Well today the entire radiator was replaced, and the fan clutch. This was after replacing the thermostats last week.

The gauge still reads at the top of the "N".

The gauge is lying. IR measurements at the radiator are 129 to 175.

Today Ron was kind enough to forward me this link, to an article written by someone with the background of an engineer who was clearly as fed up, or moreso, than I am. Here it is: http://www.efsowell.us/ed/jag/coolin...atureGauge.pdf

I'll cut to the chase. I connected a 50 ohm resistor in place of the sender. This is supposed to split the N dead center in a properly working gauge system.

On mine, depending on whether the car is running or not, it reads either 1/16" above the N (ignition on, motor off) or halfway between N and H (motor running).

Clearly, according to the linked article, this means one of two things:

1) The gauge is improperly grounded, in a major way
2) The gauge itself is faulty

The other clues are that:

1) If you shut the car off, then turn the ignition to on, the gauge comes up about 1/8" cooler than the instant before, when the car is running

2) If you turn the climate blower on, the temp gauge reads a little higher (1/16") then goes back down again when you switch the climate blower off.

This problem started happening virtually overnight. That would seem to rule out drifting resistors in the gauge (there are two) and the gauge itself going haywire.

But it's maddening as all get-out, having now spent over a thousand dollars chasing this fault, worried about the cooling system which is now literally all brand new, only to have the $%&@! gauge still read "edging towards hot".

At least I now know for certain that the car ISN'T overheating. Now I know that "N" is really in between N and H.

What do you think about taking a ground wire directly off the screw terminal of the temp gauge and running it to a ground point on the chassis behind the cluster?

Lastly, does anyone know WHERE the ground lead (that comes out of the instrument cluster connector) actually is grounded in the car? That point might be corroded....

 

After reading your article I'm tempted to spring for the remote temp. sensor tool I've seen in my auto supply store. At $100 plus dollors it's a bargan.
KBO mate.

 

I've learned more in the past three weeks about the XJS cooling system and temperature gauge system than in the previous 17 years I've owned the car.

Regarding the above-referenced article by the scientist who was obviously equally frustrated, I proved that one of his postulates was dead wrong.

He asserts that each temperature gauge has a 50 ohm and 220 ohm resistor inside.

My 1984 gauge, manufactured week 39 of 1983, has a 330 ohm instead of the 220.

My 1989 gauge, manufactured week 13 of 1989, has a 430 ohm instead of the 220.

Both have the 50 ohm as stated.

How do I know this? I took them apart, and looked. (Yes; I can read color code.)

What does this mean? Well, it means there are at minimum three versions of the temperature gauge (220, 330, 430).

There could have been some mandate from Jaguar management regarding either a large number of "false hot" warranty issues, or maybe "false cool" readings resulting in engine frying (also a warranty issue perhaps). These changes in the gauge which is in essence an ohmmeter, must've been running production changes, which is quite interesting.

Or maybe the sending units were found to have a greater degree of variation and the resistance changes were meant to compensate.

Anyhow, I think I've got my Indiana Jones sleuthing degree of Jaguar production changes for the temperature readout system, at a cost of three weeks and many Franklins. On the upside, I have a brand new cooling system and a gauge (finished up using the 1989 one) that reads about 3/16" below the "N" during normal operation on a cool February day. The gauge is also calibrated to read "H" at an engine sensor input of 15 ohms, which is in line with a true overheat, and read the center of the "N" at a temperature you would expect to see on a summer day with the A/C on (50 ohms).

Having the cluster out again during the course of the troubleshooting, I took Doug's suggestion and added an auxiliary ground lead from the bar at the cluster top, to the steering column mount bracket behind the gauges. This didn't do anything, but at least I now know the cluster is grounded at least twice.

Thus ends the tale of the hot-reading gauge. New OEM Jag sender, replacement 1989 gauge, appropriate 15 ohm offset resistor, new radiator, new viscous fan clutch, new thermostats, new hoses, drained account, satisfied synapses.

 

And a satisfied reader of this informative saga.

All the best,

 

Lou,

Don’t know why you say I assert all XJ-S gauges have the same R2 value. All I say is my spare gauge had that value. Moreover, I don’t use that value anywhere in my write-up. All I do in the Appendix is derive the form of the equation, which would be the same regardless of the value of the resistances.

Accepting what you report, i.e., the gauges have different values of R2, I would assume that the gauge maker selected a resistor in that position that would calibrate a particular gauge to meet some specification. Perhaps this is necessary to compensate for minor variation in the coils. I’m led to this conclusion because the senders part numbers are the same (to my knowledge) for all the XJ-S 5.3 L engines.


Ed Sowell
1976 XJ-S
www.efsowell.us/myjag

 

Ed,

It is great to hear from you and to be in contact, because I had no way of reaching you until now. The fine and meticulous article you authored did not indicate any contact details, so I'm appreciative of your post.

I have nothing but gratitude for your serious study of the gauge and sender in the XJS. I relied upon it heavily for my work, and thank you very much for putting it out on the web where Ron put up a link to it.

