996 C2 cabrio roof issues

Hi all, I'm in the middle of trying to sort out a cabrio roof issue on my 98 911. After the hard top was taken off from the winter, the roof didn't want to come up. I had to do an emergency raise of the clamshell by turning the Allen screw in the potentiometer of the clamshell drive anticlockwise to raise the clamshell. I've ruled out a few things and have replaced both side flaps are neither were in good order. I've set the pot2 on the clamshell to 6.15Kohms (As close as I can get to 6.2Kohms) and PIWIS reports voltages within the ranges needed. The clamshell can be raised and lowered with PIWIS but it only does an inch at a time, before it stops with "Function Inactive, end point not reached". Go back and perform the drive link test gain and it moves another inch and so on. The mechanics must be OK, but something is stopping it. Checking out the input values, all seems to check out, compartment lid open/closed, flaps up/down, handbrake, engine lid, latch, etc except now I've had to raise the roof manually, by releasing the pressure in the pump, it's locked and all input values are as expected, apart from the roof thinks it's down when it's up and I can't raise the rear windows, only the fronts. Anyone know where the roof position is measured from on the early 996? Is this something to do with the round blanking plates in the rear cards above the speakers? There is also a procedure I've seen about resetting the latch mechanism on the roof, so all the cogs line up, but this is only supposed to be needed after an emergency open, which I didn't do, only a clamshell emergency open, which I've replaced the parts for with a kit. The roof itself operates, but stops about half way up or down, presumable because it can't detect the position. The POT1 values, which I think are the roof position ones are 0.86, 1.37, 2.00, 2.45, 3.16, 3.41. A calibration halts with "Function Inactive, end point not reached".

If I was a betting man I'd put my money on moisture in the roof ECU and I'd be giving that a good blast with the hair-dryer.

Cheers, I'll take it out, pop the lid off and have a look for moisture ingress. I took the cabrio deck bay carpet out though and it was pretty dry, but you never know.

I had a problem with mine where the rear windows would go down but not come back up. Everything window related was fine. The only thing I could trace it to was the roof ECU. Left the car indoors for a day or 2 in the dry and everything went back to normal. ECU looked fine, but knowing how much condensation I get in the car over winter I could only put it down to moisture in it.

There is a way to fool the rear windows into rising, simply insert your finger into the roof latch recess overhead (just behind the mirror) on second press of the buttons the read windows rise up, even with the roof down. The roof can also be re-set easily, press and hold the roof button on the instrument console for 10 seconds and then turn off the ignition. You can hear the mechanism resetting itself to the zero position.

Yeah that wasn't the issue with mine.

What did you find out? I'm having a similar problem. I was out for a drive. It started to rain and my roof wouldn't go up. I had to manually put it up. When I got home I found the parking brake switch faulty. I replaced it and checked the fluid. The fluid was low so I topped it up. I tried to use the switches on the dash and the gear above me tries to move a little but that's all. If I extend the latch out and then hit the button it will fully close. But it still won't open. I hooked up my PIWIS to calibrate and I'm getting Function inactive, end position not reached. I metered the 3 micro switches in the latch and they are all testing fine. The gears are moving good too. I'm going to check the pots in piwis next.

With it being wet where I live, I've left it under the car port with the rear windows down. I also get "Function inactive - end point not reached", very quickly after trying any drive link test. All seems to function independently, all be it, an inch at a time. The input signal/status of everything seems OK, other than the top thinks it's down when it's up. My plan is to check the control unit for obvious damp ingress or solder damage, plus the POT1 which is a device on the LHS hydraulic ram and tells the control unit if the roof is up or down. I'll also check and top up the fluid as I've opened the bleed valve a few times now to raise the roof by hand. Both flap assemblies have been changed and the compartment lid POT2 has been set, but something stops it from working. Will keep you posted as the weather perks up over the next few days.

Did you check the microswitches in the latch and on the windshield? Mine are good in the latch but I haven't checked on the windshield side.

In PIWIS it tells me the latch is locked or unlocked correctly, so they must be working. Start looking at the input values and spot the one which is wrong for your current roof situation. I.E. is it reporting the flaps, roof position, latch, clamshell position correctly?

At the front of the roof and in front of the latch assembly is a plastic fork .. make sure this isnt broken off .. also with the tester look at the values for roof open and press this fork up and down .. its connected to a micro switch and there should be a reading for in in values . To be honest though this shouldn't cause the issue you have .. it doesn't work off the button at all i take it ? With Piwis you are forcing a function but even so it will still shut down if it sees a safety issue .. These are normally the CTL in the wrong place .. the engine lid open .. the rear side flaps in the wrong place .. What yours is doing atm doesn't make sense though .. a control unit fault is possible but there's normally a reason for that and its mouse damage or its water .. very rare for these to fail though . I'm going to have to ponder on this for a while i think .. possible option is you me your piwis at my house going forward .. can't guarantee it will be anything other than a waste of a journey as there's little i can do at home other than check values and actually see what the roof is doing .. It might be worth the trip though on the off chance as i don't charge anything

My gears had to be recalibrated as per the PDF I got. Did that then pulled out the rear gear and adjusted the pot 6.2k ohms and still am getting end point not reached. All the values are in spec on the PIWIS. If I hit the close button or open button, the locking latch cycles, opens, closes, opens, etc. I just can't get this thing to calibrate.

