Chevy Impala SS Forum banner
21 - 40 of 61 Posts

·
Registered
Joined
·
6 Posts
WP sensor feeds the PCM, which includes the fans.

Head sensor feeds the gauge only. The OP apparently rewired his car to use a single sensor on the WP. Odd modification considering having a sensor in the head is exactly where I would want to take my gauge readings. The heads overheat first if there is an issue. A total loss of coolant may not even be picked up from the WP sensor until it's too late.

As for your issue, perhaps it's better to start a new thread. Sounds like you may need a completely different set of troubleshooting steps, including telling us if the Check Engine light is on, and if you tried scanning for any codes. (hint: a CEL will run the fans constantly). I don't think we should hijack this thread. The OP hasn't even commented back yet after a week of us posting suggestions.
I just started a new 1 and thanx . Looking foward to talking to you.
 

·
Registered
Joined
·
109 Posts
Discussion Starter · #23 ·
Z09B4U Thanks for clarifying the grounding issue. That makes perfect sense. Those sensors come with thread seal pre-applied, so that may be contributing. It should be pretty straightforward to troubleshoot, I'm hoping to have some time this weekend.

Thanks again.

sod
 

·
Registered
Joined
·
1,677 Posts
To summarize a test light from battery positive to the sensor case(wrench area) would be the easy first test. No light no ground.

The next would be to check the sensor pin to case with a ohm meter. Then compare the sensor to water pump readings. You can not damage this sensor with a ohm meter test. Even after years of doing this daily I have sent myself on a dead end chase by probing the wrong pin.(usually unlabled connectors with multiple pin rows)

I hope these tests will prevent a unnecessary sensor change but my instinct says the problem is sensor, ground, or sensor plug related.

This is a instrumentation circuit that has a low power signal. Sensor to connector, sensor to ground, and perfect wire are all very important. A less than perfect circuit will allow a fan motor or headlight to work but will not be good enough for a gauge circuit.
 

·
Registered
Joined
·
109 Posts
Discussion Starter · #25 ·
To summarize a test light from battery positive to the sensor case(wrench area) would be the easy first test. No light no ground.

The next would be to check the sensor pin to case with a ohm meter. Then compare the sensor to water pump readings. You can not damage this sensor with a ohm meter test. Even after years of doing this daily I have sent myself on a dead end chase by probing the wrong pin.(usually unlabled connectors with multiple pin rows)

I hope these tests will prevent a unnecessary sensor change but my instinct says the problem is sensor, ground, or sensor plug related.

This is a instrumentation circuit that has a low power signal. Sensor to connector, sensor to ground, and perfect wire are all very important. A less than perfect circuit will allow a fan motor or headlight to work but will not be good enough for a gauge circuit.
I was able to both test for ground & test resistance over the weekend.

I got over 12 volts when connecting from battery positive to the sensor case (the same voltage I got when testing the battery). So it looks like I have ground. I also have continuity from the sensor case to the waterpump (ground to ground).

I then tested resistance & I'm getting accurate readings there as well starting at over 1200 ohms & then decreasing as the car ran. I watched it drop from 1200 to 1000 to 499 while the car heated up before concluding the test.

Seems like all the tests check out, but still nothing on the gauge. FYI, the gauge pegs to the left (colder) from it's sitting position when the key is turned on. It still goes full hot briefly when the car is started, which, according to the FSM, gets grounded out at startup.

I'm stumped...what next?

Thanks for all the help.

sod
200725
200726
 

·
Registered
Joined
·
176 Posts
Did you happen to see my previous post(s) about the gauge needle spinning on the shaft? It got a bit derailed by others trying to tell me they would fire me if I didn't diagnose the sensors first, as per the FSM. I get it. I wasn't trying to tell you NOT to check the sensors. I was trying to tell you that I think it's the gauge because I've worked on many 9C1 gauge clusters. And I know that there it's never supposed to read below "C" (regardless of sensor issues), even if the sensor is shorted to ground or open. I know you did some custom wiring, so I bowed out at that point.

So if you want my help now, then please reread my post(s) first.

The focus needs to be shifted to the gauge itself next. Which is what my instincts initially told me was your problem.

Been working with 9C1 Caprices for 25 years. Instinct goes a long way.
 

