Mini cooper mpi 1300 (mems rover ecu) 1998 trigger setup

[retro]

I was just wondering in the chart that Adam posted, what "link injector timing" was.
The value you enter in the software is stated as BTDC crank degrees correct? 410 BTDC I can see on the chart for cylinder 1. (or 320 Link injector timing?)

Got the car wired up yesterday and I'm running into the exact same issue as the OP with no cam 2 trigger voltage at all. (VR cam 2 trigger) Removing the sensor and waving it over a trigger wheel shows voltage on the scope (4-6 volts)... So not a wiring or trigger setup issue

Hmmm wonder what the heck is going on with these MPI cam triggers (I need to find a pic of how the cam triggers as I didn't build the engine)
Unsure if its tooth on the cam, or a window, slot etc (but I get basically NO change in voltage on cam 2 trigger when cranking)
 

[Adamw]

2 minutes ago, retro said:

I was just wondering in the chart that Adam posted, what "link injector timing" was.

The injector timing would be set to start of injection and 410degBTDC to achieve the timing shown in my chart.   

[Adamw]

14 minutes ago, retro said:

Hmmm wonder what the heck is going on with these MPI cam triggers (I need to find a pic of how the cam triggers as I didn't build the engine)
Unsure if its tooth on the cam, or a window, slot etc (but I get basically NO change in voltage on cam 2 trigger when cranking)

The cam sensor fits in the old mechanical fuel pump hole at the back of the engine, the MPI camshafts have special lobes on them for this sensor.  So possibly if someone has fitted a SPI or non-efi camshaft you will be missing these lobes.  You might be able to confirm with a snake camera. 

 

[retro]

Cheers for the replies Adam. Yes they had the cam made and copied off an MPI camshaft. It has those lobes and I just got a call from the owner that they are around 5.9mm-6mm apart (not super scientific measuring method but putting in drill bits/punches in the back there and seeing what fits the tightest between the lobes). (he's SURE they copied exactly off the MPI cam, but we don't have one here to measure the factory gap between the lobes.)

The factory flywheel gap for the triggers is around 4.8mm width (he's actually running custom a 36 tooth crank trigger with the same gap to the trigger teeth.

the peg on the trigger sensor(s) is around 2.7mm diameter.

So sounds like a 1mm gap for the crank a VR trigger... I was getting +0.8-1volt at the crank trigger while cranking (but that was 36 teeth at the crank moving over the sensor) at crank speed. So no issues with the crank trigger.

So that means the CAM has around a 1.6-1.7mm gap either side of the lobes which also seems a bit big especially when the cam is spinning at half speed of the crank.

All I got on the scope for the cam was 0.1-0.2volt noise at times while cranking... nothing representing a pattern at all. I THINK at times we MAY have got 0.2volts pattern for those lobes (but just mixed in with all the other 0.1-0.2volt noise its not really anything we can trigger off.

[Adamw]

Its a bit of an odd setup, I assumed the cam sensor just pointed at one of the lobes rather than in the middle but have never seen one in real life.  I wonder of they have made the cam out of cast iron or something Can you check if the sensor actually reaches in far enough?  

You could try cranking with the spark plugs out to see if higher cranking speed achieves a useful waveform.  

[retro]

Thats cranking with the spark plugs out and grunty battery, it would at LEAST be getting 250rpm at a guess. (I could prob work it out from the trigger scope)
"I assumed the cam sensor just pointed at one of the lobes rather than in the middle but have never seen one in real life."

Hmmm yeh maybe it does point at one lobe, but why have two lobes seems odd....
I'd have to get in there and have a look myself (car is at the guys place 20 mins away) Then I could see if it pointed at 1 lobe and how far away it is.

" I wonder of they have made the cam out of cast iron or something"
Hmmm I would have thought that be just fine for a VR trigger? (no idea what the cam blank is made from )

 

[Adamw]

1 hour ago, retro said:

" I wonder of they have made the cam out of cast iron or something"
Hmmm I would have thought that be just fine for a VR trigger? (no idea what the cam blank is made from )

Sorry, I did write more about the cast iron originally then decided to leave that theory out since it was a bit random until some of the more basic alignment and crank speed stuff was confirmed.  I must have missed deleting some of it. 

But to give a little more context so that un-deleted comment is less confusing - grey cast iron which is a fairly common camshaft blank material can have some very unusual magnetic properties, I remember when I used to do a lot of magnaflux you could get some parts that just wouldnt demag no matter what you tried and you could have multiple north and south poles adjacent to each other etc.  

[retro]

hmmm a possibility I guess, but the factory cam is prob a chilled iron blank?
But the big question is, if the sensor does point between the lobes (which the owner says it does)
Should we be picking up a signal as the lobe wipes past the side of the sensor (on each side) I'd assume YES provided the gap isn't too big (and its about 1.5-1.6mm)

The odd thing is we are picking up basically nothing at all at this stage..

