Electromotive XDi-2 ignition system problem

I have used Electromotive ignition systems since the late 1990's with good results. My current XDi-2 is not currently working. When I turn the ignition on the upper LED remains red and the lower status LED remains green. When I crank the engine over the upper LED remains red and the lower status LED remains solid green when it should flash. There is no spark. According to section 6.0 Diagnostics, this means XDi-2 does not see the trigger wheel. The 3/8 inch chisel point mag sensor had a resistance of about 613 ohms measured at the J1 connector pins. I replaced this sensor with a new one from Clewett Engineering. The lead length was only 3 feet so I had to build an extension. The results were the same except that I measured the replacement sensor resistance at about 630 ohms.

I purchased a used XDi system on ebay but it also did not see the trigger signal with either the original sensor or the replacement sensor,Can anybody help?
Bill

 

I've never used an Electromotive system, but here are a few questions/comments that are hopefully helpful:

1. Is this a new install that has yet to work or is this a previously working system that went bad?
2. What is the air-gap from the sensor to the trigger wheel? A brief look at the manual showed a max air-gap ranging from 0.025" up to 0.080" depending on the trigger wheel diameter.
3. How is the crank trigger mounted? From the description it sounds like the 3/8 chisel sensor can be damaged with a set-screw type mount.
4. What is the diameter of the trigger wheel you're using? If I'm reading Table 2 on sheet 12 correctly, different diameter trigger wheels favor different style crank trigger sensors:
https://electromotive.com/wp-content/uploads/2015/05/XDI-2-v1.3.pdf
5. Maybe it would be possible to get your hands on one of the "crank trigger simulators":
https://electromotive.com/our-products/crank-trigger-simulator/

I hope this helps.

-David

 

Hi David. Thanks for you input.

1. This is a system that previously worked. We saw 350 bhp on the engine dyno with open stacks.
2. The air gap is currently set at 0.032 inches.
3. The crank trigger is held by a pinch bracket fabricated out of 1/2 inch aluminum plate.
4. The trigger wheel is 5-inches in diameter and mounted to the back side of the damper.
5. I spent a good bit of time looking for an Electromotive crank sensor simulator, their part number 150-10001, but was unsuccessful. I believe the crank sensor signal is a 5V square wave. I measured a 120 Hz signal with my Fluke multimeter when cranking the engine over with both my original sensor and the new replacement sensor. A signal seems to be getting through to the ECU.

I found and purchased a used XDi system and plugged it into my harness, at the main connector, J1. While a little different, it also did not see the crank sensor signal. This leads me to think that it is not an internal crystal failure as some have reported as it is unlikely both systems would have the same failure. I'm thinking it could be a pinched wire or grounding issue. All my system grounds are terminated at a single point. Maybe this can be improved with a heavier ground wire?

Good questions David.

 

Do you have an oscilloscope? If not can you borrow one as a score is pretty important to a job like this and you can buy something functional for under $100

Are you checking the signal at the ECU pins?

I posted this on another XDi thread...littery any aftermarket ecu will run spark at least as well as the XDi, so it's pretty easy to just replace it with something modern and supported.

 

How old is your unit? I was told a couple of weeks ago that Electromotive is no longer in business. If so, are spares still available?

 

How old is your unit? I was told a couple of weeks ago that Electromotive is no longer in business. If so, are spares still available?

 

Thanks mk e.
Somewhere I have a 50 MHz scope that I have never used. I'll see if I can dig it out and see if it still works. I got it about 20 years ago from a guy who used it for tech school.
I was checking the pickup signal on the mating side of J1, the main connector.
Richard Clewett suggested a Performance Electronics PE3-IG as having equivalent functionality to the XDi-2. He says I should be able to use my existing trigger wheel and coil packs.

Hi Yoric.
My XDi-2 was an early one, so I think it's about 10 years old. We updated the firmware to the latest available WXD 0605000 on 6/26/19. New spares are apparently not available.

 

If you have signal at the main connector a wiring issue seems unlikely. A scope will confirm the shape of the signal is good and clean.

Again, literally ANY ecu that you can buy these days will replace the electromotive and read all the same sensors and work with the coils if it comes to that. Hopefully though it's just the sensor or some other wiring bug.

 

If I understand the website correctly, I believe you have a variable reluctance sensor. I’ve only used a Hall effect sensor, so I’m not going off personal experience here, but I believe VR sensors output is proportional to both the air gap and the RPM. A smaller air gap gives a larger amplitude and higher RPM gives a higher amplitude.