·
Registered
Joined
·
109 Posts
Discussion Starter · #27 ·
Did you happen to see my previous post(s) about the gauge needle spinning on the shaft? It got a bit derailed by others trying to tell me they would fire me if I didn't diagnose the sensors first, as per the FSM. I get it. I wasn't trying to tell you NOT to check the sensors. I was trying to tell you that I think it's the gauge because I've worked on many 9C1 gauge clusters. And I know that there it's never supposed to read below "C" (regardless of sensor issues), even if the sensor is shorted to ground or open. I know you did some custom wiring, so I bowed out at that point.

So if you want my help now, then please reread my post(s) first.

The focus needs to be shifted to the gauge itself next. Which is what my instincts initially told me was your problem.

Been working with 9C1 Caprices for 25 years. Instinct goes a long way.
Thanks Tom,

I have read your posts, and I think swapping out the gauge may be my next step. I've had the cluster out on numerous occasions resolving intermittent digital speedometer display issues, as recently as last month (the display is working now) so I'm familiar with pulling the cluster & swapping gauges. I just need to dig thru some boxes & find my old spare cluster.

I'll keep everyone updated & thanks again for everyone's help. The good news is that I'm still learning a lot after owning this car for over 20 years.

sod
 

·
Registered
Joined
·
176 Posts
Sounds good.

Unfortunately, in order to swap the entire cluster, you have two options:

(1) The FSM way: remove bezel, then remove entire lower dash assy. Quite invasive.
(2) The alternate way: remove bezel. Then remove the torx screws holding the plexiglass cover around the perimeter, then you can remove the cluster via the bezel hole.

But... keep in mind that if the gauge on your spare also reads below "C", it's messed up too. Every 9C1 cluster I've worked on has this problem (except a low mile unit). Civvy and Impala clusters, not so much.

If you don't want to reclock the needle on your existing cluster as instructed, and your spare cluster has a good temp gauge, you can also swap those too. The gauge is a removable module from the cluster. If you are comfortable digging in that deep.
 

·
Registered
Joined
·
1,677 Posts
storm9c1 I try to read all of your posts. You do not post much but what you do always seems relevant, useful and backed by experience.
I have no doubt your comments on this thread are no different.
In the internet and on this forum there are examples of 9C1 temperature gauges malfunctioning and the solution has been that the needle has become detachehed from the pivot rod.
With out you and others pointing this out people might miss this problem.

Where we differ is your insistence that a gauge that reads under C is damaged and needs to have the needle removed and re calibrated.

If you turn on the key into the run position, car cold (sitting for 12+ hours) and the needle sits below cold, the gauge itself is the problem. Even if the sensor is bad, open or shorted, the gauge should never sit below cold.
The OP reports that it reads below C. This is hint #1. And this is something specific to our cars. Why? Because the state of reading below C cannot not electrically exist with our gauges.
I actually agree with your second statement. In a working circuit a cold sensor will "pull the needle" up towards C.

Where we disagree is that any temperature gauge that falls below C needs to be modified.

200738


For sale on Ebay:

200739



200740



200741


These are all examples of people selling clusters with temperature gauges below C,
I too have functional temperature gauges that can drop below C BUT read correctly when the circuit is functioning.(Key on)

Please wait a hour before commenting. I have a second post to make and it will take some time to post.
 

·
Registered
Joined
·
1,677 Posts
200743

200744

200745

200746


What I see reflects my experience. The needle can be anywhere when unpowered BUT if the gauge circuit works the gauge shows the correct temperature.

200747

200748


200749


Between 9:10 and 9:30 the ignition is turned on and you can see the needle move up.
Below C at 10:10 then above by 10:30
Seems to function and definitely can read above and below C,


I think Ghetto Wagon is a member of this forum.


To be clear I understand that temperature gauge needles can spin on the pivot.

I do recommend that every thing else should be investigated BEFORE touching any gauge needle.
I do not believe that the temperature needle falling below C is a indicator of a damaged temperature gauge.
 

·
Registered
Joined
·
1,677 Posts
Sorry for the detour sod.

When you remove the gauge cluster you can test the cluster to sender circuit at the cluster plug. I just push something like a sewing needle or pin in from the back if my meter probe will not touch the contact As the FSM says it is connector C2 row A4 a dark green wire. You should get the same resistance as at the sensor.
Light Electrical wiring Composite material Cable Wire


Grille Automotive lighting Hood Motor vehicle Automotive tire

Second picture courtesy of 95wagon
 

·
Registered
Joined
·
1,677 Posts
Sometimes the pins on the cluster are damaged and do not make contact with the circuit board. Sometimes the pins are dirty.
Musical instrument accessory Audio equipment Hardware programmer Electronic instrument Circuit component


Sometimes the contact in the connector loses tension and does not make proper contact with the pin. Sometimes the contacts are dirty.

