servo control circuirt

[sahilkit]

hi everyone

i am sahil from india and i want to build a servo i have all the parts

except the control circuit used in the servo , i know its working

and so far i have been able to build up to pulse comparator part of the ckt

on my own and i need to know more about the drive part of circuit

so if anyone of u have any info on this or circuits or anything please post

it



sahil :)

[Dan Baldwin]

The control circuit of a servo consists of;

a one shot (monostable multivibrator) that is triggered at the beginning of the pulse from the receiver.The period of this one shot is controlled by a pot that is attached to the output shaft of the servo. The period will be 1 MS to 2 MS over the range of rotation you are looking for.

a gate circuit (pulse comparator) that sets one output if the pulse from the receiver is longer than the one shot pulse, and sets a different output if the pulse from the receiver is shorter than the one shot

a pulse stretcher; The pulse that comes out of the gate circuit is very short, and it needs to be made longer to drive the motors.

a motor drive circuit that usually consists of an H-bridge consisting of 4 power transistors that can drive the motor either direction. Only two of these transistors will be turned on at a time.

Of course the easiest thing would be to buy a servo, and the next easiest would be to use an IC that is designed to drive a servo.

Dan

[Comatose]

I'd say that unless you have a line on integrated servo control ICs, it'd be easier to use a small microcontroller and end up with a digital servo. If you're taking the time to build one, you might as well build the niest one you can. No point (IMHO) to spend all that time and end up with something functionally equivalent to a $10 servo.

[Dan Baldwin]

I'd say that unless you have a line on integrated servo control ICs, it'd be easier to use a small microcontroller and end up with a digital servo. If you're taking the time to build one, you might as well build the niest one you can. No point (IMHO) to spend all that time and end up with something functionally equivalent to a $10 servo.

Which brings up a question that I've had for quite some time; If all you have top do is replace the normal servo IC with a cheap microcontroller (the PIC 16F684 has 10 bit analog in, 10 bit PWM out, will drive a full bridge and sells for $1.56 in single qty), why are digital servos so expensive, in fact why aren't all modern servos digital.


Dan

[lazy-b]

Guys: Why keep re-inventing the wheel, somebody have already done it.

http://www.mstar2k.com/

Mr. Gordon target price is USD 20 for servo controller + cheap servo USD 10= USD 30 ........a commercial digital servo cost USD 100.

Maybe be we need just to write email to Mr. Gordon.

[sahilkit]

thanks for the replys

[mike50]

Guys: Why keep re-inventing the wheel, somebody have already done it.

http://www.mstar2k.com/

Mr. Gordon target price is USD 20 for servo controller + cheap servo USD 10= USD 30 ........a commercial digital servo cost USD 100.

Maybe be we need just to write email to Mr. Gordon.
re-inventing??? If you go to the microstar web site it appears that Mr. Gordon hasn't quite invented it yet either.

Mike

[JimDrew]

There are PIC micros that have 10 bit a-d converters, PWM input/output, and a h-bridge driver control in one unit. Cost in single quantity is about $5.00. These are perfect to run at 8Mhz, with a switching of 20Khz for the motor control. You can determine the motor speed by using EMF, and can even detect a jam (motor stall). You just need a few resistors for MCLR and pull-ups, and one each of a dual P-Channel and dual N-Channel MOSFETS. Total cost to build is less than $10... about $3.00 w/PCB in 1,000+ quantities.

[duhaas88]

does anyone have a diagram or a schematic that shows how this works??

[Mr.RC-CAM]

Here is the data sheet to the NJM2611 (NE544) IC: http://www.njr.co.jp/pdf/be/be06014.pdf

There are other servo IC's out there. They are all similar in function.

RC-CAM

[MatC]

The reason digi servos are so expensive:
Analog servos are extremely cheap to make, have had their R&D paid for years ago, they use very cheap components and sell by the million. As soon as you develop something new, you increase the cost which decreases the market significantly, which increases the cost.... and before you know it, something that costs £6 to buy ends up costing £20 to make even though it's only marginally more expensive in terms of components.
Another factor is that digi servos strain the motor and gears more (they drive the components harder in order to increase performance), hence those components have to be more expensive.
In a few years the market will come down and be sensible... 10 years ago a servo was £10 (£6 now), and a mini servo was £45 (£6-£15 now).
I don't think we'll ever see 100% digital servo sales, because they will always be _slightly_ more expensive to make, and people will still want a £6 servo rather than a £6.25 servo, even if it's not as good.
The mstar2k.com stuff looks interesting :)

[duhaas88]

Is there any other way to determine the position of the servo motor other than using a pot. I'm working on a new project that I'm using small dc motors to drive gears and need 360 degree rotation, but don't have room to use a servo. I have a small dc motor that is setup with a pic to drive it, but I need a encoder or someother way to accurately determine the position.

