Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Opto sensors on a geartrain (Tamiya Double Gearbox)

Status
Not open for further replies.

blueroomelectronics

Well-Known Member
I've switched over to the Tamiya Double Gearbox (from the Tamiya Twin Gearbox) for a robot platform I'm working on. It's a wider and slightly more spacious gearbox the twin version. Here's the optical solution I'm testing using a matched pair of LTE 302 & LTR 301 photo pair. The Dual gearbox also has three indents in the blue gears, makes it easy to drill out to a 3mm hole.

**broken link removed**

**broken link removed**
 
have you thought about using the sensors on old mice? those are much better for that job. those big chips you want to use see too much. mouse sensors have small beams that are lot more accurate (probably good enough to read the gers themselves if positioned good).

Good luck ;)

PS mice are cheaper also + you get good cable and a ball to play with (or give to children/pets).
 
Since these will sold as kits in the future I limit myself to digikey parts. I'm building a hand wired prototype as we speak. The LTE parts are cheap too. less than $1 per pair.

Sometimes you need to keep the tolerences low so the kits can be built by beginners without too much trouble.
 
Just did a test, the blue tamiya gears are IR transparent! Well that means paint will work and the holes are unnessary. There is a 3 volt differential with black paint on 1 side and if you apply black paint to both sides its 4v (almost full swing) Perfect for the comparators.
**broken link removed**
 
Last edited:
now THATS a solution nothing can beat... there's nothing better than a transparent plastic gears for that use ;) Lucky strike man! :p I wish i could do it as easy as that... but it's hard to find IR transparent metal chain gears these days...

PS what happened to the old one? or is this Mark2 already?
 
They are blue but IR shoot right through them.

I was unhappy with the reflective sensors on the top mounted controller. Looking at the Tamiya model line I came across the Double Gearbox, a slightly wider & slightly shorter with more ratios too. Smaller gears and a ratio of 114.7 to 1 as the chosen ratio for Mongoose will make it a little speedier and still keep the torque.

The real clincher is a front mounted controller (using the four screw mounts on the front of the gearbox) and a simple bracket means the controller faces forward and other PCBs can be stacked in front of it. I'll draw the gearbox and controller later. This leaves the entire top free for a solderless breadboard on a simple chassis that supports the rear caster & battery holders.
 
Last edited:
I've used it for a couple of years. The more you use it the faster you can draw with it, and it's actually fun to use.

There is a free version called Google Sketchup.
 
very cool work man.
only question is, how do you make those gearboxes? buy one and use a measuring tape or something? i know i draw much of the stuff like that (servos, car windscreen wiper motors etc). takes a hell of a lot of time though...
1 day per part is totally possible... (not a 9h day, but 2-3 h a day or so).

but i must admit, this is fun work, no doubt. especially if one can render them into lifelike parts afterwards ;)
 
bloody-orc said:
very cool work man.
only question is, how do you make those gearboxes? buy one and use a measuring tape or something? i know i draw much of the stuff like that (servos, car windscreen wiper motors etc). takes a hell of a lot of time though...
1 day per part is totally possible... (not a 9h day, but 2-3 h a day or so).

but i must admit, this is fun work, no doubt. especially if one can render them into lifelike parts afterwards ;)

Digital Calipers. If you draw it accurately it will all fit together in the end.

I've built some test hardware, so far it's working as expected.

Here's a crude schematic of the prototype Mongoose
**broken link removed**
 
Neat. I'm looking at something similar (but magnetic) to see if I can't add some rotor position feedback to a brushless motor so I can work more reliably at lower speeds than with sensorless commutation. I was planning on going through the hole on the transmission case opposite of the motor and pinion gear. It's not looking good though since I'm not sure if the pinion gear is ferromagnetic enough and how to align the gear with the rotor. It would probably be easier for me to install a 2nd pinion gear attached to an encoder to mate with the original...

https://www.tracgear.com/product/rc4wd/custom/transmission/R2/g2/1.jpg
 
Last edited:
That's what I was looking at. Some of the problems are:
-can I mount it onto the inside of the dust cover that covers the hole opposite of the motor and pinion gear?
-can keep it's distance stable?
-will the gear steel and dimensions work with the hall sensor?
-can I reroute wires outside somehow?
-will I be able to make it correspond to rotor position (to commutate the brushless motor since I'm trying to avoid sensorless feedback if possible and don't believe I can add a hall sensor or encoder or any motor-mounted feedback)

THat pinion gear picture was just representative since you could see the insides. THis is what it actually is:
**broken link removed**
The motor and it's pinion gear go through one side (through the square hole) and a dust cover goes over the other side. The motor mount has that curved hole so that the motor position can be adjusted to make varying sizes of pinion gears mesh with the internal gears. For some reason the dust cover side also it. Perhaps I could modify the dust cover so it could also be adjusted to adjust for the distance.

It's all pointless anyways if I can't find a way to get the gears to correspond to rotor position to commutate the motor since anything that can detect rotor position can give speed and direction info as well. The motor is supposed to have 12 rotor poles and 14 stator poles which would make tight tolerances required to differentiate between the three phases since they are so close together, especially since the pinion is 15 tooth so one tooth pass doesn't exactly represent the pass of a single pole. I could always get the robot to do open-loop startup commutation and "learn" the rotor positions relative to each tooth during every power-up period. It would allow it to run at low speeds and startup under heavy loads and most importantly...no calibration! The only limitation is whether there are enough teeth on the gear to properly define the rotor positions.

My own little update- scrapped the tooth sensor. I can't get enough resolution from the gear teeth to differentiate poles not to mention I have to calibrate it every time it powers up. Sensorless doesn't work too well for slow speeds with lots of start/stops and is terrible for starting up things under load (all of which are present in my robot). It's an outrunner motor so a motor- a mounted encoder of any kind is pretty unreasonable since the entire motor case is spinning with very little stationary external parts to mount it onto. I think a $20 absolute magnetic encoder shaft coupled to the motor shaft entering the gearbox from the opposite side will work well. I just have to get the shaft coupler and motor shaft tip to mate together without a set screw (like a square slot or something) since I can't get inside the gearbox to tighten the set screw that fastens the coupler to the motor shaft.
 
Last edited:
Bill,

In you schematic above, RB4/5 are connected directly to 5V and I don't understand what you are doing with RA0.

As far as sketchup goes, are you saying there is a commercial version?

Mike.
 
Pommie said:
Bill,

In you schematic above, RB4/5 are connected directly to 5V and I don't understand what you are doing with RA0.

As far as sketchup goes, are you saying there is a commercial version?

Mike.
Opps lol, RB4 & 5 go to the collectors. RA0 uses one of the LEDs as a 1.2v reference to measure battery voltage. Mongoose is run from 4xAA NiMH. There will also be a pair of LEDs on RA6 & 7 plus a pot on RA1. Both the LEDs, POT & RA0 reference can be disabled via jumpers / dip switch.

As for Sketchup, yes there is a pro version. About $500 US.
**broken link removed**
 
Last edited:
I've put an L298 on the underside, I'll use PTCs instead of resistors and two small 1.5A bridge diodes to act as snubbers.
**broken link removed**
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top