Water temp gauge woes

Welp, took my car on it's 4th maiden voyage since i've had it yesterday.

Still runs like a top, wideband is working as far as i can tell. (Connect your grounds, moron!) It's amazing how a car with a refreshed tight suspension feels compared to one with everything shot. (Escort)

Anyways, i have three little issues to correct.

1) Boost controller seems to be content to just give me ALL the boost
2) Leaking coolant from the temp probe fitting on the hose adapter

And this is the one the thread is about:
3) The water temp gauge's reading seems directly influenced by amount of electrical draw on the car. With the lights off, it seems to read correctly. If i turn the headlights on, it jumps by at least 40 degrees. What could cause this?

Bad ground on the temp sender? I grounded to the valve cover where there was already a ground. Should i run it back into the cabin and ground to the barrier strip i've got set up as ground?

Are you sure you are getting enough juice from the alt to run everything?

I should be... this is the only electronic device that's acting up, i'd think if i was low on juice all three of the Prosports would be E36 M3ting a brick. Well, at least the two with electronic sending units.

When the car is running i'm measuring a touch over 14v.

I'm sure it could use a replacement of every ground in the thing, they're all 21 years old at this point. If i turn on every light at once, there's a real quick change in engine note.

It's worth mentioning that my oil pressure sender was grounded to the barrier strip rather than in the engine bay, and i'm not getting any problems out of that one. May answer my question in itself?

I'm going to say somehow the dash lights are trying to ground through the sender circuit (assuming the gauge has an illumination light). Don't laugh, that's the kind of weird stuff that happens with ground problems. Disconnect the gauge illumination and try it again, if it works normally now [Jamie] 'well there's your problem' [/Jamie].

It could also be somehow your wire from the gauge to the sender has become a ground for the lights. Make real sure there's not a bare splice touching somewhere etc and be certain the engine is well grounded to the body of the car and also to the battery.

If it's electrical and on a car the odds are it's a ground problem. Especially if it's doing something strange. Hot side problems manifest as blown fuses or smoky wires. Ground problems are the really weird things you didn't know electricity could do.

Curmudgeon wrote: I'm going to say somehow the dash lights are trying to ground through the sender circuit (assuming the gauge has an illumination light). Don't laugh, that's the kind of weird stuff that happens with ground problems. Disconnect the gauge illumination and try it again, if it works normally now [Jamie] 'well there's your problem' [/Jamie]. It could also be somehow your wire from the gauge to the sender has become a ground for the lights. Make real sure there's not a bare splice touching somewhere etc and be certain the engine is well grounded to the body of the car and also to the battery.

The gauge illumination is built in and integrated in the power/ground circuit for the gauge itself. If i disconnect it, the gauge itself will have no power. Can't test like that, bummer.

No bare splices, every connection was soldered, taped, then shrink wrapped.

Hrmm... looks like i'm playing with a fluke again today.

Can you pull the bulb on the gauge. That would effectively disconnect the lighting circuit.

Toyman01 wrote: Can you pull the bulb on the gauge. That would effectively disconnect the lighting circuit.

I think there's like... 10 LEDs in the thing. If they aren't hard-wired in, i can give it a shot.

92CelicaHalfTrac wrote:

Toyman01 wrote: Can you pull the bulb on the gauge. That would effectively disconnect the lighting circuit.

I think there's like... 10 LEDs in the thing. If they aren't hard-wired in, i can give it a shot.

I was afraid of that. A lot of the new gauges are coming with LEDs. Is the gauge illuminated at all times or only when the lights are on. If it's always illuminated then the ground problem probably isn't in the gauge. Something in the lighting circuit is trying to ground through the gauge like Mr. C said. Check all the lighting grounds. Not only the dash lights, but the signals and headlamps as well. It could also be feeding through the lighting circuit from something else. Good Luck.

Toyman01 wrote:

92CelicaHalfTrac wrote:

Toyman01 wrote: Can you pull the bulb on the gauge. That would effectively disconnect the lighting circuit.