Maybe the signal amplitude was previously marginal and an aging starter or battery bumped it into non-functional?

Mark’s idea to scope the sensor is a good one.

If the amplitude looks a little weak, maybe try decreasing the air gap. Another idea to increase the RPM for testing might be to disconnect the fuel pump and pull the spark plugs so the starter isn’t fighting the compression (obviously it won’t start but maybe you’ll see blinky lights on the XDI to see if you’re on the right track).

Hope this helps.

-David

 

How many wires on the crank sensor? Relutor will be 2, hall will be 3 as a general rule.

If 3, generally 1 is power, 1 gnd, 1 signal, maybe confirm that with a meter.

If 2, relutors are sensative. Yes to the rpm and air gap, but also the core relative to the trigger tooth width if the trigger wheel has square teeth....which I learned the hard way.

 

Thankis guys. I dug out my never used by me oscilliscope. It's a 150 MHz Tektronix 2445B. There are two channels. I'll try to figure it out how to use it tomorrow.
My mag sensor is a 3-wire. There is a twisted pair for signal and a shielded ground (foil wrap with a bare wire). It is my understanding that Mag sensors consist of an iron core with a coil around it for a 2-wire (+/-) signal and a permenant magnet at one end. The sensor is noticeably magnetic in that it will easilly support its entire weight when attached to an iron body. Presumably, the iron tooth on the trigger wheel induces a pulse as it passes the end of the pickup. There is no power to the sensor, it's passive.

 

Hi guys.
I believe this schematic represents my mag sensor. What is not shown is the wire shield. I don't think the shield is attached to anything at the sensor end but is terminated inside the ECU.

Image Unavailable, Please Login

 

Ahh, the oddly built connector from the other day. Yes, a reluctor is like a generator, as the tooth passes a magnet a current is created, + as the tooth approaches, - as it moves away or the other away around if the wires are reversed. But the key it the voltage is 0 when the center of the tooth is at the center of the sensor. The ecu sees + and gets ready then when it sees 0 or more accurately when it sees + turn to -, it triggers. If the tooth and trigger are a mismatch there is dwell are 0 and timing is inaccurate, obvious on a scope.

I think I have the same scope you're describing.....it's easy to use.

 

mk e
It turns out the 2445B is a 4-channel scope. I spent some time looking it over and going through the operator's manual. I have four Tektronix P6137, 10X probes that appear to be in good condition. Eventually, I plugged in scope and let it boot up. No error messages were displayed so all looks good.

1. To view the wave form quality, I think I just use one of the probes connected to one channel (CH1) and connect the probe to one of the sensor leads. Do I then connect the ground clip on that one probe to the other sensor lead?

2. Once connected, I need to turn the motor over to see the wave form and view any noise or other unexpected features.

3. Do I need to set a trigger examine the wave form?

4. Do 10x probes amplify or attenuate the measured signal by a factor of 10?

5. Do you have a recomendation for a manual V/Div setting?

6. Not sure about the autorange setup yet.
Bill

 

1 yes
2 yes. There is probably a record feature so you don't need to crank and look
3 no
4...you know, I'm not sure I've ever figure out what that means non has it ever mattered to me
5 try 1v/div while cranking...or 0.5, there won't be a lot of voltage at cranking rpm
6 maybe that will set the time for you? It should crank at like 200rpm, times 60 teeth, but divide by 60 seconds so 200 teeth were sec so scale time for no more than 1 sec full screen. That will show you the whole pattern but maybe too small, so 1/2sec full screen might work better.

Good luck!

 

4 googled it, a 10x probe reduces the voltage so the scope can read a larger range. All the probes I have are 10x, but 1 has a switch for 1 or 10. My scopes both want 10x probes as standard so all the numbers are right, no need to convert anything. I vaguely remember there is a setup screen to pick 1x or 10x but default is for 10x.

 

mk e,
2. Unfortuneately, I don't see anything that looks like a record function, so I may have to crank and look.

 

And video

 

Somewhere on the scope you will see “trigger” options. “Auto” means the scope continuously acquires and displays waveforms. “Normal” means that an acquisition will be displayed every time the trigger condition is met. “Single” will capture and display the first time the trigger condition is met.

For the condition I would probably select “rising edge” or similar and put the trigger level at 0.5V or something like that. See what you get.

