|
|
Lessons learnt.
This section is a fast forward of the 3 months we enjoyed creating 'drishTi'.
A (really) little bit of history and then some of the invaluable lessons we learnt.
Before we start, it is necessary to tell you about that strength which has brought us to a level where we feel bold enough to speak out in the world. The person who has spent days (and nights!) with and on us. The fire we have in us today is from the spark Shankar sir lit in us and fanned it over the four years we spent with him at PESIT (our UG college in Banaglore, India). He tried to teach us what it means to work, what research means and most of all what it means to be passionate.
The runs to his room (we called 'aDDa') during every free minute and the haunts during holidays have taught us immense lessons. All our previous works were born and nurtured in that room, which we can easily claim to know as the back of our hand- and may be even better than Sir himself.
Those episodes are our cherished earnings, which I doubt we'd ever share with anyone else. But what we want to share is what he taught us or rather the bits we managed to learn and understand (we still recount his words and learn something new each time). The result of which is this robot.
Lets start....
|
|
THE BIG Question- What do we do? |
|
|
THE day. 19th
November 2005. Prakshepan is the Robotics and Rocketry Club of PESIT. On the day of its launch, Mr. HNLN Simha from L&T talked about Robotics and Unmanned Aerial Vehicles. He brought out the point that the problems in that field were truly multi-disciplinary. It was decided there that the project should have something to do with robotics.
|
|
Who talks to the circuits? |
You
decide to use a laptop to control your robot because you have to do a lot of
image processing. You need some way to get at least six or seven signals out
of the laptop into TTL compatible voltages. Unfortunately, the model of
laptop you have for the purpose does not have a parallel port. You try
looking for adapters from Parallel port to USB. Find quite a few. But they
are made mostly for printers and do not allow you to address the port at the
pin level. What do you do? Well, look at another port on the laptop to bring out the signals. Against popular opinion, we chose the sound port. Yep, the port that you plug your headphone into to listen music. How do you get 6 signals out of there? – you ask… Quite simple, if a particular frequency is present at the output, it represents a 1 in some bit otherwise it’s a zero. For this you need to have a set of filters in hardware to decode your bits. But hey…if you put your frequencies far apart you wont get as many outputs...because sound card can only produce up to 15kHz with good fidelity. If your frequencies are too close, then resolving using a filter is a problem. Catch 22 – perhaps not. |
|
|
First Encounter with Harmonics We made the filters to decode the bits and it was working pretty well too. But suddenly something went wrong. All the filters started detecting even if there was no output from sound card! Turned out we had put the laptop for charging. The harmonics from the power line and the transformer in the rectifier was picking up noise and causing the problem. Moral of the story : Don’t charge laptop when you run the robot.
|
|
|
Motor Repair!!! We had started with two motors which had nearly the same characteristics when we started the project. But gradually, there was a drift to the right. When we checked, the right motor seems to be running slower than the left, even if they were connected directly to battery with no electronics in between. Problem had to be in the motor. What to do? Simple. Open it up. We opened up the motor and found that one of the carbon pieces, which acted as the brush for the motor providing contact to armature, had nearly corroded. Lucky for us, we had a spoilt motor(no we didn’t spoil it!) with a good brush. We replaced the brush and the motor was back to normal. Moral 1 : Don’t be afraid to try things out, if you have no knowledge regarding the outcome, you can still be right in half the cases! Moral 2: (Important) Do not wear white clothes while taking a motor apart. Just because you don’t study lubrication as a part of your I Semester Electrical Engineering, doesn’t mean there is no grease inside a motor!
|
|
|
Birds of a flock… We designed a IRF9530 MOSFET drive using 741 op amps and it seemed to work well for a while. But when we duplicated the drive for other motors and plugged new MOSFETS in…BOOM…fireworks (this was before we put fuses). If we plug in the older MOSFETS, it works! So where is the problem? Upon careful examination ( and I mean reaallyyy careful ), we thought of reading the batch numbers of IRF9530, they were different! Turns out our circuit was sensitive to the variation in the threshold voltages across these two batches of MOSFETs. Now if we were doing a test bench for International Rectifiers, our circuit was excellent. But for our purpose, its BAD electronic design. We changed from 741 to LM324, that gave use better saturation voltage which seemed to solve the problem. |
|
|
Coupled Motors? – Second Encounter with Harmonics We were usually testing only one motor at a time till this time. We assumed if it worked for each motor independently, it would also work for all motors together…WRONG Assumption… When we connected the all motors together, turning one ON turned at least 2 or 3 motors ON. Thus defeating the purpose of having different control channels. What was happening? Turns out the when the motors turned, they put up harmonics on the power line, this caused the other filters to fire. When we were only testing with one motor, we couldn’t observe it because other filters were not driving anything. So we knew the problem, but what was the solution?
|
|
|
Always carry Spare MOSFETS… The number of spare MOSFETS to be kept ready for a demo is given by Rohith’s Estimate RR: (No. of MOSFETS on board)*(Importance of Event)/Availability of MOSFET at the Demo venue If its your final demo, then this number works out close to 213.9865 but you can round it off to the nearest tens and buy just 10!!! |
|
|
Do not paint the wheels We got carried away as things started working and decided to decorate the robot. We painted it the body jet black, body blue and so on. We painted the tyres as well. We left it do dry overnight. Next day…The paint on the tyres made the robot skid as it lost traction. We had to file away the paint to get to the old state. Moral : Function first…beauty later.
|
For the more technical lessons with explanation please see 'The Neurology' and 'drishTi's drishTi' sections.