So when closed, PIWIS reports roof closed, latch closed, compartment lid closed, both flaps down, handbrake on, speed under 3mph (or whatever it is), plus all other pre requisites before the roof can be moved, such as engine bay lid closed? I'm in the same state as you, but my roof says it's currently down when it's actually up. I'd be steering towards the control unit for your problem. Maybe pop the lid off it and inspect for damp ingress or dry solder joints? I must get around to having another look at mine. I plan on checking the control unit, fluid level and having a look at POT2 on the left hand hydraulic ram, as it's currently reporting roof down when it's up.

It's completely dry in there. I never drive it in the rain or wet roads. It was working until the parking brake switch failed and I did an emergency close and haven't gotten it to work since. I'm convinced it's just out of alignment.

Quite a few forums I've been reading lead to dry solder joints in the control unit being a popular fault when people get this far in troubleshooting. Might be worth popping the lid off to check? I've left mine with all the rear decklid carpet and trim around the top and back of the rear seats removed, to keep easy access for now. It's wet today, but I hope to have another crack tomorrow. Art this stage, I reckon mine is down to either the POT1 roof position device on the left ram, the control unit or low fluid.

Some info ...

Very helpful. Many thanks! :thumb:

Big time. I night find away of modifying the speed restrictor signal so I can close the roof up to 28mph.

Handbrake signal on is pretty easy as its either an earth or not .. i think its an earth. Speed signal and i have no idea what it should be but i would disconnect the wire .. cut it basically in a position that i can solder it back together .. then connect a voltmeter to the incoming wire and drive the car .. If you get a 12 volt and static at 12 volts then it needs an ignition feed .. If the voltage moves up and down then it will be a PWM signal .. you can't reproduce that but as these are an earlier car i feel it might not use this type of signal . If no voltage then do the same but on ohms . Obviously these cars are not designed to have the roof operated when moving so at speed you would cause damage .. that's my disclaimer anyways . On a separate note .. i would love to know what the readings are on the speed signal side so happy to help anyway i can .

The smartop ECU you can buy allows it to be operated at 28mph.

Ahh i didn't know there was such a thing .. Checking the speed signal and if possible cheating it would be cheaper though

Well, the fluid was fine, but I topped it up anyway. The control unit had no signs of moisture on the board and I couldn't see any dry solder joints. I corrected a few minor things like a retaining clip on one of the arm joints had made a break for freedom, so I fished it back into daylight with a magnetic grab and refitted it. My issue seems to be with the clamshell motor. With PIWIS it refuses to open the clamshell, all you hear is the relay click then almost immediately you get the message - "Function inactive, end point not reached". Lowering it will lower it by an inch then stop immediately with the same error. Go back, do it again and you get another inch of movement before it quits almost immediately with the same error. Rinse and repeat until it gets to fully closed. I presume the clamshell POT needs setting to 6.2Kohms then refitting with the clamshell open fully, not closed fully? My issue seems to be when I set the POT and go to line it up with the motor, the shaft is slightly out for where the retaining bolt holes are, I then have to force it one way, to get the bolts in, or turn the shaft and POT slightly. I'll have another crack during the week and I'll see if the motor drives both ways from a 9V or 12V supply. I'm just not sure if it's some safety feature stopping the process as the "Function Inactive - End Point Not Reached" occurs almost immediately no matter what task I give the clamshell or calibration. I know the mechanical side of the clamshell drive is fine as I can raise it all the way using the old (bent from emergency lift of the shell) drive shaft, with a set of mole grips. I'm determined to sort this and don't want to just resort to using the hard top all the time.

This would be far easier if i could see the car but i'll work with what we have ! As per image below .. is this micro switch doing its job ? CTL Pot .. possible .. something doesn't sound right there .. the car sees an issue with the CTL operation it will shut down for safety reasons .. a closed Ctl And the roof wont open due to the possibility of damage .. It must know the CTL is open .. this is that micro switch and the pot on it to measure the opening position of it . I assume you have no fault codes at all ? there is also the issue of the roof saying its shut when its open ? .. cant remember off hand .

