Tuesday, September 16, 2014

Diagnosing a speedometer failure on a 1999 Mercury Cougar

(Update) Bottom line: the VSS signal is too low for the instrument cluster. It needs at least 7.2V on the VSS signal wire at the instrument cluster for the speedometer to function. When the instrument panel is unplugged, the VSS puts out enough voltage. But when instrument panel is plugged in, it loads down the VSS signal voltage. Read on for the whole story. 


Original post
I just bought (Sept 2014) a 1999 Mercury Cougar that had a non-functioning speedometer. From what the previous owner told me, the speedometer stopped working maybe 6-8 months ago. It turns out that the last oil change sticker on the window was within about a thousand miles of the current odometer reading. The sticker was dated January 2011. Somebody has a poor estimation of time. The digital readout on the left side of the instrument cluster would sometimes read 1 or 2 MPH when set to average speed. But the speedometer needle was stuck at 0 MPH.

The first thing I did was check the OBD-II reading on the vehicle speed sensor. There was no code for a VSS problem. I can plug-in my bluetooth OBD-II and have the speed read out on my phone's Dash Command app. So this tells me the problem was not with my VSS.

Next, I checked the connections to the instrument cluster. When I unplugged the instrument cluster and plugged it back in and turned the key to on, the needles went to full and back to zero. The speedometer gauge itself seemed to be fine. So the problem was between the VSS and the gauge. The VSS to PCM was good because I could get data out of the OBD-II port.

The VSS signal goes to pin 9 on the white connector. It's a white/blue wire on my car. To test for signal there, I jacked up one front wheel, put it in neutral, and connected my volt meter between pin 9 and ground (frame). I rotated the tire by hand and saw the wire getting 12 volt pulses. That told me that the signal was getting to the instrument cluster.

The problem appears to be between the plug and the gauge. It sounds like I need a new instrument cluster. Ford part number 10849 for $338.46. For now, I'll just use my phone's app to tell me my speed.

While I was in there I decided to replace a mini instrument panel bulb that failed. The tiny bulbs are number 2721. For some reason, they aren't listed in the parts catalogues at auto parts stores.

I found an instrument cluster on ebay from a junkyard for $28 and ordered it. Unfortunately, its speedometer doesn't work either. It's not the gauge itself. It's somewhere in the electronics on the board. Does every 1999 Cougar speedometer fail?

Update 28 August 2015
I decided to try one more instrument panel, so I ordered one online from a junkyard. I was hoping, but getting ready for disappointment. The speedometer didn't work on it either. Sigh. I figured it was somewhat unlikely that all three had failed. So I did some more looking online and found a helpful post about the instrument panel loading down the VSS signal. Now I can't find that post to link to it.

From my calculations, my VSS signal at pin 9 on the white plug has an output impedance of 8.7 k ohms. That's too high to drive the speedometer. In case you're curious and want to check my methods and math, I rotated the tire until the VSS had a high signal of 11.34V without the instrument panel attached. I then connected a 3.95 k ohm resistor from the white plug's pin 9 to ground. That loaded the VSS signal down to 3.53 V.

I was getting about 4.9 V when the instrument panel was connected. Apparently the PCM is happy with a lower VSS signal voltage. That's why the OBD-II app would show the speed, but the speedometer wouldn't.

So I took a variable resistor and connected it in series between 12 V and the instrument panel pin 9. I'd scratch it across the 12 V terminal to get a bad square wave signal and see if the speedometer moved. I turned the knob until I found where the cutoff was. The voltage on pin 9 had to be above 7.1 V on my panel or it wouldn't move the speedometer. I then measured the variable resistor and it was 3.9 k ohms.

The speedometer operates by frequency, and the voltage of that signal needs to be above about 7.2 V for it to work. 

Now I need to figure out if a new VSS will have a low enough output impedance to drive the circuit. If not, then I need to install a signal amplifier. But first, I need to use my 3.95 k ohm resistor directly on the output of the VSS to measure the output impedance and make sure the problem is not somewhere else in the circuit.

I'll update this blog on the results. Feel free to leave a comment if you found this helpful.

Update 5 September 2015

I put a new vehicle speed sensor on it. Done. The speedometer works now.

Sunday, September 7, 2014

Rebuilding a W5M33 AWD Manual Transmission, Part 1

Finding a Project Car

When I was looking in the classifieds for a car with AWD and a manual transmission, I saw a 1997 Eagle Talon TSi AWD available for $2000. The reason for the $2000 price tag was the transmission. He was driving on the freeway and the car started vibrating. He took it to a shop and they said there were metal fragments in the gear oil.

My daughter was looking for a car at the same time, so I thought if she bought it, I could rebuild the transmission for her. I paid a mechanic friend to trailer it home. The transmission worked in every gear, so we drove it onto the trailer and parked it in my garage.

I looked for months for a transmission to come up on ebay, or the local classifieds. This was taking too long, so we just traded cars. Don't worry. She came out on the better end of the deal with a nice 2001 Outback with a manual transmission and AWD.

Rant about jack placement on unibody cars

Ok, can we please learn that we don't put the jack on what looks like a frame rail on unibody cars? That's not a frame rail, and if you try to lift up the car with the jack on that part, it will bend the body. Here are some pictures of the damage.




Initial Assessment

I found that when I jacked the car up, if I raised the passenger side, gear oil would drain out the driver's side. The driver's side seal was dislodged. The gear oil was black like it was burnt. There were fine aluminum fragments in the oil that made it look like metal flake paint. Removing the transmission was more difficult than average, but not as bad as a Taurus. Once I figured out how to get the transfer case off, the transmission could separate from the engine.



