Finally got my '92 300SE back in one piece. I've got an issue with the HVAC and I'm trying to run it down...

When I reinstalled the controller unit in the dash, as soon as you start the car, the radiator fans come on at stage 1 (low speed). Disconnect the controller, fans shut off. This condition exists whether or not the controller is on or off.

Switch the controller to LO on both dials (max cooling), and you get no A/C compressor. HVAC fan runs on high speed as expected, but no A/C compressor.

There is pressure in the A/C system.

Ran a diagnostic from the controller (press Rest for > 5 seconds). ECT sensor (sensor 6) shows HI, even when car is stone cold. Value does not change as car comes up to temp (about 85c indicated). Refrigerant pressure shows 5 bar - about 72psi (sensor 7) which I think should be sufficient to allow the compressor to engage.

Disconnect one sensor at a time at the thermostat housing - no change in behavior.

From this, I make the following assumptions:
1) A/C compressor is "off line" because the controller thinks the car is overheating.
2) Fans are running (even if at low speed) for the same reason.

Here's the rub - I got the car back from my mechanic in August after he did a head-gasket replacement. The car was in pieces when I took it to him as I was in mid-stream doing a more-or-less full interior replacement. I'd gotten too busy to finish the interior work, so I let him have it to do the head-gasket.

The HVAC didn't exhibit this problem before I took it in to him.

I've seen some harness degradation at the connector to the A/C clutch. I pulled that part of the harness loose (includes all the wires to the fans, plus one wire to the coil and another unknown connector) and stripped the cover off of the compressor leg all the way back to the main harness bundle (about 12") and found no degradation at all - wires seem very pliable and insulation solid. Used some hi-temp electrical tape on the exposed wires and recovered the harness exposed and reinstalled.

I don't think it's a harness thing, at least as far as I could tell.

The harness that works the injectors, etc has already been replaced.

I do have a spare HVAC controller which I can plug in to test the problem, but I'm wondering...

If the ECT sensor is shorted out, shouldn't disconnecting the sensor allow the HVAC system to reset (no longer seeing overheating)? Or does the absence of a sensor reading translate to overheating?

I've got a little more debugging to do before I take it back to the mechanic and start with the "it wasn't doing this when I gave it to you" dance...

Any suggestions?

Bump - surely someone has some input on this????

High ECT will disable the compressor on the 210, and I would expect the same of your 140. Do you know if you have one or two ECT sensors? The 210 uses a signal from the instrument panel for the ECT value, with one sending unit.

There are 3 ECT sensors mounted on the thermostat housing. One has 4 pins, and AllDataDIY clearly identifies that as the ECT for the engine computer. The second has a single pin, and the third has two pins.

Digging through AllDataDIY reveals that the errors logged by the HVAC control for an error relating to the ECT refer to component 10/8 (I think). However, when you dig up the location of that item, it's not listed - grrr. I think I've been down this road before. (I haven't checked to see what errors are stored by the control unit).

I've gotten the resistance values for the ECT from the diagnostic charts. When I have a minute, I'll compare those values with the values for the two sensors and see if either one of them is faulty.

I'm sure that the single-pin sensor is reference to ground. I'll hazard a guess that the double-pin sensor is reference between the two pins.

I also answered one of my earlier questions while digging through the error codes for the HVAC control - it does log short or open circuits as errors. So, if the sensor is connected but shorted to ground, that's an error. Disconnecting the sensor entirely will also be an error, and I'll bet that both errors result in the same behavior - namely no A/C.

Thanks for your input. I'll let the board know what I find out.

Regards

Single pin sensor is temp gauge................
Your problem is the 2 wire, blue sensor/circuit..that is temp input to the Control Panel [ N/22] and if it is open or disconnected , the CP Defaults to High Fan to protect over-heat condition in the event of a bad sensor.
Unplug that sensor and bridge those wire with a resistor of approx 500 ohm...fan shoud stop if diagnosis is correct.

