I have an 87 e30 that has an issue with the speedometer. It usually doesn't work, but occasionally it does. It's not an exact correlation, but it seems to be more likely to work when the car is REALY cold, I am going really fast, or if the engine is running very hot (only happened once when my head gasket let go). I figure there is a short somewhere but I am not sure how to reproduce a speed signal to test the wiring to and from the sensor and to test the function of the gauge itself. Any ideas?
Start by checking the signal. If it's good, then use it to test the gauge. If not, well, fix it 
Testing wiring to/from is just continuity. The signal isn't important.
I have treated the wires for continuity and the seem to be ok. I suspect the signal issue is in the gauge cluster so I was hoping to send a test signal into the cluster to see if the speedo moved. I just don't know what the real signal would look like and I don't want to burn anything up trying. From the pick up configuration I am guessing it is a pulsed 12v signal. Is there a way to test that or reproduce a fake pulse?
Let me start by saying that I know nothing of BMW speedometers.
Now, that being said, I had a similar issue with one of my 911s. It uses a magnetic sender that plugs into the side of the transmission. I replaced the sender with no luck. I also had the speedometer rebuilt. Still didn't work right. I decided to try cleaning the magnets on the ring gear.
I jacked the rear of the car up, removed the sender and used a pencil eraser wrapped in a shop rag to clean the junk off the tiny magnets. I think there were eight. Find it by feel. They're small. Just wipe it a little, remove the pencil and move to a clean spot on the rag, turn the rear wheel an eight of a turn and wipe then next magnet.
Worked like a charm and I would have save several hundred dollars had I just done this first.
I wouldn't be surprised if BMW uses a similar system.
You can almost kind of see them through the hole:
itsarebuild wrote:
I have treated the wires for continuity and the seem to be ok. I suspect the signal issue is in the gauge cluster so I was hoping to send a test signal into the cluster to see if the speedo moved. I just don't know what the real signal would look like and I don't want to burn anything up trying. From the pick up configuration I am guessing it is a pulsed 12v signal. Is there a way to test that or reproduce a fake pulse?
It's probably pulsed 12v or 5v, that's fairly common. You might be able to check it with a an analog voltmeter at a slow speed, you'll see the needle twitches more easily than with a digital one. A more powerful 'scope would give you more options, probably letting you see the frequency.
Once you know you have a good signal, you have your test signal source