After I got the transmission out, I saw the reason the seal was dislodged. The output bearing on the front differential was destroyed. The front differential was also broken. The cost for the front differential case alone is about $290. The rebuild kit is $350.

I'm into this transmission about $950 in parts alone. A new transmission is about $2500 from the dealer.

As a core, this transmission would have very little value. If I was a professional rebuilder, I probably wouldn't re-use this case. But it looks like I'll be able to re-use this one. 

Disassembly

Disassembling the transmission

The gear oil was blackened and had fine aluminum flakes and steel fragments.

Transmission with cover removed
Transmission numbers. This one is a W5M33 late model AWD with a 57 tooth differential gear.


Is this a junk yard transmission? Nope. The VIN matches the car.

Removed viscous coupling
At this point, I was stuck. I couldn't figure out why the transmission wouldn't separate until I read about a bolt in the side holding a gear. Once I removed that bolt, I was able to separate the transmission and inspect for damage.




Here's what I found.

  • The front differential case was broken.
  • Both output bearings on the front differential were bad.
  • The speedometer gear was shredded. 
  • The case had damage because the differential wasn't being held in place by the bearings.
  • The magnet and magnet holder were missing. 
It appears that this was a hastily rebuilt transmission. 

Case damage and damaged speedometer gear. Bearing race still there. To remove the race, I got the center race from the other output bearing and put six ball bearings in, and spread them out. Then I was able to put it on my press and easily press out the race.
Showing more case damage under the 1-2 gear assembly. The magnet and magnet holder were missing.
First-Second Gear Assembly, Disassembled. Note the three-piece synchronizer for second gear.
To disassemble the gear assembly, I couldn't have done it without my press and large bearing separator from Harbor Freight.


To disassemble the differential, I had to drill out the back side of the pin holding the shaft because it wouldn't budge. Once I did that, I could press the pin out. Since the differential case wasn't going to be re-used, it didn't matter if I damaged the case further.
Front differential internal parts. Notice the pin next to the shaft has been drilled. I still need to smooth and polish the concave washers on the bottom left, or get new ones.

Rebuilding

I used a press and a bearing remover to disassemble the first-second gear assembly (after removing the lock ring). I cleaned off all the parts in solvent.

By the way, coconut oil is wonderful on solvent burns. I double-gloved, but got a hole through both and didn't realize it. My pinky finger got a little red. It was fine the next day. Hey, don't underestimate the power of prayer in gaining inspiration on what to do.

I polished the journal surface with 2000 grit sandpaper. I put new synchronizers in and new bearings on. The new races will be installed in the case.
When I tried to put the parts in the front differential, the gears wouldn't fit. At this point, I thought I had the wrong case. The part is non-returnable. Here I am, stuck with a $290 wrong part, I thought. At this point, I got frustrated and discouraged more than usual. But read on.

Here are some pictures comparing the two differential cases:
Broken front differential case, left. Right: MD770937


I used a die grinder to try to help get the pin out. I ended up drilling the pin from this side.




After looking around, I found that I got ripped off on the case (sorry Mitsubishi). It was $286.90 from the Mitsubishi dealer, but the exact same part number was $156.80 at tascaparts.com.
Or $161 at dodgepart
Or $168 from ordermopar.com.
Or $170.80 from moparpartsstore.
Or  from $229.52 from discountmitsubishiparts.

Without a VIN (I didn't want to go home to get it), the Mitsubishi dealer wasn't sure he was looking up the right part. So I checked the classifieds and found two 1997 Mitsubishi Eclipse GSX Turbo AWD cars, both with manual transmissions. Those cars have VINs like 4A3AL54F8VE011111, (where only the last five digits were different between the two cars.).

So I had the dealer look up the VIN from the Turbo AWD 5-speed Eclipse and indeed, the MD770937 was the correct part for the turbo AWD. My transmission was definitely an AWD model, so it looks like somebody put the wrong differential in.

For front wheel drive models, it looks like the differential case number is MD746800, and is available at discountmitsubishiparts for $247.58.

Could the wrong differential case have contributed to the failure? Maybe.
Could the lack of a magnet have caused the failure? Probably.
Could the failure have been caused by abuse like dumping the clutch at 5000 RPMS to try and spin the wheels? It has happened before on these cars.
Could the failure have been caused by lack of oil? Possibly, but only if the garage the previous owner took it to had added oil. It was full of nasty black oil with aluminum and other metal in it when I got it.

Now that I think I have all the correct parts, I'll continue this adventure in a future post.

Here's the parts list so far:

MR954819 spacer set for the differential - $37.28
MN178199 pin for diff  - $4.15
MD748538 differential gear kit - $170.42
MN178197 differential washers - $4.71
MD731948 nut - two @ $8.92 each
MD770937 differential case - $286.90
MD741725 oil guide - $6.60
MD955485 magnet - $2.60
MD727258 magnet holder - $2.55
MD749150 speedometer gear - $16.00
MR983368 transfer case output seal - $8.60
MD731793 transfer case input seal - $8.39
BK361BWS rebuild kit - about $350

Update Jan 10, 2022

I'm back on it. I took a long break on this project. I cleaned up the transmission in my parts washer and found that it was put it together with clear household silicon. 

Part 2

Check out my video for part 2 on YouTube: 

https://www.youtube.com/watch?v=Zj8xCEH4r6g