Welcome! Here are the website rules, as well as some tips for using this forum.
Need to contact us? Visit https://heatinghelp.com/contact-us/.
Click here to Find a Contractor in your area.

Temp differential for steam

Dave_61
Dave_61 Member Posts: 308
Hello All,
We have an Ecobee thermostat for our one pipe steam. I know there are some thermostats where one cycle per hour is recommended for steam. 
Our Ecobee has temp differential instead. I’m trying to balance fuel (oil) efficiency with comfort. We typically set our temp for 68 during the day and 64 at night. What would be a good differential for the thermostat? I believe ours is set for 1.5 degrees presently. 
Thanks 
Dave

Comments

  • EBEBRATT-Ed
    EBEBRATT-Ed Member Posts: 15,455
    Smart stats don't generally work well with steam. Your just going to have to play with it to get what you want.

    Some do not recommend any set back. You might try less set back. Generally longer run times are preferred. Shorter run times = more stable temp control but can cause short cycling of the burner
  • bburd
    bburd Member Posts: 909
    A 1.5°F differential should provide a relatively long cycle, which is needed for steam systems. You want it to run long enough on a normal call for heat (not recovery from setback) to provide even heat throughout your home, but not long enough to cause wide temperature swings. You can experiment with larger and smaller differentials to see what works best. 

    Bburd
    Dave_61
  • Dave_61
    Dave_61 Member Posts: 308
    Thanks. Typically on a call for heat (not from setback), our boiler runs about ½ hour to satisfy the thermostat.
  • Hap_Hazzard
    Hap_Hazzard Member Posts: 2,846
    That's just right. It needs to run long enough to make it worth the startup costs. If the cycle is too short, all it does is heat up the pipes, so you end up with a nice, warm basement, but you're freezing your buns off upstairs.
    Just another DIYer | King of Prussia, PA
    1983(?) Peerless G-561-W-S | 3" drop header, CG400-1090, VXT-24
    bburdethicalpaulDave_61
  • dopey27177
    dopey27177 Member Posts: 887
    For your needs a 2 degree set back will do you just fine. You will not notice any difference to the heating in your home, you will save a little on the fuel bill and the system on the real cold days will get a small rest and allow any condensate that leaves the radiators slowly to drain out of the radiators.

    Jake
    Dave_61
  • Dave_61
    Dave_61 Member Posts: 308
    Jake,
    When you say setback, are you referring to turning down the heat at night? I was asking about temp differential. So, if the thermostat is set at 68, a 2 degree differential would mean (to me) that the boiler would not come on again until the temp got down to 66. Just trying to balance efficiency with comfort. Thanks
  • Jamie Hall
    Jamie Hall Member Posts: 23,170
    Dave_61 said:

    Jake,
    When you say setback, are you referring to turning down the heat at night? I was asking about temp differential. So, if the thermostat is set at 68, a 2 degree differential would mean (to me) that the boiler would not come on again until the temp got down to 66. Just trying to balance efficiency with comfort. Thanks

    That understanding of setback is what most of us would agree on -- turning the set point temperature down during certain parts of the day. What you are referring to -- a differential between turn on and turn off temperatures in the thermostat, in this case -- would be referred to as differential or, in the controls world, as dead band.

    The primary difference, however, is the length of time which is involved, and the impact of the two in terms of overall heat demand is somewhat similar.

    I might add that in most control schemes dead band is to be minimized or avoided, as it conceals or at best confuses deliberate alterations with uncontrolled variations in the controls.

    And easier example to understand, perhaps, is in the steering of a vehicle. Considerable effort is expended in power assist steering systems to ensure that the vehicle responds evenly, smoothly, and consistently to a change in steering wheel angle. However, suppose that there is a dead band -- a difference between on and off positions -- of the wheel. Nothing would happen, let us suppose, for the first quarter turn either way from dead centre. This would be quite analogous to your dead band differential -- but would be amazingly difficult to drive!

