UPDATE: Rocka Pics! Suggestions for SMALL object detection sensors and drive motors

WBA ConsultingWBA Consulting Posts: 2,918
edited 2013-05-02 - 23:39:06 in Robotics
I am working on a new low-cost and small robot platform (named Rocka) that, for sake of discussion, can be visualized as a miniature Stingray. (~4" wide)


My original plans were to outfit it's face with a set of Ping sensors or Sharp IR sensors for object detection. However, as tiny as the Pings/Sharps are, they would be driving the size of the robot larger than my intentions. I have looked at the Ping kit with the Servo to use one ping, but the standard size servo would be taller than my goal. I could mount a ping on top of a micro servo, but the videos I have seen give me the impression that having to swing a ping around slows down your robot. (maybe I am being mis-lead by the videos?) I haven't got to the point of adding the IR sensors to my daughter's BOE-Bot, so I don't know how well that truly works as compared to Pings. (My Stingray can travel around the house all day without hitting a single thing or getting itself stuck, so I am impressed with the Pings)

My original plan for drive motors was to modify some micro sized servos for continuous rotation. After searching a bit online, it appeared that the Naro servo that Parallax sells could not be modified, but I of course had to try it myself. The process went rather smooth, but apparently the pot feedback is necessary for the servo to function, so I killed the one I had. I have found several other micro servos, but cannot find any proof of anyone accomplishing this. From what I have found, it seems to be due to people using small gearmotors and H-Bridges instead for this small of a platform. I am trying to avoid an H-Bridge due to the extra real estate it would require on my platform. I have thought of using a mini breadboard and setting up an L293 H-Bridge on it, but my goal is a "simple" solution. When looking at gearmotors, the prices/styles range all over the place and whenever I find a low cost setup, they require mods or are too large. I can use a standard size servo, but it kills my robots real estate.

Any suggestions for small object detection solutions or drive motors for this size of robot?
436 x 402 - 15K


  • kwinnkwinn Posts: 8,612
    edited 2013-03-01 - 13:26:23
    Perhaps you could use one ultrasonic transmitter and two receivers in fixed positions to emulate bat sonar. The prop should be fast enough to detect the time difference between the received signals. Not sure if the ping parts would be suitable though.
  • ZootZoot Posts: 2,227
    edited 2013-03-01 - 13:54:43
    Devantech sells a tiny sonar unit, the SRF10:


    Otherwise you might want to use a single transducer solution.

    For motors, SolarBotics sells lots of tiny, tiny gearmotors:
  • TtailspinTtailspin Posts: 1,326
    edited 2013-03-01 - 18:47:02
    Try the IR sensors, they are small, and can detect stuff just as well as the Ping(((
    It looks like you have some room on the front of Rocka to hang a Ping(((,
    Then you could have both IR and Sonic detection, and still have room for the Rocket Launching System...

  • Duane DegnDuane Degn Posts: 10,290
    edited 2013-03-01 - 19:58:45
    I know erco has converted one HobbyKing's HXT900 into a CR servo.

    The problem with the micro servo is the pot is also an axle. You'd need to add a secondary pot to adjust the zero point. I've taken about several of these servos myself and I've be concerned about the strength of axle/pot to hold a wheel but I'm not sure it would be a bad idea.

    I've used a bunch of HXT900 servos in various projects and they've worked very well (and they're nice and cheap). I intend to try them as CR servos sometime.
  • W9GFOW9GFO Posts: 4,004
    edited 2013-03-01 - 20:12:22
    These 120:1 gear motors are cheap and work well for a small robot.

    168 x 200 - 6K
  • rwgast_logicdesignrwgast_logicdesign Posts: 1,464
    edited 2013-03-01 - 21:23:17
    Take a look at the Dagu Black Duck Motor, it may fit the bill. Its a preconverted micro servo, i think its a micro not 100%
  • al1970al1970 Posts: 64
    edited 2013-03-01 - 22:51:54

    I have converted a micro servo to continuous rotation. It is a pain but can be done.

    1) Pick a micro servo that has a full gear. Some only have a 1/2 gear.

    2) You have to cut off the small metal part on the pot that stops it from going around 360. You have to be careful since the pot is also used as an axle.

    3) Cut the wires from the pot and solder them to two resistors (in my case they were 4.7k). In the servo I used there was no room inside the servo so I had to glue them to the outside of the case.

    I don't know why people think they need to keep the pot in the servo. Two resistors work fine and most of the time they keep the motor at a null point. Even if the motor does move a little you can always null the motor in software.

  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-02 - 22:55:15
    Thanks for all the feedback. To sum up my further research based upon these posts, here is where I am:

    Object Detection:
    Looks like I will be trying simple IR like used for the BOE Bot. I just can't seem to find any ultrasonic or IR modules that fit my size goal.