I think there's like... 10 LEDs in the thing. If they aren't hard-wired in, i can give it a shot.

I was afraid of that. A lot of the new gauges are coming with LEDs. Is the gauge illuminated at all times or only when the lights are on. If it's always illuminated then the ground problem probably isn't in the gauge. Something in the lighting circuit is trying to ground through the gauge like Mr. C said. Check all the lighting grounds. Not only the dash lights, but the signals and headlamps as well. It could also be feeding through the lighting circuit from something else. Good Luck.

It's always illuminated.

Do you think it's trying to ground through the gauge? Or through the sender?

Is there any direct connection between the lighting circuit and the gauge or sender. It sounds like there shouldn't be since the gauge is always lit. I would probably try a redundant ground from near the sender to the gauge and see if that solves the problem. Then make sure the dash lights if any are well grounded. If the entire dash is grounded through a tiny wire it will act like a resister as the load increases causing all kinds of problems. A quick dirty check would be a jumper wire with a couple of alligator clips and ground it back to the battery. From there I'm going to say check the voltage on the gauge circuit and the lighting circuit. A voltage drop in the gauge could be causing the gauge to act wonky. It's hard to nail 12V problems down to one thing.

How good is the engine ground strap? If there is not a good ground from the block, then grounding your temp sensor to the valve cover may not be the best bet.

Toyman01 wrote: Is there any direct connection between the lighting circuit and the gauge or sender. It sounds like there shouldn't be since the gauge is always lit. I would probably try a redundant ground from near the sender to the gauge and see if that solves the problem. Then make sure the dash lights if any are well grounded. If the entire dash is grounded through a tiny wire it will act like a resister as the load increases causing all kinds of problems. A quick dirty check would be a jumper wire with a couple of alligator clips and ground it back to the battery. From there I'm going to say check the voltage on the gauge circuit and the lighting circuit. A voltage drop in the gauge could be causing the gauge to act wonky. It's hard to nail 12V problems down to one thing.

Voltage on the gauge circuit is putting out about 14v while running, it's healthy.

There is actually a direct connection between the lighting circuit and the gauge, but not the sender. Turning on the parking lights switches the gauges to red illumination so you don't blind yourself at night. That said, i have all three gauges wired together, and then into the same wire, so if it were a problem there, all three would be doing wonky stuff.

On the "problem gauge," basically i've got the gauge, which has power/ground to barrier strips (red/black). Then there's the "night mode" signal wire (orange) tapped into the parking lights (like the others.) Then there's the actual signal wire from the sender (green). The sender has it's own ground, to the valve cover.

What the gauge reads is dependent solely (or should be) on the signal sent through the green wire, and the signal source is the water temp sender. It won't do anything without the sender hooked up, besides light up and look pretty. (tested and verified) Am i wrong in thinking that the problem lies with the sender? Whether it's a bad sender, or the ground for that just sucks?

I hate electrical stuff. I'm so newb at it.

T.J. wrote: How good is the engine ground strap? If there is not a good ground from the block, then grounding your temp sensor to the valve cover may not be the best bet.

Good question and i have no idea.

Worth mentioning that the engine note would change for a quick moment even before i installed all this crap, so i would guess that the engine grounds are pretty crappy.

92CelicaHalfTrac wrote: On the "problem gauge," basically i've got the gauge, which has power/ground to barrier strips (red/black). Then there's the "night mode" signal wire (orange) tapped into the parking lights (like the others.) Then there's the actual signal wire from the sender (green). The sender has it's own ground, to the valve cover.

Disconnect the "night mode" wire and see what happens. That would isolate the gauge and the sender from any other electrical circuit.

Also try running the sender ground wire directly to the battery.

Figured it out. Bad ground. Re-grounded it to the radiator support and all is well. Now to figure out this boost controller that seems to have been made in the fiery depths of hell.

I remember the porsche 944 had similar grounding issues. One of the "fixes" was to add a gound to the alternator's mounting to the body of the car

Glad it's fixed! Grounds look like they don't do anything, in reality they are more important than the power side of the circuits. My first encounter with weird grounds was an MG that was melting heater control cables. Somewhere along the line, the original engine to body ground cable got tossed. So everything on the engine was grounding through the heater valve cable, then to the heater control on the dash and then finally to the body of the car.

