Hi,

This is just a general brainstorming. I'm not sure I'll do it, it all depends on how complex it gets, and how destructive it gets for the Lawnmower in question.

Anyway, I got my hands on an robot-mower of exotic make (there is no make visible anywhere AFAIK), and got permission (and encouragement) to "hack" it from the owner. As it's native program is pretty crap, it always looses track of it's "virtual" wires.

Anyway, I have no idea if I can pull it off, but I love to give it a try , mostly for fun, if it cuts grass afterwards, well , that's a bonus.

What I like to try is to see, if i can replace the program in the lawnmower with a homebrew program. Mostly as a fun/learning kind of thing. I don't want to destroy the mower, but I was explicitly told, if that happens, well, no biggy.

I have experience with PIC and AVR programming, so I know how to work with MCU's, allthough I am by no means an expert.

I opened the case, and it contains a STC5410AD microcontroller on a circuitboard, in a lqfp-32 socket.

Now this is the thing I am trying to figure out. I see 3 options.

1. Get a programmer for the STC5410AD. I partly like the idea, but then again it's not really the kind of MCU I am fan of, containing 512 bytes of sram. I was aiming a few KB (1 at the least). But then again it has 10kb of flash, which is ok. I found a site here on how to program such beasts. http://grauonline.de/wordpress/?p=44

2. Somehow (without soldering to the mowers circuit), attach leads to the lqfp-32 socket, attach the leads to a second circuit, and put my own MCU (I have some spare Atmegas laying around). This I like, as I have a development enironment for AVRs thats working just fine, and I am comfortable with. However how to I get a "inverse" lqfp-32 adapter, I have no idea.

3. Totally bypass the circuit on the mower, and replace with my own circuit. This one feels a bit like a last resort. This way I need to redo the motor drivers, battery loading circuit, and whatnot.

4. Perhaps I should see if there are (semi) pin compatible AVR's out there as well. I haven't looked yet, well, we'll see....


Any suggestions, ideas, experience in this type of thing? I like option 2 most, but I don't really see how....


Thanks
-DaC

 

No comments from anyone?

 

Can you post a photo of the mower and another of the circuit board?

 

If everything is built into one PCB, then building your own controller would be a better idea, unless you want to spend some time probing the circuit to figure out what each connection to the main uC is doing at what point......

But pictures would help a lot in helping you make that assessment....

 

Ok, pictures coming.. First one of the complete mower. A few circuit pictures afterwards.

 

 

it always looses track of it's "virtual" wires.

Wired or wireless border?

No where on the mower is any type of identification? Yipes.

joe

 

The circuit

The circuit from another angle, to be able to read the motor drivers

1 of the two daughterboards. (funny-enough called robot2 and robot3)
This seems to be the wire detector.

 

Trying to downscale the pictures... hold on...

 

Wired or wireless border?

No where on the mower is any type of identification? Yipes.

joe

You need to put a wire around the lawn, to act as a border. I posted a picture, with what I think are 2 of the four onboard "wire-detectors".

Not much identification on the whole robot... It's pretty much green-generic, when the case is on.. Directly from China, before any "minting" I suspect.

 

If everything is built into one PCB, then building your own controller would be a better idea, unless you want to spend some time probing the circuit to figure out what each connection to the main uC is doing at what point......

But pictures would help a lot in helping you make that assessment....

Ok.... I was hoping for replacing the MCU, with some sort of "plug", that fits in the MCU socket, and this plug would help me connect the circuit board to mine.. But mabe that was a naive idea, as I never heard of such a adapter plug, that plugs in a mcu or processor socket.

 

Ok, sorry for that temporary mess of huge pictures... I think its fine again now, at least from my browser.

So, replacing the PCB has been coined. Replace with blank, and build, or do something arduino based is then a question I have on my mind.

 

Ok.... But mabe that was a naive idea, as I never heard of such a adapter plug, that plugs in a mcu or processor socket.

They sometimes refer to them as "Emulator" adapters, like these >> http://www.emulation.com/catalog/of...mulator_adapters/wired_one-to-one-socketable/

BUT, replacing the whole control board may not be a bad idea, since they only really have a couple of H-Bridges to control the drive motors, probably a mosfet to turn on both cutting blades (I see 2 motors for the blades), maybe a couple of analog channels for each "wire" detector, (Which I believe are set up as some type of pickup coil to detect a certain frequency emitted by a controller box connected to the ends of the buried wire), or they could just have a digital output that toggles high or low, etc... so, for the "Fun", I would build a whole new controller board for whatever uC you decide to use, but salvage as much of the through hole components to reuse (such as the H-Bridges, but then again those seem like they need a lot of external components compared to an LMD18201T H bridge) and mosfets....... unless of course the battery charging circuit is also part of the main board, which would complicate things a bit depending on the type of battery used....... so these could all be points to ponder in helping you decide the right route.....


How many wires go to each "wire" detector daughter boards?? just curious in if they have it setup as a dual detector with 2 outputs to let you know which way the "wire" is oriented under the sensor.....

