This is a shot of the completed MARV. In the photograph below you can see that the 12V SLA has been mounted as well under the top plexiglass.

The robot is now truly ‘plug and play’, and below is a shot of the modular ICSP cable, which is about 2 feet long,

Overall, I feel a bit disappointed that I do not have the time to implement on-board PID control over the wheel velocities, interface the quadrature encoder feedback with the QEI interface of the MCU or finalise the Linux Command Application.

However, this is all discussed in detail in my thesis and that takes precedence over a simple demonstration. C’est la vie, and here’s a final shot of the robot…

And a couple video clips of surface testing of the final demonstration routines - move along w1, w2, w3 and then spin on the spot in both directions and stop.

Smooth surface test:

Rough, cobble like surface test:

I have taken MARV from a simple chassis to what it is currently in 10 days exactly, although a couple days were wasted waiting for components to arrive via the post.

The project box currently looks like this:

At this point in time, I will be diverting my time to work on the thesis rather than take the robot any further. This is a video of simple open loop motor control with it just spinning as a first test:

A shot of MARV, myself and my mini-makeshift lab…

Had a bit of a scare today, when my ICD2 programmer simply refused to work. Just when I had given up, MPLAB uploaded new software to it and it is back to working properly…I hope!

Progress is now going rather slowly, as I am waiting for a few bits and pieces to arrive on Tuesday and although today is a ‘Bank Holiday’ Monday, I’m hoping that all the usual companies will ship today

The only progress made today has been this wiring:

I intend to fully move the Motor Controller breadboard inside the project box before working on any further code development. The reason is that I need to see how the robot actually works on a surface and am hesitant to do so with wires attached.

Once I got the ICD2 back up, I cleaned up some of the code since it has now been ported over to the 18F4331 and tested some more PWM output on my scope,

I hit a serious problem earlier on in the day. I had failed to notice that the PWM5 port on the 18F1330 was sharing the PGD pin and as such after each flashing of the MCU, the 3rd motor would not function as the programmer pulled the line active low.

While I let this issue rattle around inside my head, I thought it was a good time to test the Hammond project box circuitry. It astonishing the silly mistakes one can make without realising but all is well, and as you can see it is functioning nicely below:

I may abandon the idea of mounting the Sensor Controller upside down on the lid of the box, and mount it 90degrees to the left of the Motor Controller.

A bit of thinking later and I decided that I was going to shift over to the 18F4331, the same MCU that is being used in the Sensor Controller. It just makes more sense to use two ‘identical’ MCUs.

What attracted me to the 18F1330 is that it can up its Fosc to 32MHz from the internal 8Mhz via internal PLL circuitry - this feature is not available on the 18F4331, and for the moment all the code is running at 8MHz.

I have also been able to verify my CRC-CCITT direct routine and it can process a 25-Byte payload (i.e. a string of 25 ‘chars’) in 2ms. I cannot even see the trace on the ’scope if I try to visualise the time taken to process a single byte.

Many hours of arduous troubleshooting, and one 18F1330 dead later on, I have now got the same ‘base-code’ running on both MCUs. Another great feature of the 18F4331 is its QEI interface…ooh!

I have now worked 32 hours on the robot with a 3-hr power nap in between. I hope that I won’t need too long to recuperate from this…!

This was the first ‘all’ motor test for MARV. There is a problem as my mini-PSU is not able to handle the current draw; therefore, I have ordered a 0-20V/5A bench PSU from Maplin with a couple other chargers as I am now looking to finalise the mobile platform itself.

Close up of the Motor Driver <-> MCU interface…I need to find a way to manage this a bit better.