I'm pretty sure the microswith is working, as I can get the clamshell closed input value when fully lowered. Also, if I remember correctly, the central locking also beeps the horn if the clamshell catch is not fully closed. I'm going to test the motor in both directions today. I did have the input value telling me the roof was down when it was up, but this was after manually upping it by opening the bleed valve. When I opened it from the key yesterday, with the clamshell full open, it worked until half way then stopped. When the weather is better, I need to get out and write down a few things. Probably the most important is checking all input values again with the roof up and down. This way, I can keep a clear record of what is working and what isn't. With your document, I have a pretty good idea on how it all works now, I just still can't confirm if POT2 calibration on the CTL motor is done with the shell fully up or fully down. I might try both again, but pretty sure I've done this. It just seems odd that I can't raise the clamshell via PIWIS and lowering it stops after a very brief period.

Well, the motor seems strong and works in both directions for so many turns. It's possible that the POT is faulty on it, but it does show resistance going up and down as I move it, so it must be doing something. I've bit the bullet and seen a spare, identical control unit from the US on Ebay, which even after duties are paid, is £80 to my door. Can't really go wrong at that price as UK ones are £175 upwards, so I can easily punt out with my other Cayenne parts. Whilst I'm waiting for that, I'm going to have a proper look at the POT1 on the left hand hydraulic ram, which gauges roof position. Obviously, if that is wrong, it may stop the clamshell from operating as the roof needs to be either fully up, or fully down (positions 1 or 6) on that POT, for it to allow clamshell up or down. Getting there. Like all these sort of issues, by the time you finally crack it, the entire system has had a full on health check and service! :grin:

Great info. All my switches are good. But the finger switch seems to be backwards. Saying it's open when it's closed. I am going to pop the cover off and see what it looks like. I believe the part number is 996.624.122.00

YPVS let me know if that works. I'm starting to think mine is faulty. The motor does spin when I put voltage to it but the pot doesn't move when I apply the power. But the pot's resistance adjusts but it doesn't move the lid at all. The top cable doesn't move at all.

If the POT on the clamshell motor isn't changing resistance as the motor turns, is it getting driven properly? I had to buy a repair kit for mine after opening the clamshell using the emergency procedure. The kit was the square drive spline with the plastic teeth thing which drives the POT, a spring and retaining clip. It was about £50 and a genuine Porsche part. I'll check my microswitch on the clamshell drive when I get a minute. It could well be a reverse switch, as the spoiler has one of these. Closed circuit until the switch is pressed, then it breaks the circuit. You'd like to think, as the only sensor on the clamshell drive is the switch telling the control unit that it's closed fully, that this position is where you would set the POT at 6.2Kohms. I still can't confirm for sure if it should be set with the compartment lid fully open or fully closed.

I love this info. I directly applied voltage to the drive motor. The spline gear is moving fine but the square centre pot isn't moving at all. When I checked the resistance after, it was still at 6.35k ohms. But if I manually move the lid, the lid moves fine without any binding and the connection where the motor goes spins. So I think the drive mechanism is fine. Although the top cable where the microswitch doesn't move at all. Maybe that only moves when the motor is driving it? Do you remember the repair kit part number you bought? Maybe I need that?

The compartment lid latch cable should definitely move when you drive the black gearbox. You'd think this would work the other way, by moving the lid itself. I use the old, damaged square shaft to manually drive my lid gearbox with a set of mole grips. If you don't have one, they look about the same size as the square shaft which drives household door latches on interior doors. You could certainly modify one to prove the drive is OK. If not, pop the black gearbox open. I guess it's a sort of rack and pinion setup, much like the spoiler drive system.

There are 2 repair kits, pre 2003 and post. Pre https://www.ebay.co.uk/itm/183896934521?hash=item2ad11c8879:g:E5gAAOSwrtFbUJ~i Post https://www.ebay.co.uk/itm/192602959382?epid=1516398578&hash=item2cd807ee16:g:E5gAAOSwrtFbUJ~i I also had a quick look at the clamshell latch cable driven microswitch and you are correct. Mine shows open circuit when the switch is operated and closed circuit when the switch is not pressed.

logically you would instal the CTL potentiometer with the CTL fully open ( calibration 4 ) and at 6.2 K .. use the below image plus the previous one i posted on the positions . I have however been unable to find anything proving this .. it depends on whether the Pot resistance goes up or down .. if down then yes ( voltage increases ) .. if up then no and it's the opposite and is fitted with the CTL closed . It sounds to me like the potentiometer might not be working .. the control unit needs to see an altering value .. you can test this with an ohmmeter on the pot .

If you have a problem with the latch cable, I have a brand new one i bought from Porsche but did not need (thank you so much supposed Porsche Indy technician...not).

How did this happen?, ah must have pressed the quote button]

Aha, you must be psychic! Yes, my plan for tomorrow, when my 9v batteries have charged over night, was to turn the motor anti clockwise a few revolutions and see if the POT resistance increased or lowered, as anti clock is raise CTL and clockwise, lower CTL. I figured, higher resistance would equal lower voltage, so your helpful doc should confirm whether the POT should be calibrated with the CTL fully up or fully down and locked. Cheers for the document. That saved me a job!