Electrical wiring Wire Machine Cable Electronic component
 

·
Registered
Joined
·
176 Posts
Where we disagree is that any temperature gauge that falls below C needs to be modified.
Respectful disagreement creates critical thinking, which I always encourage. That's how quality stuff gets done. I just don't respond well when people start talking down to others in the process.

(1) Let's take unpowered clusters out of the equation. It's an unsprung gauge, so it can go wherever it wants with no power applied.

(2) I thought I determined years ago that as soon as the cluster receives power, the gauge climbs to "C" because it is biased internally by a driver circuit in the cluster. I use my 9C1 cars in the winter and I never recall them reading below "C", even on the coldest days...as soon as you turn on the ignition, the gauge moves to "C". So there's definitely something more going on here. Maybe it needs to see some resistance from the sensor to move it to "C". Maybe no matter how it works, if the sensor is hooked up, perhaps the minimum reading is "C". But if unhooked (open circuit), perhaps it reads below "C". That could be what I'm wrong about. That's fine. I'm not perfect. Just trying to help.

It's been a number of years, so perhaps I need to review this. Next chance I get, I will repeat my original tests with my spare clusters. If I am wrong, I have no problem admitting it.

Furthermore, I was also going off the following photo the OP posted:


This pic appears to be with the sensor wire shorted to ground. But... even though the gauge is reading hot, it is not pegged and resting on the plastic edge. This is another indication that the needle has spun on the shaft. If you take a nominally functioning cluster apart, remove the silk screen plate, and then short the temp wire to ground, it will actually come to a rest far past the "H" position. His is not doing this. His is coming to a rest before the edge. Indicating that it's at least about 30 degrees off. I use these clues to tell me it's not the wiring or the sensor, but rather the gauge. It's just really hard to explain all of this here. My posts would be 10 times longer if I explain every nuance I see that leads to my conclusion.

That said.........If it does turn out that I am wrong, and that it's possible for the gauge to read below "C" under normal circumstances, then perhaps we can amend the test procedure as follows: (1) Take the cluster apart. (2) Remove the silk screen plate. (3) Key off, ensure the needle is resting far below the "C" position off the scale. (4) Power up the cluster. If it moves decidedly to the same position below "C" every time it powers up, and continues to move as the engine warms up, then the sensor and circuit is working. But the gauge likely needs calibration (ie: needle reclocking). On a cold engine, sensors and wiring working, the gauge should sit at "C", so that's what I would go for. I imagine we both agree on that. It's how to diagnose and test the calibration that needs amended. Even if the needle is reclocked, the bulb test will still bang it out of calibration quickly. Been there, done that.

After I do a new round of testing, I'll be glad to look into adding a resistor to the cluster to prevent the violent pegging during the bulb test (further aggravated by evaporation of the original fluid dampening material in the gauge motor).
 

·
Registered
Joined
·
1,677 Posts
storm9c1 what has triggered my posts are these two quotes. I thought I understood what you were trying to say but I was worried they would be taken out of context.
the gauge should never sit below cold.
state of reading below C cannot not electrically exist with our gauges.
(1) Let's take unpowered clusters out of the equation. It's an unsprung gauge, so it can go wherever it wants with no power applied.
You and I know this but if I wasted some time I can find examples of owners that do not understand this. I posted the images for those who have never paid attention and think they have a problem when the gauges are not always at the same spot when the key is out.

(2) I thought I determined years ago that as soon as the cluster receives power, the gauge climbs to "C" because it is biased internally by a driver circuit in the cluster.
The FSM shows the gauge has power and ground so this is possible. But this is where we are both bench racing. Fuse #11 feeds connector C2 row B6 witch powers some "idiot lights" and the temp and fuel. My memory is that the temp gauge is a two coil system. Like the fuel gauge one coil is always powered and pulls left. The sensor is what pulls right.

Maybe no matter how it works, if the sensor is hooked up, perhaps the minimum reading is "C". But if unhooked (open circuit), perhaps it reads below "C". That could be what I'm wrong about. That's fine. I'm not perfect. Just trying to help.
Again we are on the same page. The things I try to do with my posts is prevent people from buying parts("parts canon"), doing destructive testing(wire piercing) and testing that wastes time when normally the common diagnostic approach suggests a structured approach. This includes cleaning electrical connectors and testing wiring.

(1) Take the cluster apart. (2) Remove the silk screen plate. (3) Key off, ensure the needle is resting far below the "C" position off the scale. (4) Power up the cluster. If it moves decidedly to the same position below "C" every time it powers up, and continues to move as the engine warms up, then the sensor and circuit is working. But the gauge likely needs calibration (ie: needle reclocking). On a cold engine, sensors and wiring working, the gauge should sit at "C", so that's what I would go for. I imagine we both agree on that.
Calibrating electronic and pressure gauges paid my bills for years.
I am just going to simplify things and skip the theory. Different products require different procedures but many things are the same.
1) Adjust needle to Zero. This is for systems that have a mechanism. This temperature gauge does not.
2) Use a known 100% signal to calibrate at 100%
3) Check at 2/3 and 1/3 with appropriate signal.

The FSM and a unlabelled temperature gauge make this a challenge.

The question is WHY is the gauge important? We want to see "normal" operation point and the "maximum" safe temperature.

With no special tools and a very firm belief that everything else was "perfect" I would reset the needle at "the same position it always was at operating temperature"
With more tools I would simulate 55 ohms (260 deg F)and set the needle to HOT.

My point is below C, at C, or slightly above C is not as important. A gauge's greatest percentage error is near zero. The simple way to say this is the gauge is more accurate in the top 1/3 (66-100%) and the gauge is less accurate in the bottom 1/3(0-33%) and at it's worst just above zero.

These gauges are crude and show a wide band of OK operation. Drivers that care "get a feeling" for what is normal for different operating conditions.

I just don't respond well when people start talking down to others in the process.
I never intend to upset anyone. If I have I apologize. What I do try and do is make my explanations simple enough that a wide audience can understand what I mean. The internet is huge and there are people reading this that the english language is not their best understood one.
 

·
Registered
Joined
·
176 Posts
Thanks. Fair enough. None of us are perfect. Let's move on.

I've began testing some spare clusters. So far, every one of them does have a bias voltage on the temp gauge with just the +/- power hooked up. Nothing else. The new question is, should the bias be at "C" or lower. I will test that next with a potentiometer. I will try a few tests including using 55 ohms on "H" to calibrate. Stay tuned.

I will try to take a video or still pics of my findings if I can.
 

·
Registered
Joined
·
176 Posts
Further findings:

A5 and A11 both need +12VDC! If you leave off A11 and only hook up A5 (+) and A6 (-) power, the speedo lights up and the volt gauge works. But the temp gauge will just flop arbitrarily. It may rest below "C". Or anywhere else. As soon as you hook up A11, it activates the bias voltage on the temp gauge and it rests at the "C" position (silk screen plastics removed, it electronically resets the same position and has no peg to fall on).

Some of the pics above from the internet (esp with clusters out of the car) show 2 wires connecting to the cluster. This is enough to light up some stuff. You actually need a third (A11) +12V wire to get the temp gauge working.

Findings so far:
  • Connecting A4 (temp gauge input) via a 1K resistor to ground results in an approximately 1/16" deflection to the right.
  • A 1.5K ohm resistor (100 degrees F according to FSM) results in barely any deflection at all.
  • Anything above 1.5K ohms (including infinity, open circuit) will just read "C".
  • 55 ohms reads slightly past H. So it would be pegged at that range. Anything below 55 ohms will just peg the needle harder (ie: the bulb test).

This tells me that the gauge is calibrated to see a range of ~1.6K to 0 ohms. Anything above 1.6K gets thrown away. Anything below 55 ohms pegs the gauge.

Ergo, I stand by my original assertion that the gauge should never read below "C". Maybe slightly below is OK but not grossly.

I did notice something else odd with one 9C1 gauge I tested that no longer had any dampening. Once the needle rests at its bias position at "C", I can push it above or below "C" with my finger and it will spring back to "C" quite decidedly as long as cluster voltage is applied. But sometimes it doesn't reset to the exact same position. It can vary by 1/8" and throw off the readings. It seems to have quite a bit of arbitrary slop. Or perhaps its just mechanically worn out. I will try to demonstrate that too. But this would really throw off any type of calibration. I have about 8 temp gauges (removed from cars), but only 3 work to use in these experiments. The other 5 were cannibalized over the years to fix issues with cars and experiment with dampening.

I will set up a better jig and take video of the process that leads me to these conclusions. It may take me a few days because I have other responsibilities to attend at the moment.
 

