Running cold?

[CaptainXL]

 

[AtlWrk]

I agree it can depend on how the needle was mounted, but I also believe these gauges are fairly accurate. For the sake of this thread...can we also agree that there is a constant varying voltage coming from the temp sensor and the gauge moves accordingly? We all drive these almost daily and the needle reads the same when hot, it's not some arbitrary voltage it's an actual thermistor and changes voltage with temp changes. Sending a changing voltage to the ECM then sending a voltage to the cluster.

When the coolant temp rises, the gauge rises,.... not cold, warm, hot, but incrementally. The thermostat according to the service manual is normal when opening from 188-206 deg....195 is fine. One tick to the left is fine, 2 or more is not good. 210 deg is straight up, I don't want my gauge reading 210 when my coolant is 195, I prefer to know what the temp is, especially when it starts getting too hot.

Back to the quote, the odd display is not linear for sure, it's setup like a half-sweep gauge and this is used widely. Helps to monitor changes like when the coolant starts to rise past normal, easier to monitor the smaller increments.

Sort of. The idea of accuracy is largely moot when referring to the cluster temperature gauge. The PCM monitors the ECT sensor resistance and converts it internally to an OBD-compliant digital value. This is what your scan gauge gets when it requests coolant temp information. However, the cluster never sees this data nor the sensor voltage. Instead, the PCM decides how it wants to massage the numbers then sends periodic instructions to cluster as to where to place the gauge needle.

PCM sees 190 degrees and tells the cluster to point straight up to what we have labelled as 210. Sees 215 degrees? Exact same command from PCM to cluster, straight up. Dipped down to 180? Ok, lets shift it down a tick or so. The intent is not to show you the actual temperature, it is to show you a 'status' of the temperature.

This doesn't make the temperature gauge useless for diagnosis. The PCM and cluster are capable of putting that needle in the exact same spot every single time...IF it wants to. Therefore if it's suddenly not in its normal position (whatever that may be for the individual vehicle) then it is because the PCM has decided that what it is sensing from the thermistor warrants moving the needle on the cluster and commands as such.


(don't get me started on the oil pressure "gauge"...)

 

[CaptainXL]

The intent is not to show you the actual temperature, it is to show you a 'status' of the temperature.

That's right.

The system is all wonky. The way it seems to me is that basically the engineers never really spent a great deal of time thinking this through or testing it. It's not a matter of keeping costs low. The signal from the coolant temp sensor could just as easily been split off and sent to an analog dash gauge. At the time of the vehicles introduction the 4.2L was a new engine. It had a strike against it due to the fact that it never was run long enough on the stand to stress many of the parts. I even doubt the thermostat was ever replaced during trial runs. And due to downsizing at GM during the mid 2000's customer service and communication between dealerships and the revision section of the engineering department lessened. Basically the problem never received enough attention during a tumultuous time. Well that's my take on it anyway.

 

[AtlWrk]

That's right.

The system is all wonky. The way it seems to me is that basically the engineers never really spent a great deal of time thinking this through or testing it. It's not a matter of keeping costs low. The signal from the coolant temp sensor could just as easily been split off and sent to an analog dash gauge. At the time of the vehicles introduction the 4.2L was a new engine. It had a strike against it due to the fact that it never was run long enough on the stand to stress many of the parts. I even doubt the thermostat was ever replaced during trial runs. And due to downsizing at GM during the mid 2000's customer service and communication between dealerships and the revision section of the engineering department lessened. Basically the problem never received enough attention during a tumultuous time. Well that's my take on it anyway.

I would tend to believe it's largely about showing the driver enough to convince them they're getting adequate information without giving them too much as to risk misinterpretations or automotive hypochondria. In other words, if they gave us no temp and oil pressure gauges people would notice and say "my, how lacking the instruments are in this car. how cheap or how can I feel secure that everything is fine?" But if they gave us truly accurate gauges they would have customers panicked about 12psi idle pressures or 215 degree coolant temperatures. Solution: give them the gauges but make sure what is displayed gives em the warm and fuzzy (i.e. lie).

FWIW, there would be zero physical changes needed to give us an accurate temperature gauge. The PCM has the right info--it's just programmed to tell the cluster to lie to us.

To GM's defense, 99% of the time you don't need to know if it's 195 degrees or 210 degrees as long as it's somewhere in that range. So that's what they gave us.

 

[gmcman]

To not make one huge quote, the last 3 posts I agree with for sure. I don't want to say the gauges read incrementally and exactly as the sensors transmit their signals, just that it's not in 3-4 step intervals. You can clearly see the stepper motors moving the needles in small "steps" for lack of better terms and they surely aren't smooth and precise like a dedicated gauge would be.

As far as the oil pressure gauge is concerned....I definitely agree to that.

I'm very curious now as to how "incrementally" these gauges respond to the sensor data. I think I'm going to take a couple DMM's and along with the scan tool, with the cluster gauge in view, and record all of them together and record on video how they interact.

Mainly for educational purposes, as I'm curious myself.

 

[CaptainXL]

My guess is its an algorithm with a few conditional statements. Probably wouldn't make much sense unless you looked at it from a few perspectives. There is probably one loop for startup, another for fully warmed up, etc... could be a few others as well. Its definitely stepping. Probably designed to not use much CPU cycles. It purposefully keeps the sample rate low. The engineers are always trying to shoehorn as much usefulness into the programming without wasting money on more expensive pcm components.

 

[Old Salt]

A little over a year ago my 2003 was getting poor mileage and the power was low. Very poor torque on acceleration and on the hiway it had to downshift on even the tiniest hill. Took it to my mechanic and they spent a LOT of time trying to figure it out. Cleaned injectors, and numerous other things trying to cure it. Finally discovered that the cat was clogged. Took it to an exhaust shop and got a replacement put on. Problem cured. The old one was full of crap on top of the inlet screen.

 

[The_Roadie]

FWIW, there would be zero physical changes needed to give us an accurate temperature gauge. The PCM has the right info--it's just programmed to tell the cluster to lie to us.

I LOVE your term "automotive hypochondria" BTW. Watching my scangauge on a lot of mountain towing trips, I can see a steep dropoff in coolant temp when coasting downhill using compression braking. But it isn't reflected in the gauge until a lot later. My conclusion is that the gauge is severely non-linear as noted, with a large range of temps being displayed as 210, needle straight up. And it's ALSO time-smoothed (filtered) same way as the fuel level is, so sloshing doesn't appear quickly on the gauge. All to keep from alarming the "Kettles" as I picture the average driver who cares nothing about the details of their vehicle's systems.

View attachment 28234