    A properly designed and calibrated thermostat in a system (and it is part of a system) should have little or no differential in responding the space temperature (no differential is almost impossible to achieve, but it can be made very small). I regret to say that most modern digital thermostats do have a dead band or swing, and that it can be quite large. A few of the best have the necessary computation power to minimise that; the old mercury T87s can, in fact, be adjusted to achieve that, but they do depend on being correctly set up and adjusted -- which was almost never done.
    Br. Jamie, osb
    Building superintendent/caretaker, 7200 sq. ft. historic house museum with dependencies in New England
  • Dave_61
    Dave_61 Member Posts: 308
    That was a very helpful analogy. My thermostat has an adjustable differential. It is set from factory at 0.5 degrees. It can be increased to 2 degrees. Are you recommending leaving at 0.5 degrees? That would mean boiler would come on at 67.5 degrees.
  • Jamie Hall
    Jamie Hall Member Posts: 23,170
    Not necessarily. There is a balance to be had between perceived comfort, suggesting a smaller dead band, and system efficiency and, in some cases, evenness throughout a house or building. For example, some people are sensitive to small changes in temperature, and they might want a small dead band. On the other hand, having longer on and off cycles minimizes the number of starts, and particularly with cold starts (most steam systems and hot water systems) this can increase overall efficiency. Then again, in some installations some radiators etc. may not get heat as soon as others, and a small dead band might result in the system shutting off before that back bedroom got any heat at all.

    One size most definitely does not fit all!
    Br. Jamie, osb
    Building superintendent/caretaker, 7200 sq. ft. historic house museum with dependencies in New England
  • Dave_61
    Dave_61 Member Posts: 308
    Jamie, Thanks! Ours is set at 1.5 degrees. When the house is at temp (68) on a 30-40 degree day outside, the heat comes on about once every 2-3 hours and runs for 30 minutes before reaching temp. All rooms seem comfortable. I think I’ll leave it there for now and see how we do when it gets much colder. 
  • Hap_Hazzard
    Hap_Hazzard Member Posts: 2,846
    Some thermostats refer to this as "swing"—the number of degrees the temperature must fall before a call for heat is made. 30 minutes is a good recovery time. Anything shorter might leave you with a warm downstairs, cold upstairs situation. (Some people actually prefer this because their bedrooms are upstairs and they prefer having it cooler when they're sleeping.)
    Just another DIYer | King of Prussia, PA
    1983(?) Peerless G-561-W-S | 3" drop header, CG400-1090, VXT-24
  • dabrakeman
    dabrakeman Member Posts: 538
    On my Ecobee as the default I use a 0.5F differential. This is because Ecobee has some nice adjustability in the program scheduling that allows you to have it run at 0.5F or 1.5F just depending upon what comfort schedule you use for that part of your program. "Away" setting adds an additional 1°F degree to the existing 0.5 differential, the set point is 1.5 degrees when ever the thermostat is in Away mode. When building a program schedule you can also specify whether that comfort setting is for when the home is Occupied or Unoccupied. Doesn't matter whether your are really going to be home or not but the convenient point is that if you specify it as unoccupied it adds 1°F degree to the existing 0.5 differential.

    In general smaller differential is more comfortable but can use more energy, however, if you have an oversized boiler and it is very cold out the larger swings can result in too long of cycles which may result in your boiler building pressure and tripping the pressuretrol.

    Can we assume you are having no issues (cycling on pressure) with your recoveries from the 4 degree night time setback? It would depend largely on whether your boiler is much oversized for your radiators or not.
    If you are oversized and seeing issues then another nice feature is that you have unlimited number of program steps you can put in a days schedule. Thus, you can use 30 minute program steps to stagger the recovery forcing it to never run more than 30 minute cycles. I.e. target 68F early in the morning and 1/2hr later drop it back to 64F then again target 68F for the next 30min and drop it back for the next 30min. A few such cycles should be enough to manage your 4 degree recovery even in the coldest weather. I usually start the sequence with a step that is a deeper setback so as to prevent the 1st recovery cycle starting when the boiler was already on which would results in a cycle longer than 30 minutes. Note you can also go into your settings and set a "minimum run time" which is good to prevent these smart stats from being too smart and giving you a worthless 5 or 10 minute cycle. I set my minimum at 15 minutes.

