UPDATE: This project is now complete. Further details can be seen at GhostRoachMouseBullRam project page. You can still read below to see my adventures and problems encountered this project.
Well, at heart I'm a cat taunter and despite b4short being able to make my cat curious, she's still too afraid to get very near. On the other hand Mr. Wiggles is too small and fragile. So following lumi's advice ("what does a cat loves more than a mouse") I finally decided to make one (for starters, if it goes well I might end up making more, I mean the more... the tauntier, right?).
So I had a crappy shaving machine kit that was on my crossairs for scavening for a while... got a couple of DC motors, one of those tipically flattened (from the nose plier :P) and a rounded one (with gearing for the 3 shaving heads :P) from the actual shaving machine. Unequal... unsuitable for my mousey...
Next day I went to the chinese bazar and bought a identical couple of the cheapest electric car toys there:
In the picture you can see one already opened up, a closed one and also a solar fan I got a while ago that I also took the chance to scavenge last night. So last night I got a couple of motors that I can use in mousey.
However the total scavenging potential tally can amount to:
2x slide switches
2x DC motors
8x small wheels
2x 2AA battery holders
2x red-pinkish leads
2x speakers
2x brackets
1x solar element
1x motor w/ fan cage
1x plastic clamp
and several screws...
then there's useless leftover plastic, and also the micro boards from the cars I can thing of them as very useful, unless I need police sirens, and even so I think I would be better off by not using them.
Regarding mousey this is what I have so far:
I also picked clean a retired pc mouse, but it proved too small to fit all the components. Now I'm,on the hunt for bigger mice...
Well the closest person I have to a boss just gave me an old PS/2 mouse telling me "I don't even want to know" so I did not tell him, but I'll tell you. It seems appropriate and I think everything will fit in nice and tidy (hopefuly). See below:
At this point yesterday, the was looking at the picture above and thinking "well a couple of hours and I'll have this ready and running on the floor" how wrong was I.... let's just say... day was starting to dawn when I decided to stop and tidy up everything and thus hiding the evidence of the mess I made... below are some photos of what I achieved. The perf board remains untested because... well it's incomplete and you know... sunrise...
So the mouse carcass need more work... and I found the pliers were not enough... hence.. I *ahem* let's just say I better buy some new soldering iron tip(s)... *cough*
This took an ungodly amount of time, and also was the messier step since I use the plier's nose inside grip identations to file away plastic mostly on the motor "windows"...
The perf board... front and back... it was harder than I thought at first. Not only did I have to think in the perf's small 2D space, I also had to think 3D regarding available height inside the mouse.
At some point I was so tired that I undid a solder point thinking it was the wrong connection just to realize it was right one, so I had to unsolder it again and solder it back in its original spot... hopefully all the connections are right and live... as for solder bridges... I may have to track them down later, but if i'm really lucky there is none (i doubt it)
By the end of the night/start of dawn after tidying everything up what I had accomplished can be summarized by the foto below:
After sleeping 4-5 hours, facing the work day and going back home there I was, again ready and this time resolved to put the mouse into running condition.
I started by making the power connections and bump switches crimping (for dupont connector).
After that I connected the battery, the connector to the perf board circuit, and toggled the switch... and voilá the nose led lit up... and no smoke... things were starting to look good, but too early to celebrate. Next followed the motor connectors crimping.
This allowed me to verify that the motors were also working as expected... my smile was starting to widen, only need to test the bumper switches now, and for that, some packing was required. Hence, I started by mounting the motors in place.
I used a material that I think is used to support furniture feet, (so it does slide or scrapes the floors) in order to provide the motor angles, fastened the motors with those plastic bracket stripes which I don't know their proper names :P After that is was relatively easy to put the rest of the stuff in, as can be seen below.
Tested the bumper switches, and HURRAY! they were working, now all I needed was to close the lid. It required a bit of force applied the right way, but overall everything fit quite well.
Now let's introduce the mouse to the floor... the first time the motors stalled right away, but that was quickly fixed by replacing the battery. The second time, one of the wheel popped right of after a couple of seconds. After re-attaching it I realized that I got a spinner, that is the bot spins in thight closed circles. The thing is the contact of each wheel with the ground is uneven, and my attempts to even it out, so far the only thing they have done is alter the spinning direction, either clockwise, or anti-clockwise.
I just need apply more engineering finess to solve this issue, despite this I'm over all happy mostly because everything fit and the perf-board connections were ok on the first go :)
And then it starts going bad...
So I was starting to try and solve the problem of spinning in place mouse... eventually I decided to change to a smaller rubber wheels that I've also scavenged from the toy cars. Those have the added advantage of fitting right into the motor axis, and being rubber. I also had to increase the angle of the motors so that the wheels actually managed to reach the ground surface. For that purpose -- bits of an ice cream stick & glue... ugly, but oh well. The problem was that it was harder to close the mouse now, but still doable. Upon turning on the bot, the spinning persisted...although it sometimes it seemed to manage to break out as the mouse velocity reduces (draining battery). Another big woe.... the reverse... disappeared... the switches are ok... so I guess that either the transistor fried or there's a broken solder join in the perf *sigh* ... I'll put everything away for today...
The good news is that the transistor is ok, so the bump switch reverse is somehow disabled by the pressure of closing the lid.
I add a video as per lumi's request :P The battery is somewhat running low, however curiously that's one of the things that makes the mousebot break out of it's spinning, alas it also makes it stop on it's track a few moments later. I must now sort the bump issue.. perhaps redo the motor mounting and tray to aleviate the pressure within.
Well...
I stopped to consider things for a bit. With free rotation both wheels seem to turn "as fast" but then again what do my eyes know? It's not possible to differentiate rotations at that speed. So I popped the chassis open and stick the multimeter leads into the motor leads and "started the engine"!!! vroomm vrooomm :-P Anyhow, the readings have shown me my right motor always at higher voltages than the left motor. I guess that would explain the counterclock spinning, instead of a mechanical build mal-adjustment. I tried to use the 100k trimming pot, but it dropped the voltage on both motors... also I still have to try blinding/shielding the "eyes" that way the voltage distribution should be "even", right?
August, 26th 2012
Since the mousebot is still not working properly I decided to dig it up today again. Opened it up, disconnected the motors and connected the mainboard with the battery. With the switch turned ON I read around 5V output for the RIGHT motor vcc PIN and (-)3.4V for the LEFT motor vcc PIN. Turning the trimpot either way change this readout very little. Previous measures directly done on the motor leads while running also shown me that adjusting the trimpot would either raise or lower the voltages but on BOTH motors at the SAME time. Following the oficial schema this below is the alteration, that is where I saw the trimpot should be placed in order to adjust differences, but oh well... if anyone can shed light where I've gone wrong... much appreciated. Many thanks in advance.
P.S. - In the original both grounds of the emitters (working as receivers) where connected directly before the 1kohm resistor.
September, 9th 2012
After having finished designing a PCB board on eagleCAD for mousey I decided to try and debug my problem before I etched said board, so I re-breadboarded the original circuit only to find the exact same behaviour that I did in my perf-boarded mousey... which led me to conclude the my board had no miss-connection problems. I was doing some wire connectivity testing while chatting on the SB about my mousey woes, when I remembered that pins 1 and pins 8 on the LM386 were tied together to maximize gain... so I though... what if I unplugged them? I did so and checked the voltages, and voilá evenly divided... however now the sensing LEDs barely reacted to light changes... suspecting something I quickly jumped into a quick search on the internet and found that pins 1 and 8 are better connected through a resistor. After that I did further testing with various pots and resistors, however at lower resistor values I couldn't make the Voltage difference equilibrate... at higher resistor values I could equilibrate them further but at a cost of losing too much sensitivity. So I still have a lot of calibrating to do... however I took mousey for a spin with a 1k resistor and well it acted a bit better... however at first run it was such a brute that it knocked over its own whiskers :P
good luck and keep going this way.