__________________
A Dalton

I tried the resistor trick - no effect. I didn't have 500 ohms, only 468, but according to AllDataDIY, that should correspond to about 85C.

I didn't reset the fault codes in the controller before hand, if that matters.

Did check them and here's what I got:
E013 - no connection to rear A/C control
E040 - ECT Sensor - Short Circuit (continuous fault)
E041 - ECT Sensor - Short Circuit (intermittant fault)
E073 - Heater Supply unit coolant circulation pump (intermittant fault)
E075 - Coolant Circulation Pump - open or short circuit (intermittant fault)

E013 can be ignored - the control unit is from a '92 600SEL, which has standard rear A/C. This car doesn't have it.
E073 & E075 have been persistent problems with the car, even after I replaced the Duovalve pump assembly. As I understand it, these errors are a system issue and routinely show up, even when there's no "real" problem.

E040 and E041 point to the ECT sensor, which when bypassed with a 500ohm resistor had no effect. However, again, I don't know if I needed to erase the fault codes first.

I've erased the fault codes and will check behavior again with the resistor.

The diagnosis tree is fairly simple here - it's either 1) The sensor, 2) The wiring, or 3) the N22 controller.

I'm pretty sure I've replaced the sensor, but the resistance values are easy to check per AllData's values.

The wiring was moved when the headgasket job was done, and I've not checked the integrity of the wires in that sheath - that is something I need to do. It is definitely possible that the insulation on that leg of the harness is damaged - it lives in close proximity to the hot motor. (I didn't check that leg of the harness the other day as I didn't yet know it was the sensor values causing the problem.)

I'll do a little more digging and report back my results.

Thanks to all that have chimed in!

Then you have an open or short in the wires going back to the N22 from the sensor, as CP is still Defaulting.
The resistor was just to take the CP out of possible Default mode and test the N22 sesnor feed loop, but it has failed the test, eliminating the sensor as suspect and leaving the suspect as the connecting wires to N22..

The fact that an "intermittent" fault is triggered would lead me to clear the codes and try again.

Can you version-code the unit so that it no longer believes that you have rear AC?

The real question is whether or not the a/c controller is telling the BASE module to turn the compressor ON or not.

IF the BASE m,module has codes about belt slippage OR clutch operation blocked, it won't try to turn the compressor on.

Can you read codes from BASE module? OR does someone have scanner that can look at live data in that module?

__________________
MERCEDES Benz Master Guild Technician (6 TIMES)
ASE Master Technician
Mercedes Benz Star Technician (2 times)
44 years foreign automotive repair
27 Years M.B. Shop foreman (dealer)
MB technical information Specialist (15 years)
190E 2.3 16V ITS SCCA race car (sold)
1986 190E 2.3 16V 2.5 (sold)
Retired Moderator

I need to make double-damn sure I got the resistor installed properly. I had to twist the leads together of 3 different resistors to get to the 468 ohms. Those wires are tiny and maybe I didn't get a good connection. (I did connect the resistors in series.)

I'm pretty sure I can pull the codes for the Base module, but I'm still gonna have at the wiring harness leg for the sensor and double check the resistance of the sensor at various temps.

As for version-coding - can't find anything like that on AllData. I installed this controller in an attempt to solve a problem I'm having with the A/C where it wants to dump all the air to the windshield when the set temp on the controller is at or near the external temp. Causes massive condensation to form. The software version that's on this controller was supposed to combat this tendency, but I don't think it made any difference. It never seemed to care performance-wise that it couldn't communicate with the rear controller...

I have the original unit which I had refurbished by Programa - I'll drag it out of the box and reinstall it this evening.

If you want to test the wires back to N22, I would just ohm them out from the sensor connector for cont. and open/shorts.

Note*

This test is assuming you have not added a bridge resistor modification across the sensor for early cut-in..that voids this test]

With sensor unplugged, go across the blu and brown wires in that connector w/Key Off.
Looking for approx 1.3K.
Now go from Brown to ground.. [ thermo housing is OK]..looking for Cont.
Now go from blu to ground..looking for 1.3 K again.
You may want to flex the harness when doing this test for intermittant short/open , considering the known harness problems on these.

