Hi Folks,

This is my first post and it's regarding an automotive project. I am not afraid to smoke some things in the testing and construction process, as I don't have any idea how to make this work and expect some smoke along the way. Maybe All About Circuits can point me in the right direction...

I have tasked myself with a goal of creating a paddle-type shifter for a kit car. The car uses an older Porsche style transaxle - the type that uses a long rod that runs from the shifter between the seats, to the very back of the car, turns back 180 degrees and then performs in-out and right-left rotation movements. It didn't work very well.

The old rod-style has since been replaced with a cable-style, but the movements at the end are the same. It works better, but an electronic version would be the best.

So, the installation would use two, three-position, push-pull devices, with the logic to understand the correct position for the required gear selection.

I am guessing that 25-30 lbs of pull-push would be enough. Progressive force would be most appropriate to emulate human arm motion. The three positions have to be positive motion (without relying on spring return). Positions should be reasonably high repeatability, perhaps software adjustable, and probably require a range of 3" of travel.

Maybe this isn't a solenoid application? Maybe a pair of quick, strong steppers? Unfortunately, I have no knowledge of positioning devices or options here.

Thanks for any suggestions you can offer.

Regards,
-- Scott

 

I would say, defiantly NOT a Solenoid application!
Whatever the force Medium, it will be linear...so think in the direction of converting the linear motion to match the desired pattern by creating diversions in it's path. Which BTW it is exactly how the human operates a manual gearshift!

 

I would say, defiantly NOT a Solenoid application!
Whatever the force Medium, it will be linear...so think in the direction of converting the linear motion to match the desired pattern by creating diversions in it's path. Which BTW it is exactly how the human operates a manual gearshift!

Thanks Quant. Okay not a solenoid. Does that mean it's a linear actuator, or stepper? I don't know of any other devices. Are there folks here that are familiar with stepper applications?

I have put up a couple of JPG images to help clarify what I am describing. I think this is the start of a detailed website. Thanks for your suggestions!

 

Hi, this may be food for thought; http://www.ece.mtu.edu/labs/EElabs/EE2304/stepper_motor_cmos/Airpax_Catalog_Idx.pdf

The part you would be interested in is called a Digital Linear Actuator. Basically a stepper motor that moves in a straight line instead of rotating.
Use one to move through the gates hooked to another that moves forward and back wards to select gears. ( the actuators would form a "T" shape)

I could sketch this up if it's not clear what I tried to explain.

Cary

 

Hi, this may be food for thought; http://www.ece.mtu.edu/labs/EElabs/EE2304/stepper_motor_cmos/Airpax_Catalog_Idx.pdf

The part you would be interested in is called a Digital Linear Actuator. Basically a stepper motor that moves in a straight line instead of rotating.
Use one to move through the gates hooked to another that moves forward and back wards to select gears. ( the actuators would form a "T" shape)

I could sketch this up if it's not clear what I tried to explain.

Cary

Thanks, Cary. I just looked at these. That's the ticket! Well, I think I understand, but, if you have the time for a quick sketch, that would be wonderful. I envision mounting the units directly on the transaxle housing with short rods (and clevis), one rod to the shift rod screw-end, one to the lever.

 

Was Googling around for components and up pops an ad for windshield wiper motors. Perfect, I thought, if I could get the crank to move in both directions. Would that be possible with the correct steppers? (geared stepper?)

The crank lever would only have to be about 2" long (pictured below). I can attach one to each axis. With the proper mounting alignment and connections, it should do the job. A small u-joint, or other arrangement, should get it done. Sound logical? Well, I don't know. Something to try.

If you know of any good sources for steppers and controllers that might work for this application, please let me know.

Thanks,
-- Scott

 

I don't know, a windshield wiper motor might work - but it might be pretty stressed, too. Don't know how long it would last in that kind of an application. They do have a fair amount of torque, but you might not be happy with the speed. No clue how rapidly you plan on shifting the gears.

Wiper motors are DC motors, so they should run both ways if you reverse the polarity on them. Don't know if they're internally grounded to the case or not, which could be a problem.

