From digest.v5.n590 Wed Nov 20 22:11:36 1996
From: M Barry Ritchey <mbritch_at_sandia.gov>
Date: 20 Nov 1996 13:59:22 -0700
Subject: Speedo Fix for '90 325iX

To 325 Owners with bad speedometer/odometers:

     Several months ago my '90 325iX started to have speedometer/odometer 
     problems in the mornings. The speedo would stay at 0 and the odometer 
     would not advance. The problem worsened with the onset of Fall and 
     cooler mornings. The speedo/odo would usually wake-up after about 5 to 
     10 minutes/miles of driving...and stay working until the next morning 
     or cold start. Here's a long winded account of my quest to fix the 
     speedo/odo.
     
     1. My first attempt at a fix was to crawl under the car and check the 
     connector at the rear differential. Maybe it was loose or the wiring 
     was frayed/damaged. Everything appeared fine.
     
     2. The next step was to post for help on the digest and pick Gordon 
     Haines' brain. Most e-mail replies (Thanks to those that replied!) 
     suggested I check for bad solder joints on the instrument circuit 
     board. I dreaded pulling the instrument cluster (later I learned it's 
     a quick job) and would pursue other avenues 1st.
     
     3. Since I didn't have a ETM, there was no easy way of tracing the 
     wiring into the instrument panel. I thought I'd start at the source 
     and tap into the two leads (parallel connection) at the differential 
     connector. I soldered some female connectors to some ~14 gauge wire 
     and then soldered wire to my standoff connections. I ran the tap 
     wiring thru the drain slots in the spare-tire well into the trunk, out 
     the trunk, and then into the interior by way of carefully closing the 
     rear door (I have a 4-dr) on the wiring. I could now have a DMM 
     located at a convenient location (the console) to monitor voltage at 
     the differential connector. A regular DMM wasn't telling me much after 
     a few days so I borrowed a Fluke scopemeter (portable LCD scope). When 
     the speedo worked, there was a nice pseudo-squarewave (0 - 8V). When 
     the speedo sleeped, there was just mV noise. I now suspected the 
     speedo or instrument board.
     
     4. Reflowed all the speedo board solder joints as described at web 
     site: (http://www.aros.net/~pinkston/). It's not that hard to pull the 
     instrument cluster, but... IT DIDN'T WORK in my case. Looking back, a 
     bad joint solder joint might be the cause if tapping the dash or 
     hitting bumps causes the speedo to jump or turn off. My speedo was 
     always steady on or off when driving bumpy roads or when the dash was 
     hit (and I did tap/pound the dash out of frustration a few times).
     
     5. I borrowed the Fluke Scopemeter again. This time I disconnected the 
     cable at the differential and just measured the voltage across the two 
     leads. It was always 8 volts DC(actually 7.9X volts). There was no 
     significant AC component. After 4 days of cold morning starts and rock 
     steady voltage at the connector, it was time to measure across the 
     transmitter leads coming out of the differential. The Scopemeter, 
     displaying resistence vs. time, showed a nice high resistence spike 
     occuring about every 2.5 ms at around 30 mph when the car was warm. 
     The switch was normally closed (~0 ohms) and was momentarily opening 
     up into the megaohm range. For 4 mornings the meter would just read 0 
     ohms for the first few miles of driving and then suddenly show the 
     opening spikes of a working sensor. The transmitter was bad!!
     
     5. To make this too long of a story shorter, I ordered the 
     differential transmitter (and o-ring) from my local dealer in 
     Albuquerque for under $35, installed it, and the speedometer and 
     odometer are working fine for two weeks now.
     
     6. I performed an autopsy on the old sensor. It's a switch (glass tube 
     encased contact rods) that's normally closed. When in the presence of 
     the magnet that resides in the sensor housing, it opens. There must be 
     a flange inside the differential fits between the sensor and the 
     magnet. This causes the switch fo close (the magnetic field is blocked 
     by the flange. The flange is slotted, which allows the magnetic to 
     momentarily act on the switch to open it. I guess my bad switch tube 
     was just sticking closed until the heat of the differential, oil 
     churning, or vibration caused the contacts to start opening (pure 
     speculation).
     
     Lesson learned: check the easy and obvious first. You don't need to 
     have a oscilliscope to diagnose this problem. You can slowly roll the 
     car and look for resistence staying low and then momentarily going 
     high to spot a good transmitter. A bad transmitter (in my case) stays 
     closed (low resistence) when you roll the car (or spin a wheel with 
     the car on jacks).
     
     One more comment: A cable driven speedometer sure if a lot easier to 
     fix and troubleshoot.
     
     Here's the part numbers for my '90 325iX:
     
     62-16-8-357-020 Transmitter ($37.28 list)
     
     33-11-1-206-166 O-Ring ($1.40 list)
     
         Hope this helps someone,
     
                                 Barry Ritchey