AutoBot

Discussions around autonomous, semi-automatic and intelligent robots and systems.

Moderators: BeligerAnt, petec, administrator

User avatar
Lincoln
Posts: 128
Joined: Sun Feb 29, 2004 12:00 am
Location: Olney, Milton keynes

Re: AutoBot

Post by Lincoln »

the pcbs have arrived!
waiting on some components from mouser and then i shall solder one pcb to start with over my easter holidays. i've ordered components for 2 but had 13 made by pcbway, (i ordered 10????)

so the hearwere of the robot itself is now being designed. i've ordered some 25D gearmotors from pololu and 12mm hex adapters to be able to use wheels off of a 1/10th scale rc car.
i shall probably use "open beam" (15mmx15mmm extruded aluminum) to construct much of the robot. im attempting to make suspension, maybe even using pneumatic pistons instead of shock absorbers.
my dads work had a big clear out and i ended up with 2, motorized 2 axis swiveling cctv cameras. which im removing the electronics and camera from to have a motorized gimbal type thing on top of the robot so it can "look" around without turning the whole robot.

i still haven't thought of a perpous for this robot. but we are going to attach a geiger counter so guess it could be used for exploring radiation levels in different areas
you should be able to view some picture here https://photos.app.goo.gl/iA0UkrSumJv5aYq62
Team RobotMad, home of the Smart robots, and very mean pots :)
Chris and Lincoln Barnes
User avatar
Lincoln
Posts: 128
Joined: Sun Feb 29, 2004 12:00 am
Location: Olney, Milton keynes

Re: AutoBot

Post by Lincoln »

urrrm oops. kinda forgot to update this page.

well we made a prototype on a piece of wood, named plank-bot. put range sensors and neopixels on and got it to drive to stay a set distance from anything in front of it.this worked realy well.
next we tried hot air soldering the large motor driver chips on to a pcb, as before it was using a external motor driver board. this took some, practice, as we melted 2 pcbs before getting it right. we also have encoders on the motors going to a fpga that feeds the data back to the main board, but in adjusting the equations in the software it is now not as good as before so we need to fix that. other than that we have just started working on driving the cctv camera mechanism. and the cad for the proper version for the robot, thats not just a piece of wood, is coming along nicely.
Team RobotMad, home of the Smart robots, and very mean pots :)
Chris and Lincoln Barnes
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: AutoBot

Post by MarkR »

That is brilliant.

I'm not sure whether you really need the fpga for shaft encoders, unless they're running really fast...

Right now I've got a working antweight flipper with Raspberry Pi + IMU, just waiting for me to get my lazy arse on and write the software.

In other news, there is a new laser range finder, the VL53L1X.

I bought some VL530X laser rangers last year, to test, and they are unbelievably awesome. But the new one is better!

https://www.pololu.com/product/3415

Max range 400cm!

This chip is so tiny that it needs a breakout board to be easily usable.

Also if you've used the ubiquitous sonar rangers, these laser things are at least 100x better in most ways. The chip inside must be incredibly clever, as according to my calculations, it's measuring differences in timing of less than one nanosecond.
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
User avatar
Lincoln
Posts: 128
Joined: Sun Feb 29, 2004 12:00 am
Location: Olney, Milton keynes

Re: AutoBot

Post by Lincoln »

we have some of the short range and the 1m range versions of theses. but those 4m range ones look good. currently the plankbot uses 2 of the 1m rage ones at the moment.

we use a fpga for the encoders, mostly because the main processor dose not have enough spare io pins that we haven't already used. and it is pretty fast and the encoder is on the motor side of the gearbox so would be possible with the processor but difficult,so a fpga was just an easier solution.
Team RobotMad, home of the Smart robots, and very mean pots :)
Chris and Lincoln Barnes
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: AutoBot

Post by MarkR »

Sorry to hijack your thread, but what do you think about this new H-bridge chip?

TB67H420FTG

It is tiny and powerful, but difficult to solder. I also don't know where to buy them.
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
User avatar
Lincoln
Posts: 128
Joined: Sun Feb 29, 2004 12:00 am
Location: Olney, Milton keynes

Re: AutoBot

Post by Lincoln »

looks pretty good. maybe if i make another antwight scale pcb then ill use that chip but it dose indeed seem difficult to get. but there is a pololu board available with it on.
i would say that the 10V minimum input would mean you have no choice but to use at least 3 cells rather than 2.
Team RobotMad, home of the Smart robots, and very mean pots :)
Chris and Lincoln Barnes
User avatar
Lincoln
Posts: 128
Joined: Sun Feb 29, 2004 12:00 am
Location: Olney, Milton keynes

Re: AutoBot

Post by Lincoln »

note to self. when probing the motor driver pins, be EXTREMELY careful.

all was going well, we were testing the new motor driving algorithm on the driveway, then we decided to put it in the grass, it doesn't move, seems the motor gearing is to low to have enough power to go though grass. which means getting new gearboxes. but this test also burnt out one of the motors so we have to replace than as well. then we notice that one of the motors would work off of a power supply but not off the motor driver. so we were probing around it, decided that the mosfet that powers it was maybe dead so replaced it no issues. but the motor still wouldn't turn. continue probing the pins.
OH now the processor has stopped, and isn't programming. ah whys it pulling an amp and is boiling hot.
turns out that a signal wire to the processor (3v) and the charge pump (15v)for the mosfet are right next to each other! so we must have shorted these two pins together when probing around. it turns out processors don't like 15v.

expensive day of testing that was. need a new processor, probably a new motor driver, a new motor, and 4 new gearboxes.
Team RobotMad, home of the Smart robots, and very mean pots :)
Chris and Lincoln Barnes
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: AutoBot

Post by MarkR »

Unlucky.

Something probably a bit less technical but similar happened to my Raspberry Pi antweight. It seems that the Raspberry Pi Zero W does not like 18v on its 5v input, it made it rather unhappy. I don't know yet whether the sd card survived, I hope so.
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
User avatar
Lincoln
Posts: 128
Joined: Sun Feb 29, 2004 12:00 am
Location: Olney, Milton keynes

Re: AutoBot

Post by Lincoln »

right, we finally put some time into fixing PlankBot the AutoBot prototype and designing AutoBot.
new cpu on the board done all working, new motor driver chip, massive pain to solder because of the large pads on the bottom that mean we have to hot air solder it. which when only doing one chip in a row of 4 means they all move and basically is just a massive pain, then just when its done the stuff falls off the other side of the board because hot air soldering is such a joy.
so eventually that was fixed.
we bough a cheep Chinese version of the motors we were using before with 47:1 gearboxes instead of 9.7:1.
ordered all the parts for the new AutoBot.
lots of 15 mm square extrusion aluminium called OpenBeam, some 3 mm steal bar, some ball bearings. also got some 1/8th scale nitro car shocks from Ebay for suspension.
AutoBot2-min.png
AutoBot2-min.png (37.39 KiB) Viewed 83197 times
AutoBot2-min.png
AutoBot2-min.png (37.39 KiB) Viewed 83197 times
AutoBot2-min.png
AutoBot2-min.png (37.39 KiB) Viewed 83197 times
Attachments
AutoBot4-min.png
AutoBot4-min.png (88.22 KiB) Viewed 83197 times
AutoBot3-min.png
AutoBot3-min.png (43.48 KiB) Viewed 83197 times
Post Reply