    The attachment photo depicts what I am talking about in the above paragraph.

  • SteamBoiler
    SteamBoiler Member Posts: 90
    edited December 2021

    On my Ecobee as the default I use a 0.5F differential. This is because Ecobee has some nice adjustability in the program scheduling that allows you to have it run at 0.5F or 1.5F just depending upon what comfort schedule you use for that part of your program.


    I am following this discussion as I just installed a Nest learning thermostat. The temperature differential isn't programmable in Nest, nor is a "minimum run time". I got the Nest purely based on reliability ratings being better for the Nest than Ecobee, however it definitely looks like Ecobee is more flexible for the power user. If set for 68F, the Nest calls for heat at 67F and stops the call at 68F or slightly below, and lets the radiators warm up the room to 69F. In general on 35F days it appears that Nest calls for heat for 25 minutes every 2 hours or so and this run time is letting all radiators in the home heat up - 2 are slower than the rest. The runtime is also short enough that I am not cutting out on pressuretrol max pressure. Our previous mercury thermostat rarely called for heat long enough for the 2 late radiators to warm up, and I was suspecting trap issues.

    Thus, you can use 30 minute program steps to stagger the recovery forcing it to never run more than 30 minute cycles. I.e. target 68F early in the morning and 1/2hr later drop it back to 64F then again target 68F for the next 30min and drop it back for the next 30min. A few such cycles should be enough to manage your 4 degree recovery even in the coldest weather.

    I posted previously about potentially ramping back slowly and gracefully from a deep setback like you have. It makes a lot of sense and given the flexibility of the new smart thermostats there should be ways for a power user to marry them gracefully to steam heat, and I love the ability to see the logs of heat call duration and frequency.
  • dabrakeman
    dabrakeman Member Posts: 538
    The ecobee also allows you to turn off the smart recovery which is basically anticipation. I like to do that because I guess being a control freak I like to just know when my boiler is going to turn on and off without some learning algorithm that I don't fully know trying to interfere (based on outside temperatures etc...).
    Yes, I know a lot of people on this forum are not fans of smart thermostats and they certainly can be a problem if not installed right or the fundamentals of system maintenance aren't taken care of first.  Before playing around with thermostat control make sure you have knowledge of your boiler sizing vs EDR of your radiators, your main venting is adequate, your pipes are insulated, your pigtail is clean, you have a low 0-3or5psi pressure gage to truly see pressure and you have your radiator venting balanced (or at least balanced to as you like it).

    Everybody's situation is different and mine is mainly driven by the desire to do setbacks coupled with having an oversized boiler.  If you are just doing set it and forget it or your boiler is properly sized to the radiation then this level of management probably isn't necessary.  I do have data over full heating seasons showing that in my case doing the setback does save gas usage vs heating degree days.

    These smart thermostats can be expensive but if you thought about changing seems to me that most power companies offer very respectable discount vouchers on the purchase of smart stats.  Worth checking yours if you didn't already.
  • PMJ
    PMJ Member Posts: 1,265
    Smart thermostats have trouble with steam because the time from fire to steam at the radiators varies so much depending how long ago the last burn was. That time can be on the order of 25 minutes in a cold start to 1 or 2 minutes in a short cycle into a hot system. So their algorithm is dealing with a variation of reaction time to actually begin adding heat to the space that might be 25 times. You can fiddle away with those things but you will never know exactly what their algorithm is doing. They will get the job done of course, and they give you a really nice graph of what happened after the fact.

