Steel Wool

I have an increasingly frequent problem with the ecm disengaging the a/c clutch with even minor throttle input. It has gotten so bad the compressor barely ever kicks in while driving around. It usually acts normal when it is cold but once it has reached operating temperature it starts acting up.

It is a 2000 civic with 171000 miles. There are no codes. I have the autoenginuity scan tool program but not the Honda specific package. I drove it around with the scan tool hooked up. The map sensor readings looked odd showing low vacuum at idle and part throttle but I think this may have been normal. I changed map sensor with no change to problem.

I verified it is the ecm doing it by jumping the pressure switch and the clutch still disengages. I also disconnected the clutch wire and measured voltage while revving and it still cuts out.

I have checked ground wires and everything seems good.

Does anyone know if the ecm uses the knock sensor to decide when to disengage the clutch? I may have a little rod knock at low idle which is noticeable after the engine warms up a little and the oil thins out. I am wondering if the ecm may be detecting this.

Or could the knock sensor be going bad and cause this? I am going to try bypassing the knock sensor and see if anything changes. I read online where someone used a small piezo buzzer to simulate the knock sensor to fool the ecm.

Any help would be greatly appreciated.

ezone

AC operating strategy should have nothing to do with the knock sensor.
It's mostly dependent on TPS input AFAIK, assuming no overheat issues are present.



The compressor is disengaged during rapid acceleration, it then is turned back on when you get your toes out of the radiator. The compressor is disengaged somewhere around 50%-80% throttle opening, depending on year, model, etc.



A poorly running engine could cause your symptoms though. A poorly running engine would need heavier throttle application to continue to keep up with traffic.
A flatulent exhaust system might also cause excessive extension of the right foot.

Steel Wool

It will do this with only with only maybe 10% throttle opening in park or in drive.

Otherwise the car runs fine

ezone

OK........
At this point I would probably get a good scanner and start looking into the engine data list.

But first, any powertrain codes?

Looking for AC request and AC clutch statuses, along with all the rest of the engine sensor operations and hope there is at least one obviously prohibiting AC operation. Could be a flaky TPS, could be a temp sensor, AC pressure switch (overcharged?), etc.... Maybe even bad wiring that opens up as a harness flexes on acceleration (grasping at straws).

You would probably need to watch all the data WHILE the problem is acting up, driving down the road.
A scanner with flight recorder or snapshot function is handy for this, or I sometimes take another person to drive while I operate the scanner.

Steel Wool

Okay a little update. Looking at sensor data with scan tool graphing along with a/c clutch status and not seeing anything really obvious.

I have been looking at tps since it seems to be throttle input which causes clutch to kick out. One thing that looked odd was that with engine off the throttle input on graph was much more responsive. The graph showed every little bit of throttle input but with engine on it was more smoothed out and. Having trouble describing it.

I unplugged tps sensor and car ran and would not kick out the a/c clutch. I replaced with new tps but of course with everything hooked up and codes cleared it still has the same problem.

Does any of this point to the ecm being the problem?

If not explicitly the ecm any idea what else is going on when the sensor is disconnected and it is maybe using stored data instead of the sensor data or what other sensors might it be ignoring?

Steel Wool

I disconnected o2 sensor cel came on but still had same symptom

I disconnected ect sensor cel came on and symptom went away

At what temperature does the rad fan come on? Sensor was reading 229 but fan was not on yet.

ezone

Maybe? Can't really tell you, I can't see any of it first hand.

Has nothing to do with it.

That's too high. May not even be accurate.


Radiator fan has its own temp switch in the thermostat housing. Somewhere around 196-203*F.

But that has almost nothing to do with the temp indicated on your scanner because THAT sensor is in the end of the cylinder head, next to yet another sensor for the temp gauge.

BUT that 229 reading is probably high enough to make the PCM stop AC operation. I'd be looking to see if that 229 reading is accurate or not and fix it.

Steel Wool

Will do. Thanks for the reply.

Steel Wool

Success. It was low on coolant which was causing the ecm to disengage the clutch. It seems odd that it would let it run at idle though.

I bought a new oem rad cap. I had to replace the radiator maybe a year ago and used parts from oreilly. Never buying non oem parts again. I already had bad luck with aftermarket condenser fan. Went back to oem on that too.

I did get some coolant leakage from around the water pump after it had been idling for a while with this low coolant condition. Will keep an eye on that.

Thanks for the help ezone!