I’d recommend you set the trigger to single and see what you get after you crank it. The waveform will remain on the screen if it triggered.

As Mark mentioned, x10 probes need to have the channel also setup accordingly. Most scopes have a “calibration” port you can clip onto and check the waveform. Usually they have a 1V or 5V square wave. Use this port to make sure everything is correct. It’s easy to have a mismatch between the scope and probe gain and you end up with a waveform that is off by a factor of 10 (and you scratch your head wondering why you didn’t get a trigger).

-David

 

One more thing… I think I would probably put the ground lead of your scope probe on the block or chassis. Connecting the probe ground lead to one of the VR sensor leads may have unintended consequences. You are effectively connecting house ground to your sensor. If you have a battery charger also connected you may end up with issues.

You can effectively make a differential probe by connecting one probe to each side of the sensor and subtracting the two channels (I think your scope probably has this option). Or just view them both separately.

-David

 

I think if the sensor is unplugged then checking either lead to chassis ground there shouldn't be a signal...I don't think? The sensor is just a coil of wire with no connection to the housing so no connection to chassis gnd. Plus if I'm remembering right, until there is a load of some kind connecting the 2 leads there is no signal, its just an open circuit...I think and you'd need to gnd the other sensor lead for that to work but it does sound safer?

My step 2 is plug it in and make sure the signal is still there and when I do that, I use a paper clip or of piece of 023" mig wire to slip in the back of the connector on either lead at the ecu....but I'm still remembering using both legs, although now perhaps 1 and gnd would also work. In the wiring diagram the 2 ecu pin will be marked +/- usually and the with the probe on the plus pin, gnd to the minus the signal should go + then - symmetric around gnd. If it goes - then + you have the sensor wired backwards. I'm still not sure sure if you'll get a signal grounding the probe to chassis, I'm not sure I ever tried but it seems like it should work I guess.

I never tried that, I like it.

I don't think I've ever had to tough the trigger setup....auto seems to work fine but all I ever use the scope for is checking ECU setup, crank and wheel sensors going in, spark and fuel coming out, very basic so I'm anything but expert with it.

 

Mark, with the VR sensor unplugged you are correct that it is 100% fine to ground one leg.

It’s probably also fine when it’s plugged in but I’m not sure what the output of the sensor sees, so grounding it while plugged in (especially if there is another ground connection through a charger or something) may change the waveform. I think I was mostly worried that the ECU grounded one of the legs already.

Normal and single are nice in that whatever triggered last will remain on the screen. If you’re working alone it’s nice for not having to be at the scope when the event happens. Also good if the waveform isn’t repeating. The downside is you have to setup the trigger.

-David

 

I got lucky and managed to persuade a friend of mine, who is an EE/CE, former college neighbor, and aerospace employee, to come over after work tonight and help me acquire and evaluate my mag sensor signal. Thank you RC.

 

We managed to get some images of the mag sensor signal while cranking. This image was taken while turning the motor over at about 120 rpm (I use a more modern geared starter).
Image Unavailable, Please Login

There is a good deal of noise on the signal but I don't think it's enough to prevent a spark event. Suprisingly, it's more of a sin wave than a square wave. The signal amplitude is about 800 mV, noise spikes are about 200 mV max, and the trigger voltage is 0.39 V. Anyone see a show stopper here?

I measured 10.4 VDC on the yellow, +12V power line to the ECU at J1, when cranking.

I have a second XDi-2 programed for my 4 cylinder, Lotus 26R build and tried it on the Ferrari. Again, it does not see the trigger signal. In total, that is 3-ECU's I have tried and none of them see the trigger wheel. This leads me to believe the ECU's are likely OK and the problem lies elsewhere.

Any thoughtst?

 

Without being familiar with VR sensors or the electromotive system, my first guess here would be we are not meeting a minimum voltage requirement. I did a little searching around and on the MegaJolt which uses EDIS (different system I know), it expects 0.5V (peak if I understand correctly, not pk-to-pk) min:
https://www.bgsoflex.com/mjl/mjl_edis_summary.html
You're just shy of that. It's a different system of course but I would assume other systems are in the ball park and I'd also assume that a normally operating system would probably have some good margin above the minimum.

If it were me, I'd tighten the air gap on the sensor and see what happens.

Also I would probably wait for Mark to weigh in... he's got a lot more experience in this department than me.