    But I think what is really the best steam heat needs the same partial fill of all radiators on every burn regardless of the starting temperature of the piping. I achieve this by running every burn until I get a make on a remote temp sensor on a remote rad fill pipe plus an adjustable timer. Sometimes it takes 25 minutes. Mostly though it is more like 7-9 minutes. In really mild weather that one burn may satisfy the stat but overshoot will be less considerably than if you burned all the way to stat satisfaction. As it gets colder and that one burn didn't satisfy I initiate the next burn an adjustable time after that same temp switch opens - which signals that enough of the heat from the previous burn had actually gone into the space. This way I maintain partially filled radiators adding heat more evenly and slowly than the standard control does. It results in a maximum of 2 to 2-1/2 burns per hour. The colder it gets the time between firings gets shorter and the number of burn cycles per call goes up. The system adjusts itself to the conditions. I haven't changed the 2 timer settings in several years now. The burner runs 210 seconds after a close on the remote sensor and the burner can't start a new burn until 240 seconds after that switch opens again. The colder it gets the more the radiators are just sort of the same all the time.

    Bottom line, dead band, swing, or temperature differential means nothing to me with regard to firing the boiler. I don't think good burner control for steam can be based just on temperature at the thermostat no matter how smart it is.
    1926 1000EDR Mouat 2 pipe vapor system,1957 Bryant Boiler 463,000 BTU input, Natural vacuum operation with single solenoid vent, Custom PLC control
  • dabrakeman
    dabrakeman Member Posts: 538
    No argument with PMJ but not all of us homeowners have his prowess with the electrical rigging of the system. My only point was that there are features of some smart thermostats that can be very easily and relatively inexpensively utilized by a homeowner to help mitigate potential drawbacks of setback recovery particularly with oversized boilers. The approach PMJ has described (and described before) is undoubtedly more "robust" for a recovery in that it is directly controlled off % steam fill which is the variable you are trying to control to avoid a 100% steam filled system and any subsequent pressure rise. It is better even than controlling off low vaporstat pressure rise directly since you really want to turn the boiler off before you get pressure rise... 30 minute cycles as I said can work and address recovery woes in many situations but for some depending largely on how oversize your boiler is or how small the entire system is even 30 minutes runtime for a 2nd, 3rd or 4th 30on/30off recovery cycle might still be too long (although generally not in my case).
  • PMJ
    PMJ Member Posts: 1,265
    No argument with PMJ but not all of us homeowners have his prowess with the electrical rigging of the system. My only point was that there are features of some smart thermostats that can be very easily and relatively inexpensively utilized by a homeowner to help mitigate potential drawbacks of setback recovery particularly with oversized boilers. The approach PMJ has described (and described before) is undoubtedly more "robust" for a recovery in that it is directly controlled off % steam fill which is the variable you are trying to control to avoid a 100% steam filled system and any subsequent pressure rise. It is better even than controlling off low vaporstat pressure rise directly since you really want to turn the boiler off before you get pressure rise... 30 minute cycles as I said can work and address recovery woes in many situations but for some depending largely on how oversize your boiler is or how small the entire system is even 30 minutes runtime for a 2nd, 3rd or 4th 30on/30off recovery cycle might still be too long (although generally not in my case).
    Agreed. Lots of things you can tweak in the smart stats do do a lot. At the same time what I have ended up doing is quite simple and doesn't need a PLC. Maybe I will draw up a hardwire version.
    1926 1000EDR Mouat 2 pipe vapor system,1957 Bryant Boiler 463,000 BTU input, Natural vacuum operation with single solenoid vent, Custom PLC control
  • johncharles
    johncharles Member Posts: 50
    edited December 2021
    I don't comment frequently, but this topic peeked my interest.

    I have a 1 pipe steam system, which is very well setup and properly sized and am going into my thrid winder with an ecobee.

    I really like the ecobee and could not speak more highly about it with regard to how well it controls steam.

    A couple of things I want to say:
    1. In my house I could never live with a large differential it starts to get noticeably cold fairly soon after the boiler shuts off so I leave my ecobee set to a 0.5 degree differential.
    2. The ecobee and the Nest are not comparable, while the ecobee offers many smart features it is much more configurable and honestly better than a nest.