[retro]

I've attached examples of trying cam pulse vs cam level as well as swapping the wires around on the sensor, lastly removing the cam sensor ad tapping it on metal.

In the second trigger log u can see there appears to be a pulse window of sorts but its only 0.2volts (mixed in with all the noise it prob wouldn't work changing the arming voltage....





 

Steve MPi Mini.pclx Trigger Scope - 2024-09-8 7;10;24 pm.llgx Trigger Scope - 2024-09-8 7;27;36 pm.llgx triger scope 3.llgx

[Adamw]

No thats definately not going to work.  0.2V is actually the limit of the trigger scope resolution, so in real life the voltage output is possibly even less than the 0.2V - anything more than 0.1V would show 0.2V due to the res.

There appears to be something wrong with trigger 1 also, there are 4 teeth much higher voltage than the rest, possibly the wheel is not running true or something.  There is also no missing tooth which isn't ideal with a high tooth count wheel, you really want to grind a tooth off to make that 36-1.  And, not 100% sure but I think the polarity of the crank sensor might be reversed too.  

 

5 hours ago, retro said:

but the factory cam is prob a chilled iron blank?

Factory cams appear to be steel.

 

5 hours ago, retro said:

But the big question is, if the sensor does point between the lobes (which the owner says it does)
Should we be picking up a signal as the lobe wipes past the side of the sensor (on each side) I'd assume YES provided the gap isn't too big (and its about 1.5-1.6mm)

Yes, I expect so - you still have a mass of iron that is moving towards and away from the magnetised coil so it should generate a voltage, but it's a bit of an unusual design that doesnt follow the usual best practices for inductive sensors where the iron mass is aligned with the magnetic poles.   

The flywheel teeth look like a similar odd design but with more iron mass involved and much higher surface speed the air gap will be much less sensitive.  Typical cam-driven distributors with VR sensors typically use an air gap of 0.25-0.4mm for example.   

[retro]

Ahh sorry I forgot to mention the cam trigger wheel was bent in that scope so it looks horrid, we have since resolved that. Didn’t notice the polarity tho I’ll double check that. 

Regarding the 36 tooth wheel, yes the intention is to remove a tooth. I didn’t want to remove a tooth before I knew where the cam trigger was (set it to multi tooth missing but not cross over the cam trigger tooth)

 Since I don’t have a cam signal we haven’t got that far.

 The custom 36 tooth trigger the guy made actually kinda mimics the flywheel design, with the sensor sitting in-between two rows of 36 teeth rather than pointing at the edge of teeth like I’d expect.. but it works as u can see in the skope even on a bent wheel haha. But it’s a steel wheel and moving much faster than a cam.

Agreed  the cam VR triggers are usually much closer air gap. 

hmmm so given a MPi ecu works with the OEM cam trigger like that. (Likely with quite a big gap) So at this stage to me it sounds like all fingers point to this aftermarket camshaft… either slot in the lobes cut too far or made of different Material. 

I’m not sure the factory has the trigger 0.4mm away from each lobe though. Especially as the cam might be able to move side to side slightly. Being the cam sensor peg is 2.8mm in diameter and the lobes have a spacing of around 5.8-6mm on this cam. (In the pictures on the net all the cams look like the spacing is at LEST 5mm which would give 1-1.5mm gap. 

Hmmm maybe I need to hit up all these UK people that messed with MPi back in the day on megasquirt haha. They prob skoped loads these things..

[Adamw]

Is this thing boosted?  If not you could connect the MAP sensor to one intake port and do map level sync.

 

49 minutes ago, retro said:

Regarding the 36 tooth wheel, yes the intention is to remove a tooth. I didn’t want to remove a tooth before I knew where the cam trigger was (set it to multi tooth missing but not cross over the cam trigger tooth)

Cam crossing over the missing tooth area is not a problem in G4X as there is a "sync tooth" setting to tell the ecu where to restart the sync test.  In G4+ and earlier this reset was hardcoded as tooth 1 after the gap so you didnt want to cross that.    

[retro]

Remember it’s Siamese port so we don’t have “one intake port” we have two runners that are shared. The map sensor from memory is in the plenum.
Or do you think it would still work?

