Pros and Cons: S3A7 Versus Raspberry Pi

raspberrypi
s3a7

#1

I’ve been thinking about the S3A7 and how a lot of its features that are geared toward more serious projects than a Raspberry Pi. I’m just getting started putting this “down on paper.”

Sure, they’re probably aimed at very different users. Raspberry Pi is hobbyists and learners. S3A7 requires higher skills, probably, but is more polished for moving something from prototype to pilot.

In some ways it’s apples and oranges. But I think I’m probably not the only person using both. I have both a Raspberry Pi 2 and Raspberry Pi Zero W in my office. (And a Raspberry Pi 3 with a big 7" LCD screen is in the mail as we speak!) I love them, both for what they allow me to do and learn, and for how much they excite my imagination.

Raspberry Pi Zero W (potato for scale)

I’m currently building a smart mirror for my bathroom, that displays updated information in the morning. (I’m not exactly breaking new ground here, there’s a whole subreddit devoted to the topic.)

Advantages

Quick Boot Time

But I’ve noticed a growing list of features that are very polished with the S3A7. The first one is how the S3A7 boots up almost instantly. There’s no Linux operating system running a file system that, on the Raspberry Pi, is fine for killing power intermittently, but not regularly. If you continue doing that you will run into file system corruption at some point. The Raspberry Pi is just not optimized for low power usage. I think most people have it plugged into the wall or a computer.

More…

I’m going to add some more advantages - hey, and maybe disadvantages - here soon. If you’re familiar with Raspberry Pi, I’d love any comments. I’m learning more day by day and very excited to expand my understanding.


#2

The biggest difference is probably the toolchain and libraries. With Synergy, you get a tested and certified toolchain. With Raspberry Pi, you’re largely on your own. As you know, things often and regularly break on the Raspberry Pi when it’s used as a standalone controller. For example, in the project for the VR camera that is being run by the Harvard professor we know, the libptp code is several years old and not maintained. It’s fairly flaky on newer versions of Raspbian. So, while it’s fine for a one-off project, it’s not suitable in it’s current state for production use. It may be fixable with more serious effort on libptp and the kernel.

Another example you know about is the video problem with lack of UVC (USB Video Device Class) 1.5 in Linux. It’s just not there in the kernel. Sadly, if you’re using a third-party video device with the Raspberry Pi, your mileage may vary.

The other issue is that you’re most likely going to use the Raspberry Pi Compute Module for embedded applications and that requires a $120 Compute Module IO Board. The RPi compute module is $30. You then need a screen of some type to enable human interaction for your project.

While Linux is definitely cool and fun to work with, the testing and development on libraries and potentially the kernel could get expensive quickly once the product nears production.


#3

Any other thoughts on using the Raspberry Pi for IoT projects? I guess the key thing is to show your IoT project and explain how close to a finished, usable product you managed to get. To be a fair comparison, let’s define the IoT project as follows:

  • there’s a touchscreen GUI interface for a human to interact with data
  • device is powered by battery
  • device can be turned on and off and reliably boot after loss of power
  • device is connected to Internet or network

I’ve built many Raspberry Pi projects that conform to those specs. It’s usually fun, tough, glitchy and a one-off. For Renesas Synergy, it’s fun, tough, reliable, with a clear path to production.

For the Raspberry Pi, it’s easy to get something going, but tough to get something working reliably, especially if you put it in the field and not in your house.

In the example I built below, the RPi is powered by a rechargeable battery using dual WiFi to connect to an external virtual reality camera using WiFi and the cloud using the public Internet.

This is the front of the project with the official Raspberry Pi touchscreen. It’s designed to take media for immersive experiences inside of VR headsets.

As you recall, I also built a series of projects to allow people to control the video and image stream with physical push buttons and use LEDs as indicators.

I also added IR control onto the Raspberry Pi to control the video and images.

My assessment is that the Raspberry Pi is wonderful for weekend projects, one-off demos, and is ideal for children. It’s very fast to get something that works, but it’s super difficult to get something to work reliably at scale.

If you recall, we were never able to show the Raspberry Pi demos to our client because I was always worried that the demo would break in the middle. We had to resort to taking a video of the demo because the demo was so fragile. The power button and clean system shutdown is a big problem. If you just pull the plug on the Raspberry Pi in the middle of a demo, it may reboot. Or, it may not.

