[zenwafle]

Looks pretty cool! Let me know if you need any help with machining any components, I have access to a 3 and 5 axis cnc routers and plenty of other prototyping tools through my university. Interested to see this develop.

 

[zenwafle]

Fine tuning of shifting dynamics can be achieved by tweaking the tracks…
In this demonstration I put a brief N in between every gear for now (except R-1 and 6-7) since I don’t have a guinea pig gearbox lying around for me to measure things and not doing so might cause gears crashing.
I suppose I need a lock to prevent unintentional shifting into R. Also some simple electronics to send CAN signals for R and N.
Where do I put the main N position tho is a question…

I have a friend with a blown SMG. I can take some measurements for you if needed. What do you need measurements of exactly? Happy to hop on a FaceTime or something similar too, just PM me.

 

[zenwafle]

Thank you so much! I’m really glad I posted my idea. I don’t have any access to a machining shop sadly. It’d be nice if I can get measurements of:
Complete Dimensions of the block that has shifting pistons (need to reverse engineer that)
Operation strokes of each piston
Force for shifting into each gear
Force for pulling out from each gear

I work full time so this project can take very long if done only by myself. I’m happy to do this as a team with members from the community.

I'd be glad to help out, this project sounds really interesting. As far as those part measurements, did you mean PN: 21542282998 as the block that has shifting pistons or the unit inside of the gearbox itself?

 

[zenwafle]

Exactly that block. Since it mates to the shifting rods I believe reverse engineering the block should suffice. We need to know the bolt pattern that bolts this thing to the gearbox, and where exactly the 4 pistons are in relation to the mounting holes. The diameter and width of the slots in each pistons is important. Also how much they can move. The black thing that guides the pistons and hold hydraulic pressure should also be measured, since it protrudes into the shifting rods.

Now looking at this picture and seeing how much the actual pistons protrudes into the gearbox I am optimistic in packaging selector disks into it.
View attachment 990176

Sounds good! I'll get measurements of all of that. My friend isn't back in town for around 2 more weeks but as soon as he is I'll grab those measurements and post them here.

 

[zenwafle]

Came up with a better design...
1. Slap 4 selector disks for each shift rod right over the trans opening.
2. Fit 4 blocks with a protrusion that mates into the selector tracks in the pockets of each rod.
3. The selector disks should be synchronized. This can be achieved by making them spur gears themselves.
4. Mate a gear rack to one of the selector disk, converting rotation to linear displacement.
5. Connect one push/pull cable to the gear rack. Maybe two for redundancy.
6. The cable will be controlled by the shifter, that simply has a linear ratchet mechanism for 9 positions.

Below is a concept pic. The grooves here resembles the shifting pattern below:
R-N-1-2-3-4-5-6-7
Not enough time today to animate it so pls use your imagination lol (SORRY)
View attachment 990201
This design should be low profile to fit within the package size of the original hydraulic block. Also much more robust than the 4 cable design I proposed earlier.

This seems like a better and more robust solution. It might be difficult to splash the gear oil upwards and still get consistent lubrication. As for material I would assume something that has a bit of flex to it but I'm unsure of the strength required. Brass? A lower carbon steel? I have a friend with a '91 mr2 turbo and that car has shift cables due to the mid engine layout and they feel decent and aren't super complex. Might be something to think about although the increased wear might be a no go.

 

[zenwafle]

I do have access to a 3D scanner and have some experience using it. Once my friend with the blown trans is back in town, I can try and get a good scan of it. The scans are difficult to get accurate however. Do you want the scan with or without the module on the top with the pistons?

 

[zenwafle]

@zenwafle hell yeah man I love ya. If possible I need 3 scans
1. The SMG transmission with the hydraulic pack on top (operational state)
2. The SMG transmission without the hydraulic pack on top. Shifting rods need to have detail
3. The hydraulic pack itself

I will get that done, most likely next week. I have some experience scanning some Ganador mirrors that my buddy has for his Subaru GC8, but it may take a few attempts to get it right as it can be finnicky. I will let you know when I have it done. What file format is the best for you?

 

[zenwafle]

I was using NX on my company PC but since I can't get anything out of it I'm planning on using Fusion360. So STL should work fine.

Sounds good, I'll try and get a good scan over to you next week. I'll probably end up posting them here anyways in case anyone else wants to mess around with them. I will also take those measurements so we can have as much information as possible. I also have access to Solidworks, but have limited experience, having only designed an S50 intake manifold flange.