    Drive Motors:
    From what I have found, I am going to be taking apart the Naro servo again to resolve a few "I wonder if" ideas that have come up. I will also be picking up some of those HXT900 servos and following this set of instructions. I would prefer to use W9GFO's mentioned gearmotors as I was already looking at them, but I was trying to avoid an H bridge. I will keep searching for a dirt cheap H Bridge setup though.

    Thanks to everyone for all of the posts, They have given me much more knowledge into small robot design.

    I am still finalizing the chassis drawing, but the prototype will probably be made from 0.2" thick MDF. However, I am looking at some 1/16" thick polystrene as I have recently learned some clever tricks for interlocking perpendicular joints of very thin material.
  • Duane DegnDuane Degn Posts: 10,290
    edited 2013-03-02 - 23:07:29
    I will also be picking up some of those HXT900 servos and following this set of instructions.

    You might want to also read erco's notes about his conversion. He'll be glad to know someone else cares.
  • Duane DegnDuane Degn Posts: 10,290
    edited 2013-03-02 - 23:15:27
    You might want to consider converting HobbyKing's metal gear micro servo to CR. It looks a lot stronger and it is pretty reasonably priced.

    It might be easier to attach a wheel to the metal final gear than the HXT900's plastic final gear. I haven't used these servos myself, but they look pretty good for the price. I'm not sure if the metal gear servo has a full final gear or not. I think the odds are high it's a full gear since it normally has a full 180 degrees of rotation.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-03 - 00:16:15
    Duane, thanks for the extra link. That one may be a better option. I am going for low cost, but also for repeatability. I want others to be able to repeat what I do with mine. That is one reason I am avoiding one-shot eBay deals, etc.

    I took apart the naro servo again and see what I need to do to finish the conversion. I know have the pot disconnected from the PCB. The lead spacing on the pot appears to be perfect size to squeeze in an 0603 resistor. Will have to try that. The naro servos aren't cheap though, but at least I know a bit more about what I am doing. I also keep going back to the cheap H-Bridge and $5 gearmotor solution. I did also just find this one at Sparkfun. According to the drawing, the size would be excellent and it is already continuous rotation. I think I missed it before because it is called a medium servo, which I interpreted as "standard" so I ignored it.
  • W9GFOW9GFO Posts: 4,004
    edited 2013-03-03 - 00:55:06
    ... but I was trying to avoid an H bridge. I will keep searching for a dirt cheap H Bridge setup though.

    Not sure what dirt cheap is but one of these will drive two motors for $2, or just $1 in large quantitities.
  • rwgast_logicdesignrwgast_logicdesign Posts: 1,464
    edited 2013-03-03 - 13:12:11
    As far as cheap and small where an h bridge is concerned, have you thought about either doing a diy transistor design or just using an l293d/e depending on current requirements. Both of these options are small and cheap, can easily be implemented by anyone who can solder and follow schematics.

    If you went that route you could use a dual gearbox from tamiya that comes with two motors for 11 dollars.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-03 - 14:24:12
    What I mean by dirt cheap comes from the cheapest h-bridge PCB setup that I have found was $15. Using the L293 would be my plan, but I can't see paying that much for something that breaks out a $2 chip. I really think that I should just brew my own or maybe have that be an option for my clean quick start layout I have going on another thread.
  • W9GFOW9GFO Posts: 4,004
    edited 2013-03-03 - 14:46:44
    Using the L293 would be my plan,

    What would be the benefit of using the L293 over the newer SN754410NE?[h=1][/h]
  • rwgast_logicdesignrwgast_logicdesign Posts: 1,464
    edited 2013-03-03 - 15:39:56
    If the motors are small enough, the l293 has flyback diodes, while the sn754410 does not but can handle more current. The flyback diodes built in to the chip just save that much more space.

    I definitely meant just throwing the chip on a PCB or protoboard not buying an overpriced module, the chip is very easy to connect with proper schematics anyone that can solder is capable of repeating the board
  • W9GFOW9GFO Posts: 4,004
    edited 2013-03-03 - 20:34:27
    Only the L293D has internal diodes. From what I have read external diodes are a good idea even when the chip has internal flyback diodes.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-03 - 21:34:00
    I should correct myself in that I meant that I may end up using an H-Bridge IC, so either the L293 or the SN754410. Either way, I think using one of them will be the cheapest way to go overall and a piece of prototyping board is all I need to make up the circuit. I still think making a tiny breakout PCB will also be my solution. Long term, I will revisit the idea of taking my clean QuickStart layout and adding the appropriate "robot" hardware to make a small, inexpensive robot board that is of the QuickStart form factor.
  • Martin_HMartin_H Posts: 4,050
    edited 2013-03-04 - 05:37:31
    Andrew, a company called Phidgets makes sensors for small robots. Solarbotics carries them and they might be worth a look:

  • ercoerco Posts: 19,779
    edited 2013-03-04 - 11:17:39
    You might want to have a look at the old TAB build a robot kit, if you haven't seen it. The PCB was the chassis, held DC motors and IR sensors and had an onboard PIC chip with several simple "behaviors". It also had a blank socket where you could plug in a BASIC Stamp 2 if you wanted to program it yourself.

