Faulty fuel gauge/sender?

[whiteswan]

Good morning,
my electronic fuel gauge is not showing any fuel level on my Jeanneau 41 DS
This has happened previously when l have filled the tank full and normally works as normal after l use some fuel.
However this time has not worked, not sure if the sender float is stuck or a electrical problem.
any advice?

Tony

[Mistroma]

There's a pretty good chance that you have the same sender as my 42DS. Readings will be different for every installation as it is affected by how long they leave the float arm and if there are any bends in it.


I'd suggest a few simple checks on your system.

1) Pull the connections from the sender and connect them together to make a dead short. The display should show an empty tank.

2) Put a 200 Ohm resistor between the connections and the display should show full tank.

3) Check the resistance of the sender and it should be somewhere between 50-190 Ohms.

Mine gives about 138 Ohms for around 95 litres and isn't very good at measuring anything over that value
It isn't too bad down to 100 Ohms which equates to about 55 litres. However, it isn't great below that and could well be empty by 60-80 Ohms.

[whiteswan]

Thank you Mistroma,
after carrying out further testing, found the Rochester fuel sender was faulty. New one on order.
 

[Mistroma]

I had assumed you had a standard Scheiber display with the usual cheap float arm and potentiometer sensor. It sounds as if you have a better system with a Reed switch or capacitance sensor. My suggestions probably wouldn't have been very relevant. I'll probably change the sensor on my system at some point as the existing one tends to stick and also isn't very linear (even allowing for the shape of the tank).

[chuckr]

I have the same issue except all of a sudden my showed full all the time.  My problem is a nasty build up of nasty stuff in the tank.  I built a fuel polisher and will polish the fuel and clean the float and if that does not do it will get a new float.  

[Mistroma]

Jun 28, 2021 18:24:28 GMT chuckr said:

I have the same issue except all of a sudden my showed full all the time.  My problem is a nasty build up of nasty stuff in the tank.  I built a fuel polisher and will polish the fuel and clean the float and if that does not do it will get a new float.  

It's a dodgy sensor in my boat. It works most of the time but sometimes sticks in one position for a while and then jumps to the correct reading. I clean the fuel tank annually and also check the sensor visually and electrically. I keep meaning to replace it with something better but it's a low priority as I always know the fuel level within a few litres.

I have noticed a steady build up of stuff in the tank since 2012 when I switched from red diesel in the UK to normal white diesel in the EU. I intially found less than 5-10mls of black material in the sump. This has increased steadily and I now remove about 30mls plus a couple of litres of cloudy fuel (which does quickly settle and clear).


I now pump out the small sump pre-launch as it has settled over winter and again a week or so after lifting out. I used to keep the tank brim full and had to do the first clean out after a few weeks cruising. I think that fuel degradation and bug growth are more of a hazard in Greece than possible condensation problems (never seen any sign of water in the tank apart from the tiny amount in the sump).

I thought that excess fuel is pumped back to the tank, then comes back through the filter again and should self-polish a fair bit. My fuel tank is certainly very hot after a long run, indicating that it has been through the engine a few times.

[timetra]

Jun 7, 2021 7:16:14 GMT whiteswan said:

Good morning,
my electronic fuel gauge is not showing any fuel level on my Jeanneau 41 DS
This has happened previously when l have filled the tank full and normally works as normal after l use some fuel.
However this time has not worked, not sure if the sender float is stuck or a electrical problem.
any advice?

Tony

Good afternoon,

I own a SO 42i and I also have a problem with my fuel gauge. Sometimes it seems to work as expected and sometimes the tank shows empty while I'm sure it's not.

After monitoring this for a while I figured out that it works correctly if my aft water tank is empty. When I fill my aft watertank the fuel gauge shows empty ... When I empty the aft water tank the fuel gauge works as normal again.

Hence I suspect some interference beween the cables of the fuel sender and the sender of the water level of my aft water tank.

Does anyone have an idea what kind of interference this could be and how this could be fixed?

Cheers,
Pierre

[Mistroma]

That is very odd. The following might be useful if your 42i has the same Scheiber display as my 42DS. I build my own board to replace a damaged Scheiber unit and it should mimic their unit closely. I used their published specs.


