Turn your Raspberry Pi into a cheap robotics platform with BrickPi

The Raspberry Pi is a small, cheap device designed to make computers accessible to students, hobbyists and others. The BrickPi is a small, cheap device designed to build on that concept by letting you use a Raspberry Pi as the brains of a programmable robot.

You can pick up a Raspberry Pi for just $35, and the BrickPi team wants to sell its robotics add-on for the same $35 price. So for about $70 you get a programmable device with support for sensors, motors, and just about everything else you need to start building robots.


The BrickPi is basically a board that you can attach to a Raspberry Pi, It includes connectors that let you add LEGO Mindstorms motors and sensors so that you can build a device that can move autonomously, detect the world around it, operate a claw, or perform other actions.

It draws power from a 9V battery, and you can even power the Rasbperry Pi using the same power source, which lets the mini-computer work as the brains of a robot even when it’s not plugged into a wall jack.

The BrickPi team comes from Dexter Industries, which is taking pledges for the project in a Kickstarter campaign which ends on June 16th. The project had a pretty low goal of less than $2000, and has already raised more than $90,000.

In other words, the team probably didn’t need to turn to crowd-funding in order to get the money together to build BrickPi devices. But Dexter Industries says they turned to Kickstarter in order to generate a bit of enthusiasm for the project and help build a community around the BrickPi.

Crowd-funding will also help keep the cost low, and the team’s already raised enough money to meet a few stretch goals including support for development in C/C++ and some hardware improvements.


via TechCrunch


  • toronado455

    How are they getting power from the 9V battery into the Pi? How long can it run on this power source?

    • John Morris

      Not long. A 9V battery has very little capacity to supply current and only few total mAh. There is a reason it is usually referred to as a ‘transister radio battery’; because it was designed for low drain applications that need more than a couple of volts back in a time when DC-DC conversion wasn’t practical. A 9V battery won’t run the Pi itself long, and these idjits are driving motors from it. Motors. In the plural.

      That decision alone tells me everything I need to know about the EE skills of this team. This is the problem with the kickstarter world. Designing a basic little i2c driven controller board with modern automated circuit design and Chinese fab is very easy to get to the prototype stage and thus raise a sack of cash from the ignorant on Kickstarter. Knowing enough to design an actual practical device still requires almost as much skill as it always has.

      • CyberGusa

        Seems you don’t understand what’s being offered!

        First, they’re not planning on selling an entire kit but just an adapter to let people use a Raspberry Pi with a LEGO Mind Storm robotics kit, both have to be acquired separately!

        Second, the point is to provide a low cost solution and they’re not charging any more for the adapter kit than the Raspberry Pi itself costs… So just about $35 each or $70 total for both the Raspberry Pi and the adapter kit.

        Third, both the LEGO Mind Storm kits and the Raspberry Pi can suffice for more than several minutes on a 9v battery and the point of experimenting is just trying stuff… Not have it last for hours! So that’s usually more than long enough to find out if a experiment works or not!

        Fourth, using options like a 9v battery is also because they’re cheap! Again, this is for experimenting and not meant for having something you can run for hours doing any actual work!

        Fifth, larger batteries also add more weight and the LEGO Mind Storm motors aren’t that powerful… So power to weight ratio has to be considered too, since it will have to be able to carry the battery weight as well!

        Besides, you can always adapt it to a larger battery pack if you need to and have a project that doesn’t actually need to be mobile, and they also sell rechargeable lithium 9v batteries with up to 500mAh (4.5WHr) that could probably power these kits for 15-30 minutes at a time and would still cost you less than $5…

      • John Morris

        I understand exactly what is being offered because I viewed the documents on their website before commenting. It ain’t rocket science, it is a simple, ripped from the spec sheet example, of a brain dead circuit to interface i2c to a couple of sensors and motors using an AVR micro. Not that there is anything wrong with that, unless the people doing the ripping don’t know how to read the spec sheet. And these guys forgot to read the spec sheet on a 9V battery to see what the recommended max current drain is, what else did they forget?

        Personally, considering the target market, I’d kinda like a little isolation and some fault tolerance. There ain’t much there protecting against wiring things wrong.

        And note that this thing is intended to drive Lego Mindstorm motors. The Lego controller requires 6xAA batteries, which is exactly the sort of power source you would expect if motors are involved. Because Lego knows what the heck they are doing. The good thing is the wide availability of 6AA battery holders that will snap onto that 9V battery connector.

      • John Morris

        Update. ya got me to look a little closer. Total fraud, that schematic can’t possibly work, the product pictured is nothing more than a non-functioning mockup to bait Kickstarter with. The AVR is running on 5V and the Pi’s bus works on 1.8V There is no level conversion going on.

