The saga continues. I've been running your test relay for a few days. A great testing tool. A/C seems to be operating a little more reliably. I'm starting to lean towards the Klima relay as part of the problem. Although, I still think my temp wheel and A/C button have problems.

The "Temp" wheel at full "MIN" splits cold air to the defrost and side vents. But slowly turning it to higher temps within the cool range, brings it out the center vent. It may be shot or out of calibration.

I opened the Klima and it looks o.k. no burnt components or broken solder joints. Along with a ton of resisters and stuff there are 2 mini relays. One normally "closed" and one normally "open".

The N/C relay is probably for the safty devices and trips open when something goes wrong. WOT, High temps, compressor freeze to save the belt.

The N/O relay is probably operated by the Buttons and closes upon activation and runs the A/C-compressor,etc. This is probably where the problem is as the test relay by-passes this and the A/C works now.

It would have been nice if MBZ would have separated these mini-relays externally where they could have been replaced as independant units. $5 instead of $125 for the whole unit.

I found a few more threads leading to Steve Brotherton's article and Odd-Jarie Krisoffersen tech write up at http://ow.no/mb/w124/tech/heating/ac-troubleshooting.html Tons of good info.

And of couse a lot of your threads with other folks on this board have helped tremendously.

Thomaspin, I contacted George Murphy to see about a new Button assembly. He's on vacation. His prices seem pretty good. Thanks for the info.

Can't seem to get the Aux Fan to function but, I think I'm doing something wrong. I'll give it another try.

__________________
1986 SDL
My Other Car is a Porsche. When it stops leaking, I'm out of oil.

Your Klima relay is very different from mine. I can't reduce the photo file size to attach it here but try this link:

http://photos.yahoo.com/bc/tabijan/lst?.dir=/87+300SDL+Klima+Relay&.src=ph&.order=&.view=t&.done=http%3a//photos.yahoo.com/

or go to photos.yahoo.com/tabijan and look in the 87 300SDL Klima Relay album. I have no reason to believe this is the original Klima relay.

Sixto
95 S420
91 300SE
87 300SDL
83 300SD

Hi, I am dealing with similar problem, inconsistent AC comp. activation!!!
I posted the test results on a thread (124 ACC) please read...
I have 1987 300DT, I have similar engine and similar klima relay as in "alan911"...
I first suspected the klima, I tested the car with new one, it does the same thing [starts and cut the compressor]...

I tested the inputs to the klima as in the CD, the compressor rpm sensor does not give the right voltage (should be > 0.3 v I have only < 0.1)...
I need advice on how to replace the rpm sensor, can I do that without discharging the system??

I am not sure about the problem of your system, but you need to run the test to pinpoint the problem..diagnosing these systems is VERY hard...do not just assume something is wrong..

Mainly you need to test the output of the Push Button unit, this is easy, you need to check the voltage on the pressure switch (red) close to the comp. I got positive voltage, means the the PBU sends signal to the klima to activate the compressor...(it is up to the klima to check and activate after that)...some cars have negative signal, either way if the signal is always there whenever you turn the ACC on (cold), this means the problem is outside the PBU..it is within the klima inputs. The klima recieve an input from several points, and if there is any signal that there is something wrong it will cut the compressor (for example: clutch slipping, comp. locked, high temp, defective rpm sensor on the clutch, kick down switch, full throttle etc...)


good luck

Sixto,

After searching for pricing on the Klima, I noticed there are a few manf. of the relay. The MB oem and one made by "Programa".

Mine has "Siemens" silkscreened on the solder side. Yours has the big box on the one side, and a lot fewer parts.

Perhaps the 2 little relays are within that box, protecting them from heat. These relays have extremely close tolerances and will be affected by heat. And could be burnt. I think the pin outs are the same and perform the same function as the rest of them.

toknow,

I read all the threads. We need someone to answer you on how to replace the rpm sensor or at least fool that one component to see if it's the culprit. ........I agree, testing all the components is a motha. But,

One interesting comment was, "if the Aux fan wasn't working, this could be the problem" but someone said, there is no connection between the Aux fan and the compressor.

Is this true ? I can't seem to get my aux fan running yet.

Along with this, as you suggested, I'm going to try to test the rpm sensor next, if I can find it. Red ?? Can you give me a description on where it is and how to test it. Or a pic. Which pin on the Klima ?

A/C is still running o.k. today with the test relay in. So its got to be one of the 3 or 4 sensor inputs.