·
Registered
Joined
·
1,677 Posts
Resistance temperature sensors usually have a logarithmic output.

Have a look at these two random curves to give an idea of the shape.
200765

200766

This table looks close to what storm9c1 is seeing:

200767

 

·
Registered
Joined
·
176 Posts
Update:

Thanks. Yeah, we could map the sensor chart for our cars as above. In fact, I've done this for other cars for engine swaps. A common example is older GM V8 engine blocks (ie: 1969 2bolt/4bolt mains) have a 1/2" hole for a threaded sensor while newer blocks (Vortec, LT1, LS1, etc) have a 3/8" hole. The 1/2" hole sensors always have a different curve than the 3/8" counterparts. I've spent hours trying to match up the curve of many 3/8" sensors to match a factory gauge on a '69 vehicle. Mapping out the stock resistance chart is interesting, especially when all you have is room temperature to 212 degrees to work with (sensor out of the car). And you plot the graph manually without the help of a computer. LOL. Not fun. Some folks just machine down the stock 1/2" sensor to 3/8". You can get away with that in some cases. I did find VDO sensors that were close to the factory curve in 1/8" which can be adapted to 3/8" with a bushing. But I digress. That doesn't come into play here. It's just an interesting. And also good for learning these things.

I just made a video on how the Caprice/Impala SS gauges work internally. It's using a cell phone, one handed. It's not stellar. But it should help everyone understand some things that I couldn't explain very well via posting. I'll post it to YouTube in the next few days. It's been years since I posted a YouTube video (long before being a "Creator" was a thing). I hope I still know how.

Stay tuned.
 

·
Registered
Joined
·
1,677 Posts
storn9c1 thanks for spending your time on this.

What you have verified is:

The temperature gauge when powered by key on will move to very close to the C lower mark when it has no connection to the temperature sender or if the temperature sender is is at less than approximately 70degF.

If one looks at the images from the videos that I posted they do show the needle returning to C with key on. From storm9c1's work we believe this would happen if the engine was cold(70degF) or even if the sensor and wiring was missing.

A5 and A11 both need +12VDC! If you leave off A11 and only hook up A5 (+) and A6 (-) power, the speedo lights up and the volt gauge works. But the temp gauge will just flop arbitrarily. It may rest below "C". Or anywhere else
This information is what led me to my mistake. When I was doing the fuel gauge calibration work the temp gauge flopped around. In the car fuse#11 (I/P indc) powers the gauges. For my fuel work I just powered C1B2 but did not power the C2A11. In the car one would have all gauges powered or all dead due to fuse#11 where as on the bench one needs two separate pink wires powered to have temp, oil, and fuel active.

As I was wrong about the temperature gauge needing the sender to move to the C mark storm9c1 gets a apology from me.
 

·
Registered
Joined
·
176 Posts
Thanks man. It's all good. I apologize too for making statements without remembering all of the information to back it up.

I think this is one of the best threads on the internet for diagnosing the temperature gauge problems (since the long defunct 9C1 forum days). And had you (and the OP) not stoked the fire, we wouldn't have learned anything (and I nearly forgot all of this stuff since I hadn't performed such a deep dive since about 10 years ago).

I think the upcoming video will be even more helpful. Hopefully.

Additionally... I learned something else today studying the schematics.

There is no bulb test pin on the cluster plug C1 or C2. The bulb test circuit only tests 2 things by the looks of it. Correct me if I am wrong:
  • Shorts the temp gauge to ground via splice S212 in the I/P harness.
  • Shorts the brake light to ground via splice S275 in the body harness.
That's all I can find. And I confirmed this by turning the key to the bulb test and only seeing the gauge peg and the brake light illuminate. Seems pretty silly that GM did it this way. I would have ran one wire to the IP and had a circuit in the board for it. But I think GM was carrying on an old design into a more modern car at the time. The bulb test function has been around for a long time dating back decades. And they were just as simple, short out various sensors to ground to light the dash lights.

The PCM (or other control modules) test the other lamps as needed (ie: airbag, ABS, CEL, etc).

So if we want to add a resistor to soften the blow of the gauge test, we would need to add it between the IGN switch and splice S212. Perhaps the dark green wire G1 in the C200 connector would be an appropriate place to cut and add one. Do you agree?

Stay tuned. I'm working on part 2 of the video today. Part 1 talks about a faulty 9C1 gauge and shows how it all works from the dash cluster perspective. Part 2 compares it to a good gauge. I will post both when done.
 
21 - 40 of 61 Posts
Top