Difficult start TDI 90 with MYP 6-speed conversion

BiggDogg

A2OC Donor
I converted my car (with the ATL engine) to 6-speed this year. I didn't change the starter. I did replace the DMF. None of the information sources I used when I did the conversion said to change the starter motor, but ever since, the starter has sounded a bit ropey, like an old Fiat 500 or 126 if you recall them when they were a common sight and sound, and there has been a lot of shaking transmitted to the body during cranking.
Recently, cold starting has become difficult, requiring up to 3s of normal-speed cranking (400rpm) before the engine would fire and run or sometimes die and need another go. Warm or hot cranking is better, but not near-instant like it used to be. Cranking has been normal speed, but rather lumpy. All this time, once the engine is running, all is normal, plenty of power and smooth delivery. However last week the engine misfired a couple of times at 2300-2500rpm on full accelerator whilst fully warmed up. I scanned with VCDS but no faults. The next morning, starting required several goes and it put the check engine lamp on with the glow plug lamp flashing thereafter too. The ABS and ASR warning lamps also lit up. Scanning with VCDS I saw a
Fault 16705 engine speed sensor G28 Implausible signal.
The ABS controller gave 01314 engine control module 013 check DTC memory intermittent and 00668 supply voltage terminal 30 008 implausible signal-intermittent. And the other controller "BORDNETZ-SG." gave 01598 drive battery voltage 07-10 signal too low.
After replacing the crankshaft speed sensor which I managed to convince myself was a cause (OMG what a PITA that job is) the starting is exactly the same, though I've yet to have another misfire. The battery was replaced in the spring and it spins the engine over at the normal speed, the terminals are clean and tight. I wondered about the "usual suspect" earth connection between the lower starter bolt and body earth so I removed the cable to check-everything was bright and clean and I cut the heat shrink sleeving from both ends to check for broken strands-again it was all like new under that sleeving so I put it back on. The result: No difference.
I am wondering now if there is another connection in the +12V or ground side which could be responsible?
Anyone got an idea?
I'm a bit reluctant to drive the car at all just now because I can imagine frying one of the controllers if the power supply to them is going all over the place, especially with the big voltage spikes you get off stater motors.
As ever, many thanks in advance.
Matt



Sent from my iPhone using Tapatalk
 
Thanks Steve. I just popped out with my head torch and checked. All 7 of the nuts on the battery connections and 2 metal-strip fuses are tight, and clean, the terminal clamps have a good grip on the battery terminals too (it is possible to tighten the nuts and have the clamps slip up the tapered terminals so the nuts are tight but the connection is loose-been there, done that).


Sent from my iPhone using Tapatalk
 
Thanks Steve. I just popped out with my head torch and checked. All 7 of the nuts on the battery connections and 2 metal-strip fuses are tight, and clean, the terminal clamps have a good grip on the battery terminals too (it is possible to tighten the nuts and have the clamps slip up the tapered terminals so the nuts are tight but the connection is loose-been there, done that).


Sent from my iPhone using Tapatalk

It is not related to the meshing of the starter gear and flywheel is it?

It is probably not affected by the swap since the flywheel and starter are as they were aren't they?

Steve B
 
It is not related to the meshing of the starter gear and flywheel is it?

It is probably not affected by the swap since the flywheel and starter are as they were aren't they?

Steve B

I wondered about that and mentioned it in my waffling write-up on the long MYP gearbox thread. The starter pinion doesn't make any grinding noises or other signs of distress like the solenoid is banging the pinion in and out in an uncontrolled way.
Nothing I have read says that swapping from the GRJ to MYP requires a different starter to mesh the pinion and flywheel ring gear correctly both radially and longitudinally. But maybe I missed something. Mike (A2 Sumo) alluded to needing to use the AMF 75 starter.

I think the next thing to do is check the battery voltage at the terminals during cold cranking and also at the starter motor terminals.
A mate of mine has suggested maybe there is a wire from the starter positive to the alternator that could be dodgy.