I am so impressed with all the help. Well dummy here skipped a step putting the pot back in. I tightened it snug. When I was supposed to put it in with 3 turns. I pulled out the unit and did the 3 turns then put power to it. I put voltage to the motor and it spun as well as the centre pin popped into place. I am going to put it all back in later on tonight. I'll post my findings.

I'm leaning towards the control unit now. I've calibrated the CTCl POT again and reconnected with the top lid up. Still won't calibrate and all input values look correct, other than the roof position. I've checked the POT on the left ram and the arm is connected and nothing is snapped. The POT values are in the correct range for the position on PIWIS. With roof up and clamshell down, PIWIS shows 0.98V for the convertible top POT and 1.24V for the CTCL POT. The stored calibration point 1 values (roof up clamshell locked) are 0.86V for POT1 (roof) and 1.57v for POT2 (clamshell). So the current readings close to the previously calibrated and stored values. Obviously I can't recalibrate as, like all other actions in PIWIS, within 3 seconds of pressing F8 it stops the action with "Function inactive - end point not reached". The only thing of concern which I still have to address is the RHS flap seems to be stuck down and didn't rise with the other when the roof was lowered, which it sometimes will do after stopping half way, either from the key of button. I'm a bit annoyed with this, as I also bought a second hand LHS one from the same Ebay seller and after fitting that one I noticed it was missing the plastic sliders which connect it to the ally rails, so have had to remove and rebuild from my own spares and refit. It now looks like both were faulty. If this one's motor is shot, at least I have spare assemblies with other faults to nick parts from, so for £50 each, for the entire assembly, there are plenty of parts there. The replacement control unit is on it's way from the US, so hopefully that will resolve my issue. I can repair the other flap in the mean time. On the plus side, my knowledge of the 996 sort top system has improved dramatically over the last few weeks, with plenty of help from DeMort and his pretty handy Service Manual pages, which I now have printed and kept in the car to reference. :thumb:

Thank you for helping me re-evaluate why i bother coming here .

The quest continues... I've repaired the duff flap and it works as it should. Everything seems to operate as it should except anything in PIWIS stops almost immediately after starting with "Function Inactive - End Point Not Reached" and the top shows it's not raised even when it is at fully up at it's POT shows a voltage in Value 1 (fully raised and locked) range. Halfway, the roof seems to show the correct description in PIWIS of roof raised, but when allowed to fall completely, it still seems to think it's raised. The thing is, both Potentiometer voltages are falling pretty much where they should be for all 6 stages of the roof and all 4 of the compartment lid. I've moved them about and checked voltages in PIWIS, so both POT's are working. Once up, I can use the button on the dash to lower the roof, everything does it's job, the clamshell raises, latch unlocks, roof retracts, but this stops half way, with both POTS at 2.51V. From there it sits with relays all around the car clicking when you try to up or down the top from there. If I manually raise it a bit by letting pressure out of the bleed valve on the top pump, and pulling the top up. I can then tighten the valve and using the button on the dash, it will close, latch the roof and lower and lock the clamshell. Very odd. I'm getting to the end of my idea train now, so may have to wait for the control unit to arrive from the US. It's cleared export customs and will be on a flight soon.

How did you make out? I took a step back. But I went at it tonight. I did the pot steps all over. I got it to 6.2k ohms. Then put the pot back on the motor. But the next step after re-installing the motor is to manually close the lid. Well I can't, it's on fully extend. I'm just letting the blue loctite dry and see if it will calibrate itself. Unless you think I missed another step. I know my problem is in motor on the lid.

New update. I'm getting somewhere but not too far. I got the lid to close and open with the button on the dash. The switch in the windshield had a broken wire. So here is the new dilemma. When the top is up the piwis shows it's closed. When it's down the piwis shows it up. The roof won't operate on it's own. I have to manually close it or open it. I do hear motor noise but nothing is happening. Do I have to release pressure on the pump?

You've checked the obvious, like fluid level in the hydraulic pump unit, I take it? If your tops says it's still up when it' down, what are the POT1 readings in the "actual values" in PIWIS? It will show the current voltage of the 2 POT's, then list the currently stored calibration voltage values 1-6 for the roof and 1-4 for POT2 (clamshell). This tells you where the POT is telling the car it's currently at. Mine is odd though. The POT is showing within the range for closed/up when it lists the roof as down. Going by DeMort's document on the last page, 1 is fully closed/up for the roof, 6 is fully open/down. The clamshell is 4 for fully up and 1 for fully down.