You were asking in another thread about schematic and simulation programs. There's a thread in the Electronic Resources forum regarding that sort of thing.
Tina-TI seems to be popular with a number of folks; available for free on TI's website.
Linear Technology's LTSpice is good and free.
I frequently use Circuitmaker Student for quick & dirty simulations for here, but it's over 10 years old and isn't supported any longer.

 

I don't know, a windshield wiper motor might work - but it might be pretty stressed, too. Don't know how long it would last in that kind of an application. They do have a fair amount of torque, but you might not be happy with the speed. No clue how rapidly you plan on shifting the gears.

Wiper motors are DC motors, so they should run both ways if you reverse the polarity on them. Don't know if they're internally grounded to the case or not, which could be a problem.

Thanks, Sgt Wookie. Good point about the grounding. It's a fiberglass body car, too. Given the rotational motion, I am thinking 60 RPM should be sufficient. Since posting the above, from what I can gather, and I am just regurgitating what I've read today, a pair of servo motors is probably the way to go. Apparently, servo motors are capable of the closed-loop positioning logic required to provide a continuous effort to reach their programmed location. Because the pair of servos will have to work together to reach the next position (i.e., shifting from 2nd to 3rd requires three discrete motions be completed in a specific sequence), servos seem like the best solution. Are they the most practical? I don't know. Can it be done with other technologies with the same level of reliability? No clue. I have a lot of investigating and reading to do. Your opinions and suggestions are greatly appreciated.

You were asking in another thread about schematic and simulation programs. There's a thread in the Electronic Resources forum regarding that sort of thing.
Tina-TI seems to be popular with a number of folks; available for free on TI's website.
Linear Technology's LTSpice is good and free.
I frequently use Circuitmaker Student for quick & dirty simulations for here, but it's over 10 years old and isn't supported any longer.

Excellent, I will check them out. Thanks again and I appreciate the assistance. I can see that you, and others, have contributed substantially and helped lots on this forum. This project is going to take some time, but I look forward to sharing the results along the way...maybe by creating a website detailing the project, it will help others as well.

-- Scott

 

Here is a sketch of what i was thinking of in my first reply.

You know the more I have been thinking about this, I have never heard of a paddle shifter for a manual transmission. Other than one for a progressive transmission, like a motor cycle has. You know one lever moved one step for one gear change.

The logic circuit for what you are attempting is going to be very complicated. In a car type shifter, there are many different patterns that are both possible and necessary. Up shifts, single down shifts(5 to 4) multiple down shifts (5 to 3) and ETC.

Usually in a car the paddle sifters are used with a automatic transmission.

Cary

 

The logic can be simplified a great deal if you use a state diagram.

Have each state transition through one common state, which is neutral.

You have to pass through neutral anyway when going from any one gear to any other gear.

So, instead of "Shift from first gear to 2nd gear" or "Shift from 5th gear to 3rd gear", it becomes:
"Shift from first gear to neutral", then "Shift from neutral to 2nd gear".
"Shift from 5th gear to neutral", then "Shift from neutral to 3rd gear".

 

Here is a sketch of what i was thinking of in my first reply.

You know the more I have been thinking about this, I have never heard of a paddle shifter for a manual transmission. Other than one for a progressive transmission, like a motor cycle has. You know one lever moved one step for one gear change.

The logic circuit for what you are attempting is going to be very complicated. In a car type shifter, there are many different patterns that are both possible and necessary. Up shifts, single down shifts(5 to 4) multiple down shifts (5 to 3) and ETC.

Usually in a car the paddle sifters are used with a automatic transmission.

Cary

I see what you were saying now, Cary. That should work, although one of the goals is to remove the long cable runs.

There's a certain amount of truth in your statement about automatic transmissions. I don't expect they'll share a lot of information, but here's a solution for manual transmissions that can be purchased today for a little over $3k...kind of pricey for the average hobbyist. I plan on giving them a call on Monday to see if they will share their general architecture (pictured below). I would like to know what's on the other side of the box.