I think an electrical issue is the root cause, and that is causing the strange sounds too.

If the starter to ring gear was mechanically not right, how could that cause the fault codes unless the current was so high that the voltage dropped really low, I'd probably have burned the starter out by now?

If the voltages during cranking look good then I am going to be completely stumped.

If the battery voltage looks good and the starter voltage bad then I will have some lost voltage to look for, hopefully finding that will lead to sorting the problem out.

Matt


Sent from my iPhone using Tapatalk
 
The problem is still ongoing, I'm starting to doubt my ability to solve this one, keep flitting between thinking it is electrical (starter circuit), electrical (glow plugs control relay), electrical (engine ECU or relay) or fuel starvation. Here goes with some waffling:

Starter circuit:
I took some voltage readings at last, +12degC, cold engine been standing for 3 days.
Voltage at battery, ignition on: 12.40V
Minimum voltage at battery during cranking: 10.40V
Minimum voltage at starter during cranking: 8.45V
Maximum volt drop from battery negative to starter mounting bolt: -0.56V
Maximum volt drop from battery positive to starter positive during cranking: -0.84V
The minima and maxima are from 5 short bursts, so it is not all that surprising that these don't add up taking into account that the voltage drops during compression and the digital meter is sampling at a frequency that isn't synchronised...the difference is only (10.4-8.45-0.56-0.84=0.55V) anyway. I reckon that the starter voltage is OK. Anyone care to differ?

Fuel:
After Steve Birchall suggested it sounded more like a fuel pipe issue in t42's "starting problems" thread, which would certainly tie up with the gearbox work if something was disturbed at the time, on Saturday I pulled off the fuel return pipe at the tandem pump and it was dry: not even a drip of diesel - I thought it could be sucking air or else suffering from a blocked fuel filter. So I refitted the hose and used a new clip.
Further on fuel:
The car had given a misfire or 2 at full power around 2300rpm then last week one day it was misfiring more than a little at full power so I changed the paper element inside the fuel filter. Inside the filter housing was horrible, about a dessertspoonful of black stuff like leaf mould in jelly in the bottom and also the paper element was black. I cleaned everything out and reassembled it, using a torque wrench set at 2Nm on the delicate drain screw having swapped the element and replaced the 3 o rings in the filter kit. I was really hopeful that this would sort out the starting issue but it is the same (it needed a fair bit of cranking to fill the filter up of course) but was the same after turning off and back on. The misfire at full power has gone so I think it probably was fuel starvation. But the start up is still the same, and seems to get exponentially worse the colder it gets, so maybe there is still an air leak, or maybe the tandem pump needs a rebuild? My old Seat Arosa TDI with the AMF engine had a sensible lift pump in the fuel tank, it seems really odd that Audi decided to do without the lift pump and that therefore the tandem pump has to suck that fuel all the way from the tank and keep it in the flow and return pipes whilst the engine is off, on basically the same engine - maybe the tandem pumps are uprated.

Electrical - engine ECU
I am trying to upload a video from my phone to Youtube to show an example of the odd tachometer behaviour on particularly long cranking sessions of spluttering before it finally fires up and immediately runs sweet. What will happen is that the tacho will show 300-400rpm for a few seconds then it will drop to zero for a second then come back up to 300-400rpm. This ALWAYS triggers the check engine lamp to light and the glow plug lamp to flash anytime the engine is running, along with a 16705 DTC in VCDS which is G28 implausible signal - engine speed sensor. The obvious thing to do would be replace the sensor, but I've already done that. It NEVER gives the 16705 whilst driving, only if the cranking exceeds 3 or 4 seconds.