Sorry if this is a little too detailed but I hope you can follow my reasoning.

I measured the voltage from my standard 0-180 (or 190) Ohm sensor fed by a 3V reference supply. The sender was on one leg of a voltage divider and the other had a 100 Ohm resistor. This meant that full (190 Ohm sender reading) produced 1.96V and empty ( approx. 4 Ohms) gave 0.12V. These figures matched the original Scheiber specification.

The wires coming from the fuel sender should therefore give somewhere between 0.12V (Empty) and 1.96V (Full) at the display panel.

There's a box beside each water tank that converts the level measurement to a voltage. It is designed to output the voltage in the following steps.

Min.____Mid.___Max.
1,00V < 1,10V < 1,20V - empty tank
2,10V < 2,20V < 2,30V - level ¼
3,20V < 3,30V < 3,40V - level ½
4,35V < 4,45V < 4,55V - level ¾
5,50V < 5,60V < 5,70V - full tank


This means that:
1) Empty water tank wires would have 1 - 1.2V whilst a full fuel tank would result in a voltage of 1.96V.
1) Full water tank wires would have 5.5 - 5.7V whilst an empty fuel tank would result in a voltage of 0.12V.

Cross-talk from adjacent wires doesn't make much sense as they both have a similar low DC voltage. It is also odd that as water tank voltage reaches max. when full and fuel tank is almost zero when empty.

It seems unlikely that the problem lies with the wiring from each tank. The sensors are widely separated and I can't see an issue with adjacent wires unless physically damaged and shorting to each other. A short between these systems might overwhelm the fuel reading making it always 75% - 100%.

Damage to the wiring is pretty unlikely without some sort of major trauma to the wiring loom.

Bottom line: The problem is likely to lie at the Scheiber board connections or more likely the unit itself.


You could try to diagnose the issue by measuring voltage across various points on the connectors at the panel.
Unfortunately, this is messy as each water tank gets power from the panel via the connector. The water tank sensors are therefor dead when the connector is removed. Similar situation with the fuel sensor.

Disconnecting to measure is tricky because the water sensor system gets power from the panel on one pin and measurements are taken on others. You won't see a voltage with the connector unplugged. Same with the fuel tank. However, you can check for shorts easily enough and I suppose that is something. I can make a diagram showing which pins to use to test if that would help.

[timetra]

Hi Mistroma,

Thanks for the info! Odd indeed!

From your explanation I understand that the fuel sender is connected to ground on one side and to the 100 ohm resistor at the other side. With a reference voltage of 3V this gives indeed ~0V .. 1,929V (my sender is a 180 ohm version).

Could it be that the 100 ohm resistor is connected to ground instead? This would give a voltage range of 1,07 .. ~3V (full .. empty).
If in this case there were a short between the output of the waterlevel and the fuel sender, a full watertank would drive the fuel sender to >= 5,5V which would make the Scheiber circuit conclude that the fuel tank is empty.

Odd, but I can't explain it otherwise.

I'll try to make some extra measurements when I return to my boat (likely this weekend). Hopefully these shed some extra light on this mystery.

Cheers,
Pierre

PS: I read your post on your self built Scheider replacement. Impressive!!!

[timetra]

Nov 22, 2021 19:34:29 GMT timetra said:

Hi Mistroma,

Thanks for the info! Odd indeed!

From your explanation I understand that the fuel sender is connected to ground on one side and to the 100 ohm resistor at the other side. With a reference voltage of 3V this gives indeed ~0V .. 1,929V (my sender is a 180 ohm version).

Could it be that the 100 ohm resistor is connected to ground instead? This would give a voltage range of 1,07 .. ~3V (full .. empty).
If in this case there were a short between the output of the waterlevel and the fuel sender, a full watertank would drive the fuel sender to >= 5,5V which would make the Scheiber circuit conclude that the fuel tank is empty.

Odd, but I can't explain it otherwise.

I'll try to make some extra measurements when I return to my boat (likely this weekend). Hopefully these shed some extra light on this mystery.

