Manual Shift, Manual Clutch Cdi

[DesignerDave 5]

Hey guys, now that I'm back with my new smartie diesel that still needs some work to get back on the road, I figure what better time to finally get an old project finally finished up?

We have plenty of CNC machining equipment in house as well as both 3D scanning and 3D CAD software. I've been wanting to build a nice little kit to allow for manual sequential shifting and clutch operation for the Cdi.

There have been maps made that improve the car from stock but I still think a manual would be even better.

What do you guys think?

Dave

[MikeT 744]

That is the problem!

[DesignerDave 5]

It has been previously accomplished with the Smartuki swaps:

http://www.smartcarofamerica.com/gallery/displayimage.php?imageid=8836

The pedals I build will be a matched set of three though.

[booneylander 37]

I've always wanted to figure this out. The clutch is easy enough to solve.

Do you know the configuration of the gear selector? Is it just a normal shift-drum like a bike, and the servo rotates the drum to get through the gears? Would you need to come up with a set of detents and a star wheel? Or is it more complex?

I think even with that sorted out, the bigger challenge is going to be the software side of things, clutch/gear operation is so entwined with power delivery and ESP, I think figuring out how to convince the ECU to give you power without having a fit will be tough. Then again someone that knows the code that well, maybe it's simple... I have no idea.

I've often wondered if an easier route might be to import an older VW Polo TDI/trans and come up with a custom cradle to support it, a la smartuki...

[tolsen 630]

I think this is another of DesignerDave's ideas that will never take off.

There is nothing wrong with the present clutch set up. No need to change. Gear changes are faster than any manual clutch pedal system provided it is set up right. Copied from a post I made about six months ago on Thesmartclub:

By the way, I am still enjoying fast gear changes and roundabout hesitation free motoring.
Can only be achieved by having fully functional good condition clutch actuator, clutch actuator correctly preloaded, clutch actuator travel and clutch bite point adjusted in diagnostics and lateral acceleration sensor zero set point adjusted in diagnostics.
Obviously not possible if you take your Smart to most dealers and independents.

[booneylander 37]

Gear changes are faster than any manual clutch pedal system provided it is set up right.

That's ludicrous.

If we were talking about a sports-oriented dual-clutch gearbox, I could agree that the electronic system can engage the next gear quicker than a person shifting, but it still wouldn't have the precision, feel, and flexibility of a real manual.

But we're talking about a single clutch smart 450. I don't care how well "tuned" you've got it....

Didn't MikeT just post acceleration times for the 453, and the true manual was significantly quicker than the true dual-clutch electronic 453... but wait you're going to tell me that you've out-tuned the new DCT too? LOL

Even if turning the 450 into a manual isn't feasible, the mental exercise is still intriguing.

Dare to dream, friend!

[tolsen 630]

Let me rephrase my statement: Gear changes are faster than most manual clutch pedal system provided it is set up right.

Regrettably, dealers and indies never get the setting right and Smart 450 owners have to suffer as result.

[tolsen 630]

Didn't MikeT just post acceleration times for the 453, and the true manual was significantly quicker than the true dual-clutch electronic 453... but wait you're going to tell me that you've out-tuned the new DCT too? LOL

The dual clutch transmission is 6 speed and twice as heavy as the manual 5 speed transmission. Less mass means faster acceleration.

[booneylander 37]

So, getting back on topic, I recalled that people have been using the smart CDI in light aircraft, so I looked it up. FlyEco seems to have it licked, as they are obviously getting full power (80hp claimed) from the engine and without any transmission/clutch/speedo etc

They say :
"Highest Reliability and Safety due to use of original MERCEDES-Engine without any technical change (with the exception of the Control-Unit, with has a higher standard (MIL) and has been build “redundant”)"

Now whether that means they are piggybacking the stock ECU with a type of "emulator" (would that be "redundancy" lol). Or whether they've just coded out the trans/clutch, or whether they have a complete standalone setup, is unclear.

In any case, these guys have a solution for the ecu-side of the equation.

Edited March 31, 2016 by booneylander

[MikeT 744]

80 HP out of the cdi engine sounds a bit much!

[tolsen 630]

Air craft probably has higher emission limit than road vehicles. Around 62 HP is max you can get out of these wee engines without emitting too much smoke.

[andy m 0]

I've always wanted to try this mod, however time is always elusive..

The clutch is straight forward, albeit limited space for the pedal

The transmission should have the shift drums internal, ratchet mechanism questionable, as previously mentioned.

From what I understand these 6 spd transmissions are a 3 spd with 2 final drives, does that change anything?

There have been several motorcycles built overseas with this engine, minus the stock trans. Not sure what they used for engine management, lets just say some of these didn't appear to have the funding for expensive stand-alone ECUs.

