Hacker News Reader

Comment by daredoes | Hacker News Reader

Would your company be willing to provide an open-source local-first websocket-based integration with https://www.home-assistant.io/ or use the new Matter + Thread standard to provide it for everyone?

If that's available, this would be my top choice as I'm actively shopping for all-electric heat pumps

View on HN

It's something we'd definitely consider. Is the desire to have thermostat control via any third party?

edit - The current plan has been Matter/thread Homekit + Google Home integration. We can expose at least temperature control via this to everyone.

(not OP, but I also use HA)

The important bit is to have an open API that can be connected to from the local network (or via Matter + Thread, etc), and via a cloud. Then anyone can develop a client. This also has benefits in UX: the latency to update is much lower than cloud based polling, since that's almost always rate limited.

The closer anything is to the bottom of the chart here, the better. When buying something as expensive as a heat pump, I wouldn't even consider something that isn't at least local polling:

https://www.home-assistant.io/blog/2016/02/12/classifying-th...

That being said, I'm not opposed to value add things via cloud services. Want to get the weather via a cloud service, even if that changes the schedule? Totally fine! But ultimate control must be local.

Agreed, this isn't a $100 thermostat that I can just pull off the wall and replace - we're talking an expensive investment with a 10+ year lifespan. That said, even for my thermostat not being bound to the cloud was an important reason I went with Emerson/Sensi (the only value-add I lose if their cloud goes down is usage reporting, everything else I can do with the native Homekit support).

Consider it is as a competitive advantage.

An API that is public, well-documented, and easy for other systems to securely integrate with would mean that customers get the integration they want. If someone builds a better way to utilize your system, then you just got value-add on your product at basically zero cost.

Amazon can hook Alexa into it. Google can hook their Home thing into it. And whatever comes next, whatever doesn't even exist today, should have an easy time integrating, continuing to add value.

Completely agree on this point. I've worked over half a decade with smart home systems, and it's just so much easier (and cheaper!) to integrate well documented, open APIs (and please make them secure, I have nightmares of all the open, unencrypted UDP based alarm systems) that is purely local. This means we would highly recommend these products even tho we made zero money off them, and customers would actually prefer them as well - it's well beyond the time of people naively buying this stuff, and they know that clouds can just "go offline".

What also can't be underestimated is the tinker community. Many people in the smart home business are tinkerers themselfs; create a great product for them and they will inevitably try to use it at their job. But this really isn't only applicable to integrators, really.

If you support home assistant then I am far, far more likely to purchase this. My wife and I are buying a house sometime over the next year and one of my higher priorities is getting a heat pump. At the same time though I also refuse to use cloud based services and instead prefer to run my own home assistant installation.

You should consider this a huge marketing opportunity- the home assistant people are obsessed with perfecting HVAC (half of these people have their own weather stations) and are also the kind of geeky first adopter who will push useful tech on all their friends. It's also not a small community, as the subreddit alone has over 222k people on it.

https://old.reddit.com/r/homeassistant/

I just preordered and I will echo this.

On track to replace my gas furnace in the next few years, this sounds awesome, but HA integration would be very highly desired.

Chris, I think there's a lot more here to uncover. Most of your competition has no sensitivity, and almost no openness, about these aspects. I think this could become a really important part of your go-to market, but also a differentiator vs other products.

I'm a VC since recently, but have been a tech guy for a long time. Feel free to ping me if you'd like to discuss it further. $HN_username at gmail.

HA user here as well, having a non-locked down API at minimum (i.e. not actively trying to stop folks using it), through to an open and documented api (would be great), through to open firmware (would be best) and or releasing your own integration for HA would likely get you those kinda customers.

If people have a preferred way of integrating their devices, like home assistant, I'd want to make that possible for people to do with Electric Air.

Re the sizing question - this is a part of the site that needs improvement. 12kbtu is 1 ton. The 48kbtu system is a 4ton system, but rated at 5F, if you're in a more temperate climate, two of these will likely have plenty of capacity for you. It's also likely that your current system is oversized.

As someone about to be in the market for something like this, propriety lockout is a complete deal breaker for me.

/2¢

The good news, at least, is that almost anything will work. You don't need full Matter support for Home Assistant folks to be happy. The bare-minimum, local REST api built on a Monday by an intern will do it. Someone will write the Python code to fully integrate it. I get that there would be some security concerns there, but you get my drift.

If worried about security concerns, just provide a common port interface (RJ-45/RJ-11) on the unit with a simple, well-documented, serial-based protocol on RS485/RS232. You can sell an add-on device to bridge that to Ethernet/simple API, and easily replace/recall that inexpensive module if there are issues.

