If you're posting a request for installation help, please provide as much info about your system as possible. For example, for help regarding connecting a smart thermostat, please include details about your system including whether you have gas/electric, heat pump, number of heat stages, A/C condenser, etc.

Humidifier and window efficiency

rlbrlb NJMember

The manual says the following:

"Window Efficiency: The humidifier frost control setting uses this efficiency rating to compute the humidity set point within the range of 15% to 50%. Values are from 1 to 7, with 7 being for the most efficient windows possible."

What does "most efficient windows" mean? Does anyone know the algorithm behind this?

I'm assuming this means the target humidity is adjusted based on outdoor temp, can anyone confirm or deny this?





  • SteveSteve Seattle, WAAdministrator
    Answer ✓

    Great question, @rlb.

    I just happen to have this graph from ecobee that dates back to January 2013. This was Version 1 of their frost control for the ecobee Smart, with 3 settings

    It demonstrates their original "maximum indoor relative humidity set point one can expect relative to temperature changes along the Y-axis." This was shared with the old email list (that pre-dated this forum). You can read some of the discussions from back then in the archives:


    Based on user feedback, ecobee updated with their algorithm and offered more granular control over the window efficiency.

    In December 2013, ecobee push a beta firmware update to a handful of users based on these curves:

    I believe these same 7 settings are the ones on the ecobee3 (and based on my experience, they work great).

    Here's the discussion from back then about those new curves:


    This is probably a MUCH bigger answer than you wanted... but the bottom line is that my best advice for finding the best setting for your house is to wait for a really cold week, and then experiment until you find the maximum efficiency number that works for your windows. :)

  • my0gr81my0gr81 coldnadaMember

    based on the 2nd graph, most double glazed glass could use 4, the best could use 5. 6 & would be for the triple glazed glass. This is all assuming that your are heating and humidifying 100% outside air, which in most high efficient houses is not the case.

  • SteveSteve Seattle, WAAdministrator
    edited December 2014

    @my0gr1: Empirically, I think that sounds about right. My windows' published ratings at the Utah house is 2.0, but that's probably optimistic. I am using setting 5 with excellent results.

  • dborndborn Montreal, CanadaMember

    This is very interesting detailed information. My house needs a DEhumidifier much more than a humidifier. Too bad the Smart (at least) doesn't have very fancy dehumidifier controls or logic.

    I know this is an item being considered but will it be backported to the Smart or only the 3? I'm trying to do this type of logic with my (infamous) perl script :) One potential problem with that is the minimum dehumidify set point seems to be 30% which may be a bit too high for very cold temps.

    I wonder what happens if I send a dehumidify value below 30 through the API. It will take it but will the (Smart) thermostat handle it properly with my current firmware..? I know I tried sending odd numbers (like 53%) and it seemed to cause some problems with at least the android app if not the tstat.

    Anybody know what's in the plans for better dehumidifier logic from ecobee?

  • my0gr81my0gr81 coldnadaMember
    edited December 2014

    @dborn remember that these charts imply relative humidity for 100% outside air, since you are heating inside air, you need to adjust for that. Only freezers and the deep North latitudes ever have constant relative humidity below 30%. If you set the minimum at 30, when indoor temperature is about 22C, it should be fine, unless you are using this for another purpose.

  • SteveSteve Seattle, WAAdministrator
    edited December 2014

    I'm tagging @nashib and @masleem on this for no other reason than I think they'd think those charts are historically cool, since they represent yet another excellent example of a product improvement that was a direct result of user feedback. :)

  • rlbrlb NJMember
    edited December 2014

    Thanks for all of the feedback here. On those charts the y axis is labeled Maximum relative indoor humidity. What other factors are determining where your target humidity sits along that curve?

    I finally got my humidifier wired into the ecobee this morning, so before that I didn't have much understanding of how the stat was trying to maintain humidity. The only humidity control I see is the window efficiency rating. I haven't found any other setpoints (unless I'm missing a menu option).

    Based on all of this info I suspect there is additional math going on, or I'm missing a setting. Can anyone shed light on this?

    Edit: Reading the manual helps :) I was working from home today, and I decided to wire up the humidifier before I started this morning. I didn't get a chance to do much else until now. I found the humidifier settings (off, on, frost control) but based on my first paragraph I still suspect there is more.

