I always learn so much in doing these troubleshooting and problem solving processes. My experiment in putting the thermostat module in the freezer and comparing that to the plunger length at ambient was based on a completely erroneous understanding of the mechanics of the design. I was assuming a linear expansion process as might be used in a thermal expansion bimetal thermostat or a gas bellows thermostat, but alas this uses a phase-change process. Wax that is formulated to melt at a particular temperature is placed in a contained space with a plunger. When the wax melts it expands greatly forcing the plunger out and when it solidifies it contracts withdrawing the plunger. No wonder I got no movement in my quick test.

Here's a drawing of the essential design.



Of course that explains why the manual states not to disassemble the thermostat in the oil filter housing until it's cooled to 60C.

Attached Thumbnails

 

__________________
Current Stable

• 380SL (diesel)
• Corvette C5
• Manx
• Baja Bug
• F350 Powerstroke
• Auburn Boattail Speedster replica

Awesome writeup, thanks Mach4.

I would think testing for a failed oil thermostat could be as simple as touching the oil cooler. If it starts heating gradually, immediately after your engine is started, vs. staying cool for some time and then going up to temp all at once, you likely have a failed thermostat. An infrared non contact thermometer would probably work well enough to diagnose a problem as well.

__________________
617 swapped Toyota Pickup, 22-24 MPG, 50k miles on swap

Who checks their oil temperature? Besides the risk of condensate build up I suspect the effective viscosity of the multi grade oils may remain too low.

It apparently requires oil temperatures in specific ranges to give the higher viscosity ratings claimed on the can. I have often wondered why too many of these engines seem to fail at lower than expected miles.

Because of their odometers I kind off attributed failures to odometer turn backs or periods where the odometers where left inoperatable for long periods of time.

Now I wonder with the large sump capacity on these cars and the generally higher operational temperatures of more modern engines. If we may not be getting in general the higher claimed viscosities of the modern oil?

If so it may be important to have that oil thermostat functioning properly. There have been failures on these engines that I suspected may have been long term low oil viscosity issues.

It would be interesting to get some temperature data from the cars oil coolers as well with a reader. With proper viscosity of the multigrades being reached the oil drag factor may drag down fuel milage a little though as well. One last thought was if the car is used for basically only short trips what viscosity is occurring with the oil when it is colder. For example a 10w 30 may not see a higher than 20 weight viscosity when the oil is cooler.

Now on the other hand since there is so much obvious oil drag on these engines when cold. Perhaps viscosity starts high and drops to no lower than say 30 at the specified temperatures. Based on hot oil always seems thinner than cold.

These engines at least I always thought where designed for a constant 40 weight single grade oil for average spring summer and fall use. Some examples I suspect At least judging by their general appearance have probably reached high total milage as well. Maybe 700k but of course there is no way to really verify if it was done on the original non rebuilt engine or not.

mach4, have you done anymore work on this and what's your plan going forward?

I am looking into the oil temps of my 85. I made a type K thermocouple dipstick probe, drove 45 min on highway, pulled over and took a measurement. It was 181F at the top of the oil in the sump, as I pushed the probe deeper, temp slowly decreased till it touched the oil pan, where it dropped to 167F. I did this multiple times so there is no mistake about it. I know you said you are convinced your oil temp measurement is actual but my measurements indicate otherwise. If I had a sensor at the drain plug where you have yours, I'd bet it will be lower than 167F since it is cooler there due to the greater mass of the drain plug compared to the pan, dissipating more heat to ambient.

I have a complete spare oil filter housing that I plan on doing some testing on. I do not want to disturb the T stat (before doing the test) and will be testing it in the housing. In figuring how to test it, I poured water and filled the cavity where the T stat is and it did not leak out, indicating the T stat is closed. The plan is to fill the cavity with fresh motor oil, heat the oil up with a PID controlled 200W heater element and observe the temp at which the oil starts to leak past the T stat, which would be the T stat opening temp. What do you think?

__________________
85 300D turbo pristine w 157k when purchased 161K now
83 300 D turbo 297K runs great. SOLD!
83 240D 4 spd manual- parted out then junked

Is this a diagram of the Mercedes oil T stat? Do you know what type of wax is in it? I recall reading it is bees wax. According to wikipedia:

Beeswax has a relatively low melting point range of 62 to 64 °C (144 to 147 °F). If beeswax is heated above 85 °C (185 °F) discoloration occurs. The flash point of beeswax is 204.4 °C (400 °F)

So, it can't be bees wax since the phase change from a liquid to a gas will be at or above than the flash point.

__________________
85 300D turbo pristine w 157k when purchased 161K now
83 300 D turbo 297K runs great. SOLD!
83 240D 4 spd manual- parted out then junked

No, that's a generic drawing of a thermostat module as might be used in any variety of temperaure applications. Wax formulations are available to melt at a wide range of temperatures depending on the application. One vendor is Astorstat who summarize their products thusly



My plans moving forward are to replace the ostensibly failed thermostat module in my oil filter housing with the presumed good module from the spare oil filter housing (I've got to swap it rather than replace it because the one on my engine is the SD style and the spare is the D style) That being said, I think I'm going to get one of the cheap chinese temp probes and drop it down the dipstick tube or possibly make a protective sheath for the intake temp probe and drop that down the dipstick tube.

I find it very interesting that you're seeing essentially the same temps I'm seeing implying potentially that that the thermostat is maybe OK??? I don't think so however because my oil cooler line showed the same essential temperature implying that the thermostat module is stuck open (as the second filter cannister I opened up had) and not a low ambient engine bay temp and then a sudden rise as the thermostat opened.

I'd hope to not have to pull the filter housing as that's kind of a pain, but we'll see.

Your testing plan seems reasonable. Please post pictures of your test setup as you do it as I'm having a little trouble visualizing exactly what you're going to do.

Attached Thumbnails

 

__________________
Current Stable

• 380SL (diesel)
• Corvette C5
• Manx
• Baja Bug
• F350 Powerstroke
• Auburn Boattail Speedster replica

Based on your oil thermostat threads, I fired up my truck and let it run while monitoring oil filter housing and oil cooler temperatures with a non-contact IR thermometer. Both were within a degree of each other from when first started, and topped out at 165 degrees or so. Granted this is measuring the housing and not the oil itself and also only at idle(albeit at 104F ambient temp).

Clearly my oil thermostat has failed as well.

__________________
617 swapped Toyota Pickup, 22-24 MPG, 50k miles on swap

If that's true, it would appear that 4 out of 5 oil thermostats have failed - yours, Funola's, mine and one of my spare's and only the second spare appears to be OK, though that's untested, just based on appearance.

__________________
Current Stable

• 380SL (diesel)
• Corvette C5
• Manx
• Baja Bug
• F350 Powerstroke
• Auburn Boattail Speedster replica

As a quick test that the T stat is not failed in the open position, I layed it on it's side as in pic, filled it with water and it did not drain down. Is that a valid test?


I was originally going to replace the water with motor oil and heat it till the T stat opens and measure the temp (at the point it starts to drain). I'm not sure if that small amount of oil will be sufficient to heat up and open the T stat. I will change the test to blocking all the ports (b,c, B,A, turbo oil line, oil pressure gauge port) in drawing below except T stat port a, fill the housing with motor oil and heat it till oil starts to drain out out T stat port a. Will that work?

__________________
85 300D turbo pristine w 157k when purchased 161K now
83 300 D turbo 297K runs great. SOLD!
83 240D 4 spd manual- parted out then junked

No, flow out of a is prevented by a one way check valve 16. You want to leave A open - when the thermostat plunger opens it pushes a piston to the left it allows oil to flow out A to the oil cooler.

The easier solution is to open up the valve by removing the retainer nut and pulling the parts out. Place the thermostat assembly and spring in a 'C' shaped test fixture that holds the thermostat under spring pressure. Immerse in your oil bath and observe the temperature at which the plunger begins to move. A lot less hassle and a lot more accurate. That's what I plan on doing here sometime soon.

__________________
Current Stable

• 380SL (diesel)
• Corvette C5
• Manx
• Baja Bug
• F350 Powerstroke
• Auburn Boattail Speedster replica

OK thanks. It's hard to visualize how it works from photos and drawings. I will have to take the T stat out and the unit further apart and study it more. It sounds like the check valve at a requires pressurized oil (from oil pump) to open. I will test the T stat by itself.

__________________
85 300D turbo pristine w 157k when purchased 161K now
83 300 D turbo 297K runs great. SOLD!
83 240D 4 spd manual- parted out then junked

Yes, oil pressure opens the one-way valve.

Once you get it open it should be clear how it works.

__________________
Current Stable

• 380SL (diesel)
• Corvette C5
• Manx
• Baja Bug
• F350 Powerstroke
• Auburn Boattail Speedster replica

I decided to set up a thermostat test bed. Here is the basic configuration. I bent up a fixture into which I placed the thermostat spring and thermostat module/piston. I then shimmed the spring to match the position (spring tension) in the thermostat housing (measure twice, bend once...I measured once) The test fixture was placed in a small sauce pan and covered with fresh engine oil (don't tell my wife)



I then added heat and measured when the thermostat module plunger moved outward. There was a lot of lag. This is an early 85C degree module, not the later 110C degree module. It opened well after it's 185F point. (Note: arrows indicate the direction of temperature change)



In the other direction, it closed very slowly and finally was closed all the way at 60C, which happens to be the temperature the manual says to not open the thermostat housing above.



I found these results to be rather interesting.

Attached Thumbnails

     

__________________
Current Stable

• 380SL (diesel)
• Corvette C5
• Manx
• Baja Bug
• F350 Powerstroke
• Auburn Boattail Speedster replica

Just did the FSM thermostat test on my 85. Started the engine from cold, ac on with raised idle, monitoring gradual rise of oil temp of dipstick probe, hand on outer oil pipe at motor mount. Felt a gradual rise in temp of the pipe as the oil warmed to around 50C (122F) when I terminated the test. I will be replacing the T stat. Maybe this will be the final piece that will help with poor mpg.

__________________
85 300D turbo pristine w 157k when purchased 161K now
83 300 D turbo 297K runs great. SOLD!
83 240D 4 spd manual- parted out then junked

I depressed the black tri star plastic piece in the center and filled it with water. It drained immediately. Does that indicate the stat failed in the open position?

I have not taken mine apart yet. Do you know if there is an o-ring to seal the wax in?

__________________
85 300D turbo pristine w 157k when purchased 161K now
83 300 D turbo 297K runs great. SOLD!
83 240D 4 spd manual- parted out then junked