Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I wonder what the ideal one-size fits all thermostat looks like.

The one in my apartment has a “feature” a lot of US thermostats now have, where you set four ordered times called wake, leave, return, and sleep and the temperature you want the space in each interval. I know very few people who actually live in a household where everyone wakes, leaves, returns, and sleeps on the same schedule every day.

I work from home and personally just want to set a temperature and have the space stay at that temperature indefinitely but this system requires that I tap through and enter the desired temperature four times, while confirming the four intervals.

I guess I’d be happier with a more programmable thermostat that I could set to behave like an old school dial thermostat.



I'm still of the opinion that a dial works best. Especially in modern homes (in Europe at least), there seems to be a school of thought that you should just leave your thermostat at the same temperature at all times - the theory being that warming up a cold house in the morning costs more energy than maintaining a stable temperature.

Anyway, my ideal setup would be to install 'smart' thermostat taps on every radiator in the house, either manually turn them down when you're not in the room or have them automatically detect activity or open windows and adjust accordingly. But each one has the authority to trigger the central boiler if needs be, instead of only the master thermostat in the living room.


> the theory being that warming up a cold house in the morning costs more energy than maintaining a stable temperature

This is only true if the heating happens quickly and the system is less efficient when heating quickly. Otherwise, this doesn't make sense from a physics standpoint. A temporarily lower temperature differential means less kWh of heat lost.


This is a whole research topic, my PhD in fact!

FWIW I run my heat pump intermittently and with locally-smart TRVs that get to call for heat centrally, and a weather compensation only flow temperature curve, and it WORKSFORME!

https://www.earth.org.uk/heat-pump-16WW-control.html


Fascinating read.

Certainly feels like I'd need PhD from it to successfully install, modify, calibrate and run the installation until its fully adjusted, for the peak comfort and minimum cost/dirty energy use.


I am putting my efforts into making that not be true, at least for something "close enough" to peak comfort/efficiency/whatevs for any nominal shortfall not to matter in real life.

The obsessives amongst us can continue to tweak if we wish...


I enjoyed reading this, thank you.


\o/


I think that maintaining a stable temperature means warm walls/floors/furniture and potentially cooler air temperature, as opposed to a cold house with intermittently warm air. Most people can feel comfortable at a lower thermostat (air) temperature if the walls etc are warm due to maintaining a stable temperature. I don't have calculations or references, YMMV.


Heating systems generally are more efficient when they need to output less power. Whether that cancels the increased heat loss seems to be a question that can’t be answered in general.


>Otherwise, this doesn't make sense from a physics standpoint. A temporarily lower temperature differential means less kWh of heat lost.

This topic comes up anytime thermostats and heating are mentioned. The physics arguments only makes sense if you don't care about comfort. Most people would rather optimize for comfort with some energy/cost savings if possible and the physics folks seem to not care about comfort at all.


"Quickly" implies higher power which will make the air around radiators warmer than a slow heating.

The losses are proportional to the temperature differential between outside and inside.

So you should have somewhat higher losses from the hotter air streams from the radiators passing the windows.

Dunno about magnitude though.


> the theory being that warming up a cold house in the morning costs more energy than maintaining a stable temperature

I've heard this theory a lot too, but it doesn't match with physics. A warm house loses more energy than a cold house, due to a higher temperature difference allowing easier heat transfer. So in most homes, with radiators and high temperature CV, it's way more efficient to just turn it off when you gone.

One exception is when you have a very well insulated house, combined with floor heating and a very efficient, low temperature heat pump. In this case, it takes a lot of time for temperature to move in the house and it's already incredibly efficient.


It does match physics if you consider other factors. Apart from the heat pump scenario, this statement can also be true when you have condensing boilers (and okay-ish insulation)

The reasoning: when you heat up the house, then your boiler needs to produce constant high-temperature water. When you keep the house at the same temperature, then the boiler produces much lower temp water and it is more efficient.

Insulation also matters because if your house has outer insulation then it means that heat transfer from the house to the environment is mostly blocked, but cross-room heat transfer is likely not (through the walls). Therefore it is better to heat the whole house than heating just a couple of rooms because if you do the latter then you'll end up heating the whole house anyway but you're using less surface area (meaning you need higher flow temperatures, meaning less efficiency).


> The reasoning: when you heat up the house, then your boiler needs to produce constant high-temperature water. When you keep the house at the same temperature, then the boiler produces much lower temp water and it is more efficient.

How does your boiler produce heat for your water in your scenario?

> Therefore it is better to heat the whole house than heating just a couple of rooms because if you do the latter then you'll end up heating the whole house anyway but you're using less surface area (meaning you need higher flow temperatures, meaning less efficiency).

Just model the other rooms as very weird wall to the outside.


I think in real life there are more constraints. For example there are people that sleep better at a lower temperature than the daily one (so leaving the thermostat at the same temperature it's a minus for them).

Regarding "what is better" from energy efficiency, I would prefer a system that "check it" because my guess is that it depends a lot based on the individual situation. I mean everybody is going crazy over "IA" but a couple of sensors and a system smart enough to adjust your usage based on your particular situation and preferences (like "eco", etc.) is an exception.


In slightly cooler climates, the answer for sleep is to open windows. This works in much of Europe, even through summer.

But of course, not really feasible in Atlanta or Phoenix. Nighttime temps are too warm.


We are talking about places that need heating here.

In general, you should either run the heater or have your window open. Both at the same time is bad news for your energy bill.


Open windows in a city often means you are invinting in pollution, both air and noise.


Exactly. Current HVAC systems have extensive filtration; some HVAC systems have HEPA filters.


Opening windows also helps to evacuate inner pollution coming from furnitures, wall paint, parquet and other chemicals...


