jspbtown

2007 CLK550 cabriolet with 60k miles.

Its hot up here in the northeast. engine temps have been totally normal. First time I got a no start "would crank but no start) it was after driving the car for about an hour, then letting it sit for about 3 hours and went out to a no start (would crank) condition. Disconnected the battery, then went online to do some research. reconnected the battery and still no start. Decided to run cold water down onto the CPS area. After about 2-3 minutes of cool water it started up and ran fine for the hour drive home.

Went to MB dealer in the wife's car the next day and got a genuine MB CPS (made in Romania). Went home, the CLK started right up and I pulled it into my shop and swapped out the CPS with no problem.

Next day head out. Drove about an hour. Parked it. Then about 2 hours later went to leave and it started right up. Drove home and parked it again for like 3 hours. Went to go out for dinner and no start again.

Took the wife's car for dinner and was thinking what else it might be. Thought maybe a fuel pump relay? After dinner plan was to come home and see if it would start and if not swap relays. Of course it started right up....:{

Put the scanner on it. No codes. Running fine.

Suggestions??????

Rudeney

There is an adaptation process in SDS for the CPS. I might try that.

jspbtown

"adaptation process in SDS for the CPS"

Sorry....no idea what "in SDS" is? Something I can do?

Rudeney

SDS = Star Diagnosis System. It's MBZ's diagnostic computer system for dealers and high-end indy shops (i.e ones that can afford it). Clone versions can also be bought from China $250-$700.

jspbtown

Well...that might be an issue. I found a process online where you increase the rpm's by about 150 every 10 seconds and then hold at 2500 for 30 seconds. Then you drive it.

Did that and took it out for a drive. Pulled it back in my shop and it sat for about 10 minutes and I tried a re-start and it sputtered and died and then would not restart. Popped open the hood...looked around and then tried again and it started.

Also tried testing the old sensor. Ran my meter across all 3 prongs in different combinations and there was no continuity. Then I took a propane torch and slowly heated the sensor up. My infrared temp sensor's batteries were dead so I couldn't get a temp reading but as soon as it started to get warm I started to get ever increasing continuity. Took some shop air to it and it slowly went back to no continuity. This was repeatable.

Significance? Heat causing this issue?

jspbtown

jspbtown

Just took out the replacement Mercedes CPS sensor that I bought. Inspected the connection and the wiring. All "looks" fine. Pliable. no cracks, no corrosion.

Took me heat gun to the new sensor and after some heat it also gained continuity just like the original sensor. The heat was not much (maybe 2 or 3 minutes of heat?) and continuity went from 20000k to about 98 (meaning almost full continuity in my understanding).

While I am surprised the second sensor is no good as well I don't think I have any other option but to try another one.

Rudeney

It's a Hall-effect sensor, so temperature doesn't make it work or change properties (at least it's not supposed to change things). It is triggered by a magnetic field. Try your test again, but instead of a heat source, pass a steel object below the sensor. I suspect both sensors will give you similar results with a change in impedance. The problem is, though, that heat makes it fail. I've never heard a definitive reason why, but I suspect it's one of two things: either the distance between the sensor and the flywheel changes things warm up and it loses its signal, or the tiny electronic circuit inside the sensor that amplifies the Hall effect signal fails with heat. If it's a distance issue, you could try to shave a paper-thin amount of plastic off the base of the sensor so it would protrude slightly more into the bell housing. if it's the electronic amplifier circuit, then there's not much that can be done except for manufacturers to make parts that are less susceptible to heat failure.

The interesting thing about this is that I have logged somewhere around a million miles on various MBZ cars and never once have I experienced a CPS problem. Yet others have them frequently on the same car, even after replacing the sensor with an good OEM unit. This makes me thing that it could be a distance issue. Maybe there are some cars with slightly warped bell housing that are more susceptible.

jspbtown

Well I went and picked up #2 OEM Mercedes sensor. I also took the time to change the coolant.

So far so good. Drove the car for about 1/2 hour (after letting it idle for about 15 minutes while I filled the coolant) and then let it sit for 1/2 hour...it started. Let it sit for another hour....it started.

Not 100% sure if its fixed. I plan on doing more driving and sitting over the next several days.