__________________
1986 SDL
My Other Car is a Porsche. When it stops leaking, I'm out of oil.

The aux fan on a 300SDL has two modes; it comes on through a resistor when there is appropriate pressure in the receiver/dryer. Jump the pins with the smaller connectors to test the fan. I think it's the brown wires and slim white insulators, the kind you can't just peel back to make contact. It also comes on at 105C based on the temp switch hear the upper radiator hose.

The fan should come on whether or not the AC compressor is working if the engine gets to 105C, and it should come on when the AC compressor is engaged. So it's a semantic thing to say that it is or isn't associated with the compressor.

My apologies if I caused the confusion.

To toknow's point, it might make more sense to test for a controller signal at the KL port of the Klima relay socket, upstream of the relays and interrupts. That would be the port between the blank port and an end port; row furthest from the windshield, second from right. The SDL gets a ground signal from the controller to engage the compressor.

The compressor rpm sensor bolts in to the back side of the compressor. I'm 99% sure it's not connected to the refrigerant circuit.

FWIW, my relay housing has the KAE brand on it.

Sixto
95 S420
91 300SE
87 300SDL
83 300SD

Oh, there are two Klima pins for the compressor rpm sensor; row closest to the windshield, two ports closest to the master cylinder. From the right they're GK- and GK+. It should be a simple matter to monitor voltage with the simple relay in place.

I feel silly. For some reason I thought the rpm sensor emitted a pulse, not a voltage. Thanks to toknow, I can now proceed with my own diagnosis.

Sixto
95 S420
91 300SE
87 300SDL
83 300SD

Well I put a meter on the GK+ and GK- and got nothing. 0.00 volts with the enging running and compressor on. what did you get ??? What should it read ??

__________________
1986 SDL
My Other Car is a Porsche. When it stops leaking, I'm out of oil.

Hi,
The compressor rpm sensor is mounted on the back of the compressor as it is indicated before...according to the CD, this sensor should generate > 0.3 v (AC), and it will increase with the increase of the engine rpm...you can test ports # 9 & 11 at klima [after jumping # 5 & 7 to have the compressor running]...mine was giving almost 0 v there...
I tested the voltage out of the compressor sensor at the sensor wires there (I had to cut off insulation off the wires)..it was giving barely 0.1v there...
If this is your readings, and according to the CD, this is the problem??? although the sensor could be ok, but the compressor shaft (inside) is moved away from it little... or the magnitic field of the sensor is weakened through use...

because I do not always trust electronics, here is another theory that I need to test: It could be that the clutch or other things wrong with the compressor and the klima cut it all the time...I need to have the voltage output from the kilma to the compressor being monitored...if the volatge stops before seeing any problem at the compressor, this is an indication of rpm sensor or other electronics...if the clutch stops first and then the voltage... the compressor is bad or clutch...you can try this also...

one more question, what is the voltage reading of the engine rpm...between terminals 1 & 2 on klima ports....the CD says > 4v (DC)...I got 10...which might not be OK...

good luck

Have you replaced the drive belt? A slipping belt will kick the Klima out to prevent a locked up compressor from stalling the engine or smoking the belt. If the belt is old and glazed, or missing a lot of ribbing, it will slip too much and the Klima will cut out. Normal operation.

My brother had trouble with the AC cutting out on him -- runs for a while, then quits until you restart. New belt cured it.

Speed sensor is a pulse, probably reads as AC voltage, but I'm not sure.

Peter

__________________
1972 220D ?? miles
1988 300E 200,012
1987 300D Turbo killed 9/25/07, 275,000 miles
1985 Volvo 740 GLE Turobodiesel 218,000
1972 280 SE 4.5 165, 000 - It runs!

Making progress.

Compressor RPM sensor @ ports #9 & 11. I get .2v AC and increases steadily with engine RPM. Seems normal. Tells me sensor may be o.k.

I was testing for DC before.... Doh ! Thanks ! I must have missed this. It is AC !

Voltage reading between 1 & 2 on ports is 6.3v DC. A little higher than what "toknow" says is on the CD. What do 1 & 2 actually do ??

I changed my belt last week. It's fine. Did this first.

So.....running out of thigs to test but making progress.

__________________
1986 SDL
My Other Car is a Porsche. When it stops leaking, I'm out of oil.

I didn't read alan911's last post and tested it myself. 0VDC at pins 9 and 11 (compressor rpm) and 1 and 2 (engine rpm). I switched to duty cycle and got a varying 3-25Hz at the compressor at idle and 1500Hz at the engine at idle. It was fun testing the engine rpm signal, which seems to be twice crank rpm, while changing the setting on the ELR resistor. At least I know the ELR system works

I'll check for AC voltage and see what I get. The compressor frequency fluctuates at idle but is steady at revs. I expect the same will be the case with AC voltages, whatever they are.

