Servo schematic

[charles0198]

Is there any DIY servo design/schematic available? PIC based could be a good one!!!

[vintage1]

If you use a PIC here is very l;ittle to it I'd say.

Just feed the pot to the D-to-A and the input pulse stream to a digital input and drive each side of the bridge outputs from some digital outputs.

The rest is, as we say, just software....;)

[Acetronics]

;) HI, Vintage

As you surely know, PID, even adaptative, control is much too weak to drive properly a R/C servo motor ... so ...

Do you know some AFFORDABLE sources for Fuzzy logic software ??? ( means really affordable :eek: ...as Microchip doesn' t sell Fuzzylab any more )

Thanks

Alain

[vintage1]

I don't understand what you are saying.

The PIC does all the pulse stretching and logic and switches on two of the output bridge devices to drive the motor. It will probably need a fair bit of software logic to be stable, but that's software...you can do all the pulse stretching, feed forward and gain reduction around neutral etc etc. in that.

[Acetronics]

:o
Hi, vintage

Things are not so simple : old IC's ( ne 543, ne 544, BA 6226 ...) were working as you describe; so why use a PIC ???

Have a look to : www.fuzzytech.com in GDR, and you will understand what I mean ( there's a downlodable demo ). This is the software nowadays used on speed control of vehicles, servomotor control ( Yess, here we are !!!) and other non-linear processes.
Also used aboard our new digital servos ...

Alain

[stevesey]

Surely no one would just implement classical PID with a microprocessor anyway? There's so much more you can add. Systems I've worked on have always had a element of "we are a long way off, set the gain to max and drive as hard as we can until we are nearly there". This doesn't seem that different from the fuzzytech approach and the diagrams in sections 2 of the water treatment plant example look pretty much like this (i.e. traditional PID with some logic inputs). Of course I doubt if we could/would use the term fuzzy logic - the applications are safety essential/critical avionics - would you want to fly in a "fuzzy" airplane!

Steve

P.S. Vintage1 never mentioned PID anyway.

[Comatose]

Acetronics, what are you talking about? Standard servos are just proportional control and they work just fine. PID would work like a charm with the proper constants.

As for why anyone would implement PID on a PIC, lemme see. Robust, quick to implement, low processor overhead, works really well for well known systems, can derive proper control constants in about five minutes...

Sure, you can implement controller switching (what stevesey's talking about) where what he's referring to is basically handing off between some control structure and bang-bang. You can add structures as you need them. Pics are great for this. They're also great for adding prefilters and postfilters in software.

Just because you CAN is not necessarily a reason to overthink and overdesign the controller. The simplest thing that gives the desired loop response and margins is generally best.

[lazy-b]

Guys: Someone already make a Digital Controller Servo, you can convert a Standard Servo to Digital Servo:

http://www.mstar2k.com/

Guys: its just look simple to make a servo......but to make a servo really follow your finger movement to that last mSEC, thats the challenge......I have make a home made servo, it works, but the problem came out when you put different load. the servo tend to Under-shoot or over-shoot, under different loading.....I guess the Software will just take care of this.

Guys: I guess, Whats more challenging is to Make a "SERVO with FEELING"...this Technology is already exist, its just expensive, this need a two way communication......"a Surgeon have made a successful surgical operation of Human Body, thru the use of Robotics, the Patient and the Surgeon is Half way around the earth......I guess, in the Future our transmitter will be made this way.

[Acetronics]

:rolleyes:
Hi, comatose

You should learn a bit from those techniques ... They give the most precise results at this time !!! Yes, "fuzzy" makes you laugh, but look at the results given.
- And dry you tears ...

Hi, Stevesey

You should be rather frightened : quite all great liners since the "707" use plus or minus simple "Fuzzy logic" ... The Blackbird or a single F16 would never fly without those devices !!!
Japanese and French Underground trains are fully controlled by such logics ... :eek: no driver at all !!!
and it works fine ...

Hi, Lazy-b

:p Great !!! I saw that from The RX thread ... I was a little bit sorry not to have found how the logics worked !!!

:D But do not misundertand me : I really think the plane makes what YOU tell it to do ... ALL you tell it !!!

Alain

[Mr.RC-CAM]

Fuzzy programming is not the only way to produce a well designed MCU based digital servo system. Standard PID programming methods will work fine. Frankly, anything used in our R/C hobby, that is done with fuzzy code, can be done with traditional programming.

Regardless of the programming method, creating a well designed closed loop servo system is a significant challenge. But something certainly within a PIC and assembly code. I would suggest a PIC16F684 since it can directly support low parts count PWM bridged motor control.

RC-CAM

[stevesey]

And with a PIC in C as well :cool: