VR Crank Trigger EMI Question

[TC951]

Hello,

So, I am fairly convinced that I will need to change my current trigger from a VR to a hall effect, but I'd like to get some expert options first.  I am also just curious for my own internal knowledge going forward regarding how much noise is considered too much...

Basically, I've finished designing and started building a loom for my G4X Extreme install into a Porsche 944 Turbo.  "Stock," This car uses a VR trigger that picks up on the flywheel ring gear for the starter, with 132 teeth.  The PCLink SW has this as an option, however I had decided to switch to a front mounted 60-2 kit.  After installing the kit I became uncomfortable with how close the sensor is to the radiator fan.  I was concerned about possible EMI causing trigger errors, or inaccurate RPM readings.  Picture below of the sensor location to the fan:

]]

 

So before going any further and finishing my loom with a VR trigger I decided to make sure that this would not be a concern.  I measured the signal with a PicoScope while cranking, and turning the high speed fan motor on/off.  In this car, the high speed will not actually operate during cranking, however it will operate on/off at idle or cruise based on coolant temperature (the fans are controlled chassis side by a temp switch in the radiator, not by the ECU and I will keep it that way).  So this was the best method I could think of to just check what effect the fan motor might have to the sensor signal.  This is also maybe "worst case" since I don't have the shield wire grounded to the ECU, since I'm not wired up to the ECU yet...  However, since the actual sensor side wiring is not that close to the fan motor wiring, I think probably most of the EMI is being transmitted through the sensor body itself rather than the actual wiring.  I only have a very basic understanding of EMI though so I don't know if that assumption is legitimate. 

So anyway, pictured below is what I measured.  Is this something I should be concerned about?  Is this amount of noise immediately noticed as absolutely no good to the experts on this board?  Or is this small enough in magnitude that it could be considered negligible?  Any insight is much appreciated. 

Zoomed in a bit.  Looks like roughly 50-60mV difference valley to valley between fan vs. no fan:

 

Thanks!

[Adamw]

Probably not really a valid test, you need the ECU connected as that will attenuate the signal and probably improve it a fair bit as the input circuit has hardware filtering etc, the shield will likely improve it further.  

I dont really have any concerns with the proximity of the fan motor.  Many cars with distributors for instance have the VR sensor sitting an inch below the rotor with HV sparks and even CDI sparks flying around in there.

Note also your crank sensor is showing the wrong polarity, so when you connect it properly later on you need to swap the +/- pins from how you had it in this test.

 

[TC951]

OK got it.  I understand this was not really a "valid" test, but since the results clearly show that the fan motor does have some effect to the sensor signal I just wanted to know how much is worth being concerned about.  However I didn't realize (stupidly) that the ECU would provide further filtering and attenuation of the signal...So yea probably hard to really give an answer based on a signal not processed by the ECU I guess. 

I'm glad to hear you aren't concerned about the proximity of the sensor.  That gives me more confidence but I think I'll still do this test again wired up to the ECU using the trigger scope function to be sure.  My loom is sheathed in DR-25 so I don't want to be modifying it for any reason after finishing and want to be 100% confident in the trigger signal integrity.

Yep I was taking a million screen shots of both +/- input to the scope to see both signals and when swapping one for the fan speed motor input I was doing it in haste as my helper that was cranking the engine over while I powered the fans needed to leave, so I accidentally left the negative lead hooked up instead of the positive.  Figured in reality it didn't matter for the test anyway b/c all I wanted to see was any EMI from the fan motor.  I'll make sure it's correct for the ECU test.

Thank you for the quick reply and for your input. 

[TC951]

So, I've captured a trigger scope log after bench testing with just the trigger 1 input to the ECU.  The scope data doesn't look so nice though.  I also had a bit of trouble getting the an RPM signal at all.  At ~160rpm cranking speed the sensor was outputting like 0.2v.  So after changing the arming voltage at 500rpm from 0.5V to 0.2V I can get a wave form but it's not great obviously.  This was at 0.89mm sensor to wheel gap.  I moved to 0.5mm and the amplitude increased to ~0.68V.  

So I have a couple questions before moving forward:

1. Am I going to struggle with such low cranking trigger voltages if I stay with this setup?  This engine uses 20w50 oil and has quite a heavy flywheel/clutch and reciprocating mass...So cranking speeds will always be low, especially on cold starts.  

2. Do I risk the missing tooth gap not being correctly measured by running a wheel to sensor gap of 0.5mm? 

3. I've attached a log of the 0.89mm and 0.5mm trigger scope logs.  Do these look acceptable?  To me they don't look super clean but I have no experience with this. 

My goal is to still determine if the fan speed effects the ECU's trigger measurement but I want to make sure I have a solid reliable trigger input with no fan speed first...

Any advice is much appreciated. 

 

Thanks

 

 

 

TriggerScopeLog_2_Cranking_0.89mmgap.llgx TriggerScopeLog_3_Cranking_0.50mmgap.llgx

[Adamw]

Yeah the 0.5mm air gap scope looks fine.  The important bit is the zero crossing which looks good and clean.  The 0.9mm air gap would be borderline in terms of voltage.  0.5mm is not an uncommon gap for a reluctor - although with high tooth counts it doesnt usually need to be so tight. 

Is the sensor centered/aligned well with the teeth?  In the photo above it almost looks like it might need to be shimmed forward a little.

 

[TC951]

I think there is some parallax in the picture, so it looks worse than it actually is but yes it does need adjusted.  I'll be adjusting that, thanks again for your help.

I'm slightly concerned in a cold ambient cold start, or with a slightly low battery capacity, getting to the 0.2V/0.3V arming voltage might be borderline.  This test was also done with the spark plugs removed...So I'll be re-installing plugs, making sure the sensor alignment is perfect, and re-checking.