> the theory being that warming up a cold house in the morning costs more energy than maintaining a stable temperature.

That's true if you completely stop heating. However if you lower the temperature by roughly 3.5C when you're not home, you'll be saving energy.

So you can for example program it to be 16C when you're out and 19C when you're in. You don't completely turn off heating indeed.


(UK) my boiler has a control with something like the wake..leave timer (it actually has six settings for a midday period as well) and there is a separate thermostat with a temperature dial. The boiler also has a button that advances it to the next time interval if you want instant on (eg if you come home early to a cold house). I find this combo of controls meets all of my needs, given that I have a fairly repeatable daily schedule.


We've used the Tado system with a central boiler and smart radiator knobs for a few years. It's worked fine and hooks up to Home Assistant and can do the things you describe. I'm sure there are some alternatives.


I have an Ecobee and I like it. It comes standard with Home, Away, and Sleep but you can put in as many or as few as you want. You can manually change the temp and you can also set how long you want your manual temp to “stick”. Either until you cancel it, or until the next preprogrammed change. I like it cooler at night, so I have it change temp around my normal bedtime, which includes if I adjusted the daytime temp because I don’t want to have to remember to change it back myself, that’s why I have a smart thermostat. It detects when I leave the house and sets it to away, because I don’t need it running as if I am at home. If I go on vacation I can set it to keep the house safe, but not comfortable, and change back around the time my flight lands.

Unless you have a crazy random schedule, or you want the temp the same whether you are asleep, awake, or not at home, or i guess if you have different temp preferences every day. Otherwise you can program in a basic schedule and just adjust manually as needed. Nothing stops you from changing the temp manually if you wake up an hour early, but if you wake up on time, then you don’t even have to think about it.


>I guess I’d be happier with a more programmable thermostat that I could set to behave like an old school dial thermostat.

I honestly prefer the older type. Ours is programable, but we just don't program it and always just set it to the temp we want. If we are feeling a little chilly on a cold day, we'll bump it up a degree, or down a degree when it's particularly sunny and everyone is feeling warm.


Isn't this pretty much what these thermostats already allow? I have a new Honeywell Thermostat which basically does what the twenty years old one it replaced does with a few added conveniences in terms of UI. It has those wake/leave/return/sleep instants for each weekday (but also adds an optional second leave/return pair), and it has an option to override the day programme to 'holiday', which is essentially an eighth programmable weekday you can activate at any time.

Your use case is possible with that. Just set the standard program to 15°C, and activate the holiday set to whatever you need whenever you want. Configure it to go to 15°C at some sensible time in the evening, so it won't go on even if you forget it.


15 C is very cold. Are you American?


15°C would be the standard setting of 'no heating' (just keep the house warm enough to avoid mould and such). The 'holiday' programme would be 21°C or whatever is desired. That would effectively turn the thermostat into one where turn on the heating by putting it in 'holiday' mode, and can turn it off by exiting that mode or just letting it revert to 15°C after 23:00 or so.


The obvious solution is a "wake time" of 8 am, "leave time" of 8:01, "return" 8:02, "sleep" 8:03. Then just set the sleep temperature to your desired temp and the remaining ones to something reasonably close, or if it doesn't automatically switch between heating and air conditioning, set it for the no-op for the season (i.e. the highest possible temp in summer, and the lowest possible in winter) for those three minutes.


> ideal one-size fits all thermostat

The round Honeywell electromechanical thermostat with a bimetallic strip, invented in 1953: https://www.honeywellstore.com/store/products/honeywell-roun...

24VAC, dead simple, and reliable. My family’s lake house has 50+ year old Honeywell round thermostats still in service.


> I wonder what the ideal one-size fits all thermostat looks like

https://www.honeywellhome.com/us/en/products/air/thermostats...


> I wonder what the ideal one-size fits all thermostat looks like.

As you go on to describe, there probably isn't one.


My wife and I worked a six-week shift work schedule for a long time. We got second-gen nest thermostats when they first came out (2012) thinking they were neat.

Nope! The smart learning feature was the biggest pain in the ass. You’d be sleeping during the day for a night shift, only to find yourself freezing because it decided no one was home.


I like the Nest, but I absolutely turned off all the "smart" features and just set a fixed temperature range and change it when desired.


You can turn off the feature that changes temp based on presence. Or disable dynamic scheduling altogether.


I ended up having to manually changing the schedule week-to-week.


Mine works OK as long as presence detection (using people's phones and Google Home) is operating correctly.

Everyone leaves, and the thermostat adjusts.

Someone comes home (or walks in front of it), and it goes back to the normal setpoint.

Sounds easy. Isn't always easy.

The reliability of this seems to be highly dependent on the phone(s) themselves simply succeeding at not killing useful processes.

Overall, I'm not entirely displeased with it. I procured it very inexpensively by buying it from one of my energy provider's online store in conjunction with a substantial rebate from my other energy provider. I'm confident that it paid for itself very quickly, and it's nice to be able to set the thermostat remotely.

My lament is that there seems to be approximately nothing I can do to improve the presence detection function without gymnastics or spending real money. What I want is a local API that I can enable and do stuff with; what I get is "Good luck! Have you tried buying Nest Protect subscription motion detectors? (Oh lol, we stopped selling those.)"

(I'm OK with the privacy and security aspects of what I'm trying to do. I'm not OK with having a connected device that I can't bend to my will. I'm even less OK with more recurring expenses. The next thermostat I buy will have local control over the LAN, but it probably won't pay for itself quite as quickly as this one did.)

[note: I've never played with the auto-schedule "learning" function at all. It always seemed like a complete waste of time, since I for one do not have a regular schedule.]




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: