A Raspberry Pi for $5! What are your decision factors?

It it is obvious that a new trend from the US is swapping over to Europe and probably the rest of the world: Black Friday. That is the day yesterday following Thanksgiving day in the United States. It is a ‘shopping’ day. Consequently, the stores are battling with huge discounts. And I use that  to fill up my inventory for the Christmas-time projects 🙂 What caught my attention yesterday Friday was this: a Raspberry Pi Zero for US$5!!!!

Raspberry Pi Zero

Raspberry Pi Zero (Picture: Adafruit)

The Raspberry Pi Zero (see https://www.adafruit.com/products/2885) is basically a cut down board of the Raspberry Pi 2: but single core, only SD card, HDMI and one USB (the other USB is for power), but in a small and tiny size (65mm x 31mm). So when we did our collective ‘Black Friday’ order with faculty members and students yesterday, I have added one of these to the shopping cart 🙂

Why? Not because of the price of the board. One big decision factor was the form factor: I see it extremely valuable to have small and tiny boards (see the tinyK20). So this board is about half of the size of a Freescale Freedoms board which I use a lot. But for many projects I need something smaller and lightweight, so this will be a good fit.

But the really key decision factor is the availability of open source software and tools and excellent tutorials available on the internet. And I order from Adafruit not because their prices are the cheapest, but because they show dedication for what they do, and have outstanding tutorials available.

I’m using my various Raspy boards not because it is the best hardware or board in the world (there are faster, more powerful and feature rich boards), but to me it has the best software and tools eco system for small Linux-based embedded projects. And there is a huge eco system with accessories too. The key factors are software/tools/tutorials for me. Not the hardware, this comes after all that.

The Raspy is extending the capabilities of my other software and hardware projects. The Raspy is a powerful mini computer. It does not replace or is intended to run hard realtime applications, so I continue to use microcontroller boards like the Freedom board running bare metal or an RTOS like FreeRTOS, in combination with the Raspy boards. The Raspy is running the Linux, cloud connection, web server, networks stack or for example openHAB, and the ARM on the Freescale Freedom board (e.g. FRDM-KL25Z) the realtime communication protocol and dealing with the ‘hard’ time aspects. But all this was for me ‘software and tools driven’: If I don’t have good software and tools for it, I would not consider it.

So my mantra is: Pick the software and tools first, then pick the hardware. I have seen too many projects doing it the other way round and failed. Of course the hardware is important, but if I cannot get the software running on it, what’s the point? Bad software and tools beside of too optimistic planning is a key failure point for me. “Life is too short for bad tools” I would say ;-).

The question is: have silicon and hardware vendors fully realized that? Atollic is right (in my personal view (see “Atollic TrueSTUDIO Lite for ARM with Unlimited Code Size“)) that the embedded software and tools market is fragmented and many software/tools offerings are not of good quality. Tools and software stack come and go, are incompatible or breaking compatibility from one version to another. On one end you get what you pay for, on the other end there are excellent free and open source options too (see my articles series for gprof and gcov). I do not only need good  and reasonably priced hardware, I even more need excellent software and tools with tutorials and documentation with a good community support. And I need stability and confidence that the software, tools, libraries will be available for the coming years.

I have ordered now that Raspberry Pi Zero and other gadgets 🙂  because of tutorials, software and tools as my decision factors. What are your decision factors?

Happy Picking 🙂

Advertisements

12 thoughts on “A Raspberry Pi for $5! What are your decision factors?

    • Yeah, I did the same :-). I’m wondering how much the true costs of that board are. Yes, there will be a huge demand for this board (>50k?), but I’m wondering what the real costs of this board are? There must be some margin too.

      Like

  1. Although it is usually best to pick processors based on available tools, there are times when hardware comes first, rather than software. The main one is very price-sensitive applications, where there is relatively little code to write or debug, but a difference of 25¢ on the hardware can make or break the product. Of course, the increased development cost of having poor tools has to be factored into the price comparison.

    Like

    • Agreed, this is especially true for lare volume products where the cost of development can be distributed accordingly. What easily gets not counted into the equation are the maintenance costs over time. If the product is a ‘one off’ product that probably does not count, but if the product needs to be be in the field and maintained for years, then the accumulated maintenance engineering costs can easily outpace the initial development costs.

      Like

      • True. I was thinking of a design-once product, not one that needs to be upgraded to maintain viability. If you are looking for a long-life product, then how stable the manufacturer is and how popular the component is may matter more than the current price for determining availability of parts on 5 years.

        Liked by 1 person

  2. Hi Erich, good article for explaining your point of view for this dichotomy of the embedded world. In one hand we have low power MCU solutions like the Cortex M0+ with the Kinetis L microcontrollers with a minimum consumption, and more and more memory and devices day by day. And in the other hand we have solutions like a Raspy Zero Cost with a huge Broadcom BCM2835, 512MB of LPDDR2 SDRAM, SD card, etc… but the focus that I want to do is in the power management that these boards needs, because a nice oportunity of having a little board is to make wearable or portable technology, then ¿What is the energy solution for a Raspy Zero Cost working on a portable solution? and ¿What is the energy solution for a Kinetis L to make portable solution? I think the same way as you do, to have great software tools and the support of the community is a big deal, but for making weareble or portable solutions the power is a key too.

    Like

    • Hi Carlos,
      Good questions :-). There are several vectors and factors, all with different weights for a decision: software, tools, costs, ease-of-use, availability, long term availability, performance (per Watt or per MHz) and of course energy consumption per MHz. To all this, the engineering costs to get it done (and maintained) needs to be factored in.
      As for wearables: I believe the near-future is that microcontrollers with best low power capabilities will be at the ‘edge’ with the sensors/etc and the ‘huge’ systems like the Raspy/BeagleBone/etc running Embedded Linux will be the ‘concentrator’ and ‘Gateway’ role. I believe over time the ‘huge’ systems will get better and better over time, as we see today 32bit ARM processors in areas which were 8bit dominated in the past. In my view, it won’t be one core for everything, but some low-end systems for the battery operated side, and some higher end system the chain up.

      Liked by 1 person

  3. To me, the RPi Zero seems like an odd decision anyway. The lack of on-board communication severely limits the available options. HDMI is nice to have but without communication you have less choices.
    Adding an external USB WiFi adapter is an obvious choice, but it also needs a hub if other USB ports are needed.
    If the designers really intended to come up with Zero as a low-power solution, then I think the whole idea is somehow flawed. The BCM2835 is not a low-power chip anyway …

    For a project involving low-power and Linux, I used a tiny board based on iMX233, which by the way I think it was a fantastic chip for the low-power needs. I don’t understand why Freescale didn’t put much emphasis on it.
    A similar format as RPi Zero but using iMX233 would really be low power. Of course, a WiFi USB adapter needs to be added for this solution too but the iMX233 is very nice because of the integrated PMIC and battery charger.
    If Freescale would support the iMX233 with the newer Linux versions, I would use it in many projects right away. The RPi Zero would be no match for such a board when it comes to power efficiency.

    Like

    • Hi Bogdan,
      I agree that this solution is not matching low power needs. I think it is more about the small size (so you can have a bigger battery :-)). But really I think it is about the price of $5. If it would be in the $30 range, probably there would be not much hype about it. But for $5, that makes it really attractive.

      Like

  4. Pingback: A Raspberry Pi for $5! What are your decision factors? | eldofrederick

  5. Pingback: From Zero to NOOBS: Starting with Raspberry Pi Zero | MCU on Eclipse

What do you think?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.