New ArduPilot boards

We're making good progress on ArduPilot, the Arduino-based open source autopilot, and Jordi's just finished an improved version of the board, along with a new surface-mount version for those who care about size and are extra handy with a fine-point soldering iron. The new version of the through-hole board corrects one error on the previous version--a diode was in the wrong way around--and adds a better layout, better connectors for the receiver and servo cables (Samtec right-angle connectors) and the GPS, and clearer labeling throughout. You can see the PCB layout above, and buy the board here. Eagle 5 files are here: ardupilotBETA.sch and ardupilotBETA.brd The components you'll need are these:

A Boarduino kit
Multiplexer and ATTiny chips
Two 220 Ohm 5v resistors and One 10k 5v resistor
A surface-mount reset switch
An extra LED
A EM406 GPS Module
A GPS connector
Three female-to-female RC cables (you'll cut them in half to make the pigtails to the RC receiver)

I'm working on the Autopilot code now, and should have a beta in a few weeks once I've had a chance to test it. I'm starting with a port of the Basic Stamp autopilot, but this board is designed to do much more, so we'll be coding in new functionality like the ability to read a barometric sensor for altitude control and read the FMA Co-Pilot (or custom IR sensors) directly, to integrate stabilization directly into the autopilot.

The surface-mount (SMD) version is functionally identical (except it has two extra analog inputs for sensors, thanks to the SMD version of the ATMega168 chip having more I/O pins), but smaller. We designed this one mostly for ease of automated manufacture--we will be offering a pre-made board with components already mounted in a few months--but if you're not too afraid of hand soldering intzy-wintzy leads, you may prefer this one, which you can buy here.

Here's the overview of the SMD board:

Here's the component list for the SMD board (Digikey parts): Capacitors:

(C1,C3): 2x 22pf PCC220CNCT-ND
(C2,C6): 2x 0.1uF PCC1828CT-ND

Diodes:

(D1)1x 1N4001DICT-ND
(D2,D3,D4):
1x LED (red): 67-1359-1-ND
1x LED (green): 67-1357-1-ND
1x LED (blue): 160-1643-1-ND

Connector:

(GPS)1 x EM406 (sparkFun)

ICs:

(IC1)1 x Atmega168, ATMEGA168-20AI-ND
(IC2)1 x 74LS157D,296-14884-1-ND
(IC3)1 x ATTiny45, ATTINY45V-10SU-ND

Crystal:

(Q1) 1x 16MHZ SparkFun

Resistors:

(R1,R4) 2x 10kOhms P10.0KCCT-ND
(R2, R3) 2x 1Kohms P1.00KCCT-ND
(R5, R6) 2x 220 ohms 311-220ARCT-ND

Switch:

(S1) Switch SMD, sparkfun

You may notice that the image of the board above appears to be missing some chips that the through-hole version has. That's because they're on the bottom of the board.

Top:

Bottom:

You know I have been thinking about the resolution issue with 300d/s gyros and the resolution of the AVR168 chips. It seems to me if we think outside the box for a moment the problem can be easily solved. It would require 3 op amps per channel but it may be worth it.
The first op amd is a unity gain buffer. The other two op amps are configured to have 0 -3.3v for an input of 0-1.5v the second op amp would drive an ADC1 input directly. The third op amd
would be biased so that it did not produce output until it reveived an input of 1.5v. This coulkd be easily achieved with two diodes in series to the input of op amp 3 the voltage drop across each would be .7v and two would be 1.4v. A forward bias would be achieved at ~1.4v hence setting the bias. The third op amp would feed ADC2 and readings would only be taken from this if ADC1 achieved an overflow condition (full 1024 bits). This should give more resolution than ever needed. :)

You need to be a member of diydrones to add comments!

Join diydrones

Comments

hoopty August 21, 2008 at 6:47am

Will do
3D Robotics

Chris Anderson August 19, 2008 at 9:26am

Huh? Because nobody responded to your comment and suggestion? That just means that either A) nobody else here is working on the same problem, or, B) you posted your idea as comment on the third page of a complete unrelated post that is three months old.

My suggestion is that you repost it as a new discussion topic, which is what it should have been all along.
hoopty August 19, 2008 at 9:20am

I guess I have been blacklisted. :(
hoopty August 12, 2008 at 11:25am

Like this
hoopty August 12, 2008 at 5:09am

You know I have been thinking about the resolution issue with 300d/s gyros and the resolution of the AVR168 chips. It seems to me if we think outside the box for a moment the problem can be easily solved. It would require 3 op amps per channel but it may be worth it.
The first op amd is a unity gain buffer. The other two op amps are configured to have 0 -3.3v for an input of 0-1.5v the second op amp would drive an ADC1 input directly. The third op amd
would be biased so that it did not produce output until it reveived an input of 1.5v. This coulkd be easily achieved with two diodes in series to the input of op amp 3 the voltage drop across each would be .7v and two would be 1.4v. A forward bias would be achieved at ~1.4v hence setting the bias. The third op amp would feed ADC2 and readings would only be taken from this if ADC1 achieved an overflow condition (full 1024 bits). This should give more resolution than ever needed. :)
Phil Wilshire August 6, 2008 at 6:04pm

Chris,
I have the right boards and I have ordered the rest of the parts now.
I will, hopefully, build one or two of the new boards as soon as all the bits get here.

Phil
3D Robotics

Chris Anderson August 6, 2008 at 2:54pm

Bill,

Good catch! We'd actually been using the right chips, but I put in the wrong link. Now fixed...

Many thanks!
hoopty August 6, 2008 at 2:34pm

Hi Chris,

Great work but I have some bad news... The ATMEGA168V-10AU-ND listed in the SMT project is a 10Mhz unit. All the picopower units as designated by the V in the part number are 10Mhz max. You will actually want a ATMEGA168-20AI-ND and it is a 20Mhz unit!
Sorry

-Bill
3D Robotics

Chris Anderson August 6, 2008 at 7:51am

Hoopty,

Sorry about that--I've fixed the link above. We're still improving the boards, so think of this one as a test platform. The final version will be out in Sept (and available pre-soldered)
hoopty August 6, 2008 at 6:34am

Chris,

The link to buy the SMT version of ArduPilot points to the thru hole design. I sure would like to get started on the SMT version. [=)'

of 3

This reply was deleted.