This is the video that discusses heat failure on the Mercedes CPS.

jspbtown

Well let it sit for about 2 hours and guess what....a no start again.

This time I had my OBII unit plugged in. I cranked and it read rpms so I think that means the CPS is working (hell....clearly I don't know right?)

I popped the hood because I wanted to see if there was any fuel pressure. Walked to my shop (2 minutes) and realized I don't have a pressure guage. So I was thinking about the fuel pump relay. I removed it and swapped it with one under the hood (the one for the starter relay. It fired right up.

Was it me opening the hood to let out heat? Was it the relay? In theory if the relay was no good or weak I should start having issues with the starter right?

What are the changes the ecm is getting too hot?

jspbtown

Took the car to the grocery. About a 15 minute drive. Went in, shopped for about 15 minutes and came out to a no start. I opened the hood for a moment, then closed it, went inside, bought a bottle of cold water, came out and poured it on the ecm. Got in car and it started right up.

I am going to go out to my local Home store and see if I can get a piece of heat reflective material to put under the ecm as an experiment. I am bringing a cold bottle of water as well.

Update: Well I drove to the home improvement store (20 minutes), parked it for about 20 minutes and it started right up. Drove home, parked it, let it sit for an hour and it started right up.

Argh.

Rudeney

It does sound like the heat is affecting the ECU and not the CPS. One thing you might want to try is unplugging its wiring harnesses and cleaning them with CRC contact cleaner.

Theoretically, the ECU should be getting some cooling by sitting on top of the intake manifold and between the cold-air intake tubes. But of course on a hot day on a hot engine, that's not much. Make sure the insulation on the underside of your hood is not loose and sitting on top of the ECU when the hood is closed.

jspbtown

When I am getting the no start the ecu unit is VERY warm. Like I can't touch it for more than a second or two. I drove around with the OBD2 scanner hooked up and the temps ranged from 199-212 or so. I don't think that's hot. It has been very warm here with high temperature indexes (105+)

I will try to clean the contact tomorrow.

I had hoped I would get a no start again after the time I poured cold water on the ecu but no luck. I will have the water with me for the next time it happens though.

Fatz

With hind sight now, I had similar heat related issues with my ecu. Intermittent P0300, P0305,P0351 and P0355. I drove it on a 400 mile trip w/o any issue. Checked into the hotel, took my bags in, came out to move the car away from the crowd and had misfires on start. Started fine the next day. On the trip home I got caught in a traffic delay while guys were working on a bridge; was creeping (5-10MPH) for an hour. Misfires started. Pulled off after a couple miles, raised the hood, put a tie wrap on cylinder 5 spark plug cable and started it - no misfire - drove home from there (300 miles) w/o any problems.

In hind sight, the ecu probably got heat soaked after my arrival at the hotel, without any airflow; when it started misfiring while creeping on the trip back home, it was again probably heat soaked and me raising the hood to tie wrap the spark plug cable probably just allowed the ecu to cool down.

I eventually had the ecu repaired ($600). Had to send it to them twice. The first time they just changed one of the two current drivers. The last time they replaced 1 or 2 logic chips.

If it is the ecu exhibiting problems triggered by heat, it will get worst. You can probably just use a spray bottle of water during your testing and to keep in the care. The water doesn't have to be cold; ambient temp is good enough. Besides, just opening the hood will probably work for now.

Rudeney

You know, my CLK550 would act a bit "unhappy" with high temps. It would chug a bit at idle. I remember a coming back from a long summer trip and getting stuck in traffic on the freeway and it was really idling roughly, almost to the point of dying, so I just kept my foot resting on the accelerator to keep RPMs up. The SL550 has been idling a bit rough lately, too, in the last few week (of oppressive Alabama heat!) but I figured it's just because it doesn't get driven enough. Neither engine threw any codes. I am beginning to wonder if a hot ECU is the cause. I may try to cool it and see if that makes it idle more smoothly.

Glyn M Ruck

Next time it cranks but won't start check the fuel pressure at the Schrader valve on the front fuel rail that links the injectors. s/b > 55psi while cranking. Sounds like intermittent fuel pump. Happens as rust particles stick to the magnets & ultimately seize the pump.

jspbtown

You know I was wondering about the fuel pump. It seems awfully coincidental that when I cooled the ecu it went from no start to start right away. Coincidental? Who knows. I can't seem to guess right these days.