Glow plug circuit
I tried measuring the glow plug voltage and got 0.31V for a fraction of second then nothing. I've checked the 12V 70A relay in front of the double-deck engine ECU and the CCCU which was fine. I'd love to check the glow plug control relay but I have no idea of the pinout - about 10 pins I think. Does anyone know it? I believe that the ATL engine uses low-voltage glow plugs and I guess the "relay" is actually a high-frequency dc chopper using power transistors. This is completely different to the AMF version. Although I have pukka Audi A2 "Current flow diagrams" for just about everything, the ATL version of the engine ECU, sensors, injectors, glow plugs etc., is something I'm missing. If only the temperature would drop below 10degC... I will get a can of freezer spray and use it on the coolant temperature sensor to try and get some more glow plug action. BTW the coolant temperature sensor reports sensible temperatures in VCDS and the gauge temperature is good too - the sensor is about 3y old. Does anyone know how to check the glow plug circuit is functioning? I can use an oscilloscope.

I'm struggling to come up with a logical order to work through this problem, need to narrow it down.

Thanks in advance,

Matt
 
If youv concerns about the starter alignment ect id suggest removing the starter and taking a look through the mounting hole as youl see if anythings amiss on the flywheels teeth or the end of the starters teeth if anything is mis matched or not aligning 100%

As you are throwing up a code for engine speed sensor , id add theres 2 the ecu uses and the one by the camshaft pulley wheel can give the same fault code, i remember reading a post from years back with simular issues and turned out to be the spigot ring, "think thats what its called " on the crankshaft , this is the magnetic wheel the speed sensor reads off and had a damaged finger possibly from a broken dipstick end floating around in the sump..

Id suggest removing both the sensors and cleaning them , as there magnetic they may have picked up metal particals on there tips
 
Thanks very much Jigsaw, the crank one is brand new. I'm pretty sure there is a different DTC for the cam position sensor but will take it out for a clean if not TOO difficult.
There is definately a dipstick tip in my sump like close to 100% of these cars surely so could be damaged like you say, but I would probably notice a problem the whole time and not associated with cold start?
I will see if I can check the waveforms from the 2 pulse sensors with a USB oscilloscope and post results.

Here's the YouTube link for a video of the tacho dropping to zero during cranking: https://youtu.be/2S44ip5YLGI

Cheers,

Matt
 
Thanks very much Jigsaw, the crank one is brand new. I'm pretty sure there is a different DTC for the cam position sensor but will take it out for a clean if not TOO difficult.
There is definately a dipstick tip in my sump like close to 100% of these cars surely so could be damaged like you say, but I would probably notice a problem the whole time and not associated with cold start?
I will see if I can check the waveforms from the 2 pulse sensors with a USB oscilloscope and post results.

Here's the YouTube link for a video of the tacho dropping to zero during cranking: https://youtu.be/2S44ip5YLGI

Cheers,

Matt

its easily removed , remove cambelt top cover , , at about 4 or 5 oclock position on the cam pulley youl see the sensor and small allen key holding it to the head..

although possibly nothing to do with your issue, but may be of help as your stuck finding the fault..

how i understood it, if one finger is slightly bent the ecu gets confused while counting the fingers and takes the reading from the camshaft sensor instead , the engine would still run like this but possibly harder to start and give miss fire symtoms possibly too..

ive one those inspection cameras on a lead for a iphone , posibly something like that could be put into the sensor hole to see all the fingers as the crank is turned by hand slowly and help eliminate this possible fault

I know you saying your sensor is brand new but even brand new parts can be faulty, id try your original again as youv test equipment and see if they output the same readings
 
