Open Road

This is a G28/11 with factory LS on the 86.5.
Currently off the car for other updates.
I was going to clean and replace gaskets.
I have eye balled internals. It looks good from the top but more metal on the plug magnets then I like.
Pics of drain plugs show about two years of seasonal track events since last gear oil change.
Any words of advice while it’s open and WYAIT?

System related, Hans short shifter and front link delete will be going on.
When looking at rear bushings, I saw there’s a tighter set available for the rear shift coupler.
The stock set always had an oval shape to permit some slop IMO.
The new option is a circle. Has anyone converted to these yet?
I switched them out today. Easy with the shop press.
End result, very tight. We’ll have to see if this of benefit when everything is assembled.

Gathering up all the special tools and knowledge to rebuild gear boxes is on my todo list.

You do seem to have more “big shards” on your drain plugs. I always see more-fine particles, more of a metal sludge.

I’ve used the split plastic bushings to take up the ‘oval slack’ in the rear coupler when still using the OE front rod.

With the Hans’ Shifter the slop reduction is very noticeable. I’ve not messed with the oval in the rear coupler when installing Hans’ bits.

IIRC there were discussions on TOS that you need a smidge of slack on the assembly.

Yeah, that was my concern after seeing how tight the new version is. The original play was probably there for a reason.
Pending results I guess it can be undone without alot of fanfare.

I bet each sliver on that plug is from each time I chirped the transmission coming out of a certain turn at the local track.
I think I finally figured that turn out.

I have no comfort touching the internals in the box.
Some of that is undoubtedly the years of indoctrination from GB’s right or wrong scare tactics.

Crumpler wrote: ↑Mon Dec 16, 2024 9:47 am I have no comfort touching the internals in the box.
Some of that is undoubtedly the years of indoctrination from GB’s right or wrong scare tactics.

I’ve sent two to GB that were rebuilt (and then shredding themselves) by others.

“Did a blind monkey rebuild these Dave?”

I’ve talked to him about the issues with those. The main thing is having a process whereby *everything* goes back together exactly as it should. That level of meticulousness is rare and it is something that must be cultivated. But, I’m pretty sure I have the process already. Next is having the correct tools and not relying on sledge hammers and pipe wrenches…

There are a few “special” tools required to replace the stub axle seals. But, they are simple enough to make or cobble together if you don’t want to pay Porsche/Snap-on.

The biggest “expense” from a tools POV is that associated with messing with the diff.

I, in turn, will be happy to send any and all transmissions (and future TT rebuilds) to you :)

How about lubricant?
I’m sure the metal I’m seeing is from the driver but I was hoping to try something else in terms of gear oil this time.
I had used RP 75w-90 max gear last time.
What say we about Swepco? 202?

Crumpler wrote: ↑Thu Dec 19, 2024 9:17 am How about lubricant?

Yes, often I wish they’d use lubricant. IRS in particular.

Oh, wait, you mean in the transaxle?

Crumpler wrote: ↑Thu Dec 19, 2024 9:17 am I’m sure the metal I’m seeing is from the driver but I was hoping to try something else in terms of gear oil this time.
I had used RP 75w-90 max gear last time.
What say we about Swepco? 202?

Seems that M1 has been reformulated in recent years and does not work as well. That, speculation until I switch to something else. Me, personally, I’m going to try the Greg Brown juice real soon.

https://www.napaonline.com/en/p/SL_SL2463

For you… while it’s out, I would figure out a means to get gear oil temperature data to the cockpit. If you are cooking your gear oil you need to know that. Chances - IMO - are that you are cooking your gear oil.

Also, as bits inside wear, waste heat production goes up.

More frequent fluid changes or active cooling might be indicated.

Factory put a gear oil cooler in GTSs.

Interesting discussion.

Noticed the change in M1 75w90 as they have taken GL4 off the label. (Was looking for a GL4/GL5 for the gear/diff side of the PDK)
Looks like I will be using Valvoline Synthetic 75w90

Thank you for the oil recommendation, I will look at that.

The heat issue is interesting. I had not associated that with extra wear but it makes sense.
I did have plans for transmission cooling but it always got pushed by bigger issues.
Agree, this is the time.
So deeper dive, I looked at stock transmission coolers but always associated them with automatics.
I found a PET diagram of what looks like a late model manual transmission with the bar cooler.
I’m fuzzy on flow, there appears to be no pump in the circuit?