 

[zenwafle]

Quick update, just grabbed the transmission and will try and grab some scans of it tomorrow. Not sure if my uni has weird hours for the holidays though.

 

[zenwafle]

I envy you so much, you got everything lol
I miss my days in uni where I had all the precision machinery in the world...

It’s so nice but finding the time to use the stuff is hard.

 

[zenwafle]

Hi! Happy new year!!!
No rush, just wanna know how is it going?

Happy New Year. Sorry for the lack of updates, our shop has been closed but is opening this weekend. I plan on uploading the scans on Sunday or Monday. Here is the transmission in my trunk, it’s heavy!

 

[zenwafle]

I now have a rough idea of making an H pattern possible but now what pattern should we go for?
given the gear layout of SMG, the following will be most simple to design on the mechanical wise.
R-2-4-6
1-3-5-7
but now there are 4 vertical channels and I wonder if that can be confusing to shift without looking at it. Also where to put central position? 2-3? 4-5?
or should we go with the porsche:
R-1-3-5-7
-2-4-6-
But again, I wonder if you may mistakenly shift from 4 to 7, for example. Does someone know how it is like to shift a Porsche 7 speed?

Currently I am looking at the Corvette pattern:
1-3-5-7
2-4-6-R
With this configuration you can shift it like a 6 speed. It is like adding an additional channel to a conventional 6 speed, and you need to push/pull extra hard (like how it is for R on a BMW manual shifter) to get into 7, and push down or lift a ring to shift into R.
I think this one is the least confusing. I mean that 7th gear ain't useful 90% of the time anyways.

That Corvette shift configuration looks to be the most viable option to me. I haven’t had a chance to drive a 7 speed Porsche but that configuration also interests me. Would the lockout be difficult to manufacture/design for the corvette pattern?

Edit: After further research and asking some friends for advice, I think the first option will be the best. It is the same pattern that Aston Martin uses in their 7 speed. According to a friend who’s driven one, it isn’t that confusing once you get the hang of it. In my opinion, it’s best to keep the mechanical aspect of it as simple as possible with minimal points of failure.

 

[zenwafle]

My university has now closed the lab until Tuesday due to some weather issues that we've had . I will scan it as soon as they reopen it and keep everyone updated.

 

[zenwafle]

That Aston Martin design looks the best to me. What materials would be the most desirable?

 

[zenwafle]

Also, if possible, can I also get a scan of the whole transmission without the hydraulic assembly? Like the pictures I posted above, just the bare transmission. I need to see where can we attack shifter brackets. I plan to make it as robust as possible so no cables. That way we have to mount the shifter on the trans, not the body.

Yes, I will also grab a scan of that.

 

[zenwafle]

The concept looks promising, now the hard part is squeezing everything into the package size of the original hydraulic unit...
I'm not worried about the conversion assembly. The bigger problem may be the main selector shaft, which is not shown either diagram above. We have 4 channels instead of 3, vertical space to spare is probably some 10~15mm... I need some ideas.

I wonder how much space the SMG stuff takes up in the cabin, not saying that should be our solution but could be worth investigating.

 

[zenwafle]

@zenwafle Hi! No worries on the scan, I still have lots of drafting to do before putting in dimensions.
Is it possible for you to check if there is an interlock in the gearbox? Like try to move each rod and see if you can move two rods at the same time. Also listen to clicks and stuff

Yes, I will do that tonight. Trying to scan it asap but just went out to the car and the engine bay smells of gasoline . One problem at a time I guess.

 

[zenwafle]

Wow this is fantastic news. I wish I could read through the patent and see that but I don't want to get you into any trouble. Do you still need me to mess around with that box in my trunk?

 

[zenwafle]

I will try to post it once I figure out how to work around my company policy. And yes I still need you to scan the box and check for interlock😂
So far the idea checks out and I can start engineering once I got dimensions. I will try the Porsche design first since it is just so much easier. The mechanism is also not dependent on shift pattern - any pattern is possible and complexity does not vary much.
Side note, I just found out there was a box sold for $200 3 months ago on Japan auction site. I wonder when that can happen again 😭

I will try and scan it Sunday, and I should be able to mess around with it tonight. Fixed my fuel leak so should be good to get the transmission there

 

[zenwafle]

Got it!
What was leaking?

It was the lock ring that holds the fuel pump into the fuel tank. The temperature has been fluctuating a lot here from below 0 degrees Fahrenheit at night to around 50 degrees during the day so maybe that was a factor.

I just reserved the scanner for Sunday so I should have the scans then.