The hackers will do what they want to anyway, and they'll implement cool interfaces/HA plugins/controllers on RPis or ESP32s for free. Someone will get enterprising and package a pre-built/programmed ESP32 unit that makes it plug-and-play for nerdy, but less capable users, and your company can avoid any liability from users using an 'unsupported' add-on.

Also, I'm in the market for a new AC/heater - so have been looking.

Your specs are in kbtu, but at least for cooling I'm used to tons; I have 2x4.5ton units, and I think that's ~120kbtu, so I wouldn't be able to get just two of your biggest units?

That would be 108kbtu, but you are likely oversized. With proper duct configuration including ACCA Manual D calcs and air sealing, you will most likely be more than fine with 96kbtu. Most systems (even in very nice houses) are just sort of thrown in rather than actually calculated, and then they are installed sloppily with a lot of leakage and only roughly resembling the design in at least a few locations. I do these calculations for a living, and would recommend presuming that will be the case until proven otherwise. Obviously get hard numbers before full commitment.

My company took GP’s suggested strategy and got an incredible community response. Really resonates with the high-end residential market, both DIY and DIFM (pros / CEDIA).

I think the desire would be a locally accessible API so that anyone could build an integration to it.

My desire would be to have the possibility of thermostat control (entirely) via first party, me. Odds are I would never even touch it, but the fact that I couldn't would be a dealbreaker for something like this.

For me the goal is 100% local control. I found out the other day that I couldn't turn my 8sleep mattress on because a guy down the street struck a fiber line. My bed was basically a $4k brick as soon as the internet went down.

Look at it this way, it saves you money in server costs and performs much better.

Hey, that happened to me too this weekend! Do you also live in Ingleside in SF?

WOW.

For me, it's just peace of mind. I need to know that if cloud servers go down, or even my internet goes down, my house still works. My Nest thermostats are currently cloud-only (the only parts of my house left that are), and I had to write a script to never update them exactly on the hour. That's when their API is very likely to 500; I'm guessing because most people have automations trigger on the hour.

I guess I can sum it up like this: I only want to deal with my own problems in my free time in my own home.

The point is to not have your company as a single point of failure for the thermostat. So yes, integrate into a wider (open) ecosystem, and that problem is addressed.

Yes, and particularly in a local/LAN/no internet manner. Local push or local pull.

I would like to be able to add it to HomeKit without it ever phoning home once. Matter + Thread help make that a possibility.

I consider Matter support to be vital for any product that’s “connected”. You can keep all the specialised features limited to your app but basic controls should be exposed via a standard protocol like Matter so I can use any Matter controller to set the temperatures and such.

I think going with Matter would give you the best bang for your buck as you would get virtually every major smart home hub system.

You would still be able to do more with your own app as matter will probably mainly support the lowest common denominator across your category.

The desire is to have any control/knowledge you'd want to have as the creator/developer. I want to be my own mechanic, and if I'm not that mechanic I want to be able to have anyone be that mechanic.

Do you want to know what the active airflow rate is? I want to know what the active airflow rate is. Do you want to know power consumption? I want to know power consumption. Do you have a way to write unique schedules/programs that get executed on the heat pump? I want a way to write unique schedules/programs that get executed on the heat pump.

Context: Full-Stack Software Engineer (worked at a few start ups including Bird Rides). Active Home Assistant user and community-run integration creator (Linked Lovelace)

The incredible part about a company like yours is the ability to do hardware at scale. There's no reality in which I'm safely and cost-effectively building my own Electric Heat Pump, or television, or [cool product here].

Companies that make it easier for me to do my own things stand out.

If you don't want that for the standard customer, fine. Provide some way to open up the device and trigger dev mode, or manually upload firmware, or OTA update firmware.

Side notes:

- a failed kickstarter I joined shipped out their original product with bad firmware and no way to do a physical firmware update without destroying the product, so they tanked instantly despite having a great hardware setup. Despite their failure, I and a few others opened up our products and manually flashed custom firmware onto it to make use of the product we bought

- I exclusively buy LG Televisions now due to their usage of Web OS (has a local-first [HA Integration](https://www.home-assistant.io/integrations/webostv/)) which used to be hackable with https://rootmy.tv. With that, my TV runs custom Linux software that controls lights just like [Phillips Hue Sync](https://www.philips-hue.com/en-us/explore-hue/propositions/e...) but free/open-source and no additional hardware. Compare that to Samsung which is like 40% ads now?

- I don't mind using a voice assistant, steal my ideas Tim Apple. I hate the part where the request goes (phone -> router -> cloud server -> router -> local server -> light bulb) instead of (phone -> router -> local server -> light bulb) or worse (phone -> router -> local server -> router -> cloud server -> router -> local server -> light bulb).