Any Ideas??

[Dan Baldwin]

Is there any other way to determine the position of the servo motor other than using a pot. I'm working on a new project that I'm using small dc motors to drive gears and need 360 degree rotation, but don't have room to use a servo. I have a small dc motor that is setup with a pic to drive it, but I need a encoder or someother way to accurately determine the position.

Any Ideas??

Absolutely. The cheapest and easiest way by far is to use a pot, but there are many methods of determining position of a shaft. Early servos used variable capacitors instead of pots, but that doesn't help you since they generally don't have full 360 deg rotation either. You could use a quadrature encoder, but the encoder only keeps track of movement of the shaft, not absolute position, so it would have to be "homed" when it is first powered up. Some encoders have a separate home output which is active at a particular position. You could use an absolute encoder which does keep track of absolute shaft position. There are some pots available that have full 360 degree rotation with no stops, but they have a "dead zone" of perhaps 10 or 20 degrees where you don't know exactly where the pot is. You could probaby use two linear output hall effect sensors set at 90 degrees and a magnet on the shaft. Resolvers give very good resolution, but are a pain to use.

Do you need continuous rotation?

Dan

[Acetronics]

Hi, Mr Cam

In that way ( NJM 26xx ) ... don't you have any replacement for the NE 543K / WE 3141 ...or even µL 914 on old varioprop servos.
Some retired persons from our club do not want to renew their servos or speed controllers for boats ...

regards
Alain

[Mr.RC-CAM]

... don't you have any replacement for the NE 543K / WE 3141 ...or even µL 914 on old varioprop servos.The obsolete NE543K is available from some IC brokers (Google will point the way). Be prepared to buy a rail at a time and to pay too much for them. A brand new servo or ESC is not such a bad idea after all. :)

I agree with Matc's comments. It is so much less risk, and cheaper, to copy and clone. Innovation is another story. To compound things, Far East Mfg's worry about every single penny on the Bill of Materials. A microcontroller would add several cents more, which would give them material costs grief. And, they would have to learn something new, which they are usually not thrilled about. These observations are from my personal experience with off shore mfg.

[dedal4o]

The obsolete NE543K..
..there is not other way to do [by the logic: with PIC, Hall sensors or even the sample POT! ] I don't belive... :eek:
It doesnt matter that is not cheap as a standart solution [jist to buy servo]. The forum name is "DIY Electronics"! perhaps... ;)
I am insatiable. I want also to made by my self. :)

[Mr.RC-CAM]

there is not other way to do [by the logic with PIC?, Hall sensors or even the semple POT.! ] I don't belive... I misunderstood. I thought Acetronics specifically needed to help a retired club member find the NE543 replacement part. But, a custom DiY solution is more interesting. It all depends on what the goal is.

[dedal4o]

I misunderstood.
that I quote to you... :) And keep in mind that my English language is poor...
But I think that the possibilities of mind is not terminated by the rules. If we want to made our oun servo...we can do it. It doesn't matter how much will cost [money, time, etc.].

[Mr.RC-CAM]

But I think that the possibilities of mind is not terminated of the rules. If we want to made our oun servo...we can do it.You are preaching to the choir. :)

[dedal4o]

You are preaching to the choir. :)
:) Yes.
Again sory... I read wat you do...The www. make our world closer. And I can say again: I glad to see you there Mr Cam!

[duhaas88]

Absolutely. The cheapest and easiest way by far is to use a pot, but there are many methods of determining position of a shaft. Early servos used variable capacitors instead of pots, but that doesn't help you since they generally don't have full 360 deg rotation either. You could use a quadrature encoder, but the encoder only keeps track of movement of the shaft, not absolute position, so it would have to be "homed" when it is first powered up. Some encoders have a separate home output which is active at a particular position. You could use an absolute encoder which does keep track of absolute shaft position. There are some pots available that have full 360 degree rotation with no stops, but they have a "dead zone" of perhaps 10 or 20 degrees where you don't know exactly where the pot is. You could probaby use two linear output hall effect sensors set at 90 degrees and a magnet on the shaft. Resolvers give very good resolution, but are a pain to use.

Do you need continuous rotation?

Dan

Yes I do need continuous rotation:)

[Dan Baldwin]

That makes it a bit tougher. If you needed 360 deg rotation, but did not need continuous rotation, you could gear a pot to get the rotation you need. The fact that you need something extremely small means that you probably won't be able to use any off-the-shelf encoders. Probably the best bet is to use two analog sensors (hall effect, capacitive, optical, or inductive) placed 90 degrees apart, and use a microcontroller to turn the signal from these sensors into actual position. It might be as simple as a small washer on the shaft that is offset so it wobbles as the shaft spins, and progressively breaks the beam of an interrupter type optocoupler.

Dan