I am struggling since some time with my compressor being cut- out.
Compressor is typically starting then after some minutes of driving it is cut out without an obvious reason. The compressor would not turn on unless the car will be restarted.

If I replace the Klima Relay with the ice cube relais- there is no problem- and the AC works well.
I had the relay replaced but without any difference. So I am sure it was not the relay.

Well if the ice cube solves the issue the compressor must be cut out by the functions the relay manages (which the ice cube does not address)
This is
A)Compressor speed sensor
B)Engine RPM input

Kick down should cut the compressor out temporarily and compressor will come back when kick down is finished. Same with the full throttle cut- out.

I have attached the electrical diagram.
Do you have a cure for my problem?
Any suggestions?

Attached Files

Klima Relay 86.pdf (468.9 KB, 529 views)

I have similar problem with one of my SDL and I suspect it is the compressor speed sensor. I reworked the Kilma and bypassed everything except the signal from the CCU, Pin 10 on the relay. I do not have any problem since.

It is similar to the ice-cube solution but the CCU still monitors the freon pressure switch.

__________________
Not MBZ nor A/C trained professional but a die-hard DIY and green engineer. Use the info at your own peril. Picked up 2 Infractions because of disagreements. NOW reversed.

W124 Keyless remote, PM for details. http://www.peachparts.com/shopforum/mercedes-used-parts-sale-wanted/334620-fs-w124-chasis-keyless-remote-%2450-shipped.html

1 X 2006 CDI
1 x 87 300SDL
1 x 87 300D
1 x 87 300TDT wagon
1 x 83 300D
1 x 84 190D ( 5 sp ) - All R134 converted + keyless entry.

An ice cube relay plugged into the Klima socket should have the low pressure cut-off feature since the low pressure switch is upstream of the Klima relay. Or am I missing something?

Did you replace it with a known good Klima relay? Did you try your original Klima relay in a similar car that doesn't have AC problems?

More likely to be a problem with the compressor speed sensor than the engine rpm sensor. Your tach works, right?

Sixto
87 300D

I had a very similar problem with my 91-2.5.I replaced the serpentine belt with a new one after cleaning all the accumulated dirt and rubber out of the grooves on all the pulleys and now have only had it happen on rare occasions like once every couple of months where before it might do it several times in one day.Cleaning the pulleys and the new belt also helped a great deal to smooth out the jumpy tensioner,i had previously replaced the tensioner shock,the new one cured the noise but there was still alot of jumping going on after cleaning the pulley grooves it is much smoother. You will be surprised how much crud & rubber accumulates in those grooves. Don

__________________
Red Green "This is only temporary,Unless it works!"

97 E300D 157000 miles
87 300TD ?141k? miles

Good point. You might also send some brake parts cleaner through the clutch gap if you have a leaky power steering pump. You might be able to floss the gap with strips of an old t-shirt.

Sixto
87 300D

Sixto: Correct.
The Klima relay has nothing to do with the low pressure cut out. The low pressure switch is in direct line from PBCU to pin 10 of the Klima Relay.

I had my 2 Klima relays remanucatured. Both won't work. It is the same problem. therefore I think it is one of the two problems: A)Compressor speed sensor
B)Engine RPM input

Not sure exactly but I believe the system checks the engine speeds versus the compressor speed at the clutch and if the belt slips there will be the cut out or if one of the two sensors fail.
My tachometer works perfectly fine.

Someone said the belt could be worn and dirt around ...I'll definitely check this out.

Thanks, very good point. I'll replace the belt and also try and clean the clutch with carb cleaner.

Thanks,
as my AC worked fine two years ago and the Relay is good. I think there is good chance that either the belt is slipping ...or the clutch is slipping ....or both.
I'll test and see what happens.
I need to install a test light for the compressor in order to 'see' when the compressor is on or off.

The compressor speed sensor is at the tail end of the compressor. You'll see the wires. Obviously it's attached to the driven plate side of the clutch rather than the pulley side.

Sixto
87 300D

From my messing with the system I've determined the Klma relay:
-Waits a moment after starting to turn on the compressor
-Turns off the compressor on kickdown / WOT
-Turns off the compressor if the belt slips
-Turns off the compressor if the coolant is too hot

The evaporator temp switch goes into the pushbutton unit.

-J

__________________
1991 350SDL. 230,000 miles (new motor @ 150,000). Blown head gasket

Tesla Model 3. 205,000 miles. Been to 48 states!
Past: A fleet of VW TDIs.... including a V10,a Dieselgate Passat, and 2 ECOdiesels.
2014 Cadillac ELR
2013 Fiat 500E.