Which leaves me to wonder... what would happen if the gear switching were left in neutral, with plugs made up to simulate the shift components being there, to allow the computer to believe its permanantly in neutral? Would the computer allow the engine to run through the RPM range under load without shutting down or going into a safe mode?

Have others done this to use in alternate vehicles, or reprogrammed the software to remove the shifting sequencing, or is it absolutely necessary to go stand-alone?

Edited August 12, 2016 by andy m

[DesignerDave 5]

Gear changes being faster with the current setup vs. shifting the sequential shifter manually? Tolsen: The best you have had is the best you know.

I'm going to speak from direct personal experience. When I raced pro Superbike, we had the sequential setup on my bike setup that it would simply cut power to unload the dogs of the transmission long enough to realize the shift completely before reapplying power. How long did this process take? I could pull out my data logger files and actually look, but it was almost instantaneous. Full power shift was so fast it was awesome. Fast enough that if I was pulling a low power wheelie coming out of the corner, the tire wouldn't even get back to the ground before I was on the power again. The clutch doesn't need to be applied if this is done properly, and it doesn't hurt the transmission at all. Best of all, it's light!

If you read up on the Ariel atom, I believe it had/has sequential dog box much like mine but with actuated shifting and although there are really "fancy" dual clutch systems out there, the truth is that the sequential dog box would shift practically as fast anyway, with much less complexity and weight. They claim 40ms shift for the Ariel setup. Dual clutch transmissions are not the answer, they are heavier, and will always be heavier because they have to have two clutches that can handle the max torque of the engine. Not just heavier in terms of the clutch weight outright, but they increase the amount of rotational mass that has to be accelerated and decelerated every time that engine changes RPM. Twice as many plates having to stay at engine RPM. The weight they add will hurt lap times more than the time they save per shift. They claim 8 ms shift time for the Bugatti transmission? That is only a 32ms difference that only applies on acceleration upshifts, since you can take all the time you want on downshifts and they aren't going to contribute to your laptime if you are braking. The extra weight is going to hurt the vehicle everywhere, braking, cornering, acceleration...

The Cdi Smart car transmission isn't a dog box, rather it's a syncro box with a sequential fork drum arrangement like a motorcycle. That shifting drum is one nice piece of billet aluminum. (I have a spare tranny all ripped apart in the shop). Put a solenoid actuated pin in place on drum detent "star" to prevent shifting into reverse (pin can only be actuated when car reads no speed) and your set, while pulling two motors off the engine and putting lighter parts in their place.

The reason the Smart doesn't shift faster than it does is the max speed the motors can get the job done. They have to deal with their own mass, as well as all the weight of the parts involved. The clutch actuator and shifting actuator do take a little time to get things finished and although it isn't bad really, it can't be as fast as a fully mechanical system when the operator really knows what they are doing. Could it be made to be faster than manual? Oh yeah, put a pneumatic actuator on that shifting drum and teach the transmission to upshift without even actuating the clutch. Automated transmissions can be made faster than human operated for sure, just not the way the Smart CDI 6 speed comes with all it's parts as Tolsen suggests.

What Andy asked is what I've wondered too: Can the car's existing engine control be made to just make power and not have any idea what gear it is in, without having to resort to standalone? I know just the guys to talk to about that, I'm going to be seeing them in the next day or so.

I hope to see this one through to the point we can do some comparisons. As these cars are getting older I'm not sure who would even care that much anymore about it. I'm doing this one for me, but can make as many as interest warrants.

Dave

[DesignerDave 5]

Update on the manual shift project: Here is the design for the custom hydraulic throwout bearing assembly.

 

The pedal and center console sections have been scanned with and without carpet as well as the whole interior floorpan. Pedal and shifter designs will be finished next.

 

I've been on the fence as to whether I would try to make this all bolt on without having to drop the transmission and the benefit of the switch is more fully realized with the extra labor involved in a direct hydraulic throwout bearing, and a direct shifting assembly that actuates the shift drum direct.

 

The transmission shifting assembly design that will be installed in the transmission itself is almost finished.

 

I hope after the car is proven others will be interested in this mod. It gets rid of the clutch arm that likes to wear out at the pivot, and it gets rid of all the electric parts that are on the transmission. It literally will be like a motorcycle manual transmission that also has reverse, and best of all, you will have a manual clutch. The pedals will be flat across them too, not that major offset it has now between the brake and throttle.

 

More to follow soon!

 

Dave

[DesignerDave 5]

Quick question for everyone!

 

If you were driving a manual sequential smart car, how would you like the shifting to be?

 

A. (Bike-like) Neutral: Half shift between 1st and 2nd (like a motorcycle), with a lockout for R (Have to press a button on shifter or console to shift into R, which is a forward shift past 1. 

