Sunday, September 11, 2011

Neato XV-11 LIDAR Disassembled... COMPLETELY!

Someone had to do it...  Might as well be me since my Neato XV-11 was paid for by the hacking contest bounty.






LIDAR board pressed into Play-Doh and ready to go under microscope


 My wives microscope I got her for Christmas...  Or Birthday...  Not really sure, just glad I did!  :)


 Back side of lens assembly, I was amazed there are spots on the filter there, must not affect performance that much.


 Same thing from a different angle.  That flat black plastic piece pointing out above the reflective filter looks like it was scraped with something.  Those are little plastic shavings still loosely attached.  **EDIT** Thinking this over a little more what I believe Neato has done is make this opening fit the Panavision sensor EXACTLY.  This ensures that even if the sensor is at a slight angle on the PCB, it will still be centered behind the lens.  This is probably press fit into the plastic and then screwed/glued together.


 Bottom of lens assembly that is screwed to the PCB below.  Looks like Neato included some spots for set screws in the bottom that were not needed under the lens and laser module.


Board that the LIDAR mounts too with optical sensor on the bottom for rotation position.


11 comments:

  1. you ROCK!

    Meanwhile... I *knew* that thing was a line scan imager!

    Some of those sensors run crazy fast frame rates, enough to support the rotational rate of the XV-11.

    ReplyDelete
  2. Yea, if you look at the datasheet for this sensor it looks like it can run a HELL of a lot faster than Neato is using it... I might try to tap into the config lines to see how Neato is setting the configuration registers.

    ReplyDelete
  3. Very nice tear down!
    Can you get any information on any of the dimensions? Primarily the spacing and angle(s) of the optical center lines of the laser and optics?
    Any information on the optics would be cool as well but it would probably be a tad harder to back track out the focal length and F number.

    ReplyDelete
  4. zimme,

    I'll see what I can come up with for optics... Not my strong point but I would like to learn! I'm tempted to remove the lens and everything but don't want to make it IMPOSSIBLE to use the thing again.

    Maybe I'll get more confidence if I figure out how to recalibrate the thing!

    ReplyDelete
  5. Hash,
    A question not strictly on the optics, but on the neato. As you know, the Neato senses the boundary magnetic strips and will not cross these boundaries. I presume it has some sort of Hall effect sensors in the front near the corners.

    Do you know what type of sensor it is, where exactly it is fitted in the Neato and how it is connected to the controller?

    I have been thinking that it may be possible to improve the "near field" detection and avoidance capability of the Neato by fitting more sensors closer to the front corners (and even the rear) and then fit tiny magnets to thinks that I don't want the Neat to bump into.

    For example, I have some chairs with think wooden legs and the Neato has trouble "seeing" these and proposal is to drill holes in the legs and fit magnets inside the wooden legs. With the right sensors, the neato will then sense the magnetic field and avoid the chair leg.

    What do you think?

    ReplyDelete
  6. Hash,
    One reason Neato might be running it at the speed they are is that the slip ring that the data from the LRF goes thru is only rated for 300rpm. It would seem that they are pushing the limits of that connection already. Has anybody seen a post where the LRF is turned on its side and operated?

    ReplyDelete
  7. Johan,

    That is a tough issue and interesting solution to the problem you have come up with! You may consider an easier fix I came across while setting up a test area for my Neato.

    Take some painters tape as an example and run it around the legs of a single chair. Only run one strip but run it parallel to the floor and at the same height as the laser scanner on the top of the Neato. To the Neato this will look like a wall since the laser will hit the tape.

    You can use some decorative cloth or anything other than tape, decorate your chairs and solve the problem at the same time!

    I appreciate the idea you have using the magnets, and it would be possible to tap into the Neato's sensors, but I believe you would have a lot of false readings and the Neato still bumping into your chair.

    ReplyDelete
  8. Shocktube,

    That looks to be the case. There is no datasheet for this particular slipring on the manufacturers website, but the LPM-04A specs are for 0 - 300 RPM.

    The whole unit rides on a rather large bearing so there is no actual load on the slipring itself. I have operated it in different orientations without noticing any problems. If you were to operate for a long period of time with the LRF turned on its side perhaps bad things would happen since it was not designed with this in mind most likely.

    ReplyDelete
  9. What kind of laser is being used in this? Im interested to know if you can crank up the wattage of the laser and get more range out of it.

    It 'd be great if someone threw Eagle files on the web of this thing...maybe it'll get people finally to start making them for the masses...

    ReplyDelete
  10. Tamir, The range is limited more by how the design works than the power output of the laser. I just posted schematics recently, and I think I may do a PCB layout in the future as well.

    As you said, if we can make it for the masses at a $99 price point I think a lot more projects would be using laser scanners!

    ReplyDelete
  11. Hi,

    Great work !!
    I wonder if you can add side brush to the neato (like ROOMBA Irobot)
    for example, taking a small computer fan engine, connect it to little brushes (that can bought) and connect it to 5v power so it will start cycling when the neato starts.

    what do you think ?

    Eldad

    ReplyDelete