 Good to know about the G4X code. Ah well if and when we get the cam trigger working I’ll still remove a tooth away from the cam sync, may as well just as a best practice.

it’s starting to sound like we might need to fab something else up else engine come apart

Question: The owner has found a hall sensor thats wider and will fit in the back of the block and appears to be the correct depth (or we can adjust it), the thinking is it would MAYBE sense the two lobes coming past (pointed at the cam)
Going to give it a try anyways as it might be slightly less sensitive to the gap than VR. Prob not ideal bolted to the cast iron block tho (temperature and vibration wise)

[Adamw]

13 hours ago, retro said:

Remember it’s Siamese port so we don’t have “one intake port” we have two runners that are shared. The map sensor from memory is in the plenum.
Or do you think it would still work?

It would still work since the siamesed cylinders are 180deg apart rather than 360.  So with the MAP sensor connected to one port the ecu will see 2 intake depressions on one crank rev, none on the next rev. 

 

13 hours ago, retro said:

Question: The owner has found a hall sensor thats wider and will fit in the back of the block and appears to be the correct depth (or we can adjust it), the thinking is it would MAYBE sense the two lobes coming past (pointed at the cam)
Going to give it a try anyways as it might be slightly less sensitive to the gap than VR. Prob not ideal bolted to the cast iron block tho (temperature and vibration wise)

 Yep that sounds fine, I was going to suggest something similar.  I forget what the A series fuel pump hole looks like but I was thinking you could possibly use something like the GS100502 offset a little to one side in a simple threaded plate.  GS100502 are M12, there are other less common ones down to about M8 too.  Something like this:

  

[retro]

"So with the MAP sensor connected to one port the ecu will see 2 intake depressions on one crank rev, none on the next rev. "

yeh I was more meaning if the code was fine with seeing the two pulses close together like that, sounds like its fine. To achieve this, the guy would need to drill and tap the manifold and move the map sensor to a runner (which is suspect might cause a bit of an lumpy map signal at idle on this engine), then block up the original map hole. At that point we may as well just try the hall effect first.

The fuel pump hole in the A series is fairly small round hole around 20-25mm diameter (just going from memory) with one tapped bolt hole.

So not sure you would get the sensor to one side (depends how wide the Cherry hall is), BUT if it had a wide enough point (rather than a fine peg) it might sense the two lobes.

I believe the guy has got a mitsi sensor (unsure what model) reckons will fit in the hole. Else I'll give that GS100502 a go. Cheers

Factory MPi crank sensor below.

Example.

[Adamw]

The GS100502 is M12 x 1 thread so in a 20mm hole you would get it offset enough to be directly on top of one lobe.  

1 hour ago, retro said:

yeh I was more meaning if the code was fine with seeing the two pulses close together like that, sounds like its fine. To achieve this, the guy would need to drill and tap the manifold and move the map sensor to a runner (which is suspect might cause a bit of an lumpy map signal at idle on this engine), then block up the original map hole. At that point we may as well just try the hall effect first.

Yeah that is why I asked if it was boosted, If NA you don't necessarily need to use the MAP for fuel calculation.  But the cam sensor would still be preferable over MAP sync if you can get something to work.  MAP sync can be more fussy if for example you need to open the throttle when cranking etc.

[retro]

Thanks for the suggestions and replies, yeh and if anything with this cam, its prob lumpy as and I might need to use TPS for Y axis yet...

Good point about throttle open when cranking..

Ok ill let the guy know about the sensor.

[retro]

Ok more testing tonight, checked the mitsi sensor the guy sourced and its a J5T23071A sensor which appears to be an EVO4-9 cam angle sensor (amongst many other vehicles)

I used the trigger 2 config your EVO 4-9 basemaps have. Wired as per the pinouts I could find online.
Supplying +5v, signal, and trigger ground.

Oddly we tried to test this on the bench with the link trigger scope and waving a screw driver across the sensor but got nothing.

So tried it in the factory cam trigger hole, setting the sensor up with 1mm clearance to the lobes.

So appears we have a pattern (scope attached) Signal stays at 3.6 volts then goes 0 volts once every 720 degrees.

Will the link support this as a trigger 2? (ideally we need to see NO then a YES for trigger 2 each time. But we have the opposite currently)
 

trigger 1 VR trigger 2 HALL.llgx

[Adamw]

This looks good.  All the ecu needs is a rising or falling edge once a cycle.  

[retro]

Yeh just confusing when it says trigger 2=YES the whole time then NO when it triggers.

I was thinking it needs a pull down resistor so it reads 0volts or NO. Then a YES when it triggers up to 3.6volts. 
 

sounds like not required as that’s how the EVO map is configured and wired (how we have it currently)

regarding how many teeth to remove from the 36 tooth. Any best practices? For example is removing 2 teeth likely to be better than 1 in terms of potential sync issues on starting or cranking, or high RPM.

[Adamw]

G4X missing tooth detection is pretty foolproof as it does a further validation by comparing to the teeth after the gap as well, but in terms of best practice, 2 missing teeth is going to be harder to miss than a single missing tooth.  And for gap location I like to have the gap pass the crank sensor either 90deg before or 90 after TDC so it is well away from any spark event where you want best resolution.  

