In this sections of the website some problems we encountered will be discussed.


The IR-photo-transistors worked perfectly and enabled the robot to 
search for the ball flawlessly.

The LDR's that measure the greyscale work properly but need several adjustments:
Don't compare the sensors with analog electronics because there's a rather big offset error between the different LDR's. It's better to do a separate A/D conversion for every LDR and add or substract the offset value with the PIC. You should also be able to adjust the gain of the LDR's (just like we did) and of the LED that brightens the LDR area (just like we did not). This way it should be possible (after a lot of tuning) that every LDR gives the same value for the same greyscale. A resolution of 10-20mV/cm is possible.

Another problem is that the LDR's have a large settling time (about 2 seconds). This shouldn't be a problem if you use the right aiming algorithm.


During the design it was mentioned by the assignment that we should be careful using the printing material, as it is very costy. Nonetheless it was never mentioned if there was a strict upper limit. This cause us a week of delay, as we made our robot a bit too big. So before you start the design, make sure to ask for an upper limit in material volume! Using inventor it is really easy to check the volume you have used so far.


Always make sure that more than one persons is fully up to date regarding the program. Close to the end of the project when stress kicked in, we managed to get only one person fully up to date, and this really means a hassle when he (or she) isn't available.


You can be sure that if you don't double, tripple and quadruple check every step you make into putting your plans into reality that there will be mistakes in your hardware. Thus it is a good advice to keep track of every milestone with at least two people, controlling the circuits, controlling the designed PCB's, and the soldering of components on those PCB's. Of course double work is ridiculous, but when someone finishes one step, someone else should really control if that step was completed the way it was expected. It would be a good idea for example to let someone who designed the circuits check whether the PCB's are designed accordingly.