  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-04 - 12:19:40
    Thanks, I'll check the phidgets out. Erco, funny, I forgot about those. I had one a few years ago that I sold on eBay. Nicely laid out platform to keep costs down. I may see if I can utilize some of those ideas.

    I did finally find an inexpensive motor driver at Pololu: http://www.pololu.com/catalog/product/2135

    I will be moving forward with modified Naro servos for my prototype though. Once complete, I will make a version for the 2135 motor driver.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-07 - 20:37:22
    SUCCESS!!! Parallax's Naro micro servos CAN be modified for continuous rotation! I just resurrected the one I mentioned killing in the first post by adding 2 SMT resistors onto the servo's control PCB between the points where the pot was originally attached. Since I already destroyed the original pot, I was unsure of its value, so I used two 10k resistors in a 1206 package. They fit perfectly between the pot terminals. After re-assembling the servo (which is becoming very easy for me) and connecting it to my PropBOE with test code, it works! I will be doing a couple more and making up a fresh thread for the steps I used.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-07 - 22:28:04
    I updated my sketch for my layout with the Naro servos. Width is down to 3.5" now. I am nearly finished with the drawing for the chassis for it to be cut on the laser, but will be adding some features to give me some options for mounting the servos either direction (hub closer to the front or rear) depending on where the weight balance truly ends up.

    668 x 420 - 18K
  • Duane DegnDuane Degn Posts: 10,290
    edited 2013-03-08 - 13:50:47
    You could reduce the size even more by using these little Nordic modules. The change would also cut you costs by a big chunk.

    The Nordic modules aren't as simple to use as XBees but they sure cost a lot less. (Note the link is to a batch of ten modules which still cost less than a single XBee.)
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-12 - 01:24:04
    Excellent point, those Nordic modules would take up much less space. I will definitely look into them. I think my second version will be based upon an H bridge an DC motors, so with the Nordic module, the overall robot won't have to be squeezed into the form factor.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-12 - 01:43:58
    Made some headway tonight with the prototype. Uses the following for this setup:

    2 GWS Naro micro servos modified for continuous rotation
    My PropBSC module
    XBee Module
    Miniature Breadboard
    IR Transmitter/Receiver Kit
    Sparkfun LiPo battery
    Lasercut MDF chasss (it's actually chalkboard precoated 0.2" MDF. Cuts and etches really nice)

    some pics are attached............still need to figure out my final configuration and whether or not I can get that pin on there.
    448 x 252 - 44K
    448 x 252 - 42K
    448 x 252 - 43K
    448 x 252 - 41K
    448 x 252 - 40K
  • mindrobotsmindrobots Posts: 6,506
    edited 2013-03-12 - 03:52:38

    These are turning out really neat!

    The bottome deck is motors and battery and the upper deck is brains and sensors?

    I can see a swarm of these guys in the future!

    Well done!
  • ZootZoot Posts: 2,227
    edited 2013-03-12 - 08:17:39
    That's adorable. I love it.
  • WBA ConsultingWBA Consulting Posts: 2,918
    edited 2013-03-12 - 10:02:53
    mindrobots: Yes, in the configuration that I am doing now, the servos and battery will be part of the lower deck with a little bit of room for a few other things if needed. The top deck will house the breadboard and XBee, so as shown the brain and sensors. I was also playing around with the idea of an upside down PCB setup where a piece of prototyping board would have socket headers for the PropBSC and the XBee on its backside and then be mounted upside down to the bottom of the upper deck. This would allow for hiding support circuitry while giving a decent amount of room for it.
    394 x 112 - 20K
  • mindrobotsmindrobots Posts: 6,506
    edited 2013-03-12 - 10:36:51
    ... an upside down PCB setup where a piece of prototyping board would have socket headers for the PropBSC and the XBee on its backside and then be mounted upside down to the bottom of the upper deck. This would allow for hiding support circuitry while giving a decent amount of room for it.

    I have an Arduino based robot that's designed like this. It's actually called a Robot Shield, I believe alluding to the way it fits onto an Arduino like a shield. I've also seen others where the expectation is to flip over the controller board and just plug it into the robot.

    Of course, the big question is... "Once you get these closer to perfected, are they going to be for sale??"
Sign In or Register to comment.