94 YJ 4.0 overheating

I will not attempt to guess history of car nor full evidence of symptoms not told.
this is the answer on all cars. water cooled, for sure jeeps

pasted : (my personal logs)

all causes of Jeep overheating and tests. (my personal journal)

first are signs:
Evidence/symptoms !??????
steam blasting from rad cap or recovery side tank.
dash guage, reads too high or is erratic.. (and can lie to you,easy see why below)
ECT bad or air in the loop it reads wrong,fix air in loop first.
The scan tool shows. erratic ECT sensor. or DTC for over heating or ECT errors.
engine power drops.
engine pings/knocks/detonation, (under load)(too hot)
Coolant boils at 265F
the coolant must never be just water, it must be 50% AF, antifreeze or it will boil.
engine stalls. (some modern cars shut down too hot via PCM controls smarter)
Only over heats at stop signs, never moving fast.(or both)
Only overheats moving fast or only pulling a huge boat up hills.?
Only overheats A/c cab HVAC turned on.?

leaks to ground. Antifreeze 50%AF leaking.seen hot engine.
leaks to front passenger toes, foot well,heater core/hoses leaking
leaks to engine oil pan, it dipsticked looks like chocolate milk.
leaks to A/T, ATF oil pan. (dip-sticked just like above) skip for manual trans (m/t)
leaks , if it leaks coolant, this is repaired first.


First are simple and cheap checks.
air in the coolant loop, not burped the correct way. makes gauge go crazy with lies.
The RAD always shows (eyes) full, dead cold cap off, does it? and no oil drops seen here.
No water in the oil pan (dip stick looks like chocolate milk shake ,bad) bad head gasket.
no steam under hood or at tail pipe end.
water pump witness hole drips coolant, the pump is bad.
on min last month the pump gasket was bad. 2008 jeep, jK. where you can not see it,directly. oops

fixes and tests :
RADiator packed in dead bugs.
RAD cap bad. $5 fix.
fan belt loose, 100 pounds up on serp belt not less. jeep is 140 lbs min. (5 bolts on PS pump)
Thermostat bad, at 7 years old (normal life span) $20 cure, and loves to fail on old cars. does.
running 30 year old coolant still?

A bad Radiator., (leaks? none, the RAD tubes must be not clogged deep inside it, IR gun tool finds this super easy.
top fails for old cars for sure 30 years old.
bad hoses leak.(cheap fix to)
water pump failures and its gaskets too.
thermostat installed backwards or its bleed hole not upward side. (engine tilts front to rear)
Rad fan dead or its relay bad. The clutch fan MUST not slip at idle.
Idle RPM is 800 or more, never 500 RPM.

Dead cold engine (safe), RAD cap twisted off.
You see coolant missing at RAD filler NECK,or not full here is wrong.
It must be full ever time (each morning)you looked
RAD side overflow tank , at 1/2 full each cold engne(morning?) test.
if was low that means there was air in the loop now.
This air pocket makes most dash gauges read wrong and the thermostat fail even a new thermostat fails in air pockets.
The ECT tests must show 180f running, hot. 225F(par say driving up hill) if electric fan, the RAD fan kicks on. near 225F or A/c turned on.
The AF boils at 265F.
The ECT does not work in air pockets,only in Collant AF.
My page covers 3 decades, from clutch fans to dual fan cars at RAT.

Test#1 for leaks,etc.
A DEAD Cold engine.
RAD cap off neck to RAD is full of 50% AF mix.
cap back on.
Start engine hood up. and let idle for 15 to 30 minutes.
The dash gauge goes mid scale, normal like it did for ages.
The fully hot (normal) engine will now produce 15PSI, full pressure and when this happens (good RAD CAP) all leaks will now happen under this natural large pressure. For sure to the ground.

Go slow and look for leaks. to ground, at the pump tell-tale weep hole leaks, any leaks from any hose.
leaks to front passenger toes.(carpet wet)
the main RAD leaks , seen.
leaks can drop/dribble or squirt. (even squirt to things fully 100% good)!!!!
If coolant or steam blast to the overfill bottle tank, stop the engine.