I wonder how similar the compressor RPM signal and the engine RPM signal are? IE could you feed the engine RPM signal into the compressor RPM feed to make the relay happy...

-J

__________________
1991 350SDL. 230,000 miles (new motor @ 150,000). Blown head gasket

Tesla Model 3. 205,000 miles. Been to 48 states!
Past: A fleet of VW TDIs.... including a V10,a Dieselgate Passat, and 2 ECOdiesels.
2014 Cadillac ELR
2013 Fiat 500E.

I just don't know how the engine signal is related to the compressor signal.

I assume both work. If it won't work the compressor won't come on at all.
But under load the compressor gets cut out.

I'll first check the belt system and the clutch.

Jeremy confirmed the engine sensor puts out 144 pulses per revolution. I don't know if that signal is modified before going to the Klima relay. I don't know what signal the compressor sends to the Klima but I'm guessing it isn't 144 pulses per revolution.

Sixto
87 300D

The klima relay is hooked right up to the output going to the tach.

It's very likely the signal from the compressor RPM sensor is different than the engine RPM. But I wonder if it is different enough to "anger" the relay and kill the compressor. If I had my car here I'd test it...

-J

__________________
1991 350SDL. 230,000 miles (new motor @ 150,000). Blown head gasket

Tesla Model 3. 205,000 miles. Been to 48 states!
Past: A fleet of VW TDIs.... including a V10,a Dieselgate Passat, and 2 ECOdiesels.
2014 Cadillac ELR
2013 Fiat 500E.

How would you test this? Can you explain?
In the diagram
pin 2 is the RPM input from the engine
pin 9 and 11 are the compressor input

I have no idea how these 3 pins work together. If you have a suggestion how to test I'll do right away. Please make a suggestion


here is some information i found on the web:
....The real tough ones are under the hood. Once it has been verified that the PBC is doing its job, a few observations should be made. First start with the belt tension and tensioner. A belt that flops is going to cause enough disruption to fail the comparator test in the compressor controller. A system that starts and then stops and never restarts until an ignition cycle is likely to have failed this test, although low voltage has also been claimed to cause disengagement that requires an ignition cycle.
The low voltage problem has been described as a problem during start-up where the voltage drops below a threshold (probably around 9v) and the current load is reduced by disengaging the clutch. I haven’t verified this, but it is an internet fix response. Other reports, including one I can verify, say that a low and rapidly changing idle can set off this comparator.
I experienced this situation on a 1995 Mercedes-Benz E300 D. The problem with the car stemmed from the over-voltage protection relay. This notorious relay plays no part in the compressor circuit but does power the idle controller that converts the induced crank sensor signal from AC to a pulsed DC engine speed signal. This signal is used by the tach, the A/C compressor controller and the EGR controller, which also controls the variable manifold length servos. While the speed signal never varied, the idle controller was losing control when the relay faulted. During this time, if the compressor engaged at idle, the speed would take a big hit and the compressor would disengage.
The A/C clutch should be checked for both gap and an oil-free state. Power steering fluid, engine oil or A/C shaft seal leaks can oil down the clutch until it slips enough at idle to set off the comparator. Try some brake clean on the clutch if it’s oiled and look for signs of overheating.
If no mechanical problems turn up after a visual inspection of the belt system, then begin electrical testing. The testing takes different approaches depending upon the year and model. The 124 and 126 models up to 1990 (and some later diesels) use the Klima relay. The V8s don’t use the compressor speed signal and speed comparison, as they have a separate belt for the air conditioning.
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). See accompanying waveforms for before and after signals (top left and bottom right, respectively).
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.
Clutch coil current should also be inductively checked off the jumper. The amperage should be less than 5a and is usually around 3a. If a controller is bad all the time, definitely check this current before replacement. Unfortunately, this is also intermittent and will burn the controller just like a fuse. If a new controller comes back, I would definitely suspect the clutch coil.
Many unnecessarily condemn the compressor speed sensor, as a very small signal is monitored. The photo shows the end of the shaft that is read by the sensor. This particular compressor locked up. The sensor shows the arc that the shaft segments follow, as the shaft was hitting the sensor before it locked up and caused the damage. Notice that the sensor is off center from the shaft and normally it is flat at the bottom. The only shaft speed sensor problems we have seen have been from rebuilt compressors where the distance from the sensor to the segments was too great.