    I think the answer to the OP's question comes down to comfort if you are comfortable with a 1.5 degree diff then go for it, I'm not honestly, and I don't really think the fuel savings are worth feeling like you need to layer up in your own house all winter.

    With that said the ecobee really "gets" radiant heat, it understands overshoot so you have to realize that on a 1.5 diff it won't actually run all the way up to 1.5 degrees, it will get your house pretty close to 1.5 degrees. This is different than most thermostats. What I mean by this is that it learns how much your system over shoots each cycle and at what out side temperatures it overshoots by how much, and will cut the cycle time appropriately so that your house will usually coast to the setpoint within about 5-10 minutes after the cycle stops.

    @Dave_61 I highly suggest trying out beestat.io it is an alternative app/site to view your ecobee and it presents things in a much nicer way, it also includes the temperature reported as a decimal instead of rounded which is really helpful to see exactly what ecobee is doing. Using beestat is how I've been able to figure out what I write below.

    @Jamie Hall, I believe ecobee is one of those rare exceptions that really understands how to keep things from over heating. As I've said my ecobee has been controlling a 1 pipe steam system for 3 seasons now. During that time we've had plenty of mild shoulder weather and also near record cold for our area (a 20 year low). I keep my house set at 70 with a 0.5 swing (no setbacks) on warmer days in the 50's the ecobee will start the boiler when the house falls to 69.5. With a traditional swing thermostat the burner would run until the house reached 70.0 at which point the burner would stop and the house would over shoot by some margin. The ecobee doesn't actually do this. On a warm day the ecobee will turn off the burner around 69.7 (a 0.2 degree swing) and it knows that it has enough heat in the radiators to carry the house up to 70.0 before the temp starts dropping again. It's fairly smart about this though and as the out side temperature drops it will run the burner longer. For example when it's in the 20's and 30's it usually leave the burner on until it reaches 69.8. In the teens outside it will keep it on longer still. The only time I've seen it actually hold the burner on until 70 was when it was -20 outside a few years ago which was a near record low for our area.

    The down side is that since it doesn't have an actual outdoor sensor and uses the internet weather of the area it can be a little mis-calibrated I notice that when the temp outside is rising fairly quickly the house will get a little warmer since the internet weather usually hasn't updated in time. Likewise in the evenings when the outside temp is falling it can occasionally undershoot the temp and leave the house a little cooler. But this is usually only by a few fractions of a degree.

    I really can't recommend the ecobee for steam heat enough, it really does all the right things when it comes to keeping your house at a super steady temp, additionally it's very good at not overshooting even on those fairly warm days. I've been very impressed over the last few years.
  • PMJ
    PMJ Member Posts: 1,265
    @johncharles,

    This is surely good news. Logically there is a great deal that can be accomplished at the thermostat with the computing power now available to learn the reactions of each system. Sounds like Ecobee has made great strides and is something homeowners can easily do. If I were going that direction I am sure I would be setting for the smallest deadband it would allow to be set.

    You obviously have experienced the overshoot problems of the standard control. I'm glad to hear Ecobee has improved on that. Such improvement was desperately needed. 

    Outdoor temp is surely an indicator of how much heat is needed and better than nothing. The internet temperature has no sensor to go bad. It does require an internet connection. Sun and wind on your structure does affect things a lot beyond the absolute temperature. What I do reacts to all those things in real time.

    Anyway thanks for your comment. So glad you are much happier with your steam heat than previously.
    1926 1000EDR Mouat 2 pipe vapor system,1957 Bryant Boiler 463,000 BTU input, Natural vacuum operation with single solenoid vent, Custom PLC control
  • Jamie Hall
    Jamie Hall Member Posts: 23,170
    You know, I suppose the internet temperature works well. In some places... not where I hang out. I have three places to care for, and there is one internet temperature source in the vicinity. One is in a valley. Very little wind. One is on a hilltop, a lot of wind. One is halfway up a hill, but little wind owing to trees. The one in the valley normally has free air temperatures from 5 to 10 degrees higher than the hilltop and hillside ones except on still mornings when it can be 5 degrees colder...

