Using a webcam to watch your bird feeders (2nd generation)

by May 1, 2020

A while back I did some experimenting and built a camera for my birdfeeder using a Ring camera, and wrote a piece about it. At the time, I thought it was a good solution. Over time, however, in part because of Ring’s poor privacy policies and a marketing strategy designed around making us really scared and paranoid about the world around us, I decided to retire the Ring gear from my house and life, and I took that piece down.

I’ve spent the last six weeks designing (and debugging) a second generation bird feeder camera setup, which I’m now using on a daily basis to monitor what’s going on out there. It’s not perfect, but I’m happy enough with it for now to describe what I’m doing for others that might want to try it.

After research, I ended up purchasing the following gear:

Arlo Pro 3 set with two cameras and a hub ($449)
Arlo Solar Panel ($70)
Arlo Pro 3 Rechargeable Battery ($50)
Arlo Pro 3 Battery Charger ($50)

I bought two cameras to also replace the ring doorbell camera on the front door. If you buy a single camera and hub (required to operate) the cost will be around $250, so you can set up a basic system for about $350 if you forego the solar panel. Be aware that the camera will drain the battery pretty fast, and you might need to change it out daily. The solar panel doesn’t fully keep it charged for me, but stretches the time between battery swaps to a few days depending on how sunny or cloudy the sky is. I’ll talk more about that in a bit.

I am pretty impressed with the build quality of the camera unit. I have some worry about how the battery connectors will hold up to long-term swapping of batteries a few times a week, but other than that, there’s some good design here. The camera is considered water resistant, not waterproof, and you can buy sleeves or mounting units with roofs if you want, but I’ve put mine out in the weather and I feel it’s unlikely it’ll be impacted. Good overall sealing of the seams. It’s easy to open the unit up to replace the battery. The charging connector for the solar panel or an electrical circuit is sealed and the connector held on by a good, strong battery. I do think the battery (4800mAh) is too small, but that’s my main issue with it. I’ll talk more about this and the challenges it causes later. Overall, I give the hardware a solid A-.

The software, unfortunately, doesn’t hold up its end as well as I’d like. I’ve found it to be somewhat buggy, including random app crashes on my iPad caused simply by viewing 7-8 videos in a row. There are other weaknesses to the software, but I’ll go into detail in the section below on weaknesses. If the hardware gets an A- here, the software gets a C-. It’s functional but it makes using the hardware harder and the overall experience poorer. The good news is software can be fixed, if they decide to. I will note that there was one update to the App after I installed the system — and the overall quality of the app went down and it crashed more, not less. As I write this, a second update to the app has just updated onto my iPad, and we’ll see whether it improves things or not.

My overall grade for this installation? B-. It’s good, not great, because of the software issues I’m seeing. There are definitely some weak spots to what I’ve built that I need to discuss. That said, I’m happy with it overall, and I’m using it on a daily basis, and I don’t feel any strong need to try to fix or replace this or throw it to the curb.

The strengths

What do I like about this setup? I get a good, clean image with good detail under most lighting conditions. It’s better than I expected. It generates videos that I can copy to the Photos app with a quick tap, and from there I can export or edit the video for use or grab screen images. I did find that when pointing into the sun it struggled with the backlighting, so when you look at the image of my setup, the camera is to the south of the feeders and pointing north, so it always has front light on the birds. Doing that made a big difference in image quality.

This camera has an extremely wide (110 degree) field of view, expected for a security camera. That implies, however that to get enough detail on the birds to be useful for identification, it has to be really close to the feeders. As you can see from the picture of my installation, it’s attached to a garden stand about six feet from my feeders. Being out in the middle of the yard implies no easy access to power, so it has to be battery powered, which was a primary reason I chose the Arlo Ultra 3 for this. There are cameras that include 4X or 10X zooms and pan/tilt capability that might allow you to mount them on the wall of your house, but I decided to stick with a less expensive unit like the Arlo.

You can see from this image — mid morning, mostly clear — there’s good detail and color, and it did a good job of freezing the incoming Bewick’s Wren in flight. I can’t ask for much more than that for a setup like this.

Ultimately, this gives me what I want: ever since I moved my home office away from where I could watch the feeders from my desk, I’ve wanted a way to do that. Now, using an app, I can fire it up and see what’s happening at any time, even if I’m not home.

This setup is built for you to use as your personal camera looking at your feeder. It’s not capable of running for extended time or streaming a feed for sharing on the Internet. To do that would take a more complicated and expensive installation, and if that’s what you want, this isn’t the system to use.

The weaknesses

This installation could be great, but the software holds it back. As I mentioned above, the app has issues and basic actions can cause it to crash. That indicates to me weak QA and an immature product, but those things can be fixed, and I hope they will.

There are some more fundamental product feature gaps and other problems as well. Here are some that have caught my eye.

One key feature the Ring system had that this one doesn’t: if I am monitoring the camera, there is no way to tell it to fire off a recording. It will only record if the motion detectors activate the camera. That means if I see something interesting going on, I can’t tell the camera to capture it, I have to hope the camera decides to. That wouldn’t bother me as much, except…

The motion detection system isn’t great at detecting birds. It seems to trigger most often when the feeder is swaying from the birds (or squirrels, or wind). So reliably triggering on bird activity is hit or miss. I will note, though, that this is not a typical use case so this isn’t specifically a complaint about how it operates — except that since I can’t fire it off manually, I can’t override the motion logic. It would be nice if the motion actions were more reliable (but I do get what I’ll call a statistical sample of bird activity over a day, and that’s “good enough” for me here), but I need a way to fire up the camera on request, and it just isn’t in the app. I really can’t think of a reason why not, either, that feels to me like basic functionality that’s missing.

