Honda EP3 with Smart Clutch Actuator
Discussion
Here I go, about to make my life more complicated...
I'd quite like my JDM Honda'd Elise (to be completed) to have a flappy paddle gearbox. I have extensive electronics ability, and I wondered about using the Smart clutch actuator to deal with the clutch, and two other actuators to move the gear selectors into position. From a power electronics and control angle, none of this is impossible. I would have to design and fabricate a few brackets, again no problem.
Does anyone have any experience, or ever heard, of modifying a manual box to run as a semi-automatic?
I'd quite like my JDM Honda'd Elise (to be completed) to have a flappy paddle gearbox. I have extensive electronics ability, and I wondered about using the Smart clutch actuator to deal with the clutch, and two other actuators to move the gear selectors into position. From a power electronics and control angle, none of this is impossible. I would have to design and fabricate a few brackets, again no problem.
Does anyone have any experience, or ever heard, of modifying a manual box to run as a semi-automatic?
If you have a dog box, and don't care how it drives (track only) and have a decent budget to replace broken parts, then it is possible to convert a conventional H pattern gear box to fully automatic control. But generally i'd say don't bother.
The best route is to get a barrel/cam sequential conversion, and automate the shifting of a dog box without using the clutch (you'll have to use that manually to get moving). This is how most of the psuedo track aelfshifters work.
With an H gate, automating the shift selection is NOT a simple task, from either a control point of view or from an actuator selection angle.
This is coming from someone who wrote the s/w for the Prodrive WRC H pattern selfshifter, and also who developed the control strategy for the gen1 single clutch Graziano transmission in the Vanquish, and who knows just how long/much effort that took!
The best route is to get a barrel/cam sequential conversion, and automate the shifting of a dog box without using the clutch (you'll have to use that manually to get moving). This is how most of the psuedo track aelfshifters work.
With an H gate, automating the shift selection is NOT a simple task, from either a control point of view or from an actuator selection angle.
This is coming from someone who wrote the s/w for the Prodrive WRC H pattern selfshifter, and also who developed the control strategy for the gen1 single clutch Graziano transmission in the Vanquish, and who knows just how long/much effort that took!
As MT infers, that is a completely pointless task if you are retaining the standard synchromesh box. Changes will be slower and robotising a manual box and clutch and making it work properly under all conditions is a mammoth task that even the major manufacturers don't get right all the time.
You could always speak to Sadev, empty your bank account and get a fully sequential box that could give you lightning fast changes
You could always speak to Sadev, empty your bank account and get a fully sequential box that could give you lightning fast changes
Emptying my bank account would probably just about get me a new knob to put on top of the existing shifter...
Yes, this is a robotising project, and given your advice I'll move my efforts to elsewhere. I imagine that several sensors are required to ensure that the actuators have actually engaged the gear properly, and that it's not got held by the synchro.
Thanks for the comments, you've probably just saved me a lot of wasted effort!
Yes, this is a robotising project, and given your advice I'll move my efforts to elsewhere. I imagine that several sensors are required to ensure that the actuators have actually engaged the gear properly, and that it's not got held by the synchro.
Thanks for the comments, you've probably just saved me a lot of wasted effort!
Surely the best option, rather than trying to develop you own hardwarre would be to get a produduction transverse DCT transmission, mate it to your engine with an adaptor plate and modified drive shafts/CVs etc. Then, make it think it is still in the car it came from.
That's easy to write, trickier to do, but not impossible, especially if you can log the CAN bus of the car it came out of.
As an easier route, you could even just(TM) hack the mechatronic unit directly, which actually wouldn't be that hard!
I would imagine someone might have done a bit of reverse engineering around the VW(BorgWarner) DCT unit by now?
http://www.evosoft.dk/diagrams/02E%20DSG%20Direct%20Connections.pdf
looks extremely convenient that the oil pressure control solenoids and the plastic unit to which that are attached seems to be linking the control / PWM driving electronics by a nice "hackable" ribbon cable....... ;-)
That's easy to write, trickier to do, but not impossible, especially if you can log the CAN bus of the car it came out of.
