|
|
|
|
|
|
Senior Member
Join Date: Jan 2001
Posts: 539
|
C...A...F...D
Hi everyone, my name is Mark......and I have Compulsive Addictive Fanatical Disorder (CAFD). It's been 2 minutes since my last fanatical "vision." In my ever-conquering quest to not be content, I have decided to tackle the following project for my little known pet…..Beast.
Project: Electronically Controlled Pnematic Shifting Mechanism (ECPSM) Background: Beast with 915 = not happy. Too short gear ratios, not strong enough, shifts like el’crappo. I researched and decided to go with a G50 for the increased strength and updated syncro mechanisms. I obtained a G50 shortened, re-geared with all new “stuff” and a Quaife TB Diff. I will be installing it shortly with a stock shifter but since my CAFD mind was idle for approximately 1.246 milli seconds on 11/20/01 at 12:03 a.m. I got a “Coors Vision” (CV) to do the ECPSM. How it will work: The design is still “in progress” so it may change as I execute. The end result will be buttons on the steering wheel (one for upshift, one for downshift) with another button for reverse elsewhere. The complete ECPSM will be controlled by a OOPic microcontroller. These are very similar to the Parallax Pbasic Stamps only they use an Object Orientated Event Driven type of programming language instead of assembly language. More info can be found at www.oopic.com and www.parallax.com . The OOPic will receive the button activity as input, process the sequence and output signals to electronic solenoids which will operate 2 pnematic cylinders. There will be one cylinder for axial movement and one for rotational movement. The system will be designed to bolt on to the existing shifting housing incase I have another CAFD episode where I decide to sell the little buggers. The heart of the system is the OOPic but the real challenge is the need for absolute positioning control and feedback from the cylinders. This will be addressed with linear potentiometers built into the cylinders. This is how a typical shift will work: Lets say you are in 2nd gear and want to shift to 3rd. To shift up a gear you hit the upshift button. Please keep in mind that the ecu can process over 100,000 instructions a second. It will actually have to be slowed down to operate properly. * the ecu receives the button press as an input and invokes the upshift procedure * the ecu looks at and determines what the current gear is (which is 2nd for this example) * the ecu looks up the appropriate sequence for a 2-3rd gear shift * the ecu activates the axial cylinder to move x amount forward and stop * the ecu activates the rotational cylinder to move x amount and stop * the ecu activates the axial cylinder to move x amount forward and stop (you are now in third gear) * the ecu will update the “current gear” to be 3rd. Each upshift and downshift possibility will have its own sequence to follow. This way it will always know the appropriate up/down/left/right movements to make. To put the trans in neutral you press both buttons at the same time. It will not put the trans in reverse unless and until the trans is in nuetral (hit both buttons to put in neutral, then hit the reverse button). There will be a LED display indicating what gear you are currently in. The system will be adjustable for the speed of the shifts. The initial version of the system will be separate from the clutch. After I get the initial system ironed out I plan to incorporate the clutch actuation into the package to have the system control that too. You will still have the clutch pedal for taking off but after that you will hit a button and the system will control everything. If anyone knows anything about this stuff please share your thoughts. I have had a very steep learning curve for the last several weeks learning about microcontrollers, stepper motors, linear actuators, pnematics, circuit control, etc. If you don’t have pertinent associated knowledge but are willing to supply drugs for medication (prescribed or not) that will be accepted too. |
||
|
|
|
|
Registered
|
UHHH?!
|
||
|
|
|
|
Registered
|
Take a look at the article in Jan 02 R&T titled "changing shift patterns".
You might want to create a "learning mode" in which the driver changes the gears, and the positioning sensors learn the proper sequence. Would it be easier to retrofit a Tiptronic?
__________________
Clark Retired, I'm now posting under my real name Chuck Moreland Day Job - Elephant Racing Basic Transportation - '86 Cab - "Sparky", '77 Targa - "The Peaper" |
||
|
|
|
|
Registered
Join Date: Apr 2000
Location: Kentucky
Posts: 539
|
What'chu talkin' bout Willis?
__________________
1969 Porsche 912 <- Sold
|
||
|
|
|
|
Hilbilly Deluxe
|
Wow Mark, you are nuts (but in a good way).