Way back in da day, Ford had a tech bulletin concerning mystery driveability problems. Turned out the TFI (thick film ignition) module grounded against RFI (radio frequency interference) through the body of the distributor. If that ground got a little corrosion and the resistance went up, the car would stutter and miss when going under or near those really big power transmission lines. The cure: run a ground wire from one of the TFI module mount screws to a good ground on the body.

Yeah, i'm pretty pleased with it.

However, now it seems that the same gauge has a few LEDs that aren't working 100% in "night mode." Sigh.

Looks like i'll be getting ahold of ProSport to see if they'll replace a gauge i bought like... 2 months ago. If not, good thing it's only $40 to replace, right?

Now i just have to figure out Satan's boost controller.

Sorry for resurrecting this thread but I seem to have the opposite issue with my Prosports gauge - namely, it seems to be reading about 20F low at normal operating temperatures. Gauge agrees with the datalogger up to about 145F and then it slowly diverges until it reads around 190F when the datalogger reports that the ECU is seeing around 210F.

Sensors for the ECU and gauge are right next to each other, plus the fan switch (which agrees with the ECU sensor temp-wise) is on the other half of the thermostat housing.

Voltage at the battery with the engine running is around 14.3-14.4V.

Bad ground? Bad sender? Bad car?

T.J. wrote: How good is the engine ground strap? If there is not a good ground from the block, then grounding your temp sensor to the valve cover may not be the best bet.

This!

Bad engine grounds mean that any and all connections to the chassis get used. Sendors, throttle cables, clutch cables/hydraulic lines. Install a good (or secondary) ground strap from engine to chassis (or battery) and all should be well

BoxheadTim wrote: Sorry for resurrecting this thread but I seem to have the opposite issue with my Prosports gauge - namely, it seems to be reading about 20F low at normal operating temperatures. Gauge agrees with the datalogger up to about 145F and then it slowly diverges until it reads around 190F when the datalogger reports that the ECU is seeing around 210F. Sensors for the ECU and gauge are right next to each other, plus the fan switch (which agrees with the ECU sensor temp-wise) is on the other half of the thermostat housing. Voltage at the battery with the engine running is around 14.3-14.4V. Bad ground? Bad sender? Bad car?

could be the gauge is reading on the "cool" side of the radiator and the ecu is reading on the hot side ....

my h2o gauge reads only about 210° - 220° after a 20 min session in 95° weather

and reads a steady 195° on the hwy

BoxheadTim wrote: Sorry for resurrecting this thread but I seem to have the opposite issue with my Prosports gauge - namely, it seems to be reading about 20F low at normal operating temperatures. Gauge agrees with the datalogger up to about 145F and then it slowly diverges until it reads around 190F when the datalogger reports that the ECU is seeing around 210F. Sensors for the ECU and gauge are right next to each other, plus the fan switch (which agrees with the ECU sensor temp-wise) is on the other half of the thermostat housing. Voltage at the battery with the engine running is around 14.3-14.4V. Bad ground? Bad sender? Bad car?

I'd be more likely to believe the Prosport over what the ECU is reading, to be honest.

Doesn't sound electrical, just sounds like two sending units that don't agree with each other. It happens.

It's 2:1 in favour of the ECU though. The ECU sensor and the separate thermoswitch for the radiator fan (also mounted on the thermostat housing) agree inasmuch as the ECU and the temp switch both match in temp reading when the fan kicks in, whereas the Prosports gauge reads low. And they're all mounted within 1" of each other.

Not that this matters anymore, the Talon has come up with a new fault (laptop loses connection to the ECU every couple of minutes) and I'm now heartily sick of it. Time to eat my losses I think.

Uh oh. I think if I can find a better ground for the gauge, the Prosport should produce readings closer to the ones read by the ECU.

Let's not dwell on how I arrived at this conclusion.