I don't have a fuel pressure tester so I will have to see if I can locate one.

Glyn M Ruck

Places like Autozone usually loan them out!

Crank, no starts are 99% CPS &/or fuel pump.
Good luck!

Glyn M Ruck

BTW ~ If the pump is not starting at all due to temporary seizure (it runs flooded with gas). You will have no pressure at all at the Schrader valve. Maybe just a dribble. But the engine requires >55psi to start reliably.

jspbtown

Some questions about the possibility of it being the fuel pump:

1. I replaced the left sending unit (and fuel pump?) last year due to a crack in the upper housing which was causing a fuel leak. I am confused about there being a right fuel pump as well? Are there 2?
2. Would it be unusual for the fuel pump to fail in the way mine has? I can drive the car all day long and it has never stalled. It runs smooth, accelerates fine, idles at stop lights...everything. It just wonp;t re-start after a period where it just sits after driving. And then...miraculously when cooled, it fires right up. Would the fuel pump show no other sign?

Rudeney

The device in the left side of the tank (that your replaced) is called the "sender unit". It contains a float to read the left side tank level and also has all the hes and wiring connections. On some cars (I think including yours) it also contains the fuel filter. The device in the right side of the tank is the actual pump motor.

This really does not sound like a fuel pump issue. When it goes out, you generally have other symptoms, like poor acceleration, stalling at high speeds, etc. Still, it is worth a try to be thorough and check the pressure.

jspbtown

That's what I thought. I remember that, what I thought was the pump, was on the right side. That was a fun job replacing the sending unit.

I am going to go clean the connections today. I am also going to see if I can fabricate a little insulator plate for the ecm. Yes...kinda backwoods engineering but I am thinking more as a diagnostic before I commit to an ecm repair.

I was also thinking about your comment regarding the hood insulator. There appears to be some almost "scuffing" on the top engine cover which caught my attention. Its toward the front of the car. The hood insulator looks intact but I might give that a better look. Weather has also broken here today so it will be interesting if that has any impact on the issue

megacrazy

Another thing I would consider is gas. I'm also in the northeast and have noticed very similar issues with my Audi. I was getting a hesitation to start but it always started on the second try. There is nothing technically wrong with the car as in no codes, no ongoing issues etc. There are different types of gas that are being sold in winter and summer so if you are running older gas that may be your problem right there. My problems all literally went away after I got another tank of gas.

As far as the ECU getting heat soaked, a lot of these chips work at +100C temperatures and get soldered at over 230C. Keep in mind that water boils at 100C so that ECU would have to be pretty damn hot. Of course there could be chip/solder defects which would definitely explain the symptoms you are seeing. Many times it's a bad solder point that allows part of the chip to lift up after it heats up enough and make contact again as it cools down. I just think you should be seeing many other issues if part of the ECU was damaged, not just a no start condition when warm.

jspbtown

I appreciate the response. I don't think I have filled the tank since this has happened and it is about 1/4 tank now (was keeping it low in case it was the pump). I will fill it up since I really don't think its the pump.

I did go out today with a little harmless redneck (no offense meant.. ) engineering. I took a thin sheet of aluminum and bent it to make a heat shield. It basically splits the space (about 3/4" between the computer and the intake manifold) so there is a 3/8" air gap between the ecu and the intake manifold. . It mounts to the stock ECU mounting bracket screws so nothing was touched on the ecu itself. It is easily removed.

I also cleaned the ecu contacts. The weather has broken here today and it is about 20 degrees cooler than it has been.

I also checked the felt-like covering on the underside of the hood and it sags a little but not a lot. I put a bunched up roll of tape on the ecu and closed the hood and the tape did not flatten.

I am keeping a bottle of water in the car. If it happens again I will check the shrader valve for fuel and try the water on the ecm again.

Keep the thoughts coming!

megacrazy

Haha if the water doesn't work you can try gatorade...might work better.

If you wanna step it up a little you can get something like this too: https://www.heatshieldproducts.com/hp-sticky-shield. It's more serious than the regular heat tape.