are not related to my question: 1) Propane tank empty; 2) Defective control module for furnace; 3) Circulating pumps not starting properly; 4) A stuck control relay, of which there are many possible culprits; and 5) No hysteresis in the thermostats.

I have experienced 1 thru 4 over the course of my ownership of the house. Last Winter, #3 and #4 were common, which led me this Summer to draw out complete electrical and plumbing schematics for the system, find and replace and/or clean problematic relays, redo much of the wiring to eliminate intermittent connections, properly phase a transformer that was reversed (not sure why the installer in 1993 or a later modification wanted two transformers in the same loop), and redo some plumbing (replace expansion tank and anti-backflow valve, add isolation valves for said tank, and replace pressure relief valves -- work in progress). One episode of #1 also happened, when the supplier underestimated my needs. The system is now working better, but the thermostats were old, contained mercury, and were a bit funky in operation. Most important, I wanted to add programmable thermostats to save on heating during the evening, absences, and in rooms not usually occupied.

BTW, I have just ordered from China some SSR's operated by 5V DC (Omron G3MD-202P modules, Banggood) and LM2596-based 5V DC from 28V AC supplies (eBay).

Regards, John

 

[tcmtech]

I see now. Its properly sized to keep up with the maximum expected loads but has little allowances fro running at the low end usage levels.

I am curious though as to how you figure you are wasting fuel running short burn cycles Vs longer ones and to what amount you think you will save by extending the run cycles?

The programmable thermostats definitely help though.

 

[jpanhalt]

I agree that starting and stopping a modern furnace doesn't waste much fuel compared to continuous running. There is a little loss during starting, but that is probably very small. However, heat is heat, and the extra heat from a longer run does benefit the whole house. The main issue mentioned by contractors and my propane supplier is wear and tear on relays and the water condensation problem.

Changing out to programmable thermostats was the driving incentive. I don't enjoy sitting and reading in 60° temperature, but sleeping at that temperature is not a problem for me.

John

 

[KeepItSimpleStupid]

John:

I think your idea is sound, but being me I'll overthink and simplify.

You have the safety asspects are a priority. Some circulator movement should occur. Not sure if your dealing with an AC. Prioritizing hot watermakes sense. Keepin condensation from forming in the flue is a good concern. Then there's cost.

The VP series has a wireless hanndheld gizmo which can be used as "where you want to be comfortable".

I think it would be cool to know which zone is the furthest from compliance and base priority that way or even fixed by time of day. Integrating occupancy sensors or even say a particular light being on could alter priority.

e.g workshop could be eliminated entirely unless occupied, but then there is freezing issues.

It really looks like the lack of hysteresis would have to be satisfied by minimum run time. I'd probably consider a minimum time per zone, possibly fixed in yor program, but variabe per zone. Time takes the plce of hysteresis.

 

[jpanhalt]

Hi KISS,

1) No A/C (log home)
2) While there are 2 circulator pumps in series, there is a cross connection outside the boiler, which is according to the manual. System works fine with absolutely no flow to the house, as is the case in the Summer. I cannot envision a failure mode for what I am doing that would be worse than no circulation. All open is fine, but basement will get warm. I turn off the furnace when temps are above freezing and I plan to be away >2 days.
3) Basement is below ground, and so far as I can remember, its zone valve has never been open, except when I first bought the house. I doubt it will freeze in Cleveland to that depth.
4) I have read, but not verified, that the new Honeywell thermostats do have a minimum run time, such as 2 or 3 minutes. I plan to do some testing on the interactions between the new thermostats and what is called an Aquastat (Honeywell) that I believe can control turn off the furnace without similarly affecting the circulators. I need to verify that the current wiring of my system allows that. I think it does. I was planning to disconnect one of the thermostats (2 wires, red and white) and simply attach a toggle switch across the contacts on the zone valve. That way, I can get better timing for events.
5) I really hadn't considered using a microcontroller until I dug deeply into the system. Now, it seems like an easy and fun solution. So long as I stick to just acting like a thermostat, all of the built-in furnace safety features remain operative.

Thanks for the input.

John

 

[KeepItSimpleStupid]

BACNET is a building automation protocal, but there are RS485 stats. Modbus stats and BACNET stats. Here's **broken link removed** a standalone BACNET stat, but definately commercial oriented. Comercial stuff seems to be based on occupancy and "shifts" and holiday schedules while residential is based on awake, leave, home and sleep with weekday and weekend thrown in.

You have override implemented in various ways. The Carrier Infinity picks like 3 hours and changes it in 15 minute intervals. it seems like reading the outdoor temperature and override is used the most. It's also nice that it either reads the system variables or you enter them BTU's.
It's nice enough to read the furnace error codes too.

The stat above says it uses PID and modulating valves. You don;t see modulating vales in home systems, but I don;t see any reason why they could not be used with the right controller. PID, I guess, is the real way to manage the temperature at say wake. The stat figures out what time it has to start turning on to get the desired setpoint when you want it. So, 72 at 8:00 AM is 72 at 800 AM and not turn the system on at 8:00 Am with a SP of 72. It may also do the reverse and try to make the lower temperature correct e.g sleep.

I never think it's a good idea to simplify at the start. It's best to understand a complex system and then simplify. Right now, you have a few goals in mind with simplicity.

You "could", in effect, get longer run-times if the valves were modulating which is basically having your cake and eating it too, but it doesn't make for a simple system.

We had a really stupid system where I worked. In theory 32 air to water heat pumps could actually use the heat from other rooms. The engineers" chose a "timer over-ride" for a heat pump system. It didn't work because it was not fast enough. I do believe the "loop temperature" changed at bit for summer and winter, but it wasn't a traditional fan coil hot/cold loop system. The "engineers" even put two heatpumps in the same room with two "different" thermostats. i.e really dumb. hey also messed up the slopes for the condensate drains.

"creature comfort", energy savings, equipment lifetime all have to be weighed. What we have found is that the temperature has to go up a bit when it rains and I will run the AC and heater at the same time in the WINTER in the car to reduce humidity and effectively increase creature comfort. It's not a concept that most people understand. Defrost and defog are different functions. The "creature" is happy with "defrost" and "Defog". The engineer is happy with heat, AC and heat+AC.