Last edited:
Hi
I too had similar starting noise since I had a myp fitted to my ATL -does not sound good but got better over time, I suppose as its run-in. However my cranking speed had gone down too.
This was due to the connection to the starter which I noticed was getting hot. I found that out when cranking to get the fuel up again after the tandem pump replacement. It surprised me how long it took to get the fuel up again. You mention you took off the return pipe to find no fuel so I think your fuel may be draining over time and you need to crank to get it up again. does the length of time it is at rest make a difference to how long you need to crank ? I just went to see how mine compares, took off the pipe, (hate these clips), some drops appeared but I heard fuel glugging down to the tank. It is the feed in pipe that needs to be checked I guess. I do not know if it is the tandem pump that stops the fuel from draining or if there is a in-pipe one way valve. A tank fuel lift (or in-line) pump would stop this happening and I guess not too difficult to fit. I have changed the fuel filter only twice but did not have to crank more than usual.
I was getting fuel starvation at low load and used Bosch pump which has a point to connect a fuel pressure gauge (or new colour DIS I hope to keep an eye on fuel supply). New pump stopped starvation issue, oil leak and better vacuum, however the ASV does not fit properly and will need to do something there.
So I think your electrical problems are only recent, and since your cranking speed is ok I put my money on your fuel system between tank and pump.
Good luck
Martn
 
Thanks all, I've not had the chance to do much more the past few weeks but had another go today. The colder weather has made no difference to the starting. The air leak into the fuel supply is tempting but the if you turn the engine off and back on with the absolute minimum delay, it is still crap, might need a few seconds of cranking, might die etc. but once running at 1000rpm it's fine. I've disconnected the plugs at the ECU and measured the G28 crank sensor resistance at 540 Ohms from the ECU plug pins. This is the same reading as I get at the 3pin grey plug near the starter. I've also tested whether the engine will run without the G28 connected and it doesn't fire at all. Jigsaw, your spigot ring theory is a good one if the bent finger confuses the ECU into incorrect injection timing only during cranking-which sounds feasible if the slower rotation speed gives a weaker signal. Does anyone know, if I drop the sump, can I get a good look at this spigot ring? I've still not checked the signal on an oscilloscope. I'm inclined to drop the sump anyway to get the dipstick end out. And if I find a finger bent, should I attempt to straighten it or is a complete strip down and replacement on the cards ? And a last point to throw into the mix which I don't think has anything to do with it but does explain the starter noise difference swapping from GRJ to MYP: I measured the bell housing length from the mating end to the starter :101mm for GRJ and 105mm for MYP. So there is about 4mm less pinion mesh length with the MYP!


Sent from my iPhone using Tapatalk
 
Last edited:
I've used a USB oscilloscope today to monitor the engine speed sensor pin 1 and 2 and got the attached results. The frequency and amplitude of the signal increases with revs. There seems to be 18 pulses per 120° of crank rotation. The gaps in the pulse train are (I guess) in line with TDC on the 3 cylinders, so counting 3 gaps gives the period for 1 crankshaft rotation. There is really quite a large (to me) variation in amplitude - possibly on a 720° period, which I don't understand. I expect that this is due to some of the spigot ring fingers getting bent by the broken plastic end of the dipstick but maybe its normal?

Does anyone know the minimum G28 crank /engine speed sensor amplitude the ECU can work with? - obviously during cranking it gets a lot less than this due to the lower speed and maybe lower battery voltage too, but it should scale to the normal idle speed I hope. I am getting about 0.32V peak-peak at 1000rpm and about 0.85V peak-peak at 1910rpm.

I've discovered that starting is really no more or less likely if you crank for 1 second or 5 seconds. It often catches immediately upon releasing the key from the start position, as if the extra voltage whilst the is engine is still spinning helps the ECU - this would be in line with marginal signal amplitude. It is a bit like a petrol car with dodgy ignition or damp in the HT circuit in this respect.
Thanks,
Matt

1000rpm.jpg
1910rpm.jpg
 
Last edited:
I think the 720deg variation must be to do with the fact that firing is every 720deg for a 4-stroke engine. My inclination is to think this signal is OK and the problem lies elsewhere. Anyone disagree?
Also I just found some relevant posts using the website search (Tapatalk never seems to dig up useful stuff). Oh, and I ordered a very cheap starter from a CAYC engine which the MYP gearbox was originally fitted to and will see if I can work out the pinion travel issue too.


