homemade servo motor

[mindoverflow]

hi,

i'm planing to build my own servo motor from ordinary dc motor, do you know about any similar project on the net? or, could you help me please.

 

[blueroomelectronics]

It's going to cost more than buying one. Basically it's an hbridge and a pot. Not sure what IC is used but it's probably specially made.

 

[HarveyH42]

It's been a couple of years, but yes there is something on the web. I think the site also sold a kit, and assembled boards. It was for pretty hefty motors though.

 

[Pommie]

You'll find a lot of information on the **broken link removed** or, if you want ready built circuitry try the **broken link removed**.

Mike.

 

[dknguyen]

**broken link removed**

 

[mindoverflow]

Thanks for you suggestions.

what i'm planing to build is a high torque servo, unlimitted angle, then i don't think it'll cost more than comercial ones, yet i have some medium-big ac/dc motors laying around (from used industrial machines).

My idea, is using a photodiode, a microcontroller and h-bridge, probabely a gearbox for precision control and higher torque. Maybe it'll not be driven as a servo (pulse width control) probabely as a stepper or fully digital (serial tx). My aim isn't really servo replication, but precise position control that's all.

This guy did some nice work YouTube - ivanblogs's Channel

I'm not an electronics guy, i'm a computer programmer, maybe that's why i'm thinking this way.

 

[Ubergeek63]

it would be better to use a stepper. you do not mention how much torque you need. steppers usually have better torque and position accuracy than servos, espescially for a given size. and they are easier to drive, to boot!

 

[mindoverflow]

higher torque is better. the actual idea i'm having is to belt drive a router head of about 10 kg. it could be less or more when applyed.

i don't have an intention to buy a steeper or whatever else kind of motor. the subject of this thread isn't wich is best, but how could i build my own. and keep in mind that i have many free industrial class motors.

 

[Ubergeek63]

so I see now... what you want then is not to build a motor but a motor drive for accurate positioning. What you are proposing to do is rather complicated and I suspect beyond your ability.

to do what you propose using a motor that you have you need to buy a $1000 rotory encoder to get your motor to have a chance to give you the accuracy you want that you could get from a $20-$100 stepper.

 

[mindoverflow]

man, why are you thinking the hard way ??

Could a photo-interuptor detect an angular motion up to some given rpm ?? i guess yes. it could be associated to a second one or something else to detect the direction.

Did you know that you could easily hack an optical mouse output to detect motion with a resolution of at least 200 dpi (goes up to more than 1500)

 

[Pommie]

You can use a quadrature encoder for feedback and some sort of endstop but the programming become much more difficult. Just stopping at a precise position is extremely difficult. Are you aware of CncZone.com, there are lots of examples of what you are trying to achieve.

Also, check out Joes Machine.

Mike.

 

[mindoverflow]

thank you pommie,

actually i don't want to look at the thing as a whole, at present i want to focus on the servo motor as a unit, i think it would be more precise to control it at the motor level than at the transmition rail or something else. i'm planing to employ 2 or 3 gears of i dunno what diameter, i guess that would be precise enought.

I would probabely pwm rpm limitting the motor uppon the sensors/µC limits.

 

[Ubergeek63]

man, why are you thinking the hard way ??

Could a photo-interuptor detect an angular motion up to some given rpm ?? i guess yes. it could be associated to a second one or something else to detect the direction.

Did you know that you could easily hack an optical mouse output to detect motion with a resolution of at least 200 dpi (goes up to more than 1500)

sure but it will not be repeatable and will not stand up to that sort of use. you are trying to get absolute accuracy for a CNC machine on the cheap... admittedly the $1000 encoder is probably overkill and you can get less expensive quadrature encoders. The problem with them of course is having a hardware software combination that can keep up with the incoming quatrature signal.

The point is that you are a programmer, not an electronics or mechanics guru. If you had access to the equipment and knowledge in both fields I would suggest a serial capable magnetic encoder IC like the http://www.austriamicrosystems.com/eng/content/view/download/1240
that costs under $6 for 10bit resolution and a serial data interface.

 

[urthlight]

I feel you

This is exactly what I was researching when I found this thread.

I have 5 dc gearmotors 5 Hbridges (home-made) and a hadfull of microchip pics and support chips and protoboards. Add a handful of linear variable resistors and optical encoders........and.....BAM a cornicopia of homade servos.

I am willing to share (collaborate).
I am mostly a hardware guy and a struggling programmer.

 

[mindoverflow]

hey mate,

till now, my project is sleeping, due to lack of time and focus on it. Also, i can't find a big enought DC motor, and can't find a quadrature encoder, i disassembled an hp deskjet printer, maybe i'll use the optical sensors found inside.

I'm in hurry now, i would share some links with you for now, and maybe i'll get back to this thread later.

Here is something great that have been done **broken link removed** take a look at old posts.

If you already got quadrature encoders, you could say you already have your servo motor. An other alternative is the use of a steeper motor as an encoder, but i'm not sure about it's accuracy, mainly at high RPM 4QD-TEC: Stepper Motor as Encoder , Homebrew 20m rig Encoder

About software, try looking for PID algorithms, it's easy, there are some code examples, if i remember well, at the microchip website, there is a PID motor control study paper, and writen program.

Good luck.