Cheers,
Pierre

PS: I read your post on your self built Scheider replacement. Impressive!!!

I just realised that my reasoning is flawed. I assume that the reference voltage of 3V of the fuel sensing circuit is only active when the fuel level is read from the Scheiber console, i.e. when the corresponding button is pressed. Likewise, the voltage that drives the water sensing circuit (Scheiber 50.BTE05000E.00 I believe) will only be active when reading the water level.

Hence when I read the fuel level, the water level circuit will not be active. So the water level circuit will only interfere in a passive manner. I still assume that the 2 circuits interfere in some way (otherwise this is very difficult to explain).

Do you know how the Scheiber 50.BTE05000E.00 circuit is built? With that info I could try to figure out how this (inactive) circuit can influence the read out of the fuel tank.

I do have a leaflet from Scheiber that explains the blackbox operation of the 50.BTE05000E.00 but it doesn't reveal what's inside.

Cheers,
Pierre

[Mistroma]

I have never seen any schematics for the that box but have always assumed it won't be very complex. If either of mine failed I'd make a replacement with a few transistors, resistors, diodes etc.

It would only take a simple transistor switch to turn on the voltage when water reaches a screw in the tank wall. The circuit would be very simple and simply output the supply voltage (probably reduced to 6V at that point).

If you repeat this for different levels and use a voltage divider the output can then be reduced to a lower level.

That means there's one transistor switched circuit outputting a voltage via a divider for each screw in the tank wall. There's a bit more required to combine these into a cascade to output the highest of the voltages when multiple transistors are being switched. I've never really given it much thought and just assumed I'd work out the detail if needed.

The 100Ohm resistor I mentioned for the fuel measurement was on my own external board. Something close to that will be on the Schieber board and not an external component on your system.

I think that I didn't bother switching the water sensor power on my Mk1 I/O board. The power drain was tiny and I was short of pins on the Arduino. I assumed Scheiber did switch power to that circuit and added it to the ESP32 version. I should have mentioned that and imagine you are correct that fuel and water aren't powered at the same time.

I'm leaning towards a fault on the board rather than the wiring.

Have you tried following:
1) Start with empty water tank
2) Disconnect fuel sensor and measure voltage and resistance at the wires (record sensor resistance as well)
3) Connect fuel sensor again
4) Fill the aft water tank to get the fault
5) Repeat checks on fuel sensor and wires to check for voltage or resistance changes and connect again
6) Completely disconnect the aft. water tank sensor box (50.BTE05000E.00)
7) Check if the fuel gauge reads correctly

You can then carry out further measurements on fuel sensor and water sensor to see if you can find the combination of connections needed to give the fault. Disconnecting wires to the screws in the tank wall will foll the system into thinking the level has dropped. That will let you measure output voltage at various levels and check the effect on the fuel reading.

[timetra]

Hi Mistroma,

I'll try to do some measurements this weekend. As the boat is out of the water and already winterized I won't be able to do some of them until the beginning of the next sailing season.

Cheers,
Pierre

[Mistroma]

You could make up some wires with push-on terminals for the interface box. I think they'd be the most common size (is that 6.3mm) but oversize ones would probably still grip well enough to work.

I brought one interface box home once and just connected wires for power and stuck the ends of the other wires into a glass of water. I confirmed it was working by just measuring the voltage out as I removed wires from the water. I didn't trim back the insulation and that prevented wires touching each other.

It would be even simpler when still fitted to your boat. Just a matter of removing the sensor wires, replacing them with your short wires and sticking them into water. You can remove the one by one and measure the output voltage to see if it matches the specification. You can also take measurements at the fuel tank.

I'm basically suggesting that you try the measurements I mentioned earlier but fool the water tank interface by using some short wires. No need to actually fill or empty the tank. The wires you are replacing only run between the interface box and the tank and any damage should be visible. That means you are not affecting any hidden parts of the wiring.

A pack of small crimp connectors and any spare wire will provide a cheap method of fooling the system into reporting a range of water levels. Just keep a note of the position on the real sensor wires coming from the tank.

