servo reliabillity

bitrootvz

I am a tech teacher in Idaho and my students and I have been working with the BOE and Boe-Bots for about 2 months. We are at the point where we are to log the durations and cout the turns to see the sequential changes. The problem is that the servos don't sequentally turn. you can run the same duration and get a different amount of rotation or you might get more rotation from one duration and less from a larger or smaller duration. Is this normal? we have tried both AC supplies as well as battery pack. Yes we used brand new batteries and moved the jumper block.

P!-Ro

I'm wondering if the potentiometer needs to be adjusted... try adjusting it and if it still doesn't work, double check your programming. If nothing else works, you could contact the parallax customer service and see if they can help. I used their service once, and they were very helpful.

By the way, Idaho rocks!!!

Mike Green

Servos are not very accurate, particularly when they've been modified to create a continuous motion servo. The feedback mechanism used to measure the position of the servo shaft is disabled. As I LIKE PI mentioned, the adjustment potentiometer may be off or may have a noisy or bad wiper connection and readjusting it may help. Follow the instructions in the Robotics with the BoeBot manual.

bitrootvz

Thanks for getting back to us. The problem is all four of our Boe-Bots have the same issue. when we type in 750 the servos don't move. this tells me that they are adjusted correctly. if we type in 650 it might turn 5.75 rotations and then if we type 650 again it might turn 5.5 rotations. 655 might turn 5.25 and then 660 might turn 6 times. I'm wondering if it is the quality of the voltage regulator.

Mike Green

It's not the quality of the voltage regulator. Servos do take a bit of current. Peak draw is on the order of 1 Amp while average current under light load is maybe 250 mA. The regulator is good for about 1A although that is limited by the unregulated supply, both current and voltage. Any excess voltage (over 5V) has to be dissipated as heat and there's no real heat sink attached to the regulator. If you're running the BoeBot off batteries or a 6V to 7.5V AC adapter, heat shouldn't be a problem. If you're running the BoeBot off batteries or a 6V AC adapter, you should set the servo supply voltage jumper to Vin rather than Vdd and that will run the servos directly off the battery or 6V AC adapter. If you're using an AC adapter with more than 7.2 to 7.5V, you can't do that since the higher voltage will shorten the life of the servo motor's brushes.

You simply cannot rely on continuous motion servos to have closely reproducible speeds. You're talking about maybe 5% to 10% differences and there's too much variation in mechanical load, wheel slippage, in temperature, in all sorts of factors to do much better than that. They're simply not designed for that kind of use. Parallax sells wheel encoder kits that mount on the BoeBot chassis and use optical sensing and the slots in the wheels to provide speed and position feedback to the Stamp.

bitrootvz

My class just left and we had one boe-bot that seemd to be working correctly. the other two still seem to run at about the same amount from 650 to 700. then they will reduce down a little but still not like they should. if we input 75 they stop. one of them will turn clock wise about an 1/8 of a turn at 749 but won't move at all at 751. Itried to re center them but no luck. Just an FYI but we had problems with one boe-Bot where it would run randomly on 13 and 15 when turned on and it turned out to be an AC power supply. when we switched to a different supply or batteries it stoped.

Mike Green

That's not a surprise. AC power supplies can vary in terms of how much current they can supply and the voltage at which they'll supply a given amount of current. Their output may vary depending on their temperature which can vary depending on the current drain over time, etc.

bitrootvz

I just went back to the two that were not working and recenterd the servos again and I think they may be working correctly now. thanks for the help.

erco

Per Mike, servos don't operate 100% consistently, and BoE-Bot's differential drive introduces a lot of variables that change constantly. Some people have had improved results by adding two wheel encoders (Wheel Watcher is one example). Of course that adds complexity to your program, weaving the monitoring of two encoders into your servo pulsout loops.

IMHO, accurate repeatable dead reckoning with a differential drive robot is one of the most difficult, yet most fundamental tasks in robotics. Of course, if it was easy, everyone would be doing it!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·"If you build it, they will come."

Whit

erco said...


IMHO, accurate repeatable dead reckoning with a differential drive robot is one of the most difficult, yet most fundamental tasks in robotics. Of course, if it was easy, everyone would be doing it!

LOL! This problem does come up a lot doesn't it?

bitrootvz, glad you got your problem worked out.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Whit+


"We keep moving forward, opening new doors, and doing new things, because we're curious and curiosity keeps leading us down new paths." - Walt Disney