In hindsight I shouldn't have said your "assertion was dead wrong" so I do apologize. Regarding your "Appendix Gauge Circuit Analysis" section, R2 is clearly called out at 220 ohms and on the gauge, but neither of my gauges have that value. Since every other resistance value in your paper is pretty much spot-on, I thought it would increase the body of knowledge on this topic to point out that we don't really know what the heck R2 is for a given gauge unless we look at it. The value of R2 is probably inter-related to the value of RC1 and adjustments were made accordingly at time of gauge manufacture. However how do we know what they were thinking? I guess it's open to speculation.

Using your table and in an effort to have my XJS read pretty close to all other "normal" ones out there, I found the following:

-The gauge from the 1984 as supplied with the car was hopeless. Even removed from the cluster, using a B+ voltage of 12.73 volts, when 50 ohms was applied as R_S, it would read halfway between N and H, about +4 on your NW (needle width) scale.

-The spare gauge from the 1989 (which I finished up using in the 1984 car) was much closer. Under the same conditions on the kitchen table, it showed about +1 at 50 ohms.

-I found Joe Bialy's statement that "voltage between 12 and 14...provided almost imperceptible changes in needle movement" to not be exactly true on my car. With the gauge in the car, switching ignition ON might yield say 2 NW less than with the engine running and charging system working. This was true for both the original 1984 and newer 1989 gauge, and also irrespective of the additional ground lead I wired in at the time the gauge change was done.

-Now, as to setting the 1989 car gauge to read at the center of the "N" (0 position) at a 50 ohm input. Even with the '89 gauge, a 50 ohm input resulted in a +1 position with ignition ON (not bad) but a +3 NW position with engine running. Since engine running is really what matters, I elected to find a sender offset resistance (which I shall call R_SO) that could be added in series to the sender to get the needle position inline with the center of the N.

After some trial and error, I found that an R_SO of 15 ohms allowed for the needle to indicate -1 at ignition ON, and +1 with engine running. Since this effectively "split the difference" between ON and RUN conditions, I selected it as the best offset value. (I should also note, the sending unit is a brand new Jaguar OEM unit, that might be a little different than the original with the car.)

-Because the resistance vs. temperature is not exactly linear, I tested the 15 ohm R_SO with a R_S value to simulate a HOT condition, according to your "Gauge Position vs. Resistance" table. I used values of 10 and 15 ohms, respectively. Here are the results:

R_SO 15 ohms + R_S 10 ohms = +10 NW
R_SO 15 ohms + R_S 15 ohms = +9 NW (just above underlining the H)

I did this to ensure the gauge was actually telling me useful readings, and not just fooling me into thinking all is well.

-Before the addition of any R_SO, the 1989 gauge and 2010 sender were underlining the N at normal operating temperatures. With R_SO added of 15 ohms, the needle now resides comfortably at around -4 NW at operating temperature, which was verified by IR thermometer as being about 190F at the sender.

-I also measured the resistance of the sender at five minutes post-operating temperature and found it to be 78 ohms. This seems in line with your own data on the senders.

-One final anomaly that I have no explanation for. When the climate control is turned on (which activates both the electric radiator cooling fan, and the in-car blower) the needle goes UP just slightly, maybe 1/2 NW. This confuses me because the voltmeter shows lowered voltage and I would "expect" that to make the needle move lower and not higher. Switching the climate control off results in a lowering of the needle back to where it was, after the normal several seconds of damping in the "down" direction.

So, at the end of this happy saga (which actually was fun in a perverse sort of way) I am running with the 15 ohms in series with the sender, and all seems right with the world. If you would PM me your email address, I would be happy to correspond with you directly should any other matter of mutual interest present itself.

Any other comments on my procedure and results greatly appreciated.

Kind regards and many thanks,

Lou

 

My theory is the individual gauges had to be trimmed to get it at N at some specified sender resistance.

Now that I think about it, I have a vague recollection of the gauge I'm running with had a different resistance (based on color code) than my spare. Next time I have the cluster out (hopefully not soon!) I'll take a look.

If you still have this gauge it might be interesting to snip the R2 resistor at one end and substitute a trim pot to test my theory.


When I was doing my testing I noticed that the gauge is kind of lazy

I'll do the test you describe next time I drive my car... probably Sunday or Monday

As I mention in my write-up at some point Jaguar specified a sender with an auxiliary inline resistor. Perhaps that was their first effort to calibrate sender to gauge, and they later decided to do it by calibrating the gauges as I've described. I'm speculating here, obviously.

Keep in mind that everything electrical in the car is, in a sense, one huge circuit. Even the ground path is in reality not necessarily at the same potential everwhere due to resistances at connections and various ground straps. Thus a change in current draw somewhere can have unexpected effects far away. This is a hand-waving kind of argument, but I believe it has a kernel of truth. Ask Toyota!

If you go to my website there is a "Send me email" hyperlink at the top.

Good talking to you, Lou.

 

Thank you, Ed! Likewise it's good to be in contact with you also. Your website is amazing!! There are so many links to so many useful things...I haven't scratched the surface yet! I'm VERY appreciative that you have posted all the write-ups and your experience and history is hard to imagine, having had the car since 1982. I've had mine since 1992 and thought that was a long time.

I don't remember seeing that in the write-up, I'll have to go back and re-read it. I would like to know what value of resistor they specified.

OK, I'm going back to your website now.

Cheers!

Lou