kirkpthompson wrote:Can you get Mindsensors to provide technical info on their daisy chain addressing scheme?
kirkpthompson wrote:have you tried all permutations of connecting the three MUXes to see if a certain one causes a problem in the chaining?
i.e. a>b works
a>b>c doesn't work
c>a doesn't work
Tried changing the sensors connected to each MUX around to determine if there is a sensor issue?
So all three have different addresses and work when they have their own NXT sensor port... hmmmm...
Basically, try to isolate if a particular MUX or sensor has "issues"
icemadsen wrote:So we don't know whether this is due to some error in LeJOS or what is causing this?
mattallen37 wrote:It sounds like a pullup issue. Three 82k's in parallel is about 27k. That may be too strong. The NXT "protection" resistors (4k7) in series with the IO lines, and the pullups, form a voltage divider when the NXT sends a LOW. The stronger the pullups, the higher the "LOW" signal the sensor will see. Probably the strongest pullups you can use with the NXTMMX is something between 27k and 41k. This would explain why two work, but not three.
Can you configure your sensors differently? Can you have two NXTMMX's on one port, and have one on another (with an analog sensor connected to the through port)? You said you were using a Lego color sensor, but what other sensors are you using?
mattallen37 wrote:I think it is just a through port, so you should be able to use almost any sensor on that port. Even though the ultrasonic sensor is weird, I think it should still work. What sensors do you need on the robot?
Users browsing this forum: Baidu [Spider] and 1 guest