  • nashibnashib TorontoEcobee Employee

    @Steve & @my0gr81 - great responses. You guys are right. There is some degree of trial and error involved.
    Also, we were listening to our very valued customers !

  • dborndborn Montreal, CanadaMember

    my0gr81 said:
    @dborn remember that these charts imply relative humidity for 100% outside air, since you are heating inside air, you need to adjust for that. Only freezers and the deep North latitudes ever have constant relative humidity below 30%. If you set the minimum at 30, when indoor temperature is about 22C, it should be fine, unless you are using this for another purpose.

    From my psychrometric chart and formulas, the current outside temperature is -3.9C and outside humidity is 91%, converted to current indoor temperature of 20.7C, corresponds to 18% relative humidity so this air is plenty dry for use with my HRV (which is warming up and feeding 100% outside air to the inside). I think we are talking about the same thing. Does Montreal qualify as "deep North latitudes"? ;) we do get down to -40C sometimes even if not for long.

    The current humidity in my house is 41% and that's with no help of a humidifier and the HRV running when needed to keep it down.

  • my0gr81my0gr81 coldnadaMember

    @dborn I lived in Edmonton for 20years with regular trips to Hay River and Yellowknife, so yeah Montreal doesn't qualify. 41% indoors @20c is plenty comfortable and shouldn't result in frosting on the windows unless:

    • overnight or away setback is too low IE. Lower than 18C causing the excess humidity to condensate on the even colder windows
    • single pane windows, or faulty insulation around the windows is causing the warm air to condensate on the glass. This is also occurs with old wood windows.

    One is an easy fix, though it means foregoing any savings from setting the temperature back. The other not so much.

  • I've been doing the trial and error over the last week with my newly installed humidifier and ecobee3. It's been fun, outside of mopping up frost and condensation off my windows in the morning. Playing around with the settings is great, and it's cool to see an example of ecobee tech support listening to customer needs and using their input to drive development.

    I'm finding that at window setting 4, it was running too much, and at 3, it wasn't running at all. In looking at the figures that Steve posted, it's hard for me to see which curve my ecobee is tracking. Like today, it's 30 f outside, my stat is set at 71, and the frost control output is to set at 25% humidity. Does that seem right? This is with window setting 4 to try to get the humidifier to run a bit more. With the extreme cold temps of 0 to negative deg f last week, I was getting frost on the windows with 25-29%, which was no good, and I backed off to window efficiency 3. But I felt it was getting too dry (20-21%), and the humidifier was never running, so I bumped it back up. Maybe I need to stick with manual control in extreme cold situations.

  • JohnLivdahlJohnLivdahl MNMember
    edited January 2015

    I can't get the frost control feature to do anything that makes sense or is useful to me. Assuming that the downloaded raw data Humidity set Point is accurately reflecting the frost control value (I verified this with support phone conversation and by correlating humidity calls with that set point and current humidity column).

    From the charts above (and industry standard auto humidity control curves), humidity should be reduced 5% for every 10degrees F outdoor temperature change at a given indoor heat set point.

    My data for window efficiency setting 4, my humidity set point only varies 5% total (20-25%) for a outdoor temperature swing of 36 degrees F (-13 to 23 degrees F). That outdoor temp swing should have seen a 18% humidity swing. And the humidity swing max and min don't even track the max and min outdoor temperature values.

    This is useless for my humidity control needs and this is one of the main reasons I justified getting the ECOBEE3. In the past I have used Honeywell's IAQ Stat and it worked perfect on my other home for frost control.

    Supports response was: "Our QA team did some test and verified the calculation using your data. We found that it is working as it should". I am not impressed with that response when you read my email to them and the phone discussion I had with support. I expect a more detail explanation of why they think this is working.

    I have had to change to manual humidity control mode. I now need to change the set point my self to prevent frost issues (Nest could do that).

    Can other's using this function please check their raw data and see if it is working as the graphs above show?


  • dborndborn Montreal, CanadaMember
    edited January 2015

    (nevermind, I hadn't noticed I already mentioned by perl script previously in this thread).

