On 12/4/2025 at 7:16 AM, Daniwl said:

Cluster unit 

Hi, I see you on another post that you said Ford replaced a piece of your wiring loom to make you fuel gauge work. Do you know the exact piece of loom as I have exactly the same problem? Many thanks, Gary.

I have the same problem, but I'm not the vehicle owner- I'm the guy who fixes instrument clusters for living.

The reason I'm here is that I see a few clever people here and hope that together we will figure this out once for good.

Questions to you guys at the end of that post if you want to skip reading.

To confirm or clarify a few things:

• yes, 2.0 has the fuel sender connected directly to the instrument cluster, unlike 2.2 where the sender is connected to BCM - I was surprised to discover that

• yes the instrument cluster announces the fuel level via MS CAN bus which would explain the engine cutting off. Particularly on ID 0x434 - first 2 bytes are 40 00 with empty tank and 43 EC with full tank. Last 2 bytes of the same frame - FB FC with full tank and F8 00 when empty. What's interesting is that if we change resistance from full to empty, the last 2 bytes change immediately, where as the first 2 change very slow, like mirroring the fuel gauge. If we change from full tank to empty, the fuel gauge drops to zero over about 40 minutes.

I really like @unofix way of thinking and diagnosing, not many people think that way which is the only way of fixing anything, but you have a few facts wrong and that's why I am here.

Apologise for the long story, but it's necessary and again- it's to help each other.

I had that cluster sent to me for the 1st time 2 weeks ago by a friend auto electrician who I trust that he knows his stuff. Despite my trust in his knowledge and skills I didn't believe the cluster is faulty (I often don't even trust myself). I powered it on the bench with the potentiometer connected, set it to full tank and left it running. There was nothing wrong with it, full tank all the time. After about 6 hours, while I was doing other jobs, I've heard the "ding dong" - low fuel level. Yeah! We're doing it, I said excited. I only did probed one node while it was in faulty state as at that stage I didn't have the schematic. I drew the schematic in Eagle Cad, determined which nodes to check with oscilloscope to figure out what has failed. I powered it up again and waited for it to fail. Several hours later it was still working (of course, typical). I set a CCTV camera on it and went home for a weekend. Got back to work on Monday, checked the camera footage- it did not fail once. Based on that 1 measurement I initially took I concluded that there could only be one scenario in which I can see that particular wave on the oscilloscope- 1 transistor shorting between C-E. I replaced that transistor and sent the cluster back. A few days later he's calling me to say it's the same, fuel frequently dropping down to 0 and engine cutting off. He sent it back, I had it running for 4 days when I did not fail once.

At that stage I questioned my eyes and my 1st measurement, I told myself that maybe my potentiometer on the bench magically disconnected from the cluster, so the cluster didn't actually fail. I was 100% convinced it's the wiring. I told the auto electrician to unplug the fuel sender from the cluster, connect a fixed resistor between 1 and say 120 Ohms and test. Day after he says the cluster is dropping to zero again. I couldn't believe so I told him to measure 4 nodes and send me screenshots from the oscilloscope. He did, and I saw it on my own eyes, so I had finally something to work with.

Why do we need those measurements- because to have the cluster to show us empty tank we can either:

• disconnect the sender

• short the wires to the sender

• short one wire of the sender to ground

• short the other wire of the sender to ground

• increase resistance of the sender above 195 Ohms

• and a few different scenarios

but now that we have these 4 measurements within the whole circuit we can tell that it is the input to the MCU shaped the way indicating the empty tank, meaning it's neither the software fault, nor the MCU fault (like internal short or something). It is also not the MCU output- let's call it a trigger of the measurement. After checking all possible scenarios (shorting transistors, lifting pull-ups/downs etc) it turned out that there are only 2 scenarios in which the wave matches what I see on his screenshot- either one transistors' emitter disconnected from the power rail, but that not only is impossible but also that's the transistor I replaced, and the only other scenario is both wires going to the sender shorted (2 and 10). Not disconnected, not shorted to ground, not resistance too high, only shorted or in other words- fuel sender giving zero ohms resistance. Knowing that I assumed that my friend is fooling me and that he probably connected the resistor at the pump plug, and the short is in wiring, but he claims the resistor is directly behind the cluster on 2 inches long wires. I inspected the path of these 2 traces from the MCU to the cluster connector and it is physically not possible that the short is anywhere on the PCB. I even looked underneath the connector (between the pins and PCB) to see if there is any loose solder ball or something (that's not possible but I have no other ideas).

That's why I stared searching the internet what I never do, to confirm if anyone solved the problem by replacing the cluster (which would 100% confirm that it is cluster and not wiring as I still can't believe it).

So, I have it running on my bench for about 8 days in total and it didn't fail once, where as in the vehicle it fails frequently. He told me it fails more often after the vehicle is parked up for a while, meaning when it's cold. So I put it into the freezer for a while but that didn't do anything. I left it unpowered for the night, powered up in the morning- still fine. I can't fix it until I see it failing on the bench.

My questions for you guys:

• did you notice any pattern of when it fails- when it's cold, maybe when it's humid (raining), when driving over potholes etc?

• does it happen more often when the tank is full or near empty?

• and lastly, at the moment, those of you who had it sorted already I assume that you had the cluster replaced- what's involved in replacing for a 2nd hand one? Are these coded to the vehicle or not and it's a straight swap? I know the mileage wouldn't match but if we can release the vehicle to the customer we can then continue working on that cluster for weeks or as long as necessary.

Thank you very much for any help, we'll all benefit from it.

Message to the moderator as "Your content will need to be approved by a moderator" - this is not advertisement, I don't state my business name, and if it's not obvious I'm not doing that for profit, otherwise I'd give up after an hour instead of spending almost 20 hours so far as this is my passion and hobby.

On 2/27/2026 at 11:24 AM, rosak said:

• what's involved in replacing for a 2nd hand one? Are these coded to the vehicle or not and it's a straight swap? I know the mileage wouldn't match but if we can release the vehicle to the customer we can then continue working on that cluster for weeks or as long as necessary.

Quick reply because I use Windows 7 and it is not compatible with the new FOC site.

Depending of the Ford model, most have the immobiliser contained within the IC. Changing the cluster will mean the engine won't start. You will need to use FORScan to reprogramme in the keys.

Hello,

The problem was fixed when Ford replaced the full wiring harness. I don’t know the specific name of the part.

Sorry

On 2/28/2026 at 9:28 PM, Daniwl said:

The problem was fixed when Ford replaced the full wiring harness. I don’t know the specific name of the part.

Thank you very much Daniwl. That would make perfect sense to me, that's what I thought since the very beginning, until the auto electrician told me that he did disconnect the wiring from the cluster and connected a fixed resistor directly to the cluster plug.

Lots of people here had the same problem, can anyone please confirm that replacing the cluster solved the issue?

I'm back at work after weekend ant the cluster still works fine..

@rosak Very detailed and helpful post, thank you.

On 2/27/2026 at 11:24 AM, rosak said:

I powered it on the bench with the potentiometer connected, set it to full tank and left it running. There was nothing wrong with it, full tank all the time.

I am also trying to bench test one, but putting resistance between the pins is not activating the gauge like it should, but if I put resistance between pin 10 (sender) and ground it works fine but goes up slowly. Tried it with another one that was definitely working and it works the same with the external ground but not with resistance between 10 and 2.

Also unusual are the signals from the pins - most IPCs have a voltage or PWM one one and the other is ground - I get 2.5Hz signal from pin 10 and 4.2V DC from pin 2. Same with the good one, so seemingly not faulty, just odd.