Re-engineering the F125, a philosophical discussion

[HPsauce]

So, we all know the F125 is a bag of worms, an electromechanical unit, totally critical to the vehicle working, placed in the most ridiculous location on the underside of the vehicle and responsible for many A8's being scrapped in the past with apparent "gearbox failure".

My brain has been saying for some time "There must be a better way", but what?
I "think" I (definitely) know how it works and (mostly) what it does, but I'm not really sure why it's exactly where it is?
Most of it seems to be signalling (to various places) what the gear shifter (in the cabin) is doing. But I don't know enough of the mechanical and electrical links from that to understand why the F125 is where it is and whether it (or parts of it) could be elsewhere or implemented differently.

I can see there's potentially a difference between what the shifter is set to (requiring) and what is ACTUALLY going on in the gearbox, but does it all need signalling from the side of the gearbox? How much is mechanical and how much controlled via ECU?

e.g. I assume there is a certain amount of mechanical linkage from the gear shifter to the shaft on the gearbox to enable/command specific functions, is this just P (R?) (N?) and D and do the ECUs know what's going on independently of the F125 or not?

Anyone interested in discussing this or am I on my own? It just seems to me that the F125 incorporates a lot of functions that just don't need to be there, especially on cars with the later PRNDS setup.

[MikkiJayne]

So the F125 is attached to the actual shift rod on the transmission, not just put there because a ZF engineer's wife once slept with a mechanic. The shift rod is connected to the valve body and at its most basic level it controls the fluid flow to the forwards or backwards bits of the valve body, or shuts them off, or locks the parking pawl in place. Those functions inside the transmission are purely hydro-mechanical. It will drive purely on that mechanical input without even having a TCU plugged in to it, although as most of us know, it will slam in to gears and only have 4th gear. Limp=mode is the TCU effectively just giving up and letting the valve body do its default things.

Does the F125 need to be down there? Not even slightly. From an electrical perspective, it would make no difference whatsoever if it were part of the shifter mechanism and inside the car. I think ZF designed the F125 as part of the transmission (since the 5HP was used in loads of other stuff including Jags, BMWs, Range Rovers) and left the manufacturers to do their own shifters, therefore the F125 was attached to something in ZFs control. The shifter is attached to the shift rod with a chunky bowden cable. Again, this was left to the manufacturers to implement, along with their chosen shifter design.

Could it be moved? Theoretically yes, the functionality could be achieved by something attached to the shifter pivot or potentially a link rod attached to the bottom of the shift lever which actuates a separate mechanism. The optimal solution for a switch would be an optical encoder, thus getting rid of the frangible metal contacts. That might need some electronics to interpret the signals, or the TCU might cope with an optosensor grounding its inputs. However, design and construction of such a thing in a reliable format is likely beyond the home gamer. There's virtually no space to play with so it would be an extreme packaging challenge. Not impossible, but not economically viable to do on a commercial basis.

I have considered an optical replacement for the existing F125, but again it is extremely tight. Alas I don't have the time or energy any more to devote the hundreds of hours such a thing is likely to need to perfect. We're pretty much at the mercy of the Chinesium clone switches now that ZF have forsaken us with original parts.

[HPsauce]

Thanks MJ, I was thinking of Hall Effect sensors rather than optical, but to achieve the same things. Really just a "Gedankenexperiment" at this stage.

I've never looked at (and never will) the mechanical side of the shift rod and valve body, but how many positions does it actually have? e.g. on a PRND321 car do all those positions move something in the gearbox directly or does it stop somehow at D and the rest is by ECU electronics? And what happens in a later PRNDS version?

Your wording suggests that the F125 is a ZF originated (or maybe just specified?) unit, if so are there versions still available from other manufacturers that could be used/modified as an alternative to the Chinesium ones?

Edit: This seller on eBay https://www.ebay.co.uk/str/bandhservices18 actually offers a reconditioning service as well as selling them! XYZ switches!

[MikkiJayne]

Mmm hall effect would also work, enclosed solid state stuff.

The shift rod has PRND for mechanical positions in the transmission. This directly controls fluid paths inside the valve body to hydraulically select a gear (ie engage clutch packs to make it go). The TCU then uses its solenoids to modify those fluid paths, so engage clutch packs smoothly by slowly ramping up pressure, or apply pressure to the clutch packs in different orders for different gears.

S, or 432 are purely done in software. The shifter cable underneath stops moving once you get to D, and the lever simply engages switches in the shifter to select different programs in the TCU. You can put a PRNDS TCU in to a PRND432 car (we've done it on Mike's Murple), and S works perfectly. Just don't select 3 or 2 as the S TCU has no idea what those mean and crashes Also some early FL2 clusters which were PRND432 from the factory don't understand the 'PRNDS' message from the TCU and just go blank.

(I have no idea about the ZF 4HP in the PF, which had PRND321, but it was probably the same or similar idea)

IIRC there are other F125s which look nigh identical to ours but with different wiring. Can't remember from which model though. If it was BMW then they will likely still be available since BMW doesn't shaft its classic owners like VAG does. Looks like that company on ebay is doing exactly the same as I do with them - open them up, clean / replace the contacts, and bolt them back together. I've never proven how long that works though so I never sold 'reconditioned' switches in case they broke again after 3 weeks. No big deal on my own car, but a bit of an embuggerance if someone paid a garage 5 hours labour to swap one.

[HPsauce]

The BMW units e.g. BMW part 24107512755 look very similar BUT have a round connector with only 5 contacts in it.
This suggests they are only doing certain functions, or in a very different way, maybe some are signalled from the stick in the cabin?
If I could get hold of a defunct one it would be interesting to take apart.

[HPsauce]

So there are several videos online of people taking the BMW units apart and it's totally clear that the 5 wires connect to the gear recognition side of the unit which just signals to the ECU.
The other side (start relay and reverse) have no wires and no contacts on the arm, though the tracks are there. So now we know, they're simpler.
In fact they're probably over complex (for historical reasons) as they're only really telling the (later) ECU if it's P R N or D.
See attached blurry still from a YouTube video:

BMWrangesensor.png

We already know the starter issue can be overcome by replacing the relevant relay in the footwell with a jumper.
Not ideal as it's not inhibited when in gear, but I ran like that for ages without a problem, and "back in the day" all cars were like that .

Reverse can probably also be implemented another way if necessary, though with rather more effort. If all else fails a switch to turn on reversing lights manually could be fitted. I think the R signal is used by the SatNav (but who cares?) but probably not the gearbox ECU, that will use the "5-wire" information presumably.
On our old cars reversing lights are not required for MOT purposes, nor are headlamp washers.

[HPsauce]

Interestingly Jaguar part 0501320754 seems to have all 9 contacts and a rectangular connector very like the Audi one.
HOWEVER a Jaguar forum shows an XK8 one taken apart with very different internals using microswitches: https://www.jaguarforums.com/forum/xk8- ... ir-270183/ a design which I've also seen on some Land Rovers! ZF part 0501 209 953.

BUT the Land Rover Discovery 2 has Audi-like internals but a short lead:

Disco2switch.png

Again a blurry still from a YouTube video.