EDIT: my markdown is showing, I should stick to lurking

Speaking of heat pumps and granular control, I recently figured out that some Midea heat pumps (usually sold rebranded to various brands) contain two interfaces to get and set detailed information:

1) Detailed diagnostics i.e. RPMs, feeds, speeds, currents, temps etc can be read from the high voltage signal line between the indoor and outdoor unit - there is an official tool (search "Dr. Smart Midea" on YouTube) that does this. If you look in the right places, there is a PowerPoint with a schematic of the PHY that you need to interface with this using a LV isolated controller

2) The outdoor unit can be driven manually using the same tool using an i2c port that is on the PCB on the outdoor unit

I have the new generation of their Dr. Smart tool, which I haven't had the time to thoroughly test or document. If there is interest I will likely try to reverse engineer and document these protocols properly.

> "I want to be my own mechanic, and if I'm not that mechanic I want to be able to have anyone be that mechanic."

> "Do you want to know what the active airflow rate is? I want to know what the active airflow rate is. Do you want to know power consumption? I want to know power consumption. Do you have a way to write unique schedules/programs that get executed on the heat pump? I want a way to write unique schedules/programs that get executed on the heat pump."

There's one feature of this diagnostic interface that I really really like. It allows you to "drive" the hardware of the unit manually, like Program Auto mode on a camera. You can manually set the frequency of the compressor and fan, and the control board will continue to enforce safeties.

I haven't tried yet, but something I really want to try to implement myself is "frequency lockout". This is a common feature on commercial variable frequency drives. I noticed that under some operating conditions, the unit vibrations resonate with the building and create unpleasant sounds - I'd like to program it to never dwell on that frequency and skip over it.

> The desire is to have any control/knowledge you'd want to have as the creator/developer.

While I was exploring this, I had this in mind constantly. I definitely had this diag interface in mind when I was selecting a unit.

Can you provide any more details on the I2C interface on Midea units?

I found some sparse documentation on Midea's (and rebadged units) Modbus-esque protocol over RS485 (aka XYE connectors), and I control it via an ESP32 on HA. From the indoor unit, the diagnostic data I could reverse engineer is limited to temp sensors (intake, indoor coil, outdoor ambient) and basic running modes. I believe the outdoor unit also has a modbus/485 that has more info, and obviously the signal wire passes some comms between the two, but Midea doesn't make this stuff public.

If I'm gonna invest in a heat pump for the long term, I don't want it committed to any software technology that may be obsolete in 10 years. I want it to use something so dead stupid simple that anything in 50 years can still operate it without a lot of work.

The simplest example I can think of is MIME (https://en.wikipedia.org/wiki/MIME). It can be embedded in basically any transport type. It can encode anything you want (binary, UTF8, etc). It is very old, stable and well-tested. It is fairly trivial to implement a limited instruction set of.

Transport security, authentication, authorization, etc, are nice, but not nearly as important to me as a future-proof interface to my heat pump. Give me the option to enable an admin port and just run a single unsecured connection with a basic dumb text protocol, and I can hack together a client in 30 minutes in any programming language. It's not fancy but it will work forever.

I agree on the point of not wanting to end up with a brick due to obsolete software. Note though that heat pumps typically last 10-20 years.

This is extremely disappointing, given their price point.

I've seen an oil furnace still running after 50 years with just basic maintenance. Did it cost 2-5 times as much in inflation-adjusted dollars as a heat pump does now? No.

https://www.evergreenenergy.co.uk/heat-pump-guides/how-long-...

I don't see why a heat pump wouldn't run for a long time. I imagine many people want to replace old heat pumps because newer ones are more efficient.

But otherwise this is quite old and widely available tech, it should be possible to replace individual components that break down.

What happens to them after 20 years?

IMO I would prefer a Bluetooth thermostat that once paired, you don't have to login again. One thing that really annoys me about home "smart" devices is arbitrarily being logged out and having to remember passwords etc. and I don't see a reason I need to be "logged in" to control my thermostat. Also I don't want it to stop working when your servers are down... (had this happen the other day with my garage door opener sigh...) If you want internet connectivity, it should connect to some hub to bridge the Bluetooth to the web account.

Or the company could at least leave it easy to hack :)

Implement with a swappable compute module?

We need an entire industry of 'appliance based Rpi's to be able to act as modules in a standardly compute plug-in-form factor (just like blades) with a unit preloaded with a config and just swap out inter connects - ich that we have an appliance hub, the solar/whatever input pushes power and the distributor determines where to push or halt the power - with a selector that tells me the input costs, and output draws, such that it can have a system to auto switch to best performance.

Honywell missed the mark on the importance of thermostats on the economy.

My dream is something equivalent to a Dumb TV + Chromecast.

Seconded.

Reply on news.ycombinator.com