 

I see what you were saying now, Cary. That should work, although one of the goals is to remove the long cable runs.

Hi, if you look at my sketch (no where near as cool as yours!) It shows the shift lever mounted to the clevis on the linear motor. no cables at all.

Could you post a picture of your transmission, where the shifter is? So we can get an idea of what your working with.

Cary

 

The logic can be simplified a great deal if you use a state diagram.

Have each state transition through one common state, which is neutral.

You have to pass through neutral anyway when going from any one gear to any other gear.

So, instead of "Shift from first gear to 2nd gear" or "Shift from 5th gear to 3rd gear", it becomes:
"Shift from first gear to neutral", then "Shift from neutral to 2nd gear".
"Shift from 5th gear to neutral", then "Shift from neutral to 3rd gear".

Sgt., that would work. I've always been on the mechanical side of things. Not too sure on the electronic side. But I'm trying to learn! Kind of hard for a old guy!

You would also need a input from the clutch, to make sure it's disengaged before shifting.

Cary

 

I see what you were saying now, Cary. That should work, although one of the goals is to remove the long cable runs.

Hi, if you look at my sketch (no where near as cool as yours!) It shows the shift lever mounted to the clevis on the linear motor. no cables at all.

Could you post a picture of your transmission, where the shifter is? So we can get an idea of what your working with.

The car is traveling at the moment, but I might have something saved already. The cables currently run from the shifter (between the seats), back to the transaxle, which is at the very rear of the car. I should have indicated it's a mid-engine vehicle. So, the engine is directly behind the "cockpit" and the trans bolted on behind that. It's a transaxle, so there is no external "rear differential". There's 7 feet of cable runs (2 cables) from the shifter to the transaxle. I want to put the servo-motors directly on the trans housing. I'll see what pictures I have and post them up.

 

You would also need a input from the clutch, to make sure it's disengaged before shifting.

I'm thinking some type of normally-closed switch should be used to kill any driver (paddle) input unless the clutch is depressed.

 

Here are some pictures of the car...

Here, you can see the side shifter mounted on the transaxle:

Here's the shifter:

Here, you can see the two cables before they are attached to the shifter:

 

Spoke with MasterShift today. They were quite open about how their system works. They have two DC, 40A, gear-motors (90-degree output shaft) behind a dual H-bridge (don't know exactly what that is) that each generate 130 Lb/Ft torque with 3" levers on the cables. The box isn't as big as it looks...10"x10.25"x4.4".

They are using position sensors to make a closed-loop system. They have speed sensor input and logic that prevents improper gear selection by the driver (programmable over-rev protection). Shifting from 1st to 2nd gear take 250ms. Shifts across gates (2-3, 4-5) take a little longer. There's a physical spring that's activated when the next gear selection will cause a gate change to assist in speeding the movement (up and down directions with preselect). There's a switch on the clutch so the driver can preselect the next gear. When the clutch is pressed, the gearshift is executed. Gear selection is stacked, so it allows downshifting from 4th to 1st, for example. When the clutch is pressed, the shift is executed, as long as it's within the correct speed range.

I asked if it would be possible to disassemble the system to install the motors closer to the transaxle. The answer was yes, with caveats. The wiring would have to be sufficient for 40A, and, whatever is used for the motor-to-shaft connections must have some flex and give. That's a lot of juice. I asked about the car's battery and alternator. They said, in most cases, stock is fine. I have some good loads on my engine already (lights, fans, and pumps) and will have to do some calculations to see if that's too much power draw.

All said, this project might be too advanced for me in terms of time available. I have also read recently about some guys that have used the MasterShift system and done some damage to their trans. Of course, it's hard to know if the damage was from operator error, poor installation, or other issues not related to the shifter. I don't want to take a chance with my transaxle. They are too expensive to experiment with. So, for now, I will keep my ear to the ground for enabling technology and techniques, but probably not move forward at this time.

Thanks for all of your comments and suggestions. I have another project using LEDs that I'm stumped on and look forward to reading your posts...

-- Scott