  • JohnLivdahlJohnLivdahl MNMember
    edited January 2015

    Has any used this frost control feature with temp in F vs C? Plotting the data I have taken, it is ignoring outdoor temp and just setting the humidity based on the indoor temp sensor. I wonder if certain weather reporting stations report the data in a format the tstat formula can't interpret? The outdoor temp shows fine on the stat?

    This seems like too big of a bug for there not to be lots of complaints.

  • zeroskatr512zeroskatr512 Reading, PAMember
    edited January 2015

    I have frost control enabled and I agree it doesn't seem to vary the max humidity per the outdoor temperature like one would expect. The humidity set-point when it was 7F outside & 70F inside was 38%. The humidity set-point when it was 30F outside & 70F inside was 42%.

    On that same note I would dispute the reading accuracy that the ecobee3 indicates for humidity. I have two other hydrometers (in a digital clock & and a Davis weather station) that read within a 1-2 % of each other and then the ecobee is usually 8-10% lower.

    Anyway, I don't think enough people realize there is actually a problem. If they see condensation buildup they probably just lower the efficiency in the frost control setting thinking it is the windows.

  • ToreadorToreador United StatesMember

    I'm guessing some of the firmware versions have different intelligence for the humidifier. What version is everyone running?

  • JohnLivdahlJohnLivdahl MNMember
    edited January 2015

    Toreador said:
    I'm guessing some of the firmware versions have different intelligence for the humidifier. What version is everyone running?

    I have had this behavior for these FW so far:,, ( I would have to test this latest version longer to hit a cold spell to see if it will respond to temp < 20F, This weekend should have over-night lows at 8 F.)

  • JohnLivdahlJohnLivdahl MNMember
    edited January 2015

    Here is a graph of my current data for the last 5 days. The Last 1.5 days is the latest FW. I don't see humidity set point to outdoor temp correlation. (Note Current Temp is indoor temp)

  • ToreadorToreador United StatesMember

    Perhaps it also uses exterior humidity or forecasts or average min temperature.

  • JohnLivdahlJohnLivdahl MNMember
    edited February 2015

    Outdoor Humidity doesn't affect indoor frost on windows. I have analyzed data now from a ECOBEE set for C and it is also not controlling indoor humidity based on outdoor temp. This functionality is broken. I don't understand why there isn't a lot of complaints. Must be the warmer than normal spell we are having.

    I have analyzed the data and it is not using any measure of outdoor temp (unless it is monthly average and that would be stupid). For frost control to work it has to be based on current temp or forecasted low in next 24-48 hours. But that was my first question to support and QA confirmed that it only uses current outdoor temp and indoor temp.

    Support is still waiting to hear back from development team on this issue.

  • briansusernamebriansusername UtahMember
    edited February 2015

    I have my ecobee3 set to frost control and my window efficiency set to 4 as well. I also bought the ecobee3 partially for the automatic humidifier control, since I never remembered to turn the humidistat in the basement down when I should (or I left it too low to be useful all the time!).

    I plotted 10 days worth of my data similar to yours and there did seem to be a correlation between the humidity set point and the outdoor temperature when I first installed everything (firmware 3585). I've been playing with settings a lot, and I don't remember everything I did, so that might skew some of the data. I had my humidifier set up as "Steam" for the first few days (and window efficiency 3, I think) and it has been in "Evaporative" since then.

    In the last week there is not much correlation in the data - and the last three days are on firmware 3957. However, I can't say it doesn't match the curves Steve posted since the outdoor temperatures haven't dipped much below 30F here in the past week, so there's not much on those curves to test with the mild weather lately!


  • tbc007tbc007 MontrealMember

    I can definitely say that Frost Control is out of whack... on a setting of 4 for window efficiency, it barely goes higher than 22% when it's -8c outside (22c indoor). This doesn't even come close to ecobee3's own description for window efficiency setting of 4 which I believe is -11c and max of 35% RH...not home right now, so I'll have to double check, but I'm pretty sure its that.

  • JohnLivdahlJohnLivdahl MNMember
    edited February 2015

    I am so frustrated with ECOBEE!! Steve can you help with your contacts?

    Examples of broken humidity control. Stat set to Window Eff 4.

    Here is the E3 forecast display:

    So 22 F is the coldest it is forecast to be in the next 48 hours then 21 F next Monday. So based on the information that forecast is used for humidity control (along with indoor temp), one could assume that 22 F would be the worst case outdoor temp to use.