        You might get away with that if the only thing doing interop was the AVR and the Pi, just don’t explicitly drive the i2c bus from the AVR side and depend on that bus working by driving it to ground to signal. If the AVR will see 1.8V as a high, and I’m pretty sure it would, you would be ok. But they explicitly show both SCL and SDA being pulled up to 5V. Plug that up and you get a dead Pi, at least according to the spec sheet for the SoC. Fix that and you have to hope no sensor you ever connect pulls up the i2c pins like the i2c specs always depict.

        Next problem is the motor controller chip they show is a standard H-Bridge. But it requires four GPIO pins for each and they only show a single MISO going in. They have the enable pins pulled up to 5V and the PWM signals going to one of the motor directional drive inputs. Not happening. The AVR selected doesn’t have enough input pins to drive three motors plus the analog sensors plus the tach signals. Total fraud.

      • CyberGusa

        Ah, then you must think this Raspberry Pi with I2C-Arduino-Slave?


        As for the Motor Controller… The SN754410NE is a Quad Half H-Bridge and not a “standard H-Bridge”!

        The pin out they show for it is exactly what TI shows for the chip and it’s been used in projects to control at least two motors per chip, and this RPi adapter has two of them!

      • John Morris

        Yes, look at that article a little more closely. Notice what is between the Pi and the AVR? A level shifter. Now look what isn’t on this snake oil product. OK, instead of relying on an aging memory, just checked the big G. Looks like the GPIO lines are 3.3V, not 1.8 as I thought… could have swore the Pi blog said 1.8V logic when they were warning people to be careful.. And the net is full of warnings to NOT do what these clowns did, which is to directly connect a Pi to 5V TTL logic. BOOM!

        The funny bit is that every current AVR will run just fine at 3.3V and plenty of motor bridge chips can accept 3.3V logic. So the only problem is they didn’t stick a 3.3V regulator on the board to drive the logic along with the 7805 to power the Pi.

        As for the motor controller, it is a bog standard H-Bridge of the sort I have wired up myself. It needs two ‘gpio’ type control lines and a pwm for each motor to control direction and speed. They support three motors. I make that as nine digital/pwm I/O lines needed. Add two lines to each for the tach signals. That comes to fifteen, right? Now add in four analog lines. I count fifteen other signals, clocks, vcc/gnd and two marked as unused. On a chip with thirty-two pins. See the problem? We have all been there, ‘cept most of us figure it out in the pencil & paper stage and either work out a way to pin share, add an outboard multiplexer of some sort or get the next bigger microcontroller.

        The schematic shown is only a placeholder. They will quietly update when they finish actually designing something.

      • CyberGusa

        Sorry, you didn’t look closely enough… The RPi adapter’s ATMEGA328P includes the shifter!

        You should bother to actually look up the chips they’re using before assuming you know what they’re doing with it!

      • John Morris

        Dude! It is an ATMEGA, the I/O pins will only exhibit one of three states: GND, darned close to VCC or tri-stated. All of the Atmel ATMEGA parts are built from the same basic building blocks. If you want 3.3Volts you power the chip at that, period, they don’t have anything in them to generate different drive voltages for the different functional blocks. But it doesn’t matter though, look at that schematic again, follow pins 27 and 28 out of the AVR chip and see those 43K pullup resisters connecting SCL and SDA straight to the +5V rail? Game over for the poor Pi. And if you somehow got the ATMEGA to run the i2c bus at 3.3V it would probably be toasted too.

        Buy yourself a breadboard, a UISP cable and some AVR chips and learn a bit about how this stuff actually works. Try http://www.sparkfun.com or http://www.futurlec.com. Shop hard and you can get everything you need to get a LED blinking for less than $20USD. Most Linux distros ship the AVR compiler and chip programming utils or you can download copies for less friendly platforms.

      • CyberGusa

        Man, you’re really full of yourself… Sorry but I already know how these circuits work and you’re clearly reading the diagram wrong.

        You already clearly made the mistake of confusing a Quad Half H-Bridge with a regular H-Bridge and you’re similarly wrong on this too!

        First, the diagram isn’t showing the SDA and SCL pin out leading to a 5v rail, it’s showing the output from the chip before it goes through those 43k resistors!

        With a set voltage and amperage, the resistors can reduce the voltage by a set amount and that’s what they’re doing! Using the resistors for a voltage drop!

        Really, you act like there’s no way to alter a voltage value from the source!

        Besides, the RPi GPIO pins include both 3.3v and 5v!!!