 

B. (Darwin system) Neutral: Full Shift below 1st and then a full shift into R with no lockouts. Would have to keep count (keep eye on gear indicator, which likely will not be the one on the gauges, at least at first)

 

C. (Sort of like OEM?) Neutral: Locked out full shift below 1st, with R after that using the same lockout process. The insures no false Neutral ever while driving, but you have to use the lockout more often unless you don't bother with Neutral much.

 

I can custom build the shift drum to suit all three depending on what the driver preferred. I have to make a new drum anyway, so I can make any pattern I want. 

 

As it is, the transmission has to shift out of 3th before it shifts the differential range from low to high, then it shifts into 4th. I'm reducing angular displacement on the drum to do this task. They will still be offset, but will be both occuring at the same time unlike now. If you ever wondered why that shift takes longer, the drum actually rotates 57% more on that shift than on all other forward gear shifts on the 6 speed. No matter what I do, that shift is going to take more effort. 

 

Dave

[DesignerDave 5]

For option D, you go with A, but hold the lockout button. R and 1 are directly next to each other. The only time you use N is when the car is putting away in the driveway warming up for a minute defrosting the windows or idling while you go pickup your pizza. (Ultimate car for pizza pickup in my opinion) - Basically running when you aren't in it.

 

The only reason I actually do think a lockout would be nice is for those days you didn't bother to pay attention to what gear you were in before that cyclist decided to swerve in front of you and cause you to slam on the brakes. With a lock out you do the same thing as you do on a motorcycle, tap down (ahead or reverse depending on how you set your cables) enough times to make sure you must be in 1st. No need to worry about being in N or R.

 

It is true with a gear indicator you can always see what gear you are in, but this way you can look at the road and still know your in 1st.

 

Thanks very much for your feedback. Very good point about rocking in the snow, this is why I ask. Then again, just having a clutch you can modulate yourself will help plenty in these situations. 

 

I'm also doing away with the key ignition. Just two switches and a button, maybe even one two position switch plus the push button start. You can't start the car without the remote no matter what anyway. When I have the concept ready I'll get your guys feedback on the design too. I'll bet you all have thought of nifty things that would come in handy that never occurred to me.

 

One more question too: Would you rather have a short length shifter like what is there, or a longer one that isn't as far of a reach from the wheel like more racing oriented setups? I'm already placing the pivot much lower than the stock location, so it will feel more natural than the piddly plastic OEM setup:

 

 

 

Edited September 22, 2016 by DesignerDave
Grammar - As usual

[tolsen 630]

Perhaps worth while checking if you can get insurance and whether Transport Canada will allow your modified Smart on public roads.

[DesignerDave 5]

3 hours ago, tolsen said:

Perhaps worth while checking if you can get insurance and whether Transport Canada will allow your modified Smart on public roads.

 

Ahh typical Tolsen post. Your right, I should look into it.

 

What do your local laws say about you messing with your EGR system, or your wheel bearings, or your alternator pulley... What would your insurance company say if you told them you figured you had a better setup for lubricating the front WHEEL bearings on your smart car? I mean, it's not like a locked wheel could ever be a liability. 

 

I think my insurance asks if I have modified the car for performance. I can honestly say no, because as you pointed out earlier the transmission already shifts as fast as possible, so clearly this mod will lower performance. I'm good!

 

Thanks though. Really.

 

 

[booneylander 37]

I think I'd want to go 6-5-4-3-2-1-N-R with no lockout. Just so it's easy to go to N if you're sitting at a long light or something, and you can go back to 1 without having to go through R. When shutting the car off I always put it in R, so it's easy that way to just tap all the way down before turning the key off, and when you start the car up again you just have to tap up one to go to N for warmup. Also gives you just a tiny extra bit of safety when you come to a light that you're not going to inadvertently drop it into R and creep backwards. I guess a dummy-light on the dash tied into the reverse lights could keep you from doing that. 

I'm not a fan of having a lockout on anything. Having to press a button to go to N or R would bug me. And as for shifter location I'm partial to low/out of the way/short throw. But I have long arms.

Make sure you talk to "ronf", sounds like those guys have already sorted out all the issues for disconnecting the ECU from the transmission operation.

Edited September 22, 2016 by booneylander

[DesignerDave 5]

Thanks for your feedback! 

 

I've almost finished designing the new drum with the pattern just like you suggested. I can make the lockout optional so if people want to add the lock to go into N or R they can.

 

After getting all the geometry input and playing with each syncro and the new detent design I can say you will know just by feel when you shift into N. I'll have a green and red light that come on when you are in N and R, obviously the car needs to know R for the back up lights, so we'll see how I have to go about that, whether I just bypass the existing wiring to the bulbs and just run my own discrete circuit, time will tell. 