Sixto
95 S420
91 300SE
87 300SDL
83 300SD

I checked again. I get 6.something volts from the engine rpm input (pins 1 and 2) whether I test AC or DC at idle. As I rev it goes down slightly to something like 5.8VAC (or DC) at ~3000rpm. Seems like VDC or VAC is the wrong thing to measure for the engine rpm input since there's a nice linear frequency regardless of voltage. BTW, pin 1 is just a ground, pin 2 is the engine rpm input but I don't know from where. The flywheel sensor, I suppose, but I don't know how it gets from the flywheel sensor to the Klima relay, though it's about a 15" spread.

I get 0.15VAC from the compressor rpm input (pins 9 and 11) at idle rising to ~.68 at ~3000rpm (counting revs by ear). Seems pretty linear. Again, more fluctuation at idle than at revs.

Lots of results, no answers

Sixto
95 S420
91 300SE
87 300SDL
83 300SD

You get the same as I do on 1 & 2 +/- 6v dc

It seems that your test on ports 9 & 11 may be o.k. at idle. Peak to peak should be less than 2v. Again, about the same as mine.

What's the ELR ??

From Brotherton's article: Note reference to miniscule signal 2nd & 3rd para.

The compressor control relay monitors the engine speed vs. compressor speed to protect the belt system from a locked compressor. It disengages to protect an overheating engine and to provide extra power in diesel models, and may possibly disengage to protect a low or erratic idling motor or an under-voltage system.

To test, we remove the relay and bridge the power to the output to engage the compressor. With it turning, the compressor speed signal should be an AC voltage with a peak-to-peak voltage difference of less than two volts (see waveform at right — less than 1v was coming from this functioning system). The engine speed signal is converted from an inductive crank sensor AC voltage to a pulsed DC signal by the ignition controller (gas models).

And....

The diesels use either the EGR controller or the idle controller ISC to do the conversion. On diesels, the controller disengages when the full throttle micro-switch is closed (pin #4 Klima). This signal should be battery voltage until full throttle. I have had a bad controller that was pulled internally a little less than 5v and would disengage.

His caveat:
Many unnecessarily condemn the compressor speed sensor, as a very small signal is monitored.

O.k. now how about a test on port #4. Which ports to connect to test the WOT signal ??. Appears it should have +13V or battery voltage until full kick-down. I know disconnecting it is kind of a test but I'd like to see if the port is actually getting the consistant 12-13 v and not an erratic voltage stream.

I've got to get a life.

__________________
1986 SDL
My Other Car is a Porsche. When it stops leaking, I'm out of oil.

Hi,
your test results regarding the engine rpm is Ok, mine is much higher than required (I replaced the engine rpm sensor recently, this might be the reason to be higher)...
Also, your compressor rpm output seems to be Ok, > 0.3 v at 7500 rpm engine speed and increasing with engine higher rpm...if it is lower at idle, sometimes the klima will cut the compressor...

Also, it seems to me that the klima compares the changing voltage of the comp. rpm to the steady voltage (@ 6v) of the engine rpm...if the comp. voltage within 5 -30% of engine voltage, it is OK, if not [lower means compressor slipping, or belt etc, if higher means higher rpm for the compressor to handle (> @3500 rpm or so)]....So the wider the gap between the two to statrt with, the more it will cut...THIS failed comparison has to happen through certain time, it could be few seconds, in order for the klima to decide to cut......(I am using voltage because this is what the CD asks to test, otherwise they are hiding other factors that the klima uses for comparison]

I am just wondering what is the problem with your system(s)...if not electronics and they work fine, except slipping or so..this means the clutch is not OK, either gap or oil inside, ot he belt mechanism...there is nothing else there

to alan, it is extremely harmful for the compressor to run on higher pressures, that is what will happen if you run your AC without the aux. fan...of course at certain point the swtich for higher pressure will cut the compressor...BUT I know that at 1986 cars the swith was not having this function, MRBZ updated it later...so be careful!!!
goodluck

toknow,

You lost me on the pressure switch and aux fan.

Here's a pic of my set-up. Don't know what is what. Connecting the leads on either one of these devices does not get the Aux fan working.

__________________
1986 SDL
My Other Car is a Porsche. When it stops leaking, I'm out of oil.