As an easier route, you could even just(TM) hack the mechatronic unit directly, which actually wouldn't be that hard!
I would imagine someone might have done a bit of reverse engineering around the VW(BorgWarner) DCT unit by now?
http://www.evosoft.dk/diagrams/02E%20DSG%20Direct%20Connections.pdf
looks extremely convenient that the oil pressure control solenoids and the plastic unit to which that are attached seems to be linking the control / PWM driving electronics by a nice "hackable" ribbon cable....... ;-)
Edited by anonymous-user on Wednesday 6th February 13:56
Yeah, probably best not to attempt your own semiautomated transmission control system then...... ;-)
As an aside, i wonder if there would be enough of a market to pick a DCT unit, and hack the controller to make it "aftermarket" usable? I bet the controllers are the same for the transverse and longitudunal versions of say the Getrag DCT etc? The only real requirement would be to ensure what ever EMS systemm it was going to be used for had a remote "torque down" capability. but assuming it were mostly track use, then i think people would put up with less than sparkling shift quality in return for paddles and the "banging the shift through in mS" approach!
As an aside, i wonder if there would be enough of a market to pick a DCT unit, and hack the controller to make it "aftermarket" usable? I bet the controllers are the same for the transverse and longitudunal versions of say the Getrag DCT etc? The only real requirement would be to ensure what ever EMS systemm it was going to be used for had a remote "torque down" capability. but assuming it were mostly track use, then i think people would put up with less than sparkling shift quality in return for paddles and the "banging the shift through in mS" approach!
Nope, if you power the moog valve, the clutch / selector attached to the valve unit will move, and with unstoppable force...............
We developed the basic S/W with a bench rig that used a small electric motor to spin the input shaft and a second one to load the output shaft(s). that way you have as little inertia as possible and the ability to cut torque almost instantaneously. Very useful in the early days of the S/W ;-)
We developed the basic S/W with a bench rig that used a small electric motor to spin the input shaft and a second one to load the output shaft(s). that way you have as little inertia as possible and the ability to cut torque almost instantaneously. Very useful in the early days of the S/W ;-)
Max_Torque, your comment makes me think. In a couple of years time, VW DSG units are going to be falling into the hackable market. Would say an Audi TT engine fit in the Elise? I can't imagine it's much bigger than the Honda transmission. How much power will the bottom end of that engine take?
I think a DSG Elise would be rather interesting. Interesting enough to bin my Honda conversion? Hmmm.
I'm quite happy to hack the CAN interface. I have heard on the GT40s website that a guy in the Netherlands has already made progress on this.
I think a DSG Elise would be rather interesting. Interesting enough to bin my Honda conversion? Hmmm.
I'm quite happy to hack the CAN interface. I have heard on the GT40s website that a guy in the Netherlands has already made progress on this.
Audi engine in an elise?
http://www.youtube.com/watch?v=zkrTutP3P2s
Been done numerous times before, and that video shows it could do with a better transmission ;-)
http://www.youtube.com/watch?v=zkrTutP3P2s
Been done numerous times before, and that video shows it could do with a better transmission ;-)
MrNickB said:
I'm quite happy to hack the CAN interface. I have heard on the GT40s website that a guy in the Netherlands has already made progress on this.
You'd probably need to buy a complete working donor car with engine and DSC box to hack the CAN. The basic messages transferred are not too complicated, but they are countered and checksum'd to prove validity. The most complicated bit would be the handshake for the engine torque down request, but i could probably get hold of a proper .dbc file to save having to work that bit out......However, i would suggest that you might want to replace the std OEM transmission controller competely, or more likely, reprogram it with your own propriety code, which would save a huge amount of effort trying to reverse engineer the original, and give a lot more flexibility overall for an aftermarket install
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff