Fuel Gauge Troubleshooting

[Rob Mayercik]

Having some fuel gauge issues with my '92, would like to get some suggestions:

A number of years back I had to have the fuel tank and pump/sender assembly replaced (top of tank and the sender itself were both rotted).

The fuel gauge read fine prior to the replacement, but it never read right with the new sender - it'd read half a tank but I'd only get 5 gallons in. The low fuel light would come on, and I could only get 10 gallons in. The original tank was 20 gallons, and I'm pretty sure I recall putting nearly that in after leaving the shop and going to the nearest gas station (the shop threw a gallon in so I could get it to a station, as the fuel in the old tank looked more like a milkshake than gasoline and couldn't be saved).

Since it was reading lower than what I perceived to be actual (rather than higher), I decided to live with it rather than drain the tank and mess with it. I rarely let it go below a half tank (actual) anyway, so between the "early warning" from the gauge and having a good feel for my MPG/range, I didn't feel all that concerned about it.

A couple years ago I swapped gauge clusters because of a failed trip odometer. The gauge in the replacement cluster read the sender the same way as the original, up until recently...

Last month I noticed the fuel gauge seemed to be stuck at 3/4, as it didn't come back up after I topped up at the start of a long run (though it did rise maybe half a needle's width after filling up and would wiggle a little bit going up/down hills). It seemed to be working a couple weeks before when I was out for a weekend of trail runs, though.

I have done Cruiser54's Fuel Pump Ground upgrade already as a first step. First key-on after doing that, the gauge climbed up to just off full. It should probably be a little under 3/4, though, since I the last time I drove it I topped it off and drove it about 110 miles, nearly all highway.

I'd appreciate some ideas at this point - should I bother rigging a variable resistor and testing from the body connector back up to the gauge itself, or focus on the sender assembly instead? Has anyone seen/heard of the gauge itself failing (I should note that the tach in the cluster I have in the truck now failed shortly after installing - it would read inaccurately or drop out altogether. I replaced the tach with the one from my original cluster, and that one works fine).


Rob

 

[RCP Phx]

My focus would be.......1- fuel sender 2- wiring 3- gauge

 

[cruiser54]

0 ohms is empty
44 ohms is 1/2
88 ohms is full

I wish I could post photos directly to this site without having to remotely store my stuff. I have an incredible library of stuff which I can post to all the other Jeep sites.....

 

[Rob Mayercik]

Thanks for the input, guys. I'll see about looking into the sender a bit more at my first opportunity.

 

[kastein]

0 ohms is empty
44 ohms is 1/2
88 ohms is full

I wish I could post photos directly to this site without having to remotely store my stuff. I have an incredible library of stuff which I can post to all the other Jeep sites.....

This is correct for 87-90 but for 91-95 it is different. 105 is empty, 5 is full, as I recall.

I've actually begun putting together a massive year-over-year XJ/MJ specific tech (parts crossover and other useful info) website but it is *not* ready for primetime yet.

Like the others in this thread I would suspect sending unit long before anything else. They fail the most. Try a 100 ohm resistor at the sender plug, if that reads (roughly) empty, try shorting the sender pin to ground, that should read slightly over full. Work from there.

Bear in mind that depending on the gauge design (I haven't checked current draw through the sending unit on these...) the resistor may get pretty hot.

 

[fenns7]

This is correct for 87-90 but for 91-95 it is different. 105 is empty, 5 is full, as I recall.

I've actually begun putting together a massive year-over-year XJ/MJ specific tech (parts crossover and other useful info) website but it is *not* ready for primetime yet.

Like the others in this thread I would suspect sending unit long before anything else. They fail the most. Try a 100 ohm resistor at the sender plug, if that reads (roughly) empty, try shorting the sender pin to ground, that should read slightly over full. Work from there.

Bear in mind that depending on the gauge design (I haven't checked current draw through the sending unit on these...) the resistor may get pretty hot.

If it might help Kastien I can probably dig out my 1996 specific fuel level sender readings from a few years back when I discovered that the 1996 only XJ's have the entire fuel pump/pressure valve/sender as a single module in the tank and the aftermarket sender used in the (rubbish & very expensive) Crown replacement reads completely wrong. My solution was to swap back the sending unit from the old module thereby restoring correct fuel readings at the gauge .... which still works after 28 years. I believe you could buy the sender as a seperate unit and missed one a few years back on ebay - went for silly money $200+ :laugh:

 

[kastein]

That actually might save me a 15 mile drive to my storage unit to get my 96 XJ hardcopy FSM out. I have the data for 87-90, 91-95, and 97-01 units already, I'm still missing 84-86 (guessing same as 87-90) and 96-only. The 96-only unit I suspect has the same resistance as 91-95 for the gauge and just adds a second resistive strip for the warning lamp sender based on some fuzzy recollections from trying to fix my 96s fuel gauge in 2009.

 

[fenns7]

So you have a copy of the 'holy grail' - a hard copy 96 FSM! I've waited 5 years for a UK version to come up on ebay or elsewhere - nada.


I'll do my best to dig them out later today - taken from multimeter readings at the time. I may have a photo of the sending unit strip too showing the p/n which confirmed it was year specific.



I do like having a 'unique' XJ but to be honest adding the RHD configuration it does make like rather too 'interesting' sometimes.

 

[Saudade]

For the OPs sending unit, Kastein has it right. This appears to be the values for '91-'96.

105 ohms empty
33 ohms half
5 ohms full

 

[Anak]

I believe you could buy the sender as a seperate unit and missed one a few years back on ebay - went for silly money $200+ :laugh:

Yowp!

I picked up one of those 10 years or so ago. Glad I did. It was much cheaper then.

My goal is to put the '96 sending "card" on a "97+ fuel pump module in a GenRight tank. I think a '96 is the only year of the earlier style XJs that may be able to run a GenRight tank.

 

[kastein]

Should be able to run it in a 91-95 as well if the only issue is returnless vs return fuel system - you'd just need to either add a return fitting to the tank and use an FPR bypass unit on the 97+ sender as well as swapping the sender card. You might be able to run a 96+ rail and injectors instead of doing that (and just swap the sender card) but I'm not sure if you'd need to tune the ECU to accommodate going from vacuum referenced to atmospheric referenced FPR and changing the injectors like that.

 

[fenns7]

Kastein,


Sorry for the delay but had to hunt thru' '000s of photos on my phone (hadn't started to index my numerous XJ related shots at that time).


The key ohm readings from the INCORRECT sender on the Crown branded "1996 Jeep Cherokee Fuel Module" were:


Full: 98.5
Mid: 555.1
Empty: 1.09


The readings from the CORRECT OE sender for the 1996 unit (and verified within +/-1% on my spare):


Full: 0.4
Mid: 28.2
Empty: 108.7


Hope that saves you a trip!

 

[kastein]

That lines up very well with the data expected - thanks. The incorrect one looks defective, 555 ohms is nowhere near midrange for the numbers given, and it looks like the rest of the numbers match 87-90 closer than 91-95, which is definitely wrong.

I actually was over at the storage unit the other day and moved 40 boxes out (long story, moving all my stuff to the shed to save money...) but didn't see any of my factory service manuals anywhere, so still not sure where that went. I'll find it one day I guess.

 

[kastein]

I finally found my 96 FSM, it was in essentially the first box I put in a 10x20 storage unit and then stacked my entire life on top of.

Confirmed: the 96 fuel gauge sending unit card contains TWO SENSORS.
Confirmed: the main sensor is exactly the same 105/33/5 ohm curve that 91-95 gauges expect. This is why 96 gauge assemblies work in 91-95 and vise versa.

New info: I was wrong on what the second sensor element is for. It's not for the low fuel warning lamp in the dash, that works off the main sensor just like it does on 91-95. It's for the ECU. In 1996 Jeep was required to comply with the modern OBD2 standard and as a result, they completely revamped the powertrain control system as many already are aware. Belt drive changed a bit, ECU is totally different, harness design is very different, fuel sending unit (thus, this thread), downstream O2 sensor added, power steering pump and its mounting to the intake manifold changed, electric cooling fan is different, fuel rail is now returnless design... the list goes on.

Anyways. Due to this they redesigned the ECU significantly, it's the JTEC (Jeep Truck Engine Controller) rather than the SBEC (Single Board Engine Controller) used 91-95. On 97 they planned to update to use the CCD bus to drive the entire gauge pack instead of individual signal wires for each, and all sensors that previously fed the gauges were either eliminated (if duplicate), or wired to the ECU so it could convert their data to a digital data packet and send it to the instrument panel. Well, 96 ECUs are 99% the same as 97 ones... so it already had a fuel sender input and they had their eye on detecting evaporative emissions system leaks in the future, which requires knowing how much fuel is in the tank. So they added a whole second sending unit strip to the sending unit, just to wire it to the ECU's fuel gauge input. This is connector C165 (later called "C3") pin C26.