As far as the thermistor goes , instead of subbing the 500 OHM resistance, just take an R value of the Sensor when cold ..it should be extremely high, so that would keep CP out of Default for High Fan. [ like 5/6K]
I only had you do the 500 b/c that is a higher R than threshold for N22.
If you have a high R factor at sensor, you do not need the sub test.
But you do need the wiring test.

__________________
A Dalton

Made the test indicated by Arthur last night.

Key off.
Pulled connector from 2-pole sensor on thermostat housing.
Resistance between wires: 29.6K ohms
Resistance between brown wire and ground: basically none (30 or less ohms)
Resistance between red wire and ground: 29.6K ohms
Resistance between poles on sensor - 486ohms
Turn key on - temp gauge in cluster indicated about 85C at this point.
No "early turn on" mod installed on this car.

Pulled the sheath on this harness leg back as far as I could (3 or 4 inches) - insulation, while a bit stiff, is resistant to any attempt to break or fracture under thumb-nail pressure.

I replaced the controller in the car with one that was original to the car that I'd had rebuilt by Programa. Pulled error codes this AM after my drive into work:
E040 - ECT Sensor - Short Circuit (continuous fault)
E041 - ECT Sensor - Short Circuit (intermittant fault)
E043 - ECT Sensor - Open or Short Circuit (intermittant fault)
E075 - Coolant Circulation Pump - open or short circuit (intermittant fault)

E013 is gone, as expected, because this controller is for a car without rear A/C.
E073 is gone, but that was an intermittant fault - it might or might not show up again.

Sensor values still show HI for the ECT sensor on the new controller.

I am 95% positive that the previous controller, regardless of where I set the dials, would not ever activate the A/C compressor.

This new controller will allow A/C. Setting the dials to LO nets cold air out of the ducts in short order.

What do we know now, based on the previous tests?

1) The controller is likely not the cause of the erroneous readings. Two different controllers, one recently rebuilt, report same (or close to same) trouble codes, and comparable sensor values.
2) The sensor itself is not the problem, as the 486ohm is within the acceptable range for 85C coolant (460 - 650ohms, per AllData).
3) Wiring values are way out of spec. (Should have seen 1.3k, got 29.6k) I'm pretty sure there should be little to no resistance to ground for the brown. The values I got were close to 0 resistance (30 or less ohms). That wire is the ground reference wire (according to AllData's diagrams) for all of the sensors in the HVAC system, so I'd think if there was an issue with that wire, possibly I'd be seeing all sorts of other problems.

One thing I have noticed is that the fans come on on Stage 1 when you first start the car. As the car warms up, they run faster, but I don't think we get to Stage 3.

What is curious to me is that the specified resistance values for the sensor are 5-8k for a cold engine (20c), and 190-220ohms for an overheating engine (120c). However, the circuit resistance is baseline at 29.6k, which confuses me as to why the controller thinks the car is overheating. I guess I just don't understand what the controller expects to see in this circuit...

I'm thinking I need to check the wiring harness further...

What say ye?

I dn;t like that 30 ohm ground... A ground wants to see Trace....

__________________
A Dalton

Please explain "a ground wants to see trace"

I was digging through my earlier posts from 2007 about HVAC issues and I noted that the intermittent code for the ETC has been around for a while. I'm thinking more and more that something bad happened in the wiring harness when the headgasket job was done.

The engine harness has been replaced, but before I got the car. This circuit does run in the engine harness, so maybe...

Trace = Less than 1 ohm.

If you have a set of jumper cables, hook one from the battery neg to the engine block, and the other from the block to a good chassis spot.
That assure both engine and chassis have ground reference to bat neg.

If that gets you a better ground reading, you know you have a poor ground..which means check all brown ground wires.

Don't be too concerned about the 1.3K I gave you, as that came off an E class rig.

__________________
A Dalton