Speed/Reference Sensor Testing Help
 

I bought my 944 about 2 weeks ago. It ran great for about a week and then I got a "no start". Been trying to troubleshoot it over the last week. DME relay tested up fine and jumping it didn't fix it. I get no spark and coil tested fine. According to the troubleshooting procedures from Clarks Garage I am down it either being the Speed or Reference sensor or the DME computer itself. I got no bounce on the tach so I am guessing it is the sensors.

I have followed the procedure to test both the speed and reference sensors and I get nothing on both sensors. The resistance test shows 0 on all leads. I grabbed an oscilloscope from work and it shows nothing either. I am thinking I got to be doing something wrong. I wouldn't think both sensors would fail at the exact same time and both be completely dead.

Here are some pictures of where I am testing and how I got the oscilloscope hooked up. Am I testing the correct leads? I should be connected to 8 and 27 on the DG. (The DG is on top and I did trace everything and it is all correct on the bracket). I am no expert on the oscilloscope, what should the settings be? I think I have it set for DC. I didn't try AC but I remember the speed sensors on my van put out AC volts. Could that be my problem.

http://forums.pelicanparts.com/uploa...1378602096.jpg

http://forums.pelicanparts.com/uploa...1378602121.jpg


Any help would be appreciated, thanks.

I have no idea how to test the sensors. Or how an oscilloscope works. But doesn't there need to be power for the oscilloscope to work right. The sensors get a pulsing from the flywheel/DME. Do you know any 944 guys in your area, try borrowing a DME. Sorry I can't help.

The sensors are hall effect, so they need power to get a signal out of them. You could back probe the connectors with everything hooked up and check between the signal and ground while cranking. I would guess the middle pin is signal, but I'm not sure.

When I had this issue on my car, it ended up being a broken wire under one of the connector boots on the harness side. Pull back the boots and check for that before getting to terribly fancy with the troubleshooting.

Clark garage has a simple procedure on this, including the contacts on the connector to check with an ohm meter. It also tells you what the value of a "healthy" sensor should indicate. If the mount itself has loosened up on the bellhousing, the likelyhood of the sensors moving beyond the 8mm clearance from the teeth on the flywheel would cause the failure you describe. Check the mount to see if it has moved. To adjust the clearance of the sensors, glue a washer that's 8mm thickness on one of the sensors. Loosen the mount assembly and move it down until the washer/sensor is bottomed on the teeth of the flywheel and then tighten the mount. Remove the sensor and remove the washer and wipe the glue off with acetone so it's clean and the reinstall.

8mm? I thought I set mine at like 2mm? Are you measuring from the flywheel to the sensor, or the pin to the sensor? It's been so long that I don't remember how I did it last.

Check the wires from the DME plug to where the sensors plug in. Make sure you have the right resistance there first. It's easiest.

speed and reference sensor testing
 

both sensors are the same part take an ohm meter and measure across center and outer terminals should have between 900 to 1100 ohms very rarely is sensor out of adjustment. gap should 0.8 mm (less than a millimeter) make sure plugs are connected to the proper sensor switch plugs and try again if you did not mark connectors. tach needle should slightly move when cranking over engine also check dme relay many have solder contacts cracked inspect with magnifying glass. try this and get back to me if still no go. regards tom

Yeah, 0.8mm sounds right.

If these sensors are original equipment, my humble suggestion is just to bite the bullet and purchase new ones. I had an issue with mine where the resistance tested fine ( I don't have a O-scope to test proper wave function) but my tach bounce was lacking during start-up. I think, due to vehicle age, the magnets simply lost their strength which caused the failure. Anyway, this host does carry them at a fair cost although the leads are a few inches longer. Additionally, I was informed that BMW used the same sensors during that time period and may be had for a much cost (please research this assumption). Good luck!

How much tach bounce is enough? I've seen mine go as high as 1,500 RPMs, and as low as just "a wiggle".

First off I get zero bounce on the tach. It doesn't move at all and I have checked that many times.

I did the ohm test described on Clark's garage first. I used my multimeter setup for resistance and I know how to use it becuase I tested the resistance on the ignition coil just fine. I had the multimeter connected to the same pins that I have the o-scope connected to in the picture above. I get 0 on both the speed and reference. Clark's also says to test the two outer pins and there should be a much higher resistance. That shows 0 too. I can't see how the bracket being out of alignment could cause the resistance test to fail. I would think that would be internal to the sensor. I could see the sensor reads being off if the gap was too great.

One thing when I first drove the car home I thought the speedometer was off. I don't remember where I read this but I think I read those sensors play a part in what the speedometer shows. Maybe the bracket is loose and it just worked itself too far out and it is no longer getting reads. I got some of the maintenance records on the car at it had a new clutch put in about 8 months ago. I am guessing that when that was done those sensors would have had to been out or at minimum disconnected. It is possible during that procedure they got messed up or not put back on properly.

The car only has 54k miles on it so there is a good chance they are the orgiinal sensors. Maybe the best solution would be to just pull them both and replace them and check clearence at the same time and adjust the bracket as needed.

I have read the nightmare of pulling the sensors and have them break off or just get stuck. Is it possible to just take the bolts out of the bracket and pull the whole assembly out and then remove the sensors for the bracket? Seems that would be a lot easier and you I could work on it on my bench and get proper angle on them when pulling. But, the sensors look long enough that they might go through the bracket and down into the bell. If that is the case the assembly might not pull off.