I looked everywhere in the FSM but could find no documentation on the resistance curve used by this second strip. My guess is that it is the same 270-20 ohm curve that 97-01 XJ sending units use, since it is wired to exactly the same hardware input on the ECU, but I have no confirmation for that data... yet.

Edit: I was wrong. The second sender ranges from 100 (full) to 1100 (empty) ohms on 96. Interesting. Just found my original sender from my old 96 in a box of parts and measured its resistance at top and bottom of scale.

Fenns7, if you need any other info out of my 96 FSM hard copy, I will do what I can.

If anyone needs fuel level sensor data - not just for XJs, I'm trying to "collect the whole set" for the entire automotive community - I decided this subject deserved its own webpage and put it here. http://w1kas.net/jeepyeartech/fuelsenders.html
If you have data to contribute to this page by all means send it to me, I prefer primary sources such as factory service manuals but will accept measurements made by other forum members as well where necessary.

 

[Anak]

Great information.

Thank you.

 

[kastein]

I was wrong, the second sending unit on 96 is 100 to 1100 ohms. 1100 is empty.

 

[fenns7]

I finally found my 96 FSM, it was in essentially the first box I put in a 10x20 storage unit and then stacked my entire life on top of.

Confirmed: the 96 fuel gauge sending unit card contains TWO SENSORS.
Confirmed: the main sensor is exactly the same 105/33/5 ohm curve that 91-95 gauges expect. This is why 96 gauge assemblies work in 91-95 and vise versa.



Fenns7, if you need any other info out of my 96 FSM hard copy, I will do what I can.

If anyone needs fuel level sensor data - not just for XJs, I'm trying to "collect the whole set" for the entire automotive community - I decided this subject deserved its own webpage and put it here. http://w1kas.net/jeepyeartech/fuelsenders.html
If you have data to contribute to this page by all means send it to me, I prefer primary sources such as factory service manuals but will accept measurements made by other forum members as well where necessary.

Thank you for the offer Kastein - anything '96 specific you can PM me when you have time would be very much appreciated.


I have to do some filming for diagnosis on the XJ this week (another story!) so I'll record the resitance movement on my spare '96 sending unit and send to you. I must admit I hadn't noticed there were actually two sensors on the unit so will check that out. Allow me a couple of days.

 

[fenns7]

Hi kastein - apologies for the delay but here's a link to a YouTube vid of sender resistence readings:


https://youtube.com/shorts/iotYQWjWLtk


P.S. tried to PM you but you've exceeded your storage on - need to delete some messages


KRs


Simon

 

[Saudade]

Don't know if you meant the video for kastein only, but the video shows unavailable "private".

The fsm does refer to 2 signals. one to the gauge and one to the PCM as kastein says.

"The fuel gauge sending unit is attached to the side of the fuel pump module. The sending unit consists of a float, an arm, and two variable resistors (tracks). These two tracks are used to send two different electrical signals. One is used for fuel gauge operation and the other is for OBD II emission requirements.

Track 1-— fuel gauge operation: As the fuel level increases, the float and arm move up. This decreases the sending unit resistance, causing the fuel gauge on the instrument panel to read full. As the fuel level decreases, the float and arm move down. This increases the sending unit resistance, causing the fuel gauge on the instrument panel to read empty.

Track 2—OBD II emission requirements: A variable voltage signal is sent to the PCM to Indicate fuel level. The purpose of this feature is to prevent a false setting of misfire and fuel system monitor trouble codes if the fuel level in the tank is less than 15 percent of it's rated capacity."

The fsm does provide the resistance levels for Track 1 (as posted above) but no measurements for Track 2. Pin 5 on the fuel pump module connects to PCM Pin C26 and is labeled "low fuel sense".

 

[fenns7]

Thanks for letting me know Saudade - I did try to PM Kastein but as I say he's 'full'. I'll change the settings - please alert me if it doesn't work as I'm a still a bit of a YouTube doofus.


Appreciate the clarification on the operation of the two tracks - I hadn't actually realised there were two separate parts until Kastein mentioned it and I had a closer look.