z driver 88t

Shift Centering Spring for early G28/11 1985 US/Canada transmission.

I've researched this extensively on Rennlist and Reutterwerk and am trying to determine the best course of action.

I have gathered Colin produced a replacement for the NLA shift centering spring that is sold by Roger for the BW transmissions.

Greg B had posted at one point that there were actually three versions of the BW shift centering spring and that "early" spring has been NLA for quite some time and required the use of a "wider" upper shift roller piece, when using the "late" spring. Failure to change to the wider roller will result in the shift spring breaking quickly...and often.

https://rennlist.com/forums/928-foru...g-broke-2.html

I have inferred that the spring Colin produced is a replacement for a "Late" spring which can technically be put into an "early" BW gearbox but as Greg pointed out, could have a failure potential unless the wider roller is used.

So has anyone used Colin's spring on the early BW gear boxes? Did you replace the "roller." Does this "roller" drop in or are other changes required too?

I am trying to decide if it would be better to get a used "early" spring (understanding it has an inherent design flaw and will fail again someday) or retrofit the gearbox to adequately use Colin's replacement design without fear of part failure.

Lizard928

Chris,

It will function fine in 85/86 boxes.

z driver 88t

Thanks for the fast reply Colin.

One quick question. Based on Greg's info the "early" spring is held captive on the lower adjusting pin by a circlip. There is a washer between this spring and the lower adjusting pin. In the "later" spring the lower washer and circlip are eliminated (the lower part of the spring is therefore dimensionally different than the early spring.

If your spring is dimensionally the same as the "later" spring, do you know if the circlip and washer are eliminated and discarded or are they retained.

It may be apparent once I have the gearbox removed and opened up. I'm just trying to minimize those moments of "now what do I do?" Thanks again.

GregBBRD

Here's what I know....and it is going to "differ" from what Colin just said....but he makes his own product and should know the problems....so I can't speak for him.

Hopefully, this will clarify the problems people encounter.

The "original" '85 spring" had a "short" ramp piece (the part that actually touches the roller) that protruded from the actual spring. The profile was similar, but significantly smaller, than the "later" spring (version 3, '09 spring). This spring was retained inside the gearbox with a circlip and had a spacer behind it. This spring operated on the "upper roller" part, which retained the "roller" with a pin that was held in place with a circlip. The important part.......The distance between the two flat parts (on each side of the roller) is 7mm.

The "second version" of this spring is very rare. I believe it was supplied in "late" '86 transmissions. I've never seen a part number, or a reference, by anyone, for it. It also has the "short" ramp piece, operates on the same roller assembly. The "base" that fits onto the lower shaft is wider and the spacer is not used. The circlip retainer was also deleated. If installed with the original spacer, there would be too much preload and the spring would break in very short order.

The 3rd version is the one that most people are familiar with ('09). It has a "larger" ramp piece, has no spacer, has no circlip to retain it....but (and here's the important part) has a completely different design upper roller part. The upper roller has the roller and shaft welded into place. And the important part.....the distance between the two flats that are on each side of the roller is 11mm.....which is 4mm wider than the '85/'86 roller! It also does not use the original "spacer" piece and will break , in very short order, if installed with that piece, from the original '85 transmissions.

So what happens with the "late" spring and the "early" roller? While it makes no difference, when the spring is centered (neutral), the wider upper roller is a must when the transmission is shifted.....or the larger ramp will be leveraged off of the sides of the narrow upper roller.....which results in the "ramp piece" being twisted on the spring...or in twisting of the actual spring.

This single detail ruined hundreds of these springs. People would install the "late" spring with the early roller.....and the spring would break in a very short period of time....the "twist" when the car is shifted would literally rip the "larger ramp" piece right off the spring....or fracture the actual spring right at the bottom of the attachment point.

Hope this helps! If needed, I can supply pictures.

z driver 88t

Greg, since all of the OEM springs are NLA, and Colin's spring seems to be the only real option, then what I think I need is the "late" roller to match it. If I understand the diagrams the roller is part of the preselector lever, correct? Do you know if an S4 preselector lever will swap into the early BW trans with no other modification? Apparently the preselector lever is also NLA, or on longterm back order but I could probably get a used one from 928 INTL.

GregBBRD

Yes, the late preselector lever (why did they name it that?) is a direct interchange. No other pieces or modifications needed. While I do not think that 928 International has any more used ones....Tom told me, today, that he still has a couple brand new ones.