![]() Keeping in mind that I know nearly nothing about the inner workings of transmissions, every car I have ever owned has been a manual shift. (VW Rabbit, Ford Mustang, 2 Toyota trucks and a 911) On every one, there have been times where it was hard to shift. Having something mechanical force it into gear would make me a little nervous. Now, having said that, one of the pointless but cool projects that is bouncing around in my head to build one day, is to take an engine from a Quad-bike and adapt it to a Kart. The thought is to use paddles on the steering wheel wired to starter solonoids to trigger the shifts. Instead of moving the shift lever up and down with your toe, the solonoids would do it for you. Using a quad instead of a normal motorcycle because the 4 wheelers normally have a reverse gear, bikes don't. Tom |
||
|
|
|
|
Crotchety Old Bastard
|
And I thought adjusting the shift coupler was a big deal.
You must have borrowed that smarty hat that Bones used to put Spock's brain back after the Amazon women stole it.
__________________
RarlyL8 Motorsports / M&K Exhaust - 911/930 Exhaust Systems, Turbos, TiAL, CIS Mods/Rebuilds '78 911SC Widebody, 930 engine, 915 Tranny, K27, SC Cams, RL8 Headers & GT3 Muffler. 350whp @ 0.75bar Brian B. (256)536-9977 Service@MKExhaust Brian@RarlyL8 |
||
|
|
|
|
|
Registered
Join Date: Mar 2001
Posts: 980
|
would it be easier to start with a tip and add switch controls?
|
||
|
|
|
|
Senior Member
Join Date: Jan 2001
Posts: 539
|
OK, you guys surely do not have the same dissorder as I do if you keep asking the question......."wouldn't it be easier if....." If I was after easy I would have bought a stock 911. Where's your sense of adventure, quest for learning, need to explore unchartered waters. geeesh
|
||
|
|
|
|
Registered
|
As with manual hand shifting, if you jammed the shifter into gear if the synchros weren't lined up or the gears speeds weren't speed synced close enough, wouldn't you wear out and bust the synchros in no time? And what about reverse?
|
||
|
|
|
|
Registered
|
As with manual hand shifting, if you jammed the shifter into gear if the synchros weren't lined up or the gears speeds weren't speed synced close enough, wouldn't you wear out and bust the synchros in no time? And what about reverse?
|
||
|
|
|
|
Registered
|
Wow, and I thought the process you went throught to get your car to the state it's in now was a big deal. Makes it look like a walk in the park compared to what you've got in mind now. Sounds like a kick though. I would think that you would want to incorporate the clutch into the system right off the bat, rather than wait to do it later. Would you be able to steal some components from the 360 Modena? I read a comparison of paddle style shifters (I think in "CAR"), and the Ferrari system was pretty far down on the wish list. Maybe you should look into it?
|
||
|
|
|
|
Registered
Join Date: Jul 2000
Location: So. Calif.
Posts: 19,910
|
Mark,
Commendable project. You might want to address engine dynamics as it relates to up/downshifts. The system used in Formula 1 cars (and I suspect the optional system in Ferrari street cars) uses an ECU, servo motors, sensors, etc. to do all this automatically, that is: On upshifts, engine torque is cut momentarily while shifting (even though at wide open throttle, WOT). On downshifting, it's a little more complicated as the ECU must know current gear selection and engine rpm, then automatically sequence the declutch, gear select to neutral. adjust throttle to appropriate engine rpm, declutch again, gear select to lower gear. This sounds complicated because it is. With your system, when upshifting, the driver controls throttle opening during shifts (close throttle momentarily). But during downshifts, how would this system double-clutch? Your system would have to briefly select Neutral (clutch pedal up) for the quick rev up before snicking it into the lower gear. Would a momentary switch work for just neutral? Then how many buttons would there be? You also have to work out the ergonomics as shifting is often done while turning. Paddle switches would solve this, but with a custom steering wheel setup. Do you have that much time on your hands? I applaud your inquisitive nature to explore. If someone didn't do this, we'd all be driving model T's. Sherwood Lee http://members.rennlist.org/911pcars |
||
|
|
|
|
B58/732
Join Date: Feb 2000
Location: Hot as Hell, AZ
Posts: 12,313
|
Mark, interesting project. Make sure you account for all UI (User Interface) events, like:
1) Upshift and Downshift buttons hit un-simultaneously (but very close together). You might want to put whatever trap you're using to capture button events into "deaf" mode until a shift completes. Or maybe don't initiate a shift until a button RELEASE. That could be neat... 2) User (thats you, bud) elects to shift manually. You'll have to update the "current gear" status in memory. I guess you'd need sensors to let you know what gear you're in at any time. That's pretty easy. 3) Like Sherwood said, downshifting will be a challenge. The pneumatic shifters on drag bikes are typically upshift only. I've even seen kits for street bikes--they unload the tranny (off throttle), engage next gear up (WHAM!), on throttle. No clutching. Unfortunately you don't have the benefit of a sequential transmission, which should make this more interesting. You might have to calibrate a delay in there...maybe wait until the sensor for each gear says "I'm in gear now" before you open the floodgates to the engine again. 4) Yellow and black "CAUTION" tape around the shift lever so your passenger doesn't get an arm broken!!! Keep us updated, this is cool hacker stuff!! Not an automation engineer, but I play one on TV.
