top of page

N62 Z3 Part 8: First Start Attempt & Starter Sadness

So, it is time to try to start the car...

Before I could even attempt to start it though I had to align the EWS and DME

At first I totally messed this up... but I was doing something rather stupid and after talking to my tuner, got it sorted out quickly

(5 seconds later)

Cool. Time to crank!

And uh, well, the first crank was a total dud. Nothing at all. Not even a single click, crank or anything. Time to do some digging

As it looked in INPA like the EWS/DME alignment was successful, the first thing I wanted to check was the starter signal output line from the EWS. On the Z3 this feeds straight to the starter. On the X5, and the way I had re-wired my setup, was this signal feeds into the IVM, which then outputs it's own line to the starter. I believe the reason it does this is because of the X5 "comfort start" feature, which means you just turn the key over into the "start" position for a second and it will engage the starter on it's own until it starts the engine (so you don't have to hold the key in the start position while it cranks)

This signal line did have 12v when attempting to start - which, along with the lack of EWS manipulation codes in the DME, meant the DME and EWS were paired successfully. So that was a big win and again a big thanks to my tuner for that! But I still had the issue of the no cranking or anything. I figured at this point I had some issue in my ignition wiring and I would come back to it later

I did want to test the starter in the meantime, so I bypassed the IVM and ran the starter signal straight to the starter. In doing so - I got a single "click"


So still no cranking, but this confirmed at least some of my ignition wiring was wrong. I had altered some of the ignition wiring and the "Start signal" lines in the DME pintout seemed to match and I moved them into the correct terminal for the X5 DME. So those seemed good. It was time to do some research...

Lets look at some diagrams together

This is the Z3 starter circuit. Red is the power to the ignition source - and ultimately the power supply of the starter line. Orange is ignition to EWS and yellow the EWS output to starter (this is the line I have confirmed working). I am not sure what to do about the ZKE (which seems to expect to see information from the DME regarding starting, though this may not be needed with my setup)

Just to understand how simple that circuit really is at it's core, below is the X5 6 cylinder starter diagram, which is the same diagram as far as routing but without all the extraneous EWS stuff

So, so simple. This is pretty much how starter circuits have worked forever - starter circuits historically have been pretty simple

Now lets check out the N62 X5 starter circuit!

Cool... super cool...

So what the hell is going on here?

We have two power supplies, as you can see at the top. A smaller power supply goes to the ignition (on the top left) and a larger supply feeds the IVM (60a) which is marked with the red line (the red line signifies the circuit that will ultimately send signal to the starter)

The orange line on the left is our ignition output - this is the thing that tells the whole system "Hey, I'm trying to start over here!!!" - you may observe that on the N62 X5, this feeds into the DME and not the EWS

The DME then has two lines (pink and blue) that feed into the IVM. The pink line is one I didn't have to think about - it runs on the engine harness. And I already ran the blue line earlier as part of my wiring changes so that part of it was good

The yellow line is the EWS output - I confirmed this output line was getting 12v when cranking, so I did not have to worry about my EWS wiring or anything upstream of this

Green feeds the starter, from the IVM

When I ran the EWS output (marked yellow in the above diagrams) directly to the starter I got a single click at starter. However, with the way I configured it originally, it did not click when ran through the starter...

I was able to figure it out pretty quickly and I bet most of you can too - I'll give you a huge hint... the start signal at DME is not the same thing!

Lets look at the X5 diagram

The DME "start signal" pin here is the input from the ignition. Easy enough

So lets compare tthe Z3 diagram

Here we see the DME (bottom of diagram below starter) gets an output from the starter. The DME pin that was marked as "start signal" on the Z3 DME pinout is actually an output, which supplies the ZKE. I don't know why it does this but in any case that's a pretty simple thing to fix. And I decided to fix it in the neatest, most splice-free way possible

So first we need to identify what needs to go where. Well - that's part of why I color coded it. We need the orange lines (the ignition outputs) to "match up" - the Z3 orange line currently feeds the EWS, the X5 N62 DME is expecting to see this ignition output as an input at one of it's pins, so it can tell the IVM (presumably through either the blue or pink lines) that it is trying to start

If we look on the Z3 diagram above, on the orange line, there is a junction (X33) mid-way through. My goal was to find this junction, de-pin the pin I need to access and add an in-line splitter, which would feed the DME without interrupting the EWS input

First step was to find the X33 junction - it's the black plug under the steering column, then I verified this was the correct wire I needed with a continuity tester

Now that I found the pin, the next step was to de-pin it from the harness. This is it loose and ready to pull out

This was the adapter I built. It's pretty simple - the black is a long wire that goes to the DME. Red wire will go into where the other terminal was de-pinned, and the white plug is where the original pin will now sit

The de-pinned pin on the harness

Pin installed in the harness

And the wiring harness plugged in (the red goes to where the pin originally plugged in, black goes to DME)

And all together!