    The current humidity set point has been 24-25% RH for the past 13 hours. Current indoor temp 65-70.2 F. Current outdoor ~29 F

    Using the above chart for Double glazed:
    One would expect current set point for RH over 35% (for 22F) or RH of 40% (for 30F).

    For early morning 2/5/15 with indoor temp < 66 F and outdoor temp -9F from the above chart one would expect maybe close to 22% RH setting. The data was 25% RH setting for this period of -9F temps (the temp was below 0 for over 9 hours and -9F for 3 hours.

    So the E3's calculated setting was higher (25% ) when it was -9F outside then when it was 30F outside with no forecast temps below 22F in the next 48 hours. -9F to 30F is about 40 F swing and should have a have a 40/2 = 20% RH swing.

    This E3 has a lower %RH for 30F then it did for -9F. When -9F should have been about 20% lower. For Ecobee to say that is working is unbelievable!!

    And Ecobee says:
    "At the end of the day the algorithm seems to be working as expected (within the accuracy of the forecast) and we don't receive a lot of complaints about frost control."

    "We are confident in the algorithm applied and that it works to maintain comfort while preventing frost/condensation. I apologize that I can't explain more about how the algorithm works."

    Here is the raw data plotted, you can see the -9F low then it warm up to 30F:

    Here is just the humidity set point and outdoor temp plotted:

  • rosenquirosenqui Ottawa, OntarioMember

    I don't have the charts and extensive data to back it up, but I agree with others here that the humidity setpoint is out of whack. Once you set your window efficiency it doesn't seem to move very much.

    In contrast, the Carrier Infinity system I had at the previous house was fantastic for humidity control. You set your humidity level in relative terms (7 or 8 different settings from very low to very high humidity) and then the Infinity Control adjusted the setpoint based on the outdoor temp. It didn't have access to the weather forecast so it couldn't see a cold/warm front coming, but at least it adapted as the current temp changed. The ecobee3 temperature adjustment almost seems to be off by an order of magnitude - adjusting the setpoint by 0.1% instead of 1%.

  • JohnLivdahlJohnLivdahl MNMember
    edited February 2015

    Here is the continuation data for what the E3 did with that forecast I posted previously. You can see the temp over the weekend was much warmer and the E3 didn't raise the humidity set point (in fact the E3 lowered it another 1% today). This frost control functionality is completely useless. You can correlate the previous graph low of -9F with the new graph.

  • SteveSteve Seattle, WAAdministrator

    Tagging @nashib to check this thread and hopefully provide some feedback.

    It really hasn't been cold enough for me to empirically test with my e3 either!

  • johnwjohnw OntarioMember

    Really interested in seeing the response to the issues in this thread - I was just about to pull the trigger on the e3 until I came across this. Humidity control based on outdoor temp is exactly why I wanted to purchase this stat, and if the e3 doesn't work as expected then I'm out. I'm sick of sweaty windows but can continue to adjust rH manually if this stat won't do it for me.

  • ToreadorToreador United StatesMember

    I can verify that frost control does not work.

  • rlbrlb NJMember
    edited February 2015

    Glad to see so much action in this thread. We've been pretty cold here in NJ during that period and I've been looking at a lot of frosty windows. I've gone over my last 6 weeks of data and have found the same thing as everyone else in terms of humidity set point not correlating to the temperature very well.

    I've also found that the thermostat is sensing humidity way higher than the set point. So my frosty windows seem to be an independent issue. This is interesting to me, as I don't think my house has a very tight thermal envelope (i.e. a lot of air leaks). Or, the thermostat isn't correctly sensing the humidity. I'll have to think this over for a while.

    This was all at WE3. I just changed it to 1, I'm interested to see what this will change (not much I'm guessing).

  • rlbrlb NJMember

    As I look more at this I'm wondering what other factors are impacting my indoor humidity. You can see a few spikes on 2/26. The latest one was probably due to boiling water for pasta. Not sure what the earlier ones were.

    Also on 2/27 you can see a huge spike in thermostat humidity along with indoor temp. The furnace was off overnight and as the house warmed up the humidity increased. However, the humidifier hasn't run since January 25.

Sign In or Register to comment.