 

it always looses track of it's "virtual" wires.

I'm still working on determining the problem here.

Have you (or the owner) checked the buried wires? Tried it with border wire on the surface? Whatever is powering the border wire is functioning properly?

Rebuilding/redesigning is undoubtedly fun. Handing the scarecrow a courage badge and the lion a ticking watch won't make the tin man happy. Forget about Dorothy and her little dog, for the moment.

joe

 

Oh no!! Please don't let this topic fizzle and die!!! Chaos51, where are you?

joe

 

Right now I am waiting to get the power supply for the "virtual-wire", so I can do some tests before I start modifying anything.

Anyway you had a point with your comment to first determine the exact problem.


But then again modifying it will be alot of fun (not just for me, the owner is quite eager as well)... I think it's not so much about fixing the mower, as a mower is just a mower, as much as the challenge to hack it. As I said, if it comes out mowing in the end, that is a definite bonus. So handing the scarecrow a courage badge may be in store after all ;o)

Anyway, I am in waiting status just now.... Probably during after the weekend the owner will hand me the power supply for the virtual wires, and after that we'll test, and after that, I pick up this thread again, as I am still undecided which road to take.

 

I’m very glad you’re keeping this thread cooking!

I’ve been working on just such a project, converted electric wheelchair drive train strapped on a mower deck. Examples are all over the internet. But they are like a pinball, flipping and bouncing from one obstacle to another, mowing in a random pattern or it’s operated from an RC car transmitter. Yuk to both.

Bosch introduced a row-to-row robot mower:
http://www.cisionwire.se/bosch-verktyg/r/snabb--logisk-och-effektiv-grasklippning,c9280113

Check out the video:
http://spectrum.ieee.org/automaton/robotics/home-robots/bosch-introduces-new-autonomous-robotic-lawnmower/

Using four IR transmitters each sending unique data streams, I have a “robot vacuum cleaner” sweeping in nearly perfect rows – inside the shop. Outdoor operation is a whole ‘nother ball of issues. ‘Course, IR won’t do, GPS is not accurate (down to half width of the mower deck or wheels,) triangulation/trilateration with RF or laser (as in surveyors positioning equipment) is enough to drive a hobbyist insane. It’s not the trilateration math that makes me crazy – I have it working great on a PIC, it’s the positioning medium and getting it to safely and reliably work outdoors. Due to my disability (acquired since starting this project) and the positioning details, the project is more or less suspended. In other words, you could say I’m the third flying monkey from the far end.

I’d be interested to know if the buried wire system is really functional. Apparently, the protective jacket fails and the wiring corrodes too easily, or the circuit is a lightning rod, or the controller poops out, or the cuttings form an insulating barrier to the detector and some of the detectors do not correctly (accurately) detect perpendicular approaches to the wire. (Always something!)

Please let us know!



joe

 

I've always wanted to build one of these on my own. Having the hardware already done for you is a big step forward. I hope you can keep this project going.

I think the first step would be to map out all the wires, sensors, motors, connectors etc and figure out the function of each component.

Next would be to decide what can be disconnected and discarded, what to keep and what to replace. I think in the long run, without original schematics, it would be easier to start from scratch and build your own controller.

How does the mover cut the grass? Is it a rotating blade? I would have chosen a scissor-style action such as in a hedge trimmer for less power demand and less noise.

 

Ok, this thread is not dead yet, I am just waiting for the power supply. I'll get it tomorrow... So maybe I post some as-is test results... In order to make clear where I am and where I want it to go. (this little project)

In the meantime I backwards engineer some of my other projects... Poof that takes more then 5 minutes to go from a spagetti wire (spagetti made out of papperclips) circuit to a circuit diagram, not even talking about doing it electronically yet.

 

They sometimes refer to them as "Emulator" adapters, like these >> http://www.emulation.com/catalog/of...mulator_adapters/wired_one-to-one-socketable/

BUT, replacing the whole control board may not be a bad idea, since they only really have a couple of H-Bridges to control the drive motors, probably a mosfet to turn on both cutting blades (I see 2 motors for the blades), maybe a couple of analog channels for each "wire" detector, (Which I believe are set up as some type of pickup coil to detect a certain frequency emitted by a controller box connected to the ends of the buried wire), or they could just have a digital output that toggles high or low, etc... so, for the "Fun", I would build a whole new controller board for whatever uC you decide to use, but salvage as much of the through hole components to reuse (such as the H-Bridges, but then again those seem like they need a lot of external components compared to an LMD18201T H bridge) and mosfets....... unless of course the battery charging circuit is also part of the main board, which would complicate things a bit depending on the type of battery used....... so these could all be points to ponder in helping you decide the right route.....


How many wires go to each "wire" detector daughter boards?? just curious in if they have it setup as a dual detector with 2 outputs to let you know which way the "wire" is oriented under the sensor.....

>>>How many wires go to each "wire" detector daughter boards??

It looks like that is 4 wires. Maybe 1 GND, one +V, and 2 Signal wires coming back...