I think you mentioned the manual for the interface box but have a copy on my web-site.
mistroma.com/Manuals/50.BTE05000E.00_FR_GB.pdf

It has a diagram showing all the connections. You would just replace no. 4-8 with your wires and measure output between 1 and 2.

[timetra]

Thanks a lot for the suggestions! I'll try to do that this weekend!
Glad that I don't have to wait until the next season to continue with this (and hopefully find the culprit).

Cheers,
Pierre

PS: my copy of the leaflet of the 50.BTE05000E.00 was most probably downloaded from your website  . Thanks for that too!

[timetra]

Nov 23, 2021 13:32:47 GMT Mistroma said:

I have never seen any schematics for the that box but have always assumed it won't be very complex. If either of mine failed I'd make a replacement with a few transistors, resistors, diodes etc.

It would only take a simple transistor switch to turn on the voltage when water reaches a screw in the tank wall. The circuit would be very simple and simply output the supply voltage (probably reduced to 6V at that point).

If you repeat this for different levels and use a voltage divider the output can then be reduced to a lower level.

That means there's one transistor switched circuit outputting a voltage via a divider for each screw in the tank wall. There's a bit more required to combine these into a cascade to output the highest of the voltages when multiple transistors are being switched. I've never really given it much thought and just assumed I'd work out the detail if needed.

The 100Ohm resistor I mentioned for the fuel measurement was on my own external board. Something close to that will be on the Schieber board and not an external component on your system.

I think that I didn't bother switching the water sensor power on my Mk1 I/O board. The power drain was tiny and I was short of pins on the Arduino. I assumed Scheiber did switch power to that circuit and added it to the ESP32 version. I should have mentioned that and imagine you are correct that fuel and water aren't powered at the same time.

I'm leaning towards a fault on the board rather than the wiring.

Have you tried following:
1) Start with empty water tank
2) Disconnect fuel sensor and measure voltage and resistance at the wires (record sensor resistance as well)
3) Connect fuel sensor again
4) Fill the aft water tank to get the fault
5) Repeat checks on fuel sensor and wires to check for voltage or resistance changes and connect again
6) Completely disconnect the aft. water tank sensor box (50.BTE05000E.00)
7) Check if the fuel gauge reads correctly

You can then carry out further measurements on fuel sensor and water sensor to see if you can find the combination of connections needed to give the fault. Disconnecting wires to the screws in the tank wall will foll the system into thinking the level has dropped. That will let you measure output voltage at various levels and check the effect on the fuel reading.

I did most of the measurements you suggested but I couldn't find anything out of the ordinary. I was planning to do some extra measurements but alas got called away before I could make them. Maybe something for next weekend.

What did I do:

I made the connections you suggested to fool the system into thinking the water tank was full.



As soon as I installed this the watertank showed indeed full:



And as expected (in my boat that is) the fuel tank immediately showed empty while in fact it was nearly full (couldn't add a picture of that as I couldn't upload more than 3 pictures in this post):



Emptying & filling the watertank was now very easy   so I did that a few times. And each time the water tank level displayed full the fuel tank level showed empty and vice versa.

I also measured the following:

Fuel sender resistance: 133 Ohm
Voltage over the fuel sender when measuring: 1,67 V
That suggests that the other resistor in the divider circuit is 106 Ohm (given a ref voltage of 3 V). Close enough to the 100 Ohm that you mentioned in your post.
As expected there is no voltage present when the fuel level is not read out.
I repeated this measurement when both the tank was (simulated) full and empty. No difference ...
So I always measured 1,67 V no matter the read out of the fuel level (full if water tank empty, empty when water tank full).

I was just about to start experimenting with partial water levels (only tried 0% & 100% so far) as I was called away.
Will try to continue next weekend.

Mystery remains for now ... All suggestions welcome!

[Mistroma]

I'm glad you managed to set it up as that will make the diagnostic process much quicker. Repeated filling and emptying would probably take you most of the winter if done for real. Of course ut may just help you reach a dead end more quickly (hopefully not).

