MG Sports Cars

My apologies if there's an excellent library article on this, I couldn't find it.

I've retained the original Smiths electric water temperature gauge in my B. But it reads low. My engine operates at around 180-190F and the needle moves off C, but not much. I suspect the old engine ran hotter. I've tried a couple of different temperature sensors, neither of which are the MG part. I'd prefer to have the needle in the N range under normal operation for obvious reasons.

Can these sensors be recalibrated? I have one from a 1972 and one from a 1969 in case there's a difference.

I have NOT tried grounding the wire to ensure it goes to wide open yet due to how the wiring was set up when I first started looking into this, but I've tried two gauges and they both act the same.

Hi Keith,

The guages can be recalibrated but this is probably not required.

Your going to need and ohm meter and a test light to check out the system.

Start with the sender. At 190'f you should have a resistance of 31 to 35 ohms between the sender and ground.
Next is the guage. It should have a resistance of 58 ohms across the terminals.
Both of these need to be measured disconnected from the vehicle. In the case of the sender it's best to measure at the guage wire so that you can confirm the circuit integrity.
Lastly is the voltage stabilizer. This is the most likely culprit. It supplies regulated power to the temp and fuel guages. If your fuel guage is out of range as well I would just change out the stabilizer and call it repaired.
The stabilizer supplies an averaged 9.0v to the guages. It does this by pulsing the power on and of at a fairly slow rate. This makes measuring it difficult. With everything connected and turned on you can use a test light to get a rough idea of it's operation.
A test light connected between ground and the power side of either guage should pulse around once every 1/2 second.

That should narrow it down. If not let me know.

Cheers
Fred

Per Rick Astley's book on MGB electrics, he gives a sender resistance check at room temperature (68*F), which may be an easier measuring/reference point.
For 1968 - 1974 cars, typical sender resistance (@ 68*F) is 800-880 ohms
For later rubber bumper cars, typical sender resistance is 930-1000 ohms.
It is Rick's observation (and mine) that there is a wide variation in sender and gauge performance (rogue devices), so above numbers are median values rather than average values.
It should be noted that the water temp gauge and fuel gauge both operate through the voltage stabilizer and have identical inner workings and only the dial face is unique - gives you another possibility to check gauge performance.

edit: It follows, that if the voltage stabilizer is the culprit, your fuel gauge performance will also be erratic.



Edited 1 time(s). Last edit at 01/31/2011 06:20PM by ex-tyke.

That will help me get started, thanks. I have noticed that the fuel gauge is possibly questionable, but there's some pre-existing damage to the tank near the sender that made me expect odd readings. That's an interesting way to regulate the voltage.

What's the normal operating temperature of the 1800 engine?

Quote:

What's the normal operating temperature of the 1800 engine?

Depends on the thermostat used.
eg. if a 180*F stat is used you should see the temp gauge at perhaps 190* - 192*F .

Keith,

If you're trying to get mismatched components to work together, you can add a resistor in the circuit to get the gauge to read mid-scale at normal operating temperatures. I don't know what the resistance should be for the MG temperature, but it's easy to find by experimenting.

In the circuit below, the total resistance is (R X S)/(R + S), where R=resister and S=equal resistance of the sender. If it were me, I'd start with a 100 ohm resistor, of at least 1/4 watt, preferably 1/2 watt and take it from there. If the reading is still too low, go to a 50 ohm. If it becomes too high, go up to a 220 ohm (or the nearest standard value).

The actual reading is not important, only a change from normal. If you adjust the resistor to give an "N" reading at operating temperature, you'll know something is wrong if it starts to rise (or drop).

I know the temperature is regulated by the thermostat, I just wasn't sure what sort of thermostat was usually found on the MG engine. I don't take anything for granted on this car! For all I know, the MG engineers decided the engine should usually run at 230F and that's where N is.

I used to have a variable resistor setup, that would have been useful for this. I'll start with the voltage stabilizer and see if it's doing what it's supposed to do first, and that's a good tip about the fuel gauge being the same part. My car was basically inoperative when I got it, so I'm debugging both a 39-year-old British car and all the modifications I've made to it at the same time!

Okay, I just tested the gauge by setting it to ground. It spiked to maximum, so that's working. The voltage stabilizer didn't flash the way I was hoping, so I swapped in the one from the parts car and saw the flashing. It could be that I didn't give the original one time to warm up, but regardless - I have a good one now.

However, the senders (both ones I tried, a GM unit from the engine and the other a random one from NAPA) both tested out as 3.8 kilohms at 63F. The GM one is about 150 ohms at 185F, the NAPA one is 200 at 185F. So there's the problem. I'll see if NAPA's Big Book Of Sensors gives resistance ranges. I don't have the original MG one anymore (it left with the engine) but I need something with a 1/8 NPT thread pitch.



Edited 1 time(s). Last edit at 01/31/2011 11:27PM by Keith.

Keith,

With the MGB gauges and sender, the 180-190 should read normal to slightly above normal. If your running a sensor from something else, thats a problem.

Go back to the MGB sensor with the smiths gauge. If it screws in to the block but bottoms out (as a smiths sender would in a NPT hole) then this might be what you need:

[bmcautos.com]

Not sure of what type sensor your LSx engine would use. We use these on V6 motors with stock gauges and sender with the NPT threads. Most likely it's the same.

-BMC.

Thanks for the info, Brian. What size NPT, 1/8? I don't have an MG sensor here, but I seem to recall it's more like a 1/2" size. The LSx engine has a hole with a 12x1.5 metric thread pitch, and I have an adapter that drops that down to 1/8 NPT. I can't make a 1/2 NPT work in that.

I did find one that measured out at 31 ohms at 195F, but when installed it had the gauge basically pinned at 185F. So I have my doubts about the accuracy of those hot numbers. What I'll probably do is just get a variable resistor and add in a parallel resistor as per Dan's suggestion using the GM sensor.

Quote:

..but when installed it had the gauge basically pinned at 185F. So I have my doubts about the accuracy of those hot numbers....

Keith, what was the ambient temperature when you conducted your test. If the ambient temp is low the stat might not fully open,
It's my observation with my own conversion, that when the ambient temp is 60F vs 80F, the temp gauge can reflect that 20F difference.

According to the engine computer, the water temperature was 185F (actually, it varied slightly between 180 and 185). That should be all that matters. Ambient temperature was around 45F.

I don't see how the ambient temperature should make any difference at all in that regard unless your coolant temp is also varying with the outside temps - which might be possible with an undercooled engine or a malfunctioning thermostat. I have neither, and I have the ability to query the engine computer for current water temps in real time.

Update: a sensor that I measured at 48 ohms @195F reads halfway between N and H with coolant temp at 185. I just stuck that one in because I had it, but there's another data point for you. I'm guessing it should be somewhere around 60-70 ohms to read N.