A few small updates
1. AC problem
I think the speed sensor on the compressor is the root cause for AC having problems. I ruled out the KLIMA relay after testing.
Basically, the compressor clutch will disengage after some time and then compressor stops and car gets warm. Turning climate control on and off has no effect - but turning car off and turning back on will restart the process and re-engage the compressor clutch.
I talked to my friend Bud and have an appointment at his shop next week to have the compressor replaced. We'll do the other items too (pressure switch, receiver drier if needed, etc) while in there.
2. Door latch adjustment
This is one that had been driving me nuts. Basically, the drivers door only won't always fully close when you slam the door. It'll catch on the latch and be open about a centimeter. I replaced the door check strap a while back and I finally got a chance to replace the latch assembly (both new in the box genuine Mercedes-Benz parts).
Replacing the latch assembly seemed to help - but not a lot. I think the root cause is the striker position. I spent a good hour with a ratchet loaded with the Torx T40 bit trying to adjust the striker and it's incredibly hard to get right. You can do micro adjustments by loosening the bolts - but sometimes turning the bolt to tighten (after the adjustment) will move the striker itself. Just a frustrating process. I got the door into a much better place - it doesn't require a lot of force to close. It's still not right... but it's "good enough" (for now).
3. Speedometer problem
I've noticed my speedometer is off. This is one I could use some help on!
I've documented some of the work done on my cluster in previous posts. The plastic gears for the odometer and trip counter are in great shape - but I did have to pop the needle off to replace that. I had a local shop calibrate it after I put the needle back on and at low speeds it seems just fine. But if you get up to 80 MPH and check Google Maps, it'll show the actual speed via GPS at 70 MPH. 95 on speedometer seems to be roughly 80MPH via GPS. The car doesn't go much fast than that
I have a bunch of spare clusters and took everything apart tonight. I got the Mercedes-Benz reference manual out too. This
124.133
chassis should be having a 2.65 ratio for the rear differential and that should be what the speedometer is matched to. The 400E/E420 and E500E both have electronic speedometers but this still has the spinning cable that gets threaded on the back. I think the cable is directly connected to the transmission but I'm not sure. Here's a view from the back.
My understanding is that there is a "K" constant and that is supposed to represent how many revolutions of that cable are needed to reach one kilometer. Since my cluster is in miles, I'm guessing the number is larger to account for miles.
Cluster in my car
The cluster in the car doesn't have any K value noted. The metal "counter" piece itself (behind the faceplate) is marked
67127 085 6
. I did notice someone wrote in sharpie "W. 980" on the plastic. I'm not sure if that's what the shop thought the K constant was when they calibrated?
Spare cluster 1
Looking at the other clusters in the shed, there's one I pulled from a 1987 300D in Tucson last year.
This one actually has the stickers with the K values.
12V K= 1015
109/17/19
When you look at the metal backing, it has the number
67127 085 8
.
I put this in the cluster and the speedometer was pretty accurate. I'd say always matching within 3 MPH of GPS via Google Maps.
However, the trip plastic numbering gears are broken in this one. When the spinning cable turns the unit, it turns a worm gear on the side. As the worm gear (white) spins, this will spin the gear attached to the odometer (green).
On the other side of the unit, there is a gear connecting the roll of numbers on the odometer and the roll of numbers on the trip. I know it's been debated, but I believe teeth on this gear (or the ones going between each tens position) can break if you push the mileage reset when the worm gear is spinning. I was thinking maybe this was broken.
I dug in more and I was wrong with that theory (at least for this case). I pulled the green gear on the side out and wound it by hand.
There is a metal rod (connected to the horseshoe reset button) which is putting tension on lever connected to a roll of gears for the trip (one gear for each 10's place). When you push the reset button, this pushes the rod down, which then pushes the lever down which lifts the gears up. That allows them all to freely spin back to 0.
This is super hard to describe - here's the best picture I can share. Basically, the red arrow is pointing to the roll of gears which are on each 10's place. The lever that lifts this roll (disengaging the teeth) is the yellow line. You can also see the spring in the middle which provides tension to the speedometer needle.
When the reset button is not being pushed, there is still some amount of tension being placed on this lever. It's just enough that the teeth are making contact with the roll of numbers.
Here's a picture where you can see the metal rod that is attached to the horseshoe piece. You can see the spring it's attached to and you'll also notice the rod broke out of the plastic. The spring attached is no longer providing tension to pull the lever down (which engages the teeth) and the gears no longer turn.
The plastic part needs to be replaced in this case. That whole assembly is delicate.
Spare cluster 2
Just for fun, I took a look at a few other clusters. Another 1987 300D cluster I got on eBay for parts was marked:
12V K=1015
109/13/19
On the metal backing, it has the numbers
67127 085 6
which actually match mine.
European cluster
This is one I got from Lithuania earlier this year and I verified it's accurate - it's in kilometers of course. This one is marked:
109/16/67
K=630 12V
The metal backing for this one is numbered
67127 058 9
. If you do the km / miles conversion, 630 * 1.609344 is close to the 1015. K=1015 is definitely the correct value for miles.
Other notes
All of the US clusters have the same hall effect sensor - part number
007-542-29-17
. The euro cluster didn't have the sensor at all- just a black cover plate. I believe this sensor is getting resistance readings from the wheel speed sensors used for ABS. For sure, I know the hall effect sensor is also used for cruise control. I remember seeing some interesting cases in the tech manuals where the sensor value can be taken into account - one I think was with the radio (raise volume slightly at higher speed).
I don't know much about this part or how to test it. I could try to buy a replacement or swap the sensor (it pulls out from the metal backing).
Next steps
I'm curious if folks here have any resources they can point me at? Sometimes the more you learn, the more questions present themself.
Is the K constant configurable? What does the shop do when a calibration takes place? For sure, the correct K value for my car is 1015.
The speedometer needle itself is attached to a shaft which is spring loaded. Before giving up and bringing it to the shop, I did my best to try and tension this to match - but it's very tricky. The cluster itself has no markings so I'm curious if the shop had to do their own research (I gave them the year/make/model) and they thought it was supposed to be K=980 (what somebody wrote on the plastic) and they calibrated it wrong. If that is the case, I should be able to take it back in and let them know it's supposed to be K=1015.
Another option - I could take the labeled unit (K=1015) which has accurate speedometer, pop the needle off, put my good plastic gears, put the needle back on, and bring to the shop for calibration.
There's always a mystery waiting to be solved