I think you did mention the fuel reading showing full previously. I assume 133 Ohms is giving 100% on the gauge. The resistance indicates it might be around 70% full. But could be anywhere between 50% and Full depending on float arm length, curvature of arm, stickiness of rheostat etc. I wouldn't be surprised to see 100% and suddenly drop to 50% after using only 20-30 litres.

It seems likely that the fuel sensor, wiring and at least some of Scheiber board is working. It is turning voltage on and off, likely to have correct ref. voltage and internal voltage divider is still working.

The fact that the fuel sensor voltage is constant with empty and full water levels makes me think that the fault lies within the Scheiber unit. I'd expect further tests with 25%, 50% and 75% will also fail to affect the fuel sensor voltage. It will be interesting to see if the gauge always reads zero or gives different readings.

Do you know if the forward water tank level also affects the fuel reading? It would point to a main board fault if both tanks cause the same issue. The wiring goes in opposite directions and only come close at the board connector. I wouldn't expect them to behave in that manner as they won't share a common circuit on the board. Some parts might be common or close enough to suffer the same fault. Probably worth testing the forward tank before removing the Scheiber main board.

It is fairly easy to remove the board as a last resort. Unplug the connectors and undo the four retaining screws. The entire board slides out, complete with display and buttons.

Careful examination might reveal the problem if it is a short between fuel and water measurement circuits. If not then you will probably need to test the board in isolation.

You don't need much initially to simply mimic the problem at home:

1. Push on connectors to connect the board
2. Some croc. clips to connect resistors
3. A 12V supply for power
4. A few resistors to check the fuel reading (e.g. 47, 68, 120, 150 Ohm or rheostat)
5. Variable power supply to mimic water tank interface box (1V - 5.7V)

The easiest method to vary power would be to buy a cheap variable supply from eBay. I bought a load of tiny boards for some projects and adjusted one to give whatever I needed.

You could also use some rechargeable AA or AAA batteries it series to get close to some of the voltages you'd need.
1.2, 2.4 and 3.6 You might need to mix a variety of types to get to 5.6V.

The other option would be to take one of the interface boxes home, feed it 12V and wire it to a bottle of water as you've already done.

[timetra]

Nov 27, 2021 18:07:16 GMT Mistroma said:

I'm glad you managed to set it up as that will make the diagnostic process much quicker. Repeated filling and emptying would probably take you most of the winter if done for real. Of course ut may just help you reach a dead end more quickly (hopefully not).

I think you did mention the fuel reading showing full previously. I assume 133 Ohms is giving 100% on the gauge. The resistance indicates it might be around 70% full. But could be anywhere between 50% and Full depending on float arm length, curvature of arm, stickiness of rheostat etc. I wouldn't be surprised to see 100% and suddenly drop to 50% after using only 20-30 litres.

It seems likely that the fuel sensor, wiring and at least some of Scheiber board is working. It is turning voltage on and off, likely to have correct ref. voltage and internal voltage divider is still working.

The fact that the fuel sensor voltage is constant with empty and full water levels makes me think that the fault lies within the Scheiber unit. I'd expect further tests with 25%, 50% and 75% will also fail to affect the fuel sensor voltage. It will be interesting to see if the gauge always reads zero or gives different readings.

Do you know if the forward water tank level also affects the fuel reading? It would point to a main board fault if both tanks cause the same issue. The wiring goes in opposite directions and only come close at the board connector. I wouldn't expect them to behave in that manner as they won't share a common circuit on the board. Some parts might be common or close enough to suffer the same fault. Probably worth testing the forward tank before removing the Scheiber main board.

It is fairly easy to remove the board as a last resort. Unplug the connectors and undo the four retaining screws. The entire board slides out, complete with display and buttons.

Careful examination might reveal the problem if it is a short between fuel and water measurement circuits. If not then you will probably need to test the board in isolation.

You don't need much initially to simply mimic the problem at home:

1. Push on connectors to connect the board
2. Some croc. clips to connect resistors
3. A 12V supply for power
4. A few resistors to check the fuel reading (e.g. 47, 68, 120, 150 Ohm or rheostat)
5. Variable power supply to mimic water tank interface box (1V - 5.7V)

The easiest method to vary power would be to buy a cheap variable supply from eBay. I bought a load of tiny boards for some projects and adjusted one to give whatever I needed.