My previous cocktail napkin plan was to run -an lines from drain plug in front to pump, exchanger, back to fill port on back. Temp sensor great idea.
Better way so I don’t reinvent the wheel?

Crumpler wrote: ↑Thu Dec 19, 2024 1:00 pm I found a PET diagram of what looks like a late model manual transmission with the bar cooler.
I’m fuzzy on flow, there appears to be no pump in the circuit?

Pump is inside the back cover. Gear pump driven from differential:

http://www.autoatlanta.com/porsche-part ... ion=306-05

Crumpler wrote: ↑Thu Dec 19, 2024 1:00 pm My previous cocktail napkin plan was to run -an lines from drain plug in front to pump, exchanger, back to fill port on back. Temp sensor great idea.
Better way so I don’t reinvent the wheel?

Trying to recreate the OE solution is disrecommened. Many thousands of dollars in parts and some of them NLA. You’d start by throwing away your A/C condenser, radiator, buying two different condensers and a three-cooler radiator and working your way to the diff making lines and hoses as you go. And then you need GTS gear pump bits.

Front drain to fill port with electric pump is as good an idea as any. Trick is: where to put the heat exchanger? Finding a spot in the rear that has actual airflow as opposed to vortices is the essence of that trick unless you add a fan and ducting. There’s room in the rear of the left rear wheel well for “stuff” and it’s an area that can have high(-ish) pressure airflow if you can vent it with ducting. On the other hand, since ‘racecar’ you can use hatch volume too.

worf wrote: ↑Thu Dec 19, 2024 1:32 pm

Crumpler wrote: ↑Thu Dec 19, 2024 1:00 pm I found a PET diagram of what looks like a late model manual transmission with the bar cooler.
I’m fuzzy on flow, there appears to be no pump in the circuit?

Pump is inside the back cover. Gear pump driven from differential:

http://www.autoatlanta.com/porsche-part ... ion=306-05

That’s actually as slick design.
Ok
I will start shopping for the aftermarket stuff.
The pumps that push oil are pricey I noticed already.

Some interesting notes on the OEM cooler design (MY 1992-1995):
- The wand (Item 8 from PET above) has specifically metered holes that spray cooled gear oil on each gear set and on the diff and as a result less gear oil is used in the transmission. The oil level overflow plug is lower on the rear diff cover.
- 1992 Early designs used a direct line (#7 above) from returning from the cooler to the transmission. When the car is not moving the gear oil in the cooler will "drain" into the transmission. At some point they changed the direct line to include some coiled loops to prevent this draining.

I know of at least to others that have used an external electric pump with fittings to the front drain plug and the rear overflow plug. I would expect some sort of metered orifice would be needed to prevent from pulling too much fluid out.

Benno wrote: ↑Fri Dec 27, 2024 11:27 pm I know of at least to others that have used an external electric pump with fittings to the front drain plug and the rear overflow plug. I would expect some sort of metered orifice would be needed to prevent from pulling too much fluid out.

Yup. Development would best be done on a bench with a plexi top cover.

worf wrote: ↑Fri Dec 27, 2024 11:49 pm

Benno wrote: ↑Fri Dec 27, 2024 11:27 pm I know of at least to others that have used an external electric pump with fittings to the front drain plug and the rear overflow plug. I would expect some sort of metered orifice would be needed to prevent from pulling too much fluid out.

Yup. Development would best be done on a bench with a plexi top cover.

If the cooling loop was filled, you would not pull any more out than you are putting back in.

N_Jay wrote: ↑Fri Dec 27, 2024 11:56 pm If the cooling loop was filled, you would not pull any more out than you are putting back in.

The proof is in the pudding. And the fill procedure. It’s fucking hard to tweak the internals when the box is back in the car.

Just pumping it into the back is not the best solution either. Ideally you want a spray on the R&P and a rail to spray at least 4th and 5th if not all but first. You also need to ensure that at high speed and temp you’re not just pushing foam around. Since in this application the purpose of oil is heat removal one shouldn’t rely on “splash” for cooling.

And data collection needs to be added.

Crumpler’s supercharged track beast should be capable of >500 ft-lbs of torque (even if not right now) and we know that that level is well into the safety margin of the non-cooled box. A fully-engineered actively cooled 928 box will shred itself into a milkshake in the 700 ft-lb range.

928 5-speed boxes do not grow on trees. If you ‘splode one it’s going to take a while and possibly $5k or more to get a replacement.

It needs to be engineered on a bench.

