Smart garage door opener image

Smart Living Things


Recently, I had to re-educate my garage door opener. Our house on a hill in San Francisco occasionally settles, and the door is experiencing different stresses than it did 5 years ago when I installed a new opener. So I had to re-teach it how to open and close the door.

Let’s think about that. I use “re-educate” because it sounds a little “Brave New World”-like. Just as does the fact that my garage door opener is actually a “smart” machine. Our house is about 30 years old, and the prior opener lasted 25 of those years. It had a set horsepower rating and two physical limit switches — one for the opening limit and one for the closing limit. About as mechanical a solution to the problem as you could get while still using electricity. It had a chain drive, and just pushed the door until it closed or pulled until it opened.

Before I bought the new opener, I had watched a couple Youtube videos to make sure I understood what I was in for, and installation ended up being a breeze. When I set it up, I noticed there were no limit switches, but it was a hot day, and the instructions were printed on a big poster that you unfolded like Lego instructions (doing a lot of those with my kid these days), so I just quickly zipped through and got the new opener working. There were a couple additional parts for checking for obstacles, and retracting the door if it sensed something when it was closing, like a bike or a kid’s leg, but the installation of those parts was part of the big poster, and worked just fine when I was done.

So, here we are five years later and the door is having a hard time with closing and opening. It stops midway like something is stuck, or one of the sensors has been tripped. After several months of spotty behavior and a couple adjustments to squeaky door wheels and tracks, I decided to Google the problem. And I started seeing things like — “if you change your door springs, here’s what you need to do to update your opener,” or “how to make your garage door opener have more power.” Now, again, going back to garage door opener #1: it had a set horsepower, and did its thing until it finally died after 25 years. With opener #2: it’s nice and quiet, doesn’t mangle bike wheels or people’s legs, and stops working correctly after 5 years.

Turns out, there’s a computer board in there that helps it to understand its environment. It doesn’t have limit switches or adjustment screws for how hard it needs to push or pull because when you run the setup, it senses those things and builds that information into how it does its job. Five years ago I was building smart products, but I didn’t really think of the garage door opener itself as smart. I was excited that we could get an app that allows me to open and close the door with my phone, but that’s just taking app presses and turning those into garage door commands. It’s “smart” in the way we think of our smartphones doing just about anything we want — “there’s an app for that” — but that could have worked with the old garage door opener as well.

The smart I’m talking about is that between garage door openers 1 and 2, the manufacturers started building PCBs and firmware/software that helped the door to understand what was actually happening around it. My guess is that this was primarily due to the need for the functionality that allows it to sense a child’s leg or bike when it’s closing, but it has much larger implications. It’s only so smart in that when I set it up, there was a step where I taught it what an obstacle under the door felt like, and so it built it’s sensitivity around how that “felt.” As the house settled and the door experienced stresses that caused it to bog down in certain spots, the door used its sensitivity settings to determine that it had hit an obstacle, and so did what it was supposed to do — stop closing. It also did this on opening sometimes, which must have been very confusing for it.

In any case, this brings me to my point: technologies like the relatively simple PCB and software in my garage door opener are the first generation of “robots-all-around-us.” But they are only so smart, and sometimes their intelligence needs to be "evolved." In my garage door opener’s case, we’ve had a couple close calls with the door being open for many hours because it didn’t close after we left (we have a sensor, but sometimes it doesn’t tell us!), but nothing stolen thankfully. And I spent several hours trying to figure out what was going on, and finally fixing it. And, while I have an app that talks to the opener, to fix it I had to download and print a few pages from that original big poster, rerun the setup routines by getting out a ladder and pressing various buttons in a very specific order on the opener, and re-educating it as to how hard it needs to push or pull to get the door open or closed. And, fortunately, I knew what I was doing because I installed it originally, and because I do this kind of stuff for a living, rather than having to pay a couple hundred dollars to have a technician come out.

But these robots-all-around-us really need to be thought about carefully when we release them into the wild. I’m not blaming my garage door opener vendor (which is why I’m not naming the brand), but, as we build these products, we really need to be thinking about their entire life cycle. How hard would it have been for the opener to notify me when it started to notice that it was not completing its open or close cycle regularly after 4 years of doing just fine. How hard would it have been to have it run a diagnostic on itself — either self-initiated with my ok, or via my app to ask me to help it do so. How hard it would have been for it to do those various button presses on its own with just one command. You get the idea.

These are the things we all need to think about when we design smart products. We’re no longer just designing apps that exist on a phone or PC. And we’re no longer just designing pieces of hardware that do their thing til they break. We’re combining software and hardware and building intelligence into things that used to be reliably dumb. And so we must think far more carefully about how these things will “live” their lives in the world, and what we, as their designers, should equip them with in order for them to live their best lives. It’s not at all philosophical — poor design decisions have real implications, with the 737Max being one very unfortunate recent example.

So, if you want engineers who think deeply, give us a call and let us help you design your smart products. . . intelligently. Contact us through our website or email me: rfreedman @

+ more

Accurate Timing

Accurate Timing

In many tasks we need to do something at given intervals of time. The most obvious ways may not give you the best results. Time? Meh. The most basic tasks that don't have what you might call CPU-scale time requirements can be handled with the usual language and...

read more