__________________
ΜΟΛΩΝ ΛΑΒΕ I don't always talk to vegetarians--but when I do, it's with a mouthful of bacon. |
||
|
|
|
|
Registered
Join Date: Aug 1999
Location: Hickory NC USA
Posts: 2,502
|
I recommend adding both solenoid advance and retract switches so the ECU knows when it is and is not in gear. This could be come critical when a driver thinks it is in gear and it is not. You may wish to incorporate a mechanism that limits the RPMs during upshifting so an ocassional miss shift or shift failure does not result in a 10,000 RPM Opps.
Furthermore, the solenoid position switches should be of a fail safe design. When we build robotic work stations, a special PLC (industrial computer) is used to monitor critical switches to insure safe operation. This may be overkill for a car, but I design things for the Opps factor. |
||
|
|
|
|
Registered
Join Date: Jun 2001
Location: Sweden
Posts: 5,911
|
Mark:
Cool project! Don't let other people demoralize you! There are at least a dozen of cars that utilize similar system, albeit hydraulicaly... To name a few: -Ferrari 360 Modena (works like a charm) -M3 SMG Euro 1-st gen. (Shifted slow, crappy...) -M3 SMG Euro 2-nd (current) gen. (Works like charm! I was passenger in one...very tight & fast shifts) -Alfa Romeo 156 Selespeed (shifting mechanism made by Magneti-Marelli, i drove one...pretty slow shifts but still perfect rev-maching and perfect blip of throttle when downshifting.) -Most rally cars around...(most mated with dog-box tough) check out "X-trac" tranny's. What you are trying to do has been done many times before (rally, BTCC, you name it) and is not impossible in any way! Check out if you can use existing mechanism from Alfa Romeo...it's bolted onto Alfa's usual H-gate shifter. Most of those systems utilize electronic throttle...so they ease the throttle a bit before slamming into another gear to save synchros. Race-boxes usually cut ignition for individual cylinders in random patterns instead (resulting in detonation). So: Check out if you can build-in some sort of electronic throttle, so it could "blip" the revs before downshifting, or lift off before upshifting. Cheap solution would be cutting ignition. If you have turbo and are really into it, you can delay ignition so mixture ignites very late, and in that way spins up turbo so you have full boost after shifiting is completed. It's called anti-lag. It's not impossible, but it's going to be lot's of work. Actually, a friend of mine is currently working on this kind of setup for his Mitsu EVO VI... with dog-box and anti-lag. Personally, i wouldn't bother...ther are lot's of 911 bolt-on sequential transmissions around...they aren't cheap tough.
__________________
Thank you for your time, |
||
|
|
|
|
Registered
Join Date: Jan 2001
Location: So. Cal.
Posts: 9,118
|
Hey Mark:
I can offer you a short cut !! Go to a junk yard & find a mid-fifties Chrysler, Dodge or Desoto They had the shift buttons right in the center of the steering wheel. Some had the unit on the dash. Just bolt one of those things right on !! A quick fix for your CAFD.
__________________
Marv Evans '69 911E |
||
|
|
|
|
Senior Member
Join Date: Jan 2000
Location: Reno, NV
Posts: 3,694
|
.................................................. ............................wow..............