I topped the fluid up after the emergency closure. So that fork/finger switch seems to be reading in the PIWIS backwards. When the roof is closed, the PIWIS shows it opened. If I push up on it with the roof half open, the roof says open which is wrong. If I let the switch down then PIWIS shows the roof closed. Something is backwards. So when the roof is trying to open or close it would do the opposite yet it's not moving.

Are you talking about the fork on the roof latch? I'm on about the POT which is on the left hand ram and driven by an arm connected to a pin on the ram. It's position is supposed to tell the control unit what position in a range of 1-6, the roof is in. PIWIS will show it's current voltage and the last 6 saved voltages for each position. Mine is showing around the expected voltages for the various roof positions, but PIWIS incorrectly says it's up or down, when the POT readings look correct. I wonder if I need to look at my latch mechanism? I never did an emergency open of Roof latch itself, only the clamshell door. Because of this, I've never set the dots on the cogs. Maybe the roof latch also reports position? Unless I'm Reading your post incorrectly, we may both need to look at the latch.

I adjusted mine in the gear and I definitely got farther along. Make sure you check the switch on the windshield part too. Now I need to see what you mean about the clamshell switch in the back.

It's not the clamshell POT which determines roof position, it's a POT on the left hand hydraulic RAM. To get to it, you really need to take off the rear interior panel to the left of the rear seats. The position of this POT is moved by an arm and a pin on the ram, so as the ram moves, the POT turns and the voltage changes. PIWIS can show current voltage of both this POT (and the clamshell POT). You can move the roof to various stages of opening 1-6 and see if the current voltage falls within the appropriate of the 6 saved calibration values, which might range from 0.8V to 3.5V or similar. One of DeMort's documents on here shows the 6 positions of the calibration values.

My pot 2 1-4 are all reading the proper values in my PIWIS. But I'll look at removing that panel and taking a closer look.

POT2 1-4 is the clamshell. POT1 1-6 is the top. I think we're at a similar point. My POT's are displaying voltages pretty much in the range for the calibrated values 1-6 for POT1 (roof) and 1-4 POT2 (shell lid). I can raise my roof perfectly from half way and can lower it to half way (the stop point is when POT1 and POT2 reaches 2.51V. After this I need to raise it a little bit by opening the bleed valve and raising a bit before tightening it back up. It will then raise and lock as normal and the clamshell comes back down and locks as it should. The fork you mention on the roof latch, does the same to me. When hanging free and pushed down by the spring (down) PIWIS sees the roof as "Convertible Top Raised". It does this even if I fully lower the roof by letting the pressure out of the pump. If I then pull the top up to half way, it works on the button, but as soon as this fork hits the top of the windscreen and is pushed up into the latch, PIWIS immediately shows "Convertible Top Not Raised". It then latches up as it should, all windows rise as they should and all values in PIWIS look correct. My POTS are reading correctly, the shell and roof can operate correctly in both directions, but stops when the POTS go higher than 2.51V even though the POT values are in the calibration ranges with the roof and clamshell in all positions. The fork on the roof appears to be reading the opposite of what you'd expect. I'm going to Google the fork issue and re-read DeMort's docs as now I understand more, something may make more sense.

Wow! You and I are at the same point. I'm going to re-read the two pdf's I have and see if there is something we are missing.

How do I release pressure in the pump?

There's a little Allen screw on the ally plate, on the right hand side of the front facing edge of the pump. This is a little bit to the left of where the clamshell screw would be. Just loosen it half a turn or so and you'll be able to lift or lower the roof manually. Just nip it back up then. It's standard thread. Anti clock to loosen. Mine didn't seem to lose any fluid at all over the many times I've done this to move the roof about, once it stops at half way down.

Got this from a member on Rennlist: Don't believe the finger switch is the issue or pump pressure .There is a top position sensor 996 613 31302 that may not be reading correctly .I know nothing about piwis but it might be one of the sensors that can be recalibrated .

Wow did I get far tonight. I did see the pot arm for the roof in the side panel but it looks like it's working. Have to figure out how to calibrate it. So when I hit the close button on my dash. The latching hook locks, the clam shell closes, and the windows roll up. When I hit open, the windows roll down, the clam shell opens up, and I hear the pump but that's about it. The roof doesn't move. Obviously back to the pot readings being backwards. So here is my readings. Pot 1 - 1.04 V Pot 2 - 1.25 V Calibration Pot 1 Value 1 - .94 V Calibration Pot 1 Value 2 - 1.45 V Calibration Pot 1 Value 3 - 2.10 V Calibration Pot 1 Value 4 - 2.55 V Calibration Pot 1 Value 5 - 3.25 V Calibration Pot 1 Value 6 - 3.53 V Calibration Pot 2 Value 1 - 1.59 V Calibration Pot 2 Value 2 - 1.88 V Calibration Pot 1 Value 3 - 2.06 V Calibration Pot 1 Valve 4 - 2.78 V

Your POT values are pretty much the same as mine, give or take .05V. The current readings of 1.04v and 1.25v also look correct for the roof to be up and locked (POT 1 value range) and the clamshell down and locked (POT2 value 1 range). If you move the roof and clamshell to a different position, you should see the POT voltages go up, moving it into the corresponding 1-6 or 1-4 calibrated range. Those values as far as I understand, are the last calibrated values of the POT's.