The video rendering in the app is much lower quality than in Photos or iMovie. I don’t understand why, but it simply doesn’t show as much detail or as good color. I end up doing a quick skim of videos and pushing the interesting ones to Photos to look at instead. Good quality rendering of your own video seems like it ought to be base functionality, but it’s not here. I’d also like to be able to configure the system to automatically dump everything off the camera to Photos, and that’s not possible.

Battery life is poor. the battery wedged into this camera simply isn’t big enough and there are few options to fix that. The solar panel helps — if your camera is located where it can get good light — but as often as my camera is triggered by the feeder, the panel won’t keep the battery charged. If this camera were being depended on as a security camera, I’d throw it out (and for the one on my front door, it’s now wired to electricity).

Arlo made a decision on battery charging that both makes sense and creates a big problem. The battery charging system uses the newest QuickCharge 3.0 charging from Qualcomm. That allows a charger to throw more power at the battery and charge it quickly, which is nice. The downside is the specs for this require that charging be started via a physical switch: you have to push a button. That means you can’t put a large (say, 20,000mAh) battery on the charging port to keep the main battery charged, because it’ll charge up the battery and stop, and charging won’t begin again until you push the charge button. So I can’t work around the small internal battery by hooking it up to a really big external battery (I tried a number of different options here, and none worked). If the Arlo battery were 10,000 or 12,000mAh, none of this would be a problem. Because it’s 4800mAh, I’m swapping batteries a couple of times a week, or every other day (at best) if I don’t have the solar panel attached.

What they seemed to have done here is take their wired camera — the Arlo Ultra — and rebuilt the insides to include whatever battery fit inside the case instead. And unfortunately, that battery is not big enough for more than simple/trivial uses. I certainly wouldn’t depend on a battery camera like this for my main security camera, and I would regret mounting it where I needed a ladder to swap batteries. The unit feels a bit like a hack and not a carefully designed product — at least as far as power systems are concerned.

Another aspect of this shows up in reliability of the connection from camera to hub. The hub is required to operate and connects to your internet (via a wired ethernet, no wifi capability). I have found that the battery powered camera can disconnect from the hub and become unavailable. I saw that on both cameras until I wired the front door to electricity — and it has been 100% reliable since. That implies there’s something about the camera running on battery — power fluctuations perhaps? — that is behind these disconnects. Running with and without the solar panel doesn’t change it, not does swapping between two batteries.

I often see this crop up immediately after a battery swap. Arlo’s support pages suggest going to the camera, opening it up and pushing a “sync” button on it and then going and hitting the sync button on the hub to get them to kiss and make up and start talking again. From a usability and reliability standpoint, I don’t think I should have to initiate this re-sync; both sides can recognize the connection problems and should initiate it on their side until the other side does as well. Again, if the camera is mounted where I need a ladder to do this, it wouldn’t take long for me to give up and toss it out. Here, it’s a minor annoyance.

It seems to me the stability and reliability of the connection is paramount for a camera designed as a security camera. My experience shows they have that for their wired cameras, but for the battery/wireless ones, it’s a work in progress and an indication this isn’t a mature system. I’m okay with my main camera at the front door’s reliability. If it acted as flakey as my feeder camera, I’d be replacing the entire system already.

For my use here on the feeder, it’s all minor annoyances, and I’m going to give Arlo time to sort all this out. That said, if there’s no progress that I can point to by fall, I’ll likely start looking at generation 3 of this setup.

Executive Summary

That seems like a lot of criticizing for something I gave a B- grade to, but it’s because the issues are very fixable and there’s a lot to like in and around these weaknesses. If you are someone thinking of installing the Arlo as a security camera and happened on this, my suggestion is that as long as you can wire it up to electricity (and you can get 30′ charging cables for it, giving you a lot of options) it’s great. Maybe consider paying more for the Arlo ultra to go from 2K video to 4K, in fact. If you want to install your security camera via battery, be aware of the reliability issues and small battery. I probably wouldn’t.

Installing it as a feeder camera? I think it’s fine, as long as you understand it’ll need some babysitting. My goal here was to come up with a turnkey, easy to install and use system that doesn’t require a nerd to get it going or operate it — that’s why I went with the Arlo instead of something like an IP or streaming camera that would require configuration, a local server or figuring out technical details.

But it’s currently a system that requires babysitting, so it’s not quite as turnkey as I’d like, or as I hope it’ll be in a few months. Here’s hoping Arlo puts some focus on fixing the software to be as good as the hardware they built for it.

Despite those issues, I recommend it, and I love the quality of the image I get out of it. I just want and expect more from this setup than I currently get because of the software bugs. And those are minor if annoying, and very fixable.

Overall, I think my new feeder setup is a good and usable one that I expect will become great as the software matures. If you’re interested in this kind of setup, the Arlo three system will do the job for you pretty well.