PDA

View Full Version : Looking for custom servo design (Paid)



nhtshot
03-03-2010, 12:28 PM
I'm in the Chicago area and looking for someone to do some servo controller design for me. This is a paid contract.

If you have experience with motor controllers and feedback systems, I really want to talk to you.

Alex
03-03-2010, 12:43 PM
[mod note: moved to proper forum]

Addidis
03-03-2010, 01:14 PM
A little more info would be nice. What kind of mcu are you wishing to use. How complicated is the work to be done. what kind of interface are you looking for (serial usb none*just program the mcu to update) I have some experience with servos. Not the best not the worst.

If you are willing to use a basic stamp for your project, I would recommend rather then paying some one for their time , buy a "what is a microcontroller (http://www.radioshack.com/product/index.jsp?productId=2117994)" kit from parallax. It cost 99 bucks at radioshack and will teach you to write the code (in basic eww) and provide a servo and serial interface board for you to learn to write this code your self. The basic idea will be easy to pick up in this manner IF your project allows for it. Its not the path to go if your looking for more then a one off project as Basic stamps cost a whole bunch by them selves. But the ability to program in basic and the nice book it comes with is effective to learn.

badcommandorfilename
03-03-2010, 03:12 PM
I'd like to help you - I know a lot about Real Time Control systems, and this sounds like an interesting project.

Unfortunately, I'm nowhere near Chicago so I doubt we would ever be able to meet in person. If you don't mind a long distance working relationship, PM me and I'll see what I can do.

jes1510
03-03-2010, 03:20 PM
Please, for the sake of the industry do not under any circumstances use a basic stamp in a commercial product.

Addidis
03-03-2010, 03:28 PM
agreed.

I Just prefer teaching to subcontracting. Really there is not enough information given to make a proper assumption of his intents :)

nhtshot
03-04-2010, 07:23 AM
I made a vague post at first because I've never been to this forum before and I wasn't sure how they look upon posts like this. Since the powers that be don't seem to mind, and were nice enough to put my thread in the right spot, I'll give some more details.

This is not a robotics project per se. The 10 mile view is that we're building a custom multi-fuel ECU for a generator. It's based on a freescale cpu:

http://www.freescale.com/webapp/sps/site/overview.jsp?nodeId=06258A25802570256D

We can buy the interface off the shelf for the fuel injectors and the ignition module. The one interface that we can't buy off the shelf is to an electronic throttle body. I was able to get a schematic for the part we're using and it's basically a servo without a control board. Inside the unit is a motor and two pots. It doesn't have any of the logic usually found in a servo though.

So, what I need is someone to design a circuit to make this throttle body work more like a standard R/C servo. I want to pass it a single input signal and have it move to the proper position, read it's own pots for feedback and not give me any trouble about it.

If you're interested, drop me an email at [email protected] and I'll give you the precise details of the throttle body in question.

nhtshot
03-05-2010, 10:52 AM
A couple quick points of clarification:

I'm not trying to build a throttle body or a controller board to run standard (r/c type) servos.

I have a throttle body bought that has the motor and potentiometers built in to it. It does not have any control logic. The pin connector has pins for the motor leads and the pots and no circuitry of any sort on the inside. I don't need help interfacing to a standard servo, I can do that. What I need is to turn a throttle body that does not have any circuitry inside it into a standard servo.

I've already found several designs for home-built servos, but they are not really intended for use in this kind of an application. The two issues that prevent any of the circuits I've found from working are as follows:

1. The throttle butterfly is on a spring. The motor, therefore, has to always be pushing to keep the valve open. This requires a very substantial amount of current that will be flowing constantly. I also have almost no tolerance for failures. The design and parts have to be able to output the needed amperage at 25-30% duty cycle so they'll have a long lifetime.

2. I cannot change the wiring or values of the pots. There are two pots inside the off the shelf throttle body that I cannot change. They are attached in opposing directions and are at 500-1.2k. I'm not familiar enough with voltage dividers, etc... to get the values on the feedback loop correct. Furthermore, since these parts are built into the unit, I have to be VERY careful in experimenting. I cannot replace them if I fry one.

So, I'm looking for someone that has enough knowledge about these kind of circuits to come up with a design that meets the above requirements with a minimum of experimenting on the actual equipment. I don't care about the interface to the micro controller. Right now, I'd be happy to just be able to drive the position of the throttle body with a pot.

I do appreciate all of the responses, but I don't think that I was very clear on exactly what I'm trying to accomplish.

zoomkat
03-05-2010, 11:19 AM
Can you remove the throttle body spring? If you are controlling the throttle body with a servo, then no need for the spring. You may want to look at "drive by wire" throttle bodys that don't use mechanical linkages for operation and control. I got involved with fuel injection when the mass air flow monitor on my truck started going bad and eventually had to be replaced with a new throttle body (on a Nissan the throttle bodys appear to be calibrated to the monitor module by maching the bottom of the throttle body). A most interesting exercise in how things work.

nhtshot
03-05-2010, 12:01 PM
No, it's basically a solid unit. I suppose it would be theoretically possible to change out some of the innards, but it's built pretty well and I'd rather not mess with it. Besides, the spring is intended as a fail-safe if there's an issue with the electronics. Run away Toyotas anyone? Perhaps they should have used the German design (the throttle body I'm working with is a VW part).

The part I'm using IS a "drive by wire" throttle body. Instead of mechanical linkages, it has an electric motor and some pots attached directly to the valve. It does not have any servo logic inside of it. The connector on it has pins for the motor and pins for the pots, that's it. I have to implement all of the servo logic outside the unit. That's what I'm here hoping to accomplish.