It all depends on if you are trying to improve the lubrication system, or just adding cooling to a transmission that otherwise is well enough designed.

Good discussion you guys. Thanks for following.
The parts are trickling in.

Happy accident that the transmission plugs are M22 1.5, common enough size that you can source adapters to 10 AN.
A nice pump arrived, looks to be high quality, compact with filter.

In line temp sensor with gauge in cockpit.
I had planned on that being pre cooler, but open to suggestions.
My plan was to fill lines and exchange and then normal fill line in case, so volume would be the same. Again open to input.
There were many options as to switch, based on temperature or manual. Everything else is manual toggle switches, so I continue that theme.
Basically, turning it on for track sessions.

I’m assuming this pump could push gear oil at cold viscosity but I don’t want to burn the pump out.

This car has always been a little of the “let’s see what happens” element. But even I was going to bench test the set up. I can reproduce all the variables except movement of the internals.

No idea about foam. It’s a good question. It was obvious that I had discharge from the vent on top, but I have no idea what amount is normal, and it probably wasn’t cleaned since it rolled out of the factory.

Curious, foam in an auto I'd Nad because the fluid is used by the hydraulic control system that would not work with the "fluid" being compressable.

In a manual, where most is splash-lubed, would a foam really cause any issue?
(Other than maybe pushing out the vent)

After watching these two videos, I get the impression that the fluid dynamics that happen in a rear diff require some sophisticated engineering or at least a lot of R&D to get right. You want to managing foaming, and heat.

Those videos are insightful.

Makes you wonder about a lot of things.
I guess I can establish, at least, that the pump circuit can keep the oil level consistent while running on the bench.

Heat testing on the bench be pretty anecdotal but hopefully prove it’s cooler with pump then without.
If I can establish those two things I’m prepared to to just go with it.

The rest I hope goes back to good factory design.
But, Obviously there was a concern if they started cooling in later model years, as performance numbers went up.
Looking at the late MY diff cover, it’s anything but smooth. But I assume the wand and delivery mechanisms compensate.

I trust that watching those videos has put to rest any doubt about the complexity of what's going on in (and around the outside of) the diff and indeed in any contraption that's more-or-less a blender half-filled with liquid.

Crumpler wrote: ↑Sat Dec 28, 2024 11:10 am In line temp sensor with gauge in cockpit.
I had planned on that being pre cooler, but open to suggestions.

Since the purpose is to cool the gear oil, one temperature sender won't tell you anything. You need two. You need one that sends fluid output temperature and one that sends post-heat-exchanger temperature. This will tell you
- the temperature drop you achieve and
- when the system fails

Crumpler wrote: ↑Sat Dec 28, 2024 11:10 am My plan was to fill lines and exchange and then normal fill line in case, so volume would be the same. Again open to input.

Ideally you want operating fluid level to be the same, but that's hard to figure out. Shooting for the nominal fill should be plenty good.

Crumpler wrote: ↑Sat Dec 28, 2024 11:10 am There were many options as to switch, based on temperature or manual. Everything else is manual toggle switches, so I continue that theme.
Basically, turning it on for track sessions.

Indeed before doing any "development" I'd want temp data from the unmodified stock system. But, that ship sailed. So...

I'd start with a manual switch. First I'd run it on the track to see what "uncooled" oil temperatures look like. It may be that you aren't over-heating the oil. You also don't want oil that's too cool.

IMO, oil temp in the 180 to 200 degree range is what you want, but you should do some due diligence on that.

Once you've acquired some data, leave the switch off on the warm up lap to heat the oil and then turn it on.

If you find that you are consistently "over-cooling" the oil then you'll need temp-switched activation.

If you find that you are indeed cooling the oil, but not enough, then you'll need to do some more engineering.

I suspect though, that toggle-on after the warmup lap and then just leaving it on for the whole session will get you reasonable control assuming you have good flow through the heat exchanger.

Ideally you'd do some calcs to size the HEX and fluid flow requirements. But, in this case, you'll be constrained on where the HEX can be located and thus on it's size.

Crumpler wrote: ↑Sat Dec 28, 2024 11:10 am I’m assuming this pump could push gear oil at cold viscosity but I don’t want to burn the pump out.

Yeah, well, if the pump will pump 75w hot oil but kills itself with 90w cold oil, you'll want to figure that out on the bench.

Also, note that it is a "Fluid Pump" and not a "Foam Pump" hence the need to have some idea of the aeration of the oil. It's possible that you may need some accommodation for air/oil separation. The pump will NOT HELP with aeration. It will make it worse.

