SO...a little while ago, I replaced my timing belt, and this procedure necessitates removing the thermostat housing. The thermostat broke apart when I removed the therm. housing, so I replaced it as well. Not long after, I noticed that the temperature gauge was taking forever (like 20 minutes of driving) to display a normal temperature, or never even got there. Then it stopped working entirely. Normally, my first thought would be a stuck-open thermostat, but it's brand new; additionally I physically verified that the engine is at normal operating temperature when the gauge is reading low, using a very high tech procedure (putting my hand on the upper radiator hose). I've been fixing cars (non-professionally) for 10 years and I know what operating temperature feels like (in other words I'm too cheap to buy a scan tool).

I found the ECT sender (NOT the ECT sensor), disconnected the connector, and jumped the connector terminals; this pegged the gauge all the way hot. So I figured, "This thing is definitely the sender, the gauge must be good, the wiring must good; so the sender must be bad." SO I replaced the ECT sending unit, and....exactly the same as before. It works a little, but it always reads way low when the engine is at operating temperature. It didn't fix jack squat basically.

I should note that the check engine light has never turned on during this whole rigamarole.

I back-probed the sender while connected and measured about 8.4 volts with the engine hot and running. It seems to me like the volts should be lower; with an ECT sender, volts should decrease as engine temperature increases, and the temperature gauge needle rises. Temperature goes up, volts go down, needle goes up. My voltage reading does steadily decrease as the engine warms up, but the the voltage itself seems high to me. Sort of makes sense seeing as the needle is too low on the temp. gauge.

Does anyone know what a normal resistance reading or back-probe voltage should be with the engine warm? Google doesn't seem to know. Btw, I disconnected the sender and measured that it's getting about 9.6 volts. Is that normal? The ECT sensor is getting about 5 volts, which is a normal reference voltage for it.

I know if the sender (or sensor for that matter) isn't tightened down enough, it won't have a good ground, but I tightened it pretty damn tight. I didn't use teflon tape, which can cause the same problem; the sender had some sealant pre-applied on the threads. I even jumped the sender's ground wire to a good engine ground; no change in either the gauge or voltage reading, so it's grounded fine, seems like.

So at this point I'm thinking I was sold a defective ECT sender ("BWD"-brand, part # WT724. I checked beforehand to make sure this is the sender and not the sensor). I have no idea what else could be causing this. Thoughts? Comments? Suggestions?

 

Should be 16 ohms hot, 275 cold.

If you are checking the voltage at the connector it should be battery voltage.

 

The way this sensor is wired is to a fuse and then there is a 100 ohm resistor inside the instrument cluster then to the sensor and then ground.

So if you have the engine cold and you read the voltage drop across the sensor it should be 8.8 volts if your battery voltage is 12.

12*275/375=8.8

If its hot it should be

12*16/116=1.65

Thats off, running should be:

Cold

14*275/375=10.2

14*16/116=1.93

 

Sorry looked up 98 2.5 not a 97 2.3

Hot: 9.7 ohms
Cold: 74 ohms

Same internal resistance of 100 ohms

 

Take these measurements for me:

Start car and read battery voltage.
Take voltage across sensor when running and cold.
Take voltage across sensor after running it for a while.

And for ***** and giggles take the voltage from the ground wire on the sensor to body ground.

If there is a voltage on that last measurement you could have a bad ground.

 

Also measure the resistances prior to taking each reading for cold and normal.

Disconnect sensor to measure the resistance, dont measure while connected.

 

So did some research on that part number and its saying its compatible with the 1997 2.3 and the 1998-2001 2.5 even though alldata/ford is showing that the resistances for hot and cold are different.

 

What are those voltage specs using the revised resistance specs you found (74 ohms/9.7 ohms)? I would do the math myself but I don't know what the "116" and "375" values represent.

I replaced the BWD piece with a Motorcraft from the dealer (before I read your replies) and it didn't do sh*t haha. It barely moves from the "cold" position, if at all.

Thanks for all the info! I just took some (very strange) measurements with the engine hot; I'll have to wait 'til morning to take the cold readings (it is almost freezing cold out btw).

1. Resistance across sender (hot): 270 ohms. Not normal; I disconnected the connector from the sender and measured across the sender terminals. The metal fitting it screws into is hot, and it (the fitting) has a good ground as well (I checked). Why such a high reading? This is a brand new OEM sender.

2. Voltage across sender (hot): a steady 8.9V, engine running. Not normal.

3. Voltage at the connector is 10.9V, engine running. Not normal.

4. Running battery voltage: 14.7V. Normal.

5. Sender ground wire to engine ground voltage: .05V. Normal.

These are pretty much the same readings I got with the BWD sender. So basically...the sender's resistance and voltage is WAY too high, and the supply voltage is too low. I don't know what to make of all this. Thanks for all your help so far.

UPDATE: Before I installed the Motorcraft sender, I put both it and the BWD sender in a can of warm water (same can at the same time) and measured their resistance; they were both like 1500 ohms or something. I just tested the resistance of the BWD sender, which is apparently good, in a pot of water on the stove; ~2500 ohms cold, ~70 ohms @ boiling. WTF?

You know, maybe the temp gauge is bad. Just because it pegs when it gets 12 volts doesn't mean it's going to work right when it gets lower voltages from the sender. If that's the case I'll probably just leave it be. But I'm still interested in what you make of all this.

 

Let me sit down and look at those numbers later tonight after work.