...Looks like somebody stayed at a Holiday Inn Express last night. Mark, you impressed me far beyond reason with the original project. I am at a loss for words now. All I can do is sit at my desk, stare at the computer screen and wonder what exactly YOU FRIGGIN' DO FOR A LIVING!!!!!! ...I worship you. HCHCP (holy cow hardcore project). And i don't bring out the "HCHCP" acronym all that often.
__________________
-- Chief Architect and Mastermind, SCWDP |
||
|
|
|
|
Registered
Join Date: Dec 2000
Posts: 169
|
Mark, you are the god of Porsche DIY. I've seen the system you describe on some in car rally footage and seeing the gear lever move on it's own is frickin freaky!
However, before the CAFD completely clouds your brain, I would suggest that you actually try a car equipped with a similar system. The descriptions given by beepbeep seem pretty close to my experience though I would dispute the Ferrari 360 system from being described as good as I know of 4 Modena owners trading their cars in because of the g'box. The main problems are clutch life during low speed manouvering and hill starts (n-a in your case since you are keeping the clutch for take offs) and slow upshifts. Although the point of these systems are supposed to be the speed of the shifts, the upshifts feel unnaturally slow and jerky when all that is involved is a pressing a button instead of depressing clutch and moving the lever. The Alfa Selespeed and the Aston Vanquish have these problems too (I've tried the Alfa 156 and the Ferrari 355F1). The best part of the system is the downshifts - perfect heel and toe everytime at any speed. The E46 M3 apparently has the upshift problem licked though it comes with a disclaimer that trans life is seriously reduced in the full sport mode. If you don't incorporate an ignition and throttle cut into your shift logic sequence, I can't see you matching the 125 milisecond speed of the M3. Another minus pint for these sytems is reliability. I've read that some magazines have had trouble with the Aston system and the aforementioned Modena owners have had bugs with theirs too. Same goes for the Alfa owners. Maybe it's due to the complexity of their software but the bugs are usually stuff like it won't go into gear, won't get neutral or reverse etc. Might I suggest a compromise in the form of the Ruf EKS clutchless system. I haven't tried it but I would think that having to operate a gear lever w/o the clutch might give you the control and the speed you are looking for. Btw, I think that paddles are a must for automated shifts - buttons on the 156 selespeed are cumbersome especially in corners. On another subject, have you solved your s/c belt slipping issue and have you dynoed your Beast yet?? Are you fully happy with your Haltech programming now? |
||
|
|
|
|
|
Recreational Mechanic
|
How about a manual variant of that new Audi CVT?
(Continuously Variable Transmission) It could be manually shifted to pre-programed positions which correspond to specific gear ratios which you could pre-program into it taylored to your driving terrain/style at a given time!! Keep on keepin' on! Nick.
__________________
P Cars: 2022 Macan GTS / One empty garage space ---- Other cars: 2019 Golf R 6MT / 2021 F-250 Diesel / 2024 Toyota GR86 6MT ---- Gone: 1997 Spec Boxster Race Car, 2020 GT4, 2004 GT3, 2003 Carrera, 1982 911SC, 2005 Lotus Elise and lots of other non-Porsches PCA National DE Instructor #202106053 / PCA Club Racing / WRL Endurance Racing |
||
|
|
|
|
Registered
Join Date: Jun 2001
Location: Sweden
Posts: 5,911
|
Audi CVT's belts can stand up only to specific amount of torque...that's why they are only mating it with weaker engines...i doubt it could stand up to supercharged 3.2L flat six.
OTH, it's possible to gear-up crank...go trough CVT and then gear it down in differential...that would bring torque down but wind up RPM (and we are not talking engine RPM). Like a transformer basicaly...if you want to transmit power trough wires that cannot stand up to high Amperes (NM) , transform (gear up) the voltage (RPM) to get it trough. So...Mark...if you really are a buff...adopt Audi's CVT (probably *the* best CVT around) to your car...it will get rid of all your problems with shift actuators and pesky stuff... Engine sounds pretty weird when driving CVT (it just screams at same RPM but car is accelerating), but it's actually accelerating faster than same car with normal transmission. Well...just to make it more interesting. ;-)
__________________
Thank you for your time, |
||
|
|
|