The factory GTS system with it's 20-ish feet of hoses and two different heat exchangers probably also has some separation functionality.

Crumpler wrote: ↑Sun Dec 29, 2024 10:03 am I can reproduce all the variables except movement of the internals.

Actually, I think you could reproduce internals movement. With no bit's hooked up to the output, very little toque is required to spin the box. Now, a mechanism that would allow you to reproduce the full 900 to 6200 RPM range might be difficult, but you could get *some* movement.

Crumpler wrote: ↑Sun Dec 29, 2024 10:03 am No idea about foam.

I think you do now? Yes? After watching those videos?

Me... after watching those videos, I think the stock diff cover could be *vastly* improved upon both in terms of internal shape for oil control and external shape for scooping up air and directing it over the back of the cover. I thought that aspect of Bank's design was really cool.

Crumpler wrote: ↑Sun Dec 29, 2024 10:03 am It’s a good question. It was obvious that I had discharge from the vent on top, but I have no idea what amount is normal, and it probably wasn’t cleaned since it rolled out of the factory.

Discharge from the top vent is obvious evidence that the box got filled with gear oil foam.

Crumpler wrote: ↑Sun Dec 29, 2024 10:03 am But, Obviously there was a concern if they started cooling in later model years, as performance numbers went up.

What kills a gear box is a) heat b) too much torque. In the second case, if you're putting 1000 lb-ft into an input shaft that breaks instantaneously at 650 ft-lbs, heat removal isn't going to help. But, if you run at 500 for an extended period of time with a system that was designed for 300 ft-lbs you will also kill the box because you're putting heat into it faster than it can dissipate the heat.

We don't know exactly why the factory added gear-oil cooling. We can speculate but we don't know for sure. Sure the theory makes sense, but we don't have data. And in any case, you beast should be putting WAY over GTS torque into the box.

For you Worf , I will also spin the box by hand. ;)

I’m not seeing multiple temp sensors in my design.
The engineers approved of IR readings however.

Dave, I showed the engineers your recommendations. They rounded for a while, but then got into some debate over Bernoulli’s theories.
I finally went to bed.
It didn’t sound like they wanted a second temp sensor.

Crumpler wrote: ↑Tue Dec 31, 2024 11:03 am Dave, I showed the engineers your recommendations. They rounded for a while, but then got into some debate over Bernoulli’s theories.
I finally went to bed.
It didn’t sound like they wanted a second temp sensor.

A good product manager knows when to override the engineers… and when to give them treats.

Bench testing the system went well enough.
Used a hot plate under the case to get oil to around 105 F.
I have videos but they are hard to focus on fluid movement.
The oil level stays constant, with hand spin and without.
I measured fluid rate by disconnecting the fill and capturing in a pan. Entire volume (around 4L) of oil cycles in roughly 20 seconds.
In the video below you can get an idea of rate and baseline foam with pump movement


https://youtube.com/shorts/FUNER3aYkss ... I3VRAXBnLJ

Crumpler wrote: ↑Sun Jan 05, 2025 2:59 pm Bench testing the system went well enough.
Used a hot plate under the case to get oil to around 105 F.
I have videos but they are hard to focus on fluid movement.
The oil level stays constant, with hand spin and without.
I measured fluid rate by disconnecting the fill and capturing in a pan. Entire volume (around 4L) of oil cycles in roughly 20 seconds.

That sounds a bit fast to allow for cooling.
The radiator or heat exchangers need to hold 6L to give it just 30 seconds of cooling time.

It does seem fast.
I guess I will see what the numbers look like in real life.

Crumpler wrote: ↑Mon Jan 06, 2025 3:34 pm It does seem fast.

it’ll be ok if you have enough airflow.

But, you won’t know how well it works, or doesn’t.

worf wrote: ↑Mon Jan 06, 2025 6:34 pm

Crumpler wrote: ↑Mon Jan 06, 2025 3:34 pm It does seem fast.

it’ll be ok if you have enough airflow.

But, you won’t know how well it works, or doesn’t.

Heat transfer takes time.

This was a common problem back in the day when people would take the thermostat out of their cars and end up overheating.

Interesting pic from Hacker showing the set up on one of his cars. All the appropriate squirters going to all the right places.

I’m not seeing an cooling there?

worf wrote: ↑Sun Jan 12, 2025 12:46 pm I’m not seeing an cooling there?

Another pic. Custom exchange with electric fan tucked away.