This page is for discovery and evidence.
At this point you can get advice on what to do next.(report all evidence first)Good and bad.
Best is start with a new thermostat in most cases and new RAD cap.
And 50%AF, fresh.

A good tool to own is a laser IR thermal gun. ($15) even Harbor Fright (sic) has them.
And use it to find thermostat housing too hot over 225f is.
Use this IR thermal gun to check all small tubes in the Radiator. (all ends, checked, hot and cold side. A super easy test, at 30 minutes idling, fully hot testing. OMG the IR gun finds out my dash gauge is a liar . (lies like a gas meter my grandpa said) (and did)
End to end RAD tubes checked. if one is cold, that tube is clogged.(hot engine tests)
A good tube scanned, is hot on one end and grows colder as you scan the tupe and is 35-50f degrees
cooler at the other end (cold side). This is a hard test, it is child's play and no burned hands.

My JK jeep ,08, the pump gasket rotted out, where we can not see it.(flange) the evidence was
seeing green Antifreeze dripping from 6sp tranny bell. ( lies to me,, but expected)
the windage factors did that + gravity !
best test is to make it leak parked, and find exact source hot and parked.

I have cold engine RAD pressure tester that found mine. At 20psi,on tools gauge. saw the pump leak.
this is not my first Rodeo. 1963 to now, direct hands on.

Jargon (ASE)
RAD = Radiator main, front of car.
ECT = Engline coolant sensor.
DTC = Scanned errors, Diagnostic trouble codes. OBD2 scan tools.(hey man I got codes they say now)
A/T = automatic transmission 50%AF
AF = Antifreeze COOLANT 50% mix , and is antiboil too.
HVAC,heating, vent and air condidtioning.
Last edited:
now focused answer, (above is for others wanting a list of all causes and all tests)
now you posted IR gun data (it's all good)
and now you learned the 30 year old dash gauge lies to you,(not surprise on any car this old)
it or the wrong 1 wire gauge sensor is bad. the FSM covers the tests to prove both.
almost all cars FSM cover how to test both'

wrong sensor (1wire)
sensor brass shell; not grounded to the block (metal to metal ,no caulk)
also, like my heep, , the fire wall ground lug to battery NEG lug, was not even connected,
the engine side stud ground must be fully tight.( my 80 jeep had factory flaw back in 1988, engine stalls only left up hill turns) the factory ring lug crimp was bad day1 new car from jeep factory USA right rear block ground.(facing it , it is left side)
so a voltmeter from battery neg top cluster ground pin over 0.5vdc is bad grounds top fix.

See the IR gun saved your engine. (using one) cool no?

The gauge is real simple and crude. just wire/sensor and gauge.
one of the 3 are bad.

the ECT sensor tests if done proves the cluster gauge sensor lies. this too proved it. if done,.
ECT is what the PCM uses for all temp measurements, and I posted its performance chart above.
on engine swaps vast cases the grounds are never put back right. fixed many I have.

For sure you want the PCM to work 100% and grounds and power feeds are first .
there are 2 sensors on the car, if one reads wrong?
did you ignore that.?
ECT and gauge 1 wire sensor. 2 sensors,
both are covered in the FSM.
too hot:
key OFF
just unplug the ECT connector with dash gauge reading wrong (160f)
measure the 2 pins on ECT and see. (not the IAT leave it alone)
1000 ohms is normal hot
400 ohms is not normal ECt that is way way to hot. I bet it is not 400.
all this is in the FSM under coolant sensor tests.
The ECT can fail too this old. or worse using the wrong sensors. for year ,etc.
That is what the IR gun finds, 2 or more fibbing sensors.
and the ohm meter too.

on 1996 up jeeps the scan tool would have told you ECT reads right. 180f, idle if 180f stat is it?
but the gauge on dash is fibbling.
chapter 8e