This might be where I am confused. You say dme plug and Clark's procedure says dme plug too. In the picture I am connected to the sensor, not the dme plug. The dme plug is the 35 pin connector that plus into the computer, right? I am think I should connect the sensors back up and move my testing to the dme plug pins. Where is the dme located? I haven't gotten that far in understanding this car to know where everything is yet.

i found the DME plug. I tested there and get a 2.2V pulse out of the reference sensor on the oscilloscope. I would conclude it is working. The resistance on the reference sensor also tested up correctly between pins 25 and 26. Not sure why I wasn't getting resistance at the sensor itself.

I still get nothing out of the speed sensor. starting to think that is where the problem is. I traced through. It doesn't look like any broken wires but might be internal somewhere. I'll check continuity next.

I am guessing that power (or ground) must be supplied on the third pins, labeled 23 and 78 on the holder (Clark's diagram) and without that connected up I was unable to test at the sensor itself. I would also assume these pin, 23 and 78 must go someplace else other than the DME plug since there is not a 78 pin on the DME connector.

I think I will order a couple sensors. swap the one and keep a spare for future. Hopefully that will do it.

My mistake, 0.8 mm vs 8 mm. I forgot the decimal point.

Contrary to a previous post, there is no power or ground to the flywheel sensors. They actually generate a pulse from either the teeth on the starter ring gear or from the tdc stud on the flywheel (at least on the 8 valve cars). The third pin is for a shield around the two signal wires.

Our bench test for sensors is to use the oscilloscope and a steel rod to generate the signal. No voltage or ground required.

I wasn't inferring that the sensor mount would affect the resistance. I was considering that if the mount had loosenend up and the sensors had moved beyond the .8mm gap, the sensors wouldn't have the impulse from the flywheel. I had a hard to remove sensor and I sprayed it with nut buster after removing the hold down bolt and left it over night. I then twisted the sensor back and forth a few times while pulling it up and it came out. You need to make sure there's a rag in the opening of the bellhousing so none of the small parts fall in. If you pull the mount and the gap is properly set to start with, then when you replace with new sensors, you'll have to reset the gap.

The speedo gets it's info from a spin sensor on the transaxle. The sensor is threaded onto the transaxle and there's a typical click-on connector on the end of it. It's possible you have a separate issue with that, either a loose connector or the sensor may be either defective or come unscrewed (not likely).

I had to replace a ref sensor recently and the I got the replacment from a supplier on ebay. It was an original Bosch unit, new, and was around $50. The ohmage value on the new ones was around 1k ohms.

Sorry about that. I took a shortcut and assumed Bosch had been sensible when designing this stuff. Normally when I see shielded wires, the ground for the shield doesn't actually connect to a pin at the connector unless there's ground in the connector already. Since the sensor plug have 3 pins, I assumed they were Hall effect sensors, but it appears from the diagram they're VR sensors. VR sensors don't need power like a Hall effect sensor does.

Lesson learned.

If the sensor are original, then absolutely it is time to replace them. Additionally, the wiring ages & corrodes (at times, beneath the sheath and not visible). So, if the reference sensor wiring is original, then I highly suggest replacing it as well.
Then you shouldn't have to worry about it for another 20+ years.

Here is a quick video of what a good tach bounce looks like:

Porsche 944 Turbo 951 Tach Bounce (engine cranking). - YouTube


Additionally, I recently put up a page about the crank-sensors:

Rogue Tuning - Crank Angle Sensors

BTW: You don't have to remove the ref/speed sensor mounting bracket unless there's a problem with it. Simply remove the 10mm (I think) bolt from each sensor; I used 1/4" socket with a universal and couple of extensions to make pretty easy. Bolt torque is 6 ft/lbs.

And be prepared to have to work the sensor out. I just did the crank sensor a little while ago, and that old o-ring was really mated into the hole.

I just bought two new sensors. Hopefully this will fix it and get me running again. I'm going to try and pull the speed sensor tonight.

At the DME plug I was getting a signal at the correct voltage on the reference sensor but nothing from the speed sensor. Which sensor will cause the tach to bounce? Or is it a combination of them both? I am going to test again tonight but when was I not getting a signal from the reference sensor when directly connected to the sensor. I only got it at the DME plug. The resistance test failed too but worked on the DME plug. Is that because I was lacking the shield? Maybe I didn't have it hooked up right.

Got them out. They slid right out. Barely had to work them at all. Just used my hands. They were probably out when they did the clutch about 8 months ago so they hadn't seized up yet.

http://forums.pelicanparts.com/uploa...1378769612.jpg

I noticed on the one being used as the speed sensor it had a dent on the top edge on one side. Clearance problem? Kind of hard to see in the image but here is a picture of it. Do you think I should check the bracket clearance?

http://forums.pelicanparts.com/uploa...1378769692.jpg

http://forums.pelicanparts.com/uploa...1378770184.jpg


The reference sensor continues to test up correctly. I was able to test it directly connected to the sensor. The reason why it wasn't working for me before was I was looking at the pins in reverse. I don' t know if my bracket is upside down but my connectors are the opposite of what is shown on Clark's page. I figured it out when I looked at the keyed part of the connector. The speed sensor continues to be dead.

http://forums.pelicanparts.com/uploa...1378769897.jpg

Both need to be working properly. (well, technically, the speed sensor could be partially working and the tach might still bounce - assuming a good reference sensor)

My old sensors tested within range but I think the magnets were just tired. The car came right back to life upon replacement.

Replaced both sensors and it fired right up.

Awesome. Glad to hear your car is fixed.

Running, not sure if I would say fixed yet. This car is far from perfect and needs sorted out. I'm starting a list of things that need done. I just started a new thread for the next issue, engine is making a clicking sound.