Unfortunately, while there's a wealth of information about interfacing with servos, there's next to nothing on building them from scratch. There are some bits and pieces out there and I've done quite a bit of experimenting, but it's obvious to me that it will take a long time for me to make this work by myself, which is why I'm here looking for expert help.

zoomkat
03-05-2010, 12:26 PM
The old race car builder of the 60's, Mickey Thompson, use to have people come to his shop and want him to build them a car to break various speed records. The first thing Mickey would tell them was "The first thing you need to understand is that speed cost money. Now, with that understanding, just how fast do you want to go?". So, just how much $$$ are you willing to invest to have somebody build your servo? While you are crunching those numbers, how much torque is required to rotate the throttle shaft? How much info do you have on how the installed throttle body motor works to move the throttle body? If it is already a "drive by wire" setup, then consider using a servo to rotate the pot that is probably connected to the accelerator pedal.

nhtshot
03-05-2010, 01:24 PM
Money isn't an issue. Within reason, I can spend what needs to be spent to make this work.

Just for clarification, here's a picture of the device in question:
http://chicagolinuxservice.com/~sean/pics/

I've also posted what I was able to get from VW as to the innards of the device. It's the gray box at the bottom right. G186,187 and 188 are the motor and pots respectively. The T6a/1 - T6a/6 are the wires in the connector on it.

zoomkat
03-05-2010, 02:59 PM
I'd get a shop manual for what ever car the throttle body came off of and study the wiring diagrams. There may well be a throttle control module that you could use to skip making any complex custom mechanical parts.

jes1510
03-05-2010, 03:04 PM
Will this be a commercially produced product? If so then your going to want a design engineer that has a PE certificate since it's an automotive product. I'm not sure what safety certificates will be needed for testing in the automotive industry.

nhtshot
03-06-2010, 08:23 AM
It's not going into a car, good suggestion anyway though.

nhtshot
03-06-2010, 08:35 AM
I'd get a shop manual for what ever car the throttle body came off of and study the wiring diagrams. There may well be a throttle control module that you could use to skip making any complex custom mechanical parts.

Unfortunately, the module is built into the ECU on the cars this is used on. To further complicate things, the ECU in question is glue-filled to prevent reverse engineering and they don't publish any schematics for it. I've uploaded in a previous email everything that I could get from VW that was worthwhile.

zoomkat
03-06-2010, 09:41 AM
With a spring close throttle, you may be able to get away with only driving the motor in the open direction with a single MOSFET. You could make a "servo tester" (numerous schematics on google)that could make use of the 500-1.2k range pots in the throttle body, and have it drive the motor control MOSFET. If money isn't a problem, you may want to go to junk yards and purchase some used throttle bodys to experiment with. You may also need to determine the voltage normally supplied to the motor and the pulse frequency it normally uses. The below link has some info as to the internals of the typical hobby servo.

http://www.seattlerobotics.org/Encoder/200009/S3003C.html

nhtshot
03-06-2010, 03:47 PM
You've got the right idea. I actually have two of the units and have one wired up exactly that way, using a couple of mosfets though because the thing can pull some serious current. (Of course, I don't hold it open, I was mostly experimenting to figure out voltage/amperage requirements)

It's an automotive unit, so it's almost guaranteed to be 12v driving the motor. I've driven it with 12v through a mosfet and it works fine. Unfortunately, the pulse frequency is a bit harder to determine. I'm trying to get my hands on one of the cars that this is used in so that I can scope the throttle interface. However, none of the local car rental shops have one and most people aren't very willing to let strangers hook electronic "gizmos", like an oscilloscope, up to their (fairly new) cars.

I follow what you're thinking about with the servo tester, just using it as a source of timing input to the mosfet. I was actually thinking something along those same lines using a 555 timer.

I understand how RC servos and their pulses work. The real question is whether that kind of pulse could actually hold the throttle at various non-full open or closed positions.

From what I've been able to gather from VW, it looks as though they drive the motor and read the pots directly from the ECU. I'm assuming that they just run the pots right into an input on the CPU and have a CPU output drive a mosfet or two or three that powers the motor. I considered that an arrangement like this might be the right way to go on the long term since it minimizes additional components. Before I could do something like that though, I'd really like to experiment with it and have it working using I/C logic.

zoomkat
03-06-2010, 06:06 PM
You may want to get a motor controller like below and see if it will work with the throttle body motor to be able to position the motor to full open. The pots in the throttle body feed the ECU with position data for numerous possible functions. As the voltage in a car can vary probably from ~12v to 14.4v, the throttle position could change just due to voltage variations in the car impacting the throttle body motor.

http://www.electronickits.com/kit/complete/motor/CKMX033.htm
http://store.qkits.com/moreinfo.cfm/MX033

nhtshot
03-07-2010, 08:05 AM
I'm not using this in a car, it will be powered by a regulated power supply, so the voltage should be pretty constant.

Either way, very nice parts. Those are exactly what I was looking for. I'll order them up and report back.

Thanks,
Sean

nhtshot
03-07-2010, 09:47 AM
Thank you very much for those links.

I went for this:
http://www.electronickits.com/kit/complete/motor/CKMX067.htm

That should give me more then enough flexibility to make this work.

Sean

nhtshot
03-11-2010, 08:37 PM
Got it... Works GREAT! That's exactly what I needed. It even included the schematic so I can learn something from it.