"(1) Turn ignition switch to RUN.
(2) Disconnect Coolant Temperature Sender connector. (a 1 pin sensor it is)
Needle goes to C. If not, go to step 3.
(3) Touch Coolant Temperature Sender connector
to ground. Needle goes to H. If OK, replace sender. If
not, repair open to gauge.
this is a classic test done for decades even back to first guages made) open short tests.
chapter 8E
open show cold (C) needle hard left
shorted shows 260f (needle pegged)
the acid test is 100 OHM resistor in hand (brown /black/brown) colors. a 3 cent resistor, in hand
prove gauge is good or bad.(or wiring to it bad)
100ohms is center of dash meter scale.


about 50 ohms is engine overheated max. 7 % max error.
Last edited:
  • Like
Reactions: HRDROKN
just got back, real data from SMP. marker of this sensors ECT, yours.

tx81 due to jeep not telling it, in near jeeps 97up 1997-2006
I asked tem this to see if the old jeeps 87 to 96 are the same ohms as 97 and is.
TX43 (yours) to TX81 (1997)

TX81 has the following resistance properties: this proves only the connector shape changed, not the actual thermister inside.

49.0kΩ - 61.0kΩ at 14°F (-10°C)

29.3kΩ - 35.9kΩ at 32°F (0°C)

17.9kΩ - 21.8kΩ at 50°F (10°C)

11.3kΩ - 13.6kΩ at 68°F (20°C)

9.1kΩ - 10.8kΩ at 77°F (25°C)

7.3kΩ - 8.7kΩ at 86°F (30°C)

4.9kΩ - 5.7kΩ at 104°F (40°C)

3.3kΩ - 3.8kΩ at 122°F (50°C)

1.6kΩ - 1.8kΩ at 158°F (70°C)

1.1kΩ - 1.3kΩ at 176°F (80°C)<<<bingo same.

640Ω - 720Ω at 212°F (100°C)
Last edited:
the same maker SMP makes the 1 prong gauge sensor for this car ,the maker shows 92 to 97 usage,wrangler./ 97 is huge OBD2 changes.
actual SMP P/N.
i just asked them for data sheet on this sensor.
most these sensors are at best 20% accurate new. some are worse. we will now learn that from the horses mouth. (makers)
and temp gun it
My 94 YJ had very much the same issue....

I had upgraded (or maybe actually a down grade) to a flowcool water pump and IIRC a colder hi-flow t-stat (not sure what temp), but anyway... took out for test drive and temp gauge started going way high. Stopped let it cool down, etc. Did all the tricks except drilling out t-stat larger (it already had a hole. Messed with it for two days! I was ready to put all stock parts back on and while sitting in my angled driveway, nose up, hood up, pondering what the f..... I heard a big gurgle from the motor. Problem gone! I'm convinced that even with my best attempts to make sure she was air free, I was failing. Hasn't been a problem since.

I want to change my thermostat back to stock'ish as I periodically throw a code 17, but a bit apprehensive to go thru that whole mess again!

17 is too cold as it warms up. we never drill thermostats ever.(its the stuff of kids on youtube doing that )_
we all know 4L are hard to burp out all air. I burp all mine a the heater hoses. (1 pulled)
  1. fill , it full at rad and
  2. burp it ,heater hose off ,cold and full. ,hose back.
  3. then run it for 10mins until rad cap spews and key off and put cap back to the rad.
  4. then let it full cool, with side tank 1/2 full so it can suck back coolant fresh to the RAD>
  5. the repeat 4 & 5 cold , cap off is it full, if not add more the the cap neck hole.
repeat until cap off cold shows cap neck full. then cap back.
then drive it and check again in the morning before first run.
now it is stable.
the trick is heater hose fills. gets out the most air. ( why not start your own post this OP post is about bad gauge)
do not run colder thermostats. or drill them ever. (bleed hole up hill and not put in backwards)( bleed = jiggle pin on some)
use OE stats only, 180 to 195, as the FSM told you, (FREE)
do not run fail safe stat.. why?
if it latches open ,the EFI goes max rich and burns up the CAT to H3LL.
use only OE rated parts and win.