I'll try to grab some data from my truck as well. To see how it compares.

In the mean time can you give me your ford factory part number? I'm gonna make sure its the same as mine.

 

So my batter voltage when running was 13.96, yours was 14.7 (both normal). So this is a 14.7/13.96=1.05x difference.

The voltage at the sender (not connected) was 10.42. 10.42x1.05= 10.94 same as yours.

This means your circuitry is correct. There is a 100 ohms resistor but I forgot to take into that the gauge is ran in parallel with the sensor so battery voltage wont be what you see at the sensor.

When the truck was running and hot I got these numbers:

93.2 normal temp
5.9 volts across

To make sure your gauge is correct I would grab a 100 ohm resistor and jump the connector with it to see if the gauge shows in the middle if it does either your sensor is bad or you are not getting contact with the coolant. I believe it is the latter. You said 2500 ohms for cold, could you have set your ohm meter to the wrong range and it was actually reading 250 ohms (which would be right) the 70 for boiling water sounds right.

For boiling water your temp would be 212 F or above. Our thermostats are 192 F. So mine being 93 ohms your would be past this 93 ohms=192 F range and would have a lower resistance of 70 ohms.

How is your heat? I thinking you have low coolant. These sensors are located at a branch that starts before the thermostat and flows by the sensors to your heater core (or bypasses it when OFF or MAX A/C).

If there is not enough coolant it wont flow coolant through these hoses (which are the highest points in the system) and the coolant will stagnate around these sensors and not show hot temps.

I would pull a heater hose from the core and fill it up there.

 

I don't blame you for thinking I'm reading the ohms wrong, but my DVOM only has one setting on the dial for resistance, and it auto-ranges. The number it displayed was ~"2500", aka 2.5 k-ohms. The resistance reading stabilized at ~70 ohms when I put it in boiling water. I redid the test to verify these numbers, and got the same results. And I have pics! The 3.35 k-ohm reading is @ room temp., the 70.4 ohm reading is @ boiling.

I also just took some readings with the engine stone cold (its about 55 degrees out): the sender's resistance was 5.7 k-ohms (yes, k-ohms), and the voltage was 10.7 with the engine running.

I checked my coolant at the radiator and it's good, burped it and everything. Should be no air in the system. That metal fitting the sender screws into gets freakin' HOT, and I get good heat from the vents. It has a brand new heater control valve, btw.

The PN for the Motorcraft sender is SW-5174, alt. PN is f7dz-10884-aa. Those PNs are straight off the box...I didn't toss it and use Google later to get the PNs.

I'm going to try your test with the 100-ohm resistor. I thought I had found an explanation for these "abormally" high resistance readings: I checked the compatibility of these senders on the O'Reilly's web site, and it seems that from '95 to '97, some 2.3's used the 2-wire sender mine has, and others used a different 1-wire sender used on earlier models (Motorcraft PN SW925). My 5174 is found on later models, with all engines. So I thought, it could be that the specs you found are for the early SW-925, since both senders were used with the same engine, in the same year. Then I saw that you have the same 2-wire sender as me and you're getting the "correct" readings. So much for my theory.

 

I'll see what the part number is on my sending unit if I can read it.

Yeah I was looking at that O'reillys compatibility too.

I figured you had the same two wire version as me because the old 1 wire style sensors were mounted on the rear right of the block (only needed 1 wire and used block as ground) and you wouldn't be near that during a timing belt change.

The two wires were introduced because the thermostat housing is plastic, thus no grounding is possible through the housing itself.

 

I did the "resistor" test, but I used a light bulb instead, which measured about 30 ohms. Lo and behold, the temp gauge needle jumped right into the middle. I guess my gauge is good.

I have a metal therm. housing. I would say it has 2 wires because the fitting is connected to a rubber heater hose, completely insulating the fitting from ground.

I got on justask.com, and this "expert" told me the old style 1-wire should work on mine, if I wanted to go through the trouble of jerry-rigging the wiring. I suppose I would have to solder the pos. wire to the terminal and solder the ground wire to the side of the sender. I'm not doing that.

He said both senders have the same resistance specs. As for the abnormally high resistance on my new senders, he posited a "mix-up" on the manufacturing side. He may be onto something; he suggested I grab one from a salvage yard. It looks like that's what I'll be doing, since it seems like the units being manufactured now are defective, both OEM and aftermarket. Maybe I should let Ford know about this somehow, because I have been reading a lot of stuff about people having "temp gauge" issues.

What do you think?

P.S. Not that it matters, but I digged deeper, and turns out the two senders were never used in the same year. The 1-wire was used until '96, and then they switched to the 2-wire in 1997.

 

Guess what...on my way to the salvage yard, the damn thing started working. And it worked on the way back. WTF? It maybe read a little low, and it took a while to get there, but it was pretty much normal. Also, I pulled 4 SW-5174's from the salvage yard, measured their resistance, and they all had like 5 or 6 THOUSAND ohms. Double WTF. Maybe if these senders sit cold for a long time, their resistance goes way up (say, in a salvage yard, or on a parts shelf), and they have to go through several hot/cold cycles and get "broken in".

I measured my sender's volts when I got back, and it was about 5 volts, right around where it should be. The resistance was about 130 ohms, when it should be about 10, but my guess is that the resistance will continue to drop as it's driven until it's within specs.

Jeez.

 

Maybe you had a bubble in the system?