That part is just the POT on the hydraulic RAM - POT1 The voltage from this lets the control unit know where the roof is as the voltage value should go up and down the 1-6 POT values as the roof is moved from fully stowed open to fully closed and latched. https://www.deroure.com/diagrams.asp?MAK=3&MDL=25&TBL=2673&SMA=0&SMO=0&ST=99661331302&SC=1

This looks very interesting for both of us. When the roof stops halfway during going down, the position it stops at is meant to show which microswitch has missed it's signal. Most seem to be in the latch mechanism. https://www.renntech.org/topic/44431-996-convertible-top-1999-cabriolet/ Mine stops at 2.51V POT1 value on opening, putting it on mine, just into POT1 Value 4 - "Convertible top at middle/rear". This looks like the 3rd example down for me and the Microswitch next to the rear of the motor in the latching mechanism. The fork on my latch definitely is back to front in PIWIS, although it does seem to latch and unlatch properly. I'm going to have a crack at removing the latch and checking the obvious this weekend.

Slow progress. Rain stopped play. It made no difference taping the plastic fork on the latch (which appears to be reading backwards in PIWIS, i.e. "Top Not Raised" when it meets the windshield and is activated, then "Top Raised" when it's fully down). I also noticed that the actual value for "Speed" in PIWIS measures "Speed > 5KM/H (3MPH)", which is odd when I'm on the drive, parked up. Apparently greater than 3MPH will stop the top working. I'm beginning to wonder if this Chinese clone is reading a few things back to front. Googling results from others shows people reading > 5km/h though. I'd love to test the flaps again, but from PIWIS any action you try to take withy the roof lower than POT Position 4, half way down, is met with a quick click of the control unit then "Function Inactive - End Position Not Reached". The frustrating thing is, the roof will raise and lower perfectly up and over half way up, including the clamshell opening and lowering/locking, so clearly the bulk of it is working. I'm going to have to find a way of trying the flaps again. The sensors seem to read them down and not up OK, but maybe they need to start raising when the roof hits level 4 of 6 in closing and when the flaps lowered microswitch is still telling it they're down and not moving? Weather permitting, I'll look at the latch mechanism tomorrow, but that all unlocks and locks fine, using the button. The only issue being PIWIS reading the plastic fork on the latch, which tells it it's hit the windscreen, is seemingly reading backwards.

You're way farther than I am. I can't get the roof to move at all.

Is it locked shut or open? If open, you might need to do the manual clamshell open thing but this tends to need the repair kit fitting, which is £50 for the early cars. Handy though, because you can use the old, twisted shaft to manually turn the clamshell gearbox with a set of mole grips. If it unlocks but doesn't move, do you hear the pump try? When mine refuses to move as it gets to level 4, all I get is clicks from relays.

Well finally! :mrgreen: After weeks of learning the ins and outs of the 996 cabriolet top system, I finally got to the bottom of it. I reset the latch as per the dot line up procedure, as my cogs were one tooth out. This didn't fix it. Double checked the flaps with a battery and both worked fine. In the end, I noticed in PIWIS that the POT2 (clamshell POT) voltage got to 2.51V and never increased from there. This was with the shell fully open and up,. Now, I had set the resistance to 6.2Kohms at end point (which I assume is fully up) as per the instructions, which got it there or there about in PIWIS apart from the upper level. My PIWIS calibration value was 2.75v at position 4 of 4, fully up, yet the real reading was only 2.51V. I took the POT off and tweaked it until it hit 2.9V (as it must be above 1v at lower limit and under 3v at upper limit). Put it back together and BOOM! The top lowers and raises as it should. Just got to tidy up the headlining and put all the trim back now. Still wont calibrate with PIWIS, but I'm beginning to wonder if these Chinese clones aren't 100%, what with backwards readings of the roof up sensor on the latch. Anyhow, I've learnt an awful lot about how these work, so thanks for the input from DeMort and others, without the knowledge, links and documentation, most of this would have been impossible! :thumb:

Congrats! So you pulled out the Allen key bolt again and re-calibrated? Mine is at 6.2k right now with the clamshell open. My gears are all in great shape in the pot 2 as well. Hopefully you can help me understand mine better. Mine clamshell seems to be working properly and the pump runs but nothing opens or closes.

So my clamshell is opening and closing but the voltages are not changing on the check drive links in the piwis. That doesn't seem normal.