The black wire was run to the DME and plugged into the correct spot as well. The end result would be a wiring diagram like below

Then I tested it and... click! Which is a good sign, it seems the circuit is working correctly now. But the question now is, why is the starter still not cranking? Hmm

Few things to investigate today but testing the starter itself will be one of the first things I'll be doing

I also did spend some time investigating the steering rack fitment as well... not critical at the moment but, it looks way better now with the O2 sensor out

Installed in place (loosely) on the splines

Hard to tell but there is some clearance...

So the next day I realized I popped a fuse at some point. Actually 3. Dang. I traced down part of the problem to being a wire on the IVM body harness - it was wired to 12v (behind a 10amp fuse that popped). It is actually supposed to be a 12v output, to the e-box fan...

I also ended up frying the DME in the process I believe. I can no longer communicate to it through INPA and I was able to trace part of the faulty circuit that may have likely fed the DME. I have sent my spare DME out to get flashed and may stop by some local junkyards looking for another spare just in case. Really hope it doesn't come to that though, but never hurts to be prepared and I could use another spare IVM too

The below image summarizes ,as succinctly as possible, what happened

So that evening I went back to the shop and worked on fixing that faulty wiring. Pardon the religious reference...

Bavarians E:36 If the IVM wants to output to an e-box fan, then so to an e-box fan it shall output...

First step was to verify this pin was indeed wired incorrectly by continuity testing it and it did confirm that this constant 12v was routed incorrectly

Luckily I used good quality electrical tape so removing it was not a big deal...

The other side of the wire that was cut in the above image goes to the IVM output for e-box fan (where I originally had 12v going in), it routes through a 10amp fuse which is the same as the E-box fan so I will run this loose wire into the E-box fan positive and then if you follow that E-box diagram the ground goes into the DME. DME Pin 19 to be particular... (note it kind of looks like the harness burned or something - it didn't, I just mark the middle pin with sharpie to make counting them out a little easier since they are small and 10 wide)

And then everything back together and a new IVM installed as well. Note the e-box harness on the left is now plugged in...

I will be spending some time going into the wiring diagrams much deeper to see if there is anything else I overlooked. Part of why I made this mistake is the pinout (basically a text document explaining what each pin does) listed a few pins as "voltage supply" and I checked a few on the diagram and they all went to positive sources. Ends up all of them actually are inputs except this one...

I'm a little annoyed because even if this "fixes" this problem, it doesn't solve the previous no-start issue. Oh and there may be a whole new issue too. Let me pull up the X5 starter diagram again

I'm not sure any of that stuff will work correctly even if I wire it correctly, due to the EWS maybe being incompatible? What I mean by this is, on an N62 X5, you turn the key for a single second to the "Start" position and then you can let the key back into position "3" or run (where the key sits when you are driving it). The DME Relay 4 (with the input in pink) inside the IVM will then keep the Starter Inhibit Relay in the open position and the starter will crank until car starts

The worrying thing here is that the only purpose of the Stater Inhibit Relay is to allow the EWS signal to pass through directly to the starter. This means the X5 EWS is outputting 12v on that line even when it is not receiving 12v input from ignition. Lets look at the Z3 diagram again

The Z3 EWS only gives out 12v on that line when actively cranking the key in the "start" position. So it seems the autostart function will likely not work

If I really wanted to make the autostart work, it may be feasible to use a relay setup while maintaining functioning EWS...

Wire the EWS "Start" output to switch a timed ~60amp "IVM relay" for ~5 seconds (long enough to start the car)

Run that ~60amp "IVM relay" output to the IVM EWS input

Then the IVM will receive the required start signal at the EWS input for up to ~5 seconds, long enough to start

Technically leaving the key cranked in the start position should also work... and if so I may simply bypass the IVM and run the EWS start output to the starter directly (like the Z3 is originally wired). Maybe it was only clicking before because the jumper wire maybe wasn't making a great connection to energize starter solenoid adequately.

I would like autostart though, but getting a first start of any type is critical first and still looking like a bit of a hurdle but I have faith I will get there

Easy to feel a little demotivated after these evets but I think the important thing is that I know why I made the mistake with the IVM wiring and I can prevent making similar mistakes in the future :)

Finally figured out the reason the starter was just clicking... it was an issue with the main supply circuit for the starter. It isn't making a good enough connection - I ran a jumper cable from one end of the starter supply cable to the other, supplied the signal with 12v I tapped from elsewhere (because the correct EWS starter output won't work now since there isn't a DME) and...

Now I will wire it back up the correct way including running a new starter supply cable and wait for the DME to get here to try to start it. I have a couple other things I would like to finish in the meantime though, so I will work on those in the coming days.

Most notably of those things I want to fix is that the oil level sensor harness does not fit on my current setup and is currently not plugged in. The reason for this is that I am using an X5 engine harness and E60 oil pan but the X5 and E60 oil pans are different (because X5 is AWD) and they place the oil level sensor in different locations. I am contemplating a few ways to do this - if that plug is intact from my other engine harness I may de-pin it from that harness and run it directly from DME to oil pan. I think it would fit if I did that - it's pretty close as it is and the DME harness does a lot of twisting and contorting around so if I can run it in a straighter path then I think it should work. If I can't de-pin it and remove it from the spare harness, I will have to get a little more creative...