[retro]

"2 missing teeth is going to be harder to miss than a single missing tooth" Thats what I was thinking also...
I was going to place the gap away from the cam sync tooth (just as a best practice and technically the G4X manual still says to do this) , but yeh I'll also make sure that's well away from TDC. I'll just line the engine up and estimate it based on where the cam and pistons are.

P.S When setting up a 3 wire idle solenoid, I've chosen AUX3/4, but then noticed the output pins are labelled backwards (open/close) from the help manual. Also the label does not change when changing the active state high/low also (probably should as it changes the function of close/open)

Have a look and try

[retro]

Update: we removed 2 teeth from the 36 tooth wheel, away from the cam sync teeth and also tried to be as far away from TDC as possible (possibly not as much as we would have liked though) once we found the offset...

It seems we needed a -147.5 trigger offset to sync the light to TDC. (odd I haven't had to use 0.5 to get a car bang on before) changing the value by 1 seemed to change the light by 2 degrees? 0.5 value was 1 degree.... (I wasn't working the light, someone else was so can only go off that)

Also in the scope attached, why does it count the 2 missing teeth as "Tooth after gap"
I thought it should read:
1: Tooth before gap
2: GAP
3: Tooth after gap

But it currently in the scope says
1: Tooth before gap (looks correct)
2: Tooth after gap (but this is the 2 missing teeth AT the gap not after it)
3: Counting timeout (appears to be the 1 tooth after the gap)

How can we tell if we have the crank VR trigger wired correctly? 
Note we are getting an RPM signal while logging and flashing no1 spark plug correctly (wasted spark) so no idea if we are 180 out of not yet... But our gap does indeed appear to be 147 degrees before TDC.

Also I notice we keep incrementing the trigger 1 error count, it incremented from 33 to 35 after we stop cranking (But i notice trigger 2 signal stays at YES)....

Should the hall trigger be set to falling edge or ALL edges (I would have through the clean falling edge is what we should be trigging off, not also the rising edge (or all meaning both right...)

removed 2 teeth crank.llgx

[Adamw]

16 minutes ago, retro said:

It seems we needed a -147.5 trigger offset to sync the light to TDC. (odd I haven't had to use 0.5 to get a car bang on before) changing the value by 1 seemed to change the light by 2 degrees? 0.5 value was 1 degree.... (I wasn't working the light, someone else was so can only go off that)

With wasted spark and a dial back type timing light the advance will be doubled, so it either needs to be set at zero or some have a "2 stroke" mode. 

 

18 minutes ago, retro said:

But it currently in the scope says
1: Tooth before gap (looks correct)
2: Tooth after gap (but this is the 2 missing teeth AT the gap not after it)
3: Counting timeout (appears to be the 1 tooth after the gap)

The trigger 1 state tells you the event it is looking for, so for example when you are in the gap it will be looking for "tooth after gap", this state will change when that event occurs which will be when the first tooth after the gap falls through 0 volts.  

 

24 minutes ago, retro said:

Also I notice we keep incrementing the trigger 1 error count, it incremented from 33 to 35 after we stop cranking (But i notice trigger 2 signal stays at YES)....

This is normal as when the ecu stops receiving teeth it is considered an error. 

 

25 minutes ago, retro said:

Should the hall trigger be set to falling edge or ALL edges (I would have through the clean falling edge is what we should be trigging off, not also the rising edge (or all meaning both right...)

Rising or falling is fine, dont use all as that would give 2 edges per cam tooth.  It doesnt need to be a "clean" edge, it has no relevance to timing or anything critical, all the ecu does is check if it received an edge or not in the last crank revolution.  

The scope looks good.  

[retro]

Thanks for confirming everything Adam, and I'm glad everything looks good.
Just to note its an mitsi EVO cam sensor and all the G4X EVO basemaps have the  trigger 2 edge set to ALL. (was only reason I was suspicious)
 

1 hour ago, Adamw said:

With wasted spark and a dial back type timing light the advance will be doubled, so it either needs to be set at zero or some have a "2 stroke" mode. 

We initially used a dial back timing light, then grabbed a normal light to check also (as I thought the wasted would mess with the dial back lights calc).

Our reference timing was 0 TDC, we would expect twice the pulses yes so one at TDC 0 and one 360 out.
So using a normal light, but changing the offset by 1 moved the crank angle 2 degrees? (so twice the pulses yes but doubling the offset???)

Using a normal light I wouldn't expect a change in 1 degree offset to move crank by 2 degrees. (then again I wasn't working the light to confirm this..)

We confirmed -147.5 degrees with a normal timing light had it bang on 0 TDC.