So what voltage does POT2 show? I abandoned calibrating mine to 6.2Kohms, as when I did, with the clamshell fully up, the voltage (2.71V) was struggling to rise above the POT 4 value of 2.75v, so the control unit thought it never reached fully up, stopping the roof lowering more than half way, as the POT value remained the same for 3 seconds without reaching it's end point (Value 4 >2.75V) Instead, with the shell fully up, I took the screw out, the POT off and turned it until it read 2.9V in PIWIS, putting it well above the 2.75v POT2 Value 4. I then refitted and it works. The only slight issue I have now is POT1 only just gets above it's 3.51V value when the roof is fully down, so sometimes it takes a couple of attempts to raise the flaps, squash the roof fully down and lock the lid. It's getting to 3.51V. I suspect the POT's get old and the upper and lower ranges on them are not as defined as they were when new.

I've got it spot on now. The control unit on there currently was the replacement one from the US. The roof worked with it, but was struggling for the roof POT to get to the voltage configured for the limit of the roof. The saved value 6 (fully down) for the roof was 3.51V and my POT was only just getting to 3.51 or 3.53V fully down. Because of this, it sometimes took a couple of attempts to do the final part and bring the side flaps up. For some reason, I still couldn't run a calibration, but looking through my old screenshots, the original control unit was last calibrated with slightly lower voltage figures, as they were formed using my car's POTS, not one in the US. Anyway, this control unit only needs a voltage of 3.45v to be fully down, so the roof now goes up and down exactly as designed. A tip for anyone tackling this issue. Firstly, check the obvious, damage to framework or clamshell lid brackets. Secondly, check the side flaps. Try manually moving them away from the end position then try powering with a 9V battery to see if they move. The electrical connector is just to the rear of them. Pop if off and use some wire on the inner two connectors. Don't over do it though, remember the microswitches wont stop the motor at travel end. Thirdly, if you have PIWIS, under "actual values" look at the saved voltages for the 2 POTS. POT1 is the roof POT and will have saved ranges from 1-6. 1 is up and locked and 6 is down and locked. They may range from 0.85v or so, up to 3.5V or so. Your current POT voltage reading is above these figures in PIWIS and is live, so will change as the roof is moved. It should be lower than your POT1 saved value when fully up and must be higher than the saved value 6 when fully down. POT2 is the clamshell POT on the motor. The manual tells you to calibrate this by setting the resistance to 6.2Kohms then put it back together with the lid up. This didn't work for me, as at 6.2Kohms on my multimeter, the voltage was only 2.51V as shown in PIWIS and the last saved calibration value 4 of the 1-4 values for the clamshell, needed to exceed 2.78V. I manually adjusted the POT2 until PIWIS read 2.9V on the POT (it can't exceed 3V), then refitted with the shell fully up. This ^ sorted it. So "Function inactive - End point not reached" was basically, on lowering the roof (as it worked fine from half way to up and then back down to half way), that the shell opened to max, but the POT never went higher than 2.51V, when the control unit wanted to see it match and exceed the stored calibration value of 2.78V. After 3 seconds of not seeing this value increase, it halts all motor action. After configuring the POT to read 2.9V when the shell was at max sorted it. I've learnt loads about how these tops work, so hope, as confusing as this might sound, that this makes sense and helps others. :thumb: I can finally enjoy the rest of the summer, before it goes up on stands over autumn into winter to remove the gearbox for a rebuild, then whilst that's out, replace the brake line that goes over it (as it's starting to week at the joint), plus sort out whatever the oil leak is from the RMS or IMS cover. Should be totally sorted then.

As we discussed, my clamshell fully open only read 2.5v. I took out the motor and set it to 2.9v, put it back in, loctite the allen bolt, and it started working. I'm glad I disregarded the 6.2k ohm reading like you said. This should be a sticky.

So my top decided to open up only halfway. I checked the pot and it was off. I set it to 2.9v. The top still got stuck halfway. So I checked the voltage and it was 2.5v. I adjusted it again and still stuck and voltage went back to 2.5v. Ideas?

So with clamshell fully up, you're turning the POT on the clamshell drive, until it reads 2.9V in PIWIS2, then putting back together. What is the voltage reading when the clamshell is fully closed? Is it going lower than the calibration value?

Closed it reads 3.9v and open is 2.9v. But if I have the bolt three turns in the clamshell won't manually move. If I screw it tight, it moves fine. I checked the hydraulic fluid. It was normal but I added a little and no change. Ideas?

Your supposed to replace that allen bolt if you've done a manual open. Did you? Maybe it's stretched? I got my overpriced kit for about £50 and all it has is the allen bolt and a couple of other little bits. I cant remember the calibrated ranges now for fully up and fully down on the clamshell, but PIWIS should tell you and so long as you're exceeding them at each limit, it should be fine. If you cant adjust it to exceed each limit, then the POT must be shot or wiring from POT to Roof ECU damaged/corroded.