I frankly "freaked" out when Colin told me he was going to weld on that piece of spring steel....I've been told, virtually since birth, that welding (heating) on spring steel would result in one being sent directly to Hell. What I learned is that the heat draws the carbon to the weld site and makes spring steel very brittle (much like welding on Chrome Moly.) We haven't talked about it, for quite some time, but Colin thought that it wasn't that big of a problem. Roger and Colin can tell you how many they have made/sold and if any have failed....although I have not heard of any issues.

Keep in mind that I'm very "old school". Colin is more "new school" and knows a lot about welding and a lot about materials (and he is also very smart). He most likely knows something I do not or uses some kind of material that works fine, when heated (welded).

Since I'm already got several transmission pieces under design, construction, or testing, I'm certain that this will also be something that I will eventually make....since I can't ever imagine myself installing one that has been welded together, regardless of how much assurance I get. I also "foresee" the "late" upper roller piece to becoming impossible to get...new or used, so I'm also planning on making a direct replacement part for the '85/'86 spring....with the "smaller" ramp piece, which will allow people with the '85/'86 spring to re-use their original upper roller.

In the meantime, I've been using the few springs/preselector levers that I have left over from gearboxes that have been disassembled. When I run seriously low on those, I'll work harder on making new parts...there are so many pieces that are needed for the 928 model, that I need to make choices about what I build first. The "market" is very small and the long term "payback" for production runs can be (usually is) years in the future.

Again, hope all this information helps!

Mike Frye

So Greg,

Do you have any of the early springs available for Chris or would you recommend that he get the spring and roller combo to replace them both?

Colin and Roger, I hope you both know I'm a huge fan of both of yours. I admire your work and efforts, but in the interest of honestly and full disclosure I have to present my experience with Colin's trans spring.

I was involved in the early development of Colin's replacement spring. In fact the original one he had to model his from was a broken one I sent him. I had a 'real' early one from Porsche and I compared it to Colin's on my bench to see if it was comparable. Admittedly the tests were not scientific, but if it had passed my initial tests I had planned to use more exacting tools (like a fish scale) to measure the spring and return rate at different points on the lever as well as the overall dimension for the measurements that matter. It didn't get that far for me.

Here is the stock spring, showing how much it would move with the pressure of 1 finger (and returned to straight, obviously).

after flex, (returns to original position)

Here is Colin's (admittedly probably version 1.0) spring which flexed more and did not return to straight even when flexed with one finger on the bench...

after flex...(did not flex back to straight)


Based on this initial test I didn't go any further and didn't feel comfortable installing it in my car. I felt it would not do the job the original was designed to do and therefore I sent it back.

As far as I know I got the last new return spring for the S3s that was available several year ago. I'm sure there are more out there, but I'm pretty sure by the time Roger got mine, he had searched far and wide and he said it was the last one to be had.

Chris, good luck with your quest. This is a known weak point in the system and hopefully you'll be able to get it worked out.

z driver 88t

Thanks for all the input guys. Hard to believe this is so complicated. In searching, I found this post by RET where he was able to repair his original spring with what seems to be a pretty straight forward and almost elegantly simple solution.





https://rennlist.com/forums/928-foru...g-broke-2.html

I have an email into 928INTL to see what they may have in terms of used or new S4 preselector levers that could let me use Colin's spring with less worry about the concerns Greg raised (hoping it works properly), but in either case I know a few guys that do metal work so when the broken one comes out I will see about attempting this sort of repair to the original spring so that I have a spare.

GregBBRD

I do not have any early version springs available.

As I mentioned, Tom (928 International) has a couple of late preselector parts, brand new.

I've got several used springs and actually still have one brand new spring tucked away (that I paid a king's ransom to own.)

And for the people that have a spring go through the gears.....I still have a couple brand new gearboxes stashed away.

I saw what "RET" did to repair his spring, before. Seems like a reasonable "short term" solution. I think that trying to retain this spring with nuts and bolts is going to be a very tough task, in the long run. The forces on the hardware, when this spring has to go over the "detent" portion has to be extreme and my guess is that it will eventually come loose.

Porsche isn't dumb. They knew that they could not weld or heat spring steel...so that is why they "staked" the original detent part into the spring and riveted the spring to the lower bracket. "Different solutions" are tough.....I've tired several different ideas for replacing the entire mechanism with something different....without any results I feel I can market.

And the original springs worked for long periods of time....I'm sure I've never seen one fail without many years and lots of miles on it....unless it was improperly installed with either the spacer or the "early" preselector lever, or was improperly tensioned.

Any "repair" or " idea" has to be very robust, for my use. The broken parts of these springs go through the gears, quite often.....turning a relatively simple failure into a disaster.

ROG100

I checked and there are 10 x S4 pre-selector levers sitting in Germany.

I provided Chris with the details on sales of the last 6 months of Colin's springs. We also sold them prior to that but digging up the information from my old system is time consuming.