Sent from my iPhone using Tapatalk
 
Gosh this is a nightmare :(
Just read thread again and thought it can still is the tandem pump. It could easily be eliminated by pulling the pressurised pipe off- not the feed-in one.
Starting is a problem because the slower cranking speed is not able to pull the diesel up. Once it starts its ok because the extra pump speed enables it to do its job.
At full load it only just manages because it uses more fuel for the amount of revs.
I'm no expert with oscilloscopes etc so I would have to eliminate everything else first :)
If the dipstick causes problems I think it would be well documented.
M
 
Thanks Martin,
I have been concentrating on the electrical side of things because of the fault codes and the tachometer behaviour but it's entirely possible that these are false and misleading.
Do you mean the fuel return pipe? (It think the pressurised feed to the injectors isn't via a pipe I can see but is via the face that bolts to the head) Assuming yes, I'll give that a go. I've not yet tried it with the engine cranking, it's a good idea. I'll try and find a section of clear hose to run to a container above the level of the tandem pump and see how long cranking takes before fuel emerges from the fuel return outlet from the pump, and if the engine fires at the same time. If I then turn off I will hopefully be able to see if fuel drains back downhill from the clear pipe, back through the tandem pump or not, though whether it should or should not do this I've no idea. Further if I try starting again with a "head" of diesel on the return I wonder if starting will be immediate.

I'll come back on the starter investigation once this "experimental" motor arrives and I've had the chance to do some more measurements.

Matt
 
I mean the fuel feed pipe that goes to the injectors. As I understand it, when you look at the pump the top pipe on the left is the return pipe and under it, pointing downwards is the feed pipe( to the injectors). On the right is the servo. The face only has the connection to the cam and the oil feed to the workings of the pump. As you say, having a head of diesel above the pump(top pipe) should start straight away. Make sure diesel stays very clean.
M
 
Not long ago I had a similar problem with my Lupo 3L. Once the engine started it was OK but wouldn't always start easily. I took the rubber fuel pipe off the fuel filter (handily in the engine compartment) and could see that the filter was not full. In the end I rebuilt the tandem pump with a kit from Darwen Diesels and haven't had any problems since. There is also a non-return valve built into the feed to the pump; if this fails, it will make the job of the pump much harder as it is designed to pump liquid, not vapour. There are different versions of the pump, so make sure that you buy the correct kit.

RAB
 
Thanks very much Rab. I’ve rebuilt the pump, a LUK 038 145 209E with a kit from EBay. I found signs of engine oil leaking from the brake vacuum pump side (where there should be a little oil) through the shaft oil seal to the fuel side. Initially I couldn’t see any way this could happen since surely the fuel side should always be at a higher pressure when the engine is running, and if anything you should get fuel leaking into the oil. But then I realised that the design of the lip seal makes it unlikely that fuel will leak into the vacuum side because the fuel pressure acts to tighten the lip seal. However, once the engine stops, actually the fuel is at negative pressure because it wants to run back to the tank. There is no non-return valve I can see on this pump, so if the shaft seal is a bit worn it could allow air and oil to be sucked from the vacuum side to the fuel side. This was a Eureka moment, though I managed to stay fully clothed. Having replaced the oil seal and all the gaskets, vane-springs and screws that came in the kit, I torqued the 4 screws that hold the pump to the cylinder head to 20Nm and 10Nm (larger 2 are tighter), reconnected the 3 hoses, replaced the clips and......found that it made no difference at all. I’m still losing fuel prime... I’m wondering if the injectors could be leaking. Before engaging a garage to test the injectors, I suppose I should check for injection mass balance between the 3 injectors in VCDS. Does anyone else have a suggestion before I admit defeat and give it to someone who knows what they’re doing?
Many thanks,
Matt


Sent from my iPhone using Tapatalk
 
Back
Top