I started working on my top again. No matter what I do with the pot the voltage never changes from 2.78v. I popped the pot off and I can adjust it with my ohm meter. Any ideas?

It gets stuck here. I put a volt meter on the pot and it only read 2.45v so I moved it up to 2.95v and it still gets stuck at the same spot.

I got it to work. I should have been reading the blue/yellow and red/black pins. I was reading the other side of the pot. I got it to work but my driver side flap is busted. I have to order another one off of eBay.

Good stuff! One thing I noticed too, was the stored calibrated values, which the pot needed to reach to exceed the upper and lower limits, were set at different voltages on the 2 different roof ECU's I tried. I never did find out how you reset or change these stored requirements, so just kept turning my clamshell drive POT with a screwdriver until it showed voltage which exceeded the stored upper limit, then put it all back together. Lower limit wasn't ever an issue, it just was not passing enough voltage to tell the roof ECU that the clamshell was fully up, so it could continue with the roof lowering program. :thumb:

You ever replace a flap? I can't get it out. I ordered a new one. But I can't seem to get it out.

Yes. It was a bit of a fiddle. The blocks at the top for the roof mounts need to come out and it needs twisting in a certain way. It was a while ago and I can't quite remember how now. I remember lot's of skinned knuckles on sharp aluminium parts though.

Unfortunately, I am a member of this club! In some surprisingly similar details, for example, Pot 1 reads 2.51v when set to 6.2k ohms. I, too, have tried another control module. And I'm using the same version of PIWIS and have learned its actual value quirks (for example, the convertible top raised/not raised values are exactly opposite, and the speed always shows >5kmh). I, too, bought the car with the top nonoperational. In January of 2021. The only difference may be that I don't get any motion, nothing at all, when I attempt to calibrate. 1/2 way then stopping would be progress. Also, my actual value for the convertible top (Pot 1) is 0.86v and the stored value is 0.80v, so I'm struggling to get that actual value to below what's stored, using my original control module—the one it was last calibrated to when, I assume, it worked. That value may change when the rams themselves press the top forward. So that has to be addressed. Regardless, I'm heading back out to the garage now to manually set Pot 2 to 2.95v and see what happens when I try to operate the top, calibration be damned. Oh, I have a Durametric license also. The entire saga is here (if posting this link to another forum isn't against the rules). I don't have words for how grateful I am for this entire thread and all your posts and updates. My efforts aren't done, but at least I'm not alone. I'm not the only one with this experience. Keith

Quotes source post: 2823237

Well that almost never happens. I was able to brute force Pot 2 it to the exact stored values: Value 1 = 1.55v, actual and stored when closed. Value 4 = 2.71v, actual and stored when open. However, the first time my latch closed, it sort of internally exploded. Wouldn't reopen, and when I manually got it off, the fork had snapped and the latch seems seized: I can't get it to move manually or on the bench. So that's what's next and (I believe) what's left. With the latch off the car, I back probed the microswitch connecter to ground the two switches signally latch unlocked, and the top compartment lid opened fully and the top opened to a point. Since I've never seen a top function correctly in person, could someone please tell me whether—as the top opens—the latch closes any? That would explain why my top stopped opening, I was bypass grounding the microswitches. Hard to tell from the YouTube videos I can find, people usually don't focus on the latch when the top is in operation. EDIT: It does appear that as the top opens, the latch hook closes at least partially in normal operation. Not mine, yet, but from more videos I found on YouTube. I was unable to replicate the expected switch positions to get the top to fully close, even though that seemed pretty straightforward. I did not check actual values. Since it's a workaround anyway, I'd rather focus on repairing the broken latch. Also, for anyone who stumbles across this, there is now an STL to 3d print the H shaped fork that depresses the top closed microswitch when the top touches the top of the windscreen. Link to page with file . It is not my design, I'm test printing one now.

When I say "brute force" I mean by detaching the motor from the compartment transmission and, using a 9v battery, gradually advancing the motor in increments. Then I'd plug it back into the control module and check actual values. Slow, but rewarding.

Quick update: I've got to repair one flap that seems to be stripped, otherwise I think everything is sorted.

Following up on that side flap with the insert issue: it had cracked. So I made one. https://makerworld.com/en/models/3232512-porsche-996-cabriolet-side-flap-repair-insert#profileId-3661479 Thingiverse - The community for Open Hardware Thingiverse - The community for Open Hardware www.thingiverse.com

With my part made out of polycarbonate and installed, the top opened completely as expected until the windows went up, then closed completely as expected until the windows went up. Then my part failed: threads gone. I sent my design to a shop to make me some out of nylon, PA12, I'll see if they hold up better. Regardless of the stable solution, The top works. Thanks in large part to this thread.

That was a design flaw, and a big one, not a material flaw. I updated the files at both those sites, and the top is now 100%. I mean, except for not having any rear window at all. Video !