No known issues to date with Colin's spring.

Lizard928

The heat from welding the pieces on the spring steel does affect the spring in the HAZ (Heat Affected Zone), but with the method I used for welding on the springs Roger has have a very small HAZ.
The spring in the middle retains the same spring force.
The HAZ becomes harder (loses spring) but does not become brittle in any sense of the word. If you push too far on the spring, (beyond its design parameters) it will bend and maintain the bend in the HAZ.

I have installed around 6 of them myself and have many miles on them with no issues with any of them.
The key aspect I went for with my design was a lack of stress risers. The thought of any metal parts breaking off and going through the transmission frankly scares the bejesus out of me.

z driver 88t

Greg - one thing, the nuts / bolts you saw on RETs repair was just the mockup. His final solution used steel rivets which he documented in his thread. I have PM'd him to see how it has lasted but not yet heard back.

I also spoke with a gentleman at Perfex Manufacturing. I don't think I am allowed to link to non-Rennlist vendors, but if you google "928 Shift Spring", it is the first thing that comes up. He is working on a similar replacement that he has tested on his track car and daily driver, but it is not ready for sale yet. Still in R&D. He voiced similar concerns over welding on spring steel.

Colin - I have talked with Roger at length and plan to place an order with him. Thanks for all your design efforts on this. I understand the return on investment for 928 parts is pretty small. This is really a labor of love for many of us. In truth I would have been more comfortable with riveted connections to avoid the weld-on-steel issue, but between the six you have installed and the 5 Roger has sold over the past six months (plus other from the past, with no reported failures) with no more OEM springs out there, I'll roll the dice.

The truth is, it seems, in this instance there is an inherent design flaw in the overall layout and there does not seem to be a perfect solution. Sort of like the IMS problem on 911s. Porsche kept trying to improve it but in reality it would be best to go a different route altogether.

Lastly, I have received multiple PMs and emails from many of you with detailed information and feedback. Thank you all for your input. I emailed Greg today that without this forum 928 ownership would be completely impossible.

So...once installed I will report back. If it grenades, you will all be the first to know.

RET

My repaired spring has been holding together so far... or maybe it is falling apart but I just haven't noticed yet....

It has been two years since the repair and quite a few miles. FWIW, I have driven the car nearly everyday since (except when it was getting repainted...). I should probably pull the transmission and take it apart to see how the spring is doing, but I am afraid that would go against my nature: I haven't yet put everything back together since getting it back from the paint shop: I only put the front spoiler on last week, the wiring for all of the interior lights simply hangs in the openings because I haven't yet sorted out the part numbers for the better VW lights, and the upholstery panel for the sun roof still sets on the garage floor, etc., etc...

The original part is an incredibly poor design: square corners of the hole in the actual spring that the cross piece is staked into, and the spring tapers from the base to the cross piece, guaranteeing that more stress is put right into the weakest area.

I could not find a new part, I didn't want a used part (already had one of those, and it was a failure waiting to happen), and I could not bring myself to use a welded spring for all of the reasons given in other posts: unless the part is properly heat treated after welding, it is no longer a spring. I do not have the proper equipment for heat treating, and the brute-force-and-ignorance methods that I have used to temper small springs from time to time did not seem to be good for this large (and critical!) part.

The production of the part was by trial and error. I made a mock up out of brass with the broken cross piece hot glued to it to take measurements while in place and made things up as I went along.

The idea was to have the new spring flex in an even arc rather than the parabola of the original part, and I did not want to fasten the cross piece in any manner similar to the original either, so I tried to give it a bearing surface across the entire width of the new spring without also setting up a stress riser.

There is little room in the case for the spring to either be longer or thicker and the cross piece can not be shorter, so several test assemblies ended with the part shown. The bolts were used for ease of disassembly from the lower machined casting as I did not want to abuse it by grinding rivets off (FYI, the original spring is held on by rivets).

I might do things differently if I had to do it again, but I am pretty confidant that the new spring won't crack, my brazing won't fail, and the rivets won't pull out. However, if something does go wrong, I hope that the added hairpin clip will prevent a catastrophe.

Bob

GregBBRD

That sounds like a great plan.

Like I tried to say....Colin is a very bright guy and has considerable welding experience. Additionally, it appears that his track record on these pieces is very good.

And like anything that is made from metal and routinely flexes.....it will last until it breaks....which is inevitable....but hopefully many years and miles down the road.

UncleMaz

Great thread z driver.

I am pretty sure I have used up any goodwill on the subject, but I would feel much better installing Colin's spring if the detent profile matched that of the early spring or the updated lever was not on backorder from Porsche.