You could also use some rechargeable AA or AAA batteries it series to get close to some of the voltages you'd need.
1.2, 2.4 and 3.6 You might need to mix a variety of types to get to 5.6V.

The other option would be to take one of the interface boxes home, feed it 12V and wire it to a bottle of water as you've already done.

I'm also glad I have the circuit in place to emulate the aft water tank level. Thanks again for the suggestion!

Indeed my fuel tank isn't quite full. The level shown by the Scheiber unit is alternating between full and 3/4. I think this is consistent with the fuel sender at 133 Ohm. I just wanted to make the point that the fuel reads empty when my aft water thank is full and shows the correct reading when the aft water tank is empty.

The level of the forward water tank has no influence on the reading of the fuel level. It is only the level of the aft water tank that has the strange effect. I also start to believe it must be the Scheiber unit. But what?

I'll try to remove the Scheiber unit next weekend. Hopefully that will reveal something ...

[Mistroma]

It was a long shot about the forward tank but same fault there would have helped to show it was the Scheiber unit. Unfortunately, lack of a fault there still doesn't prove anything.


Checking the Scheiber board will be difficult but you might be lucky and spot a tiny stray wire bridging something or some dodgy soldering.

It will be tricky if you don't spot anything and need to follow individual wires on the board, starting from the connectors.

I don't hold out much hope of a fix if you don't spot something really obvious. I think the best you can hope for is to get it working at home in a test rig. You either reproduce the fault and it must be the board or you don't and need to check wiring on the boat.

I'd look the tracks around pins 2 and 6 initially as the most likely candidates.

Pin 8 is the water tank interface supply and shouldn't be involved. It will be the same voltage for both tanks and won't vary with tank level.

I think pin 6 is the aft. tank and marked Eau 2 on the board. Pin 3 is the other water tank. I'd need to double check but think I have identified them correctly.

Pins 2 and 5 are for fuel tanks 1 and 2. I don't know if they are both active at the same time in the same manner both water tank interfaces are powered. You would probably need to test to see what happens when tank 1 is being read or trace both pins. Pin 2 shouldn't be connected to anything if you only have one tank.

[timetra]

Thanks!

Your explanation of the pins corresponds to what I can find in my docs. However if I read mine correctly it looks like pin 3 corresponds to the aft water tank (tank 1) and pin 6 to the forward tank (tank 2):


Attachment Deleted

JE on pin 8 is the water tank circuit voltage supply.
JE1 on pin 3 is the output voltage of water tank 1
JE2 on pin 6 is the output voltage of water tank 2
JF1 on pin 2 is the wire towards the fuel sender (connected to 3V via 100 Ohm resistor).

I hope indeed that I can spot something obvious on the Scheiber panel. If that fails I don't intend to "play" too much with that unit as I might create extra damage. Instead, I consider to install a switch that allows me to disconnect JE1. That should allow me to read the fuel level even if the aft water tank is filled.
I obviously assume that JE1 is somehow the culprit. I first need to confirm that of course.

[Mistroma]

You could be correct about pins 3 and 6, I was working from memory plus a note I'd made and didn't have the complete documentation to hand.

You can get the manual for the Scheiber unit here mistroma.com/Manuals/41.440A2.00NavigraphV1_0GB.pdf if you don't have copy. It has information about connections and details of power requirements, sensor outputs etc.

A switch is probably a good workaround if you can't find the problem. A home test rig might still be worth trying as it will prove the board is at fault. Only do that if you are happy to play around and careful not to short something by accident. It isn't actually that difficult if you have suitable connectors to slip over the pins. Just a matter of being methodical and taking time with connections. There aren't many, just 2 for 12V power, 2 for fuel (& resistor between them), + others for one water tank. You can bring an interface home and power it directly from the same 12V supply or use a small variable supply from eBay and adjust that to mimic different tank levels. You could try both as the latter rules out a problem within the interface box (not very likely).