Even as non-production use as a demo platform, Synergy is much more robust. I have confidence that it will turn on and work. For actual production, it seems MUCH easier to use Synergy and also much cheaper.

The primary problem with Synergy is that the learning curve is steep. For example, as I’m just starting with Synergy, I don’t know how to get the IR control and HTTP API access working over WiFi. Once I figure it out, I’ll have a much better project than I could build with the Raspberry Pi.


#4

I’ve been working on small projects with both the Raspberry Pi 3 and the S3A7. They are both wonderful, although very different. The S3A7 kit is clearly geared toward rapid product prototyping. The Pi is more of an educational tool. Your assessment above is spot on. You wouldn’t want to put a fully functional Linux board into a smart thermostat, for example. There’s considerable overhead that isn’t needed.

I have been using a “Sense Hat” https://www.raspberrypi.org/products/sense-hat/ on top of a Raspberry Pi 3 board. The sense hat adds an LED matrix and joystick (see photo) for human interaction as well as a suite of environmental sensors and an easy to use API. I’m thinking of using this as a platform for teaching junior high programming.

Nothing is cloud connected, and you have to figure out how to do that yourself. For example you can use Real VNC to access the Pi from anywhere, but that’s not what I think of as a cloud connected embedded system. I’ve clearly got some more learning to do. You can do anything you want, but it requires a great deal of research/time. I have learned a ton using the platform, and I’ll continue to use it for one-off projects and possibly as an educational tool.

What impressed me most about the S3A7 prototyping kit was how quickly you could build something that is actually useful. Using the Medium One tools, it took very little time to put together a home temperature monitor that was cloud connected, easily accessible on my phone, and reliable. Renesas/Medium One have done most of the work for you, so it’s easy and quick to write Python scripts that do exactly what you want. My next challenge is going to be to dive into the e2 Studio software to modify the prototyping board software to accept a different sensor. That will be a steep learning curve because there’s no Linux running in the background to manage everything.

They are both great systems, but they are geared toward very different markets.


#5

Dan, thanks for your great thoughts. I have that same “Sense Hat”. It’s cool and the Raspberry Pi is fun in general. I have several. However, there is this pesky problem of how to interact with the RPi when it’s network connected.

My son built some equipment for the Stanford EE dept to monitor UPS batteries, basically has a cost-saving measure since his time was cheap and the commercial connector for the UPS interface was $200+. Although the GPIO pins of the Pi are easy to access and receive signals from the UPS, the interface to adjust the system is basically ssh from Putty or Linux into the Linux shell and edit the config files with vi or pico.

It’s easy to set up crontab and have the Linux shell send out email reports, but there’s a range of issues if the RPi itself loses power. Will it come up??? Remember, this is for a UPS monitoring system.


#6

Is that a Renesas logo (“R”) in the LCD, Dan? Pretty cool. :slight_smile:


#7

The reliability question is a big one. I just don’t have a good sense of how reliable the Pi is. Maybe setting up two of them would be appropriate for a critical application? Could you somehow power them off the batteries so you’d never have to worry about grid power glitches?

The Real VNC program I mentioned is brand new to RPi. If you get the very latest updates and enable VNC server on the Pi monitoring the batteries, you can then remotely access the full Pixel desktop GUI. Would that make things any easier? You have to set up an account with VNC and it cannot be an industrial application. I only tested it on my home network from a Windows machine and it was amazing. You’d want a Pi 3, I think, because otherwise it could be slow. Here’s the post from the raspberry pi website.

Dan


#8

Yes! I’m no artist, but I’m glad it was recognizable…


#9

I think that would be a cool class. You could also lay the groundwork to show them that down the line, they could look into platforms like Synergy. It could show them a connection from what they learn in class to commercial products like real-world automobile environment sensors.

The RPi Sense hat has a five button joy stick and a bunch of sensors, including Gyroscope, Accel, Magnetometer, temp, barometric pressure, and humidity. It also has 64 RGB LEDs. Start there, then introduce them to the S3A7 and Synergy one day… :slight_smile: