Date: Tue, 15 Nov 2005 01:51:22 -0600 (CST)
Subject: motor control design final
X-UID: 105
Content-Type: IMAGE/JPEG; name="img2399.jpg"
Content-Type: IMAGE/JPEG; name="img2401.jpg"
Content-Type: IMAGE/JPEG; name="img2404.jpg"
Content-Type: IMAGE/JPEG; name="img2405.jpg"
Content-Type: IMAGE/JPEG; name="img2407.jpg"

It works off the breadboard with the small Mabuchi motor for testing. I am 
pleased as now there are no technical obstacles to assembling the 
electronics enclosure. This will allow it to be mounted on the robot. The 
remaining electronics are much simpler and not on the critical path of 
this project.

The gate of the power MOSFET has enough capacitance that the emitter side 
resistor needs to be pretty small. At first it was 10 Kohm. But then even 
at a pulse width modulation duty cycle of half a percent, the motor 
remained fully on. I tried adding a resistor to pull the 4N35 
opto-isolator down to ground faster. This helped. But the real solution 
was to replace to 10 Kohm resistor with a 510 ohm. However, at 90% duty 
cycle, the power dissipated by the resistor is about 1/4 watt. The 
resistor is rated for 1/4 watt. For safety, I'll use four of these 
resistors, two series pairs in parallel.

There are now two DPDT relays, one for each drive motor. This allows for 
the same control in reverse as in forward. It is about $10 more in parts 
for vastly increased capability over what I was considering yesterday. 
This design does almost everything the more complicated three H-bridge 
design does but at lower cost and with reduced parts count.

When wiring the breadboard, this time the wire runs will be planned out. 
The haphazard approach used for the earlier design is not maintainable. If 
anything goes wrong, it is virtually impossible to trace the problem.