 

Anyway, when the syncros are disengaged, it takes almost no effort, they practically help you, so the only resistance you have to overcome is my new star detent for half the shift, the original side ball detent no longer does anything. When you shift into gear it takes a little effort to slide the syncro in. You will be able to feel that although you did overcome the detent, you didn't actually engage a syncro, mind you the detent will do some of that work for you. It'll be fine

 

That 3 to 4th shift will be interesting. By using 8 positions (A full shift for N) I can only have 45 degrees of barrel rotation for each shift, so I can't really offset the syncro engagement of the two sets of syncros. The design of the shifting forks prevents you from having too steep of a ramp on the drum. It's actually very clever, those inserts are what allowed them to make the barrel out of aluminum and still have good life.  

 

 

[DesignerDave 5]

Quick update on this:

 

After designing everything to fit inside the transmission to make the car shift manually I couldn't help but think that no one would want to invest the amount of time and money it would take to install those parts just to have the smart a true manual car.

 

So, set that design aside, here is what I have now. Although mechanically simple and rugged, this is a geometrically complicated unit. It is a cable operated detent unit that bolts on in place of the electric shifting motor, being very similar in both size and weight.

 

To install, you remove the old shifting motor, making note of what gear the car is in. The best would be to leave it in reverse before you take the motor off, but that doesn't really matter. Then you shift this unit in the same gear and bolt it on. Run your cables, install the shifting stick, pedals and external clutch slave (changing so you can do this without dropping the engine), install dummy electrical connectors and your done. The hardest part of the whole job will likely be dealing with the frozen bolts holding on the electric clutch actuator body. At least that will be the last time they need to be touched, as the hydraulic slave is self adjusting.

 

This shifting unit has a very unique detent that changes the rotational output per shifter stroke depending on the gear. To shift into Neutral you just row down to 1st and then it's a half shift, with the shift into R being a difficult tall detent well over half way through the shifter travel, the only shift that is like this. This will be very obvious to the person shifting just by feel, and I don't think anyone will shift into R by accident. The shift from 3rd to 4th is still two shifts offset internally because the drum doesn't get changed, you just perform the action in the same angular displacement of the shifter arm that you do all the other shifts, so it will take a bit more effort to complete that shift, but the transmission is not at risk of any mashing of gears.

 

One thing that is nice about shifting the car manually is not only that can shift hard and fast when you want to, but you can actually take it really easy on the transmission when you like, shifting really slowly when it's cold for example.

 

I really appreciate comments and feedback on this project. I hope I'm not the only one that finds this idea to be worth doing. I've wanted a manual diesel smart since I first was in one!

 

Dave

[davidpaulin 1]

wouldn't it be easier to simply locate a standard manual transmission and use an adapter to bolt it to the smart engine?

[CANMAN 23]

I don't think you could find a standard manual transmission small enough to fit in the space available.

Thanks Dave for an interesting thread, I hope you succeed.

Canman

[DesignerDave 5]

Thanks!

 

To answer the question about just swapping transmissions: The transmission the car has now is a manual transmission in every way. No need to change it, it has the right ratios, right clutch, everything. Just need to take the actuators off and put "adapters" on it to allow direct mechanical control, that is what I'm doing. 

 

I'm going to finish the shifter and pedals tonight hopefully, but I'm stuck on the shifting actuator. I'm getting prices back on all the purchased parts, the external unit will automatically increase the whole price of the kit by at least $250. 

 

I'm trying to keep the cost of such a kit down because I know Smart car owners, especially the cdi group, are very conservative with their pennies (nickels now I guess?)

 

Another question for all of you:

 

Which matters more to you? Lightest and best performing package at the expense of having to drop the transmission, split the cases and install the kit or,

 

Higher cost, bolt on solution at the expense of more inertial resistance while shifting but hopefully much less labor cost to install. To be honest I've never tried to drop just the transmission alone with the engine in the car, I assume you still have to drop the engine/transmission assembly to do this, just to get access to all the bolts around the bell housing? Any time I've worked on my smarts I just drop the subframe. I don't want to assume most owners are comfortable with doing this. 

 

Once I finish the pedals and shifter I'll need to decide before cutting metal. I personally don't mind pulling the transmission and going that route, but I want to run with whatever most people would want to go with.

 

Dave 

[CANMAN 23]

Dave,

My vote is for the higher cost bolt on solution.

As you said it's not too hard to drop the subframe, getting the transmission out of the subframe is a pain in the ass.

It takes about 10 hours to get it out, you have to pull both driveshafts,

I prefer to leave the engine partially still attached to the subframe so you don't have to disconnect coolant hoses.

I disconnect the 2 outside mounts and pivot the engine and transmission down useing the forward mount as the pivot point. 

Now you can split the engine and tranny.

canman