And here is how I imagine the autostart circuit could be made to work... obviously have other things to focus on first, but curious to try this once (or if? haha) I get it running

The issue was the Z3 EWS only puts out a 12v when the ignition is in "start" but I needed it to supply 12v the entire time when it could be cranking after a single moment in "start". What I think the solution may be, as reflected in the above diagram, is to use a timed off relay - which will remain "on" for a variable, preset period of time after 12v is initially applied to it. However, all of those relays tend to be lower amperage ratings, all the ones I found were a max of 30amp and the circuit is rated for 50amp for the Z3 and 60amp for the X5. So the solution would be to use the output from this timed relay to energize a higher amp relay, which will then power the IVM starter input and eventually, the starter. The only reason I think this could potentially work is because as soon as the DME detects the engine has started, it disables the starter circuit relay in the IVM and so even if the timed relay isn't set for the perfect amount of time, it won't keep cranking the starter after the engine has started. The DME will disable the circuit completely

Anyways, I have spent some time diagnosing the original issue recently and found a couple things wrong. Totally my mistake, I decided to replace some wires I ran awhile ago with a poorer quality wire with my new wire and when re-inserting the pins, had some in the wrong location. Critically, this included the DME ground

Investigating things, with a bright shop light in hand... don't know how I missed this before, but I do know I was last working on it at night

The sub-harness plug didn't look as bad...

Because the spare DME I bought awhile ago was the wrong type (ek928_2 not ek928_4, lol) I ordered 2 of the correct type to be shipped directly to my tuner and just waiting on those to get to him and tuned so I can get back on track with starting this thing!

In the meantime, I also decided to extend the oil level sensor harness

So first I mocked up how much length I needed... just used tape, for convenience

The oil level sensor plug off the old/spare engine harness

Peeled it out of the harness a bit...

After it was cut at the correct length, I cut the sheath on the existing harness and cut the wires, with my cuts being staggered

I did the same, with the lengths inverted on the plug-side and started to put it together

With the wires connected, it was time to try to get the sheath back in place as best as possible

It's a little hard to tell but on the plug-side, I used a little harness tape to tape the sheath in place

I then pulled the sheath down from the engine harness side, to cover the joints and seal it up

I wrapped this in harness tape and then had heat shrink which I had slipped on earlier over all of it to secure it as best as possible from the elements - especially critical given where this will be sitting (just hanging underneath the car)

Sorry, it was very hard to take a picture of the black harness/tape against the bright background. But below is it all installed in place

While waiting on the DME's to arrive, I turned my eyes to another project of mine... installing these Custom Z3M M Hurricane seats I've had waiting to install for ages!

One down, one to go...

All done

And now back to the Dunklenator...

The DME's arrived so the first course of action was to finally install that 5amp fuse for the ignition line to DME

I also had to get a new multimeter, as my other one had one of the probes break. It was a very welcome upgrade and now also has a visual indicator (green light on the top) to visually indicate continuity

I then plugged the DME's in and - it did crank! Which was great, as the starter circuit was once again wired through the IVM. However it still didn't start. I attempted to communicate with the DME via INPA and kept having issues. Ends up my cable had communicated to it's last DME the day prior... I did have a spare at home though

Once I did get connected, I realized I totally forgot to align the DME and EWS and it had an EWS code as a result

I then aligned the DME and EWS again...

At this point the EWS code did not come back however it still did not start. The only codes present were those below, plus a "DSC message timeout" which isn't pictured because my phone died

Further digging revealed that the crankshaft position sensor was not sending any signal during the start attempts - this was likely the source of the issue. First thing to check was continuity between the sensor and it's pins - it gets a voltage supply from the IVM and two wires to the DME. All of the wires checked out good

Next was to check for voltage at the sensor when running. It did have voltage. I checked my own personal N62 X5 sensor just to be sure and it had voltage on the same pins

I was originally thinking it was some sort of EWS related issue because there was an "EWS OK" status that never lit up on INPA. However, on my N62 X5 this did not indicate "on" at any point before, during or after starting so I am no longer suspecting the EWS... although, my research did reveal the starting diagram for the N62 is much more complicated than I originally thought

The teal lines (on the left) are wires I was not previously accounting for. Purple is a transmission line I do not need to worry about as the EWS is coded for manual. More research did reveal that a lot of the "unknown" were already addressed at some point in one way or another, but regardless, I found the complexity of this starter diagram intriguing

I definitely understand a bit more about why a lot of people - particularly mechanics who focus on non-German cars - complain about the complexity of the electronics in these cars. This reputation, apparently rightly earned, was amplified tremendously with the release of the E65 which of course was predominantly N62 powered

I think this start attempt will need to spill into a new post as well... it is getting to be a bit too long of an update. So, onto Part 9



bottom of page