    The internet temperature is in the next valley over, and is more like the valley building.
    Br. Jamie, osb
    Building superintendent/caretaker, 7200 sq. ft. historic house museum with dependencies in New England
  • dabrakeman
    dabrakeman Member Posts: 538
    I am sure the temperature is used as a relative variable for that system in that location. I.e. it knows with time that at that location it may take X% longer for the temperature to come up to set point when the internet says it is 10deg vs when it says it is 40deg even if it is really 2 deg and 30deg. Nonetheless, you are right if your location is sometimes below and sometimes above the regional temp there will be still some error.

    Although this feature works pretty well in the Ecobee as I mentioned for my own purposes I do not use it. I turned it off when I went to the recovery control strategy I shared. If my goal was to limit my run times to less than 30min then for any recovery temperature I program in "smart recovery" would just cause the system to start as early as it needed to and thus in my case too long of a cycle. With it off I do of course get some overshoot but not more than I am content with. The Ecobee support staff has been surprisingly receptive and willing to talk about these matters. Generally you would expect just a canned ignorant answer to questions such as "why can't you just allow a user to set a maximum runtime and a minimum dwell if smart recovery is turned off"? After some back and forth they understood but of course steam users are a miniscule percentage of their user group. Maybe we need all steam heat Ecobee users to coordinate and ask the same question.
  • PMJ
    PMJ Member Posts: 1,265

    You know, I suppose the internet temperature works well. In some places... not where I hang out. I have three places to care for, and there is one internet temperature source in the vicinity. One is in a valley. Very little wind. One is on a hilltop, a lot of wind. One is halfway up a hill, but little wind owing to trees. The one in the valley normally has free air temperatures from 5 to 10 degrees higher than the hilltop and hillside ones except on still mornings when it can be 5 degrees colder...

    The internet temperature is in the next valley over, and is more like the valley building.

    I agree it is a rough approximation. It will tell the system if the cold is mild or extreme.

    I hesitate to set up on things that require an internet connection when I don't have to. The Ecobee will still run I assume without a connection. I am always thinking about how things will look with no connection.

    Re the temperature input. It turns out that the rate of steam condensation in each and every radiator is reacting in real time to the conditions in each room. Also, the amount that will condense per hour through the same partial fill level varies a lot. Time between burns is reduced as it gets colder so the same filled radiator area needs a slightly higher average temperature over the same time period to get the job done.

    Anyway, I have found that just letting the one temp sensor and the two timers maintain a partial fill level during the call is in fact a self adjusting system reacting in real time to all of the conditions actually impacting the structure. Coupled with natural vacuum on the off cycles it responds to wind and sun loads favoring the colder areas within the structure with more steam. This is by the way how the dead men controlled things originally. The damper on the fire was controlled by how fast the steam was condensing changing in real time obviously with the conditions. I am just doing the same thing but with on/off.
    1926 1000EDR Mouat 2 pipe vapor system,1957 Bryant Boiler 463,000 BTU input, Natural vacuum operation with single solenoid vent, Custom PLC control
  • dabrakeman
    dabrakeman Member Posts: 538
    Yes, it runs your program without internet connection. If using the smart recovery obviously it would not have the outside temperature information so it must default to some assumption.

    @Dave_61 sorry if we hijacked your thread a bit. You said you were going to reevaluate things when the weather gets colder. If you notice issues there are options. PMJ would probably help you out to setup a system similar to his if you were interested (@PMJ, sorry to volunteer work for you :) ). He offered for me but I am more of a mechanical guy and the spaghetti wiring at my boiler caused trepidation.
  • dabrakeman
    dabrakeman Member Posts: 538
    @Dave_61. Previous discussion on this topic took place in the following thread:
    https://forum.heatinghelp.com/discussion/comment/1640391#Comment_1640391