Fuel sender unit: Strip and repair... eventually.

I thought I should also comment a bit more on the fuel level sender and my attempts at sorting it- but its as well to hold my hand up at this point and admit that it didn't go well!  I present therefore an account of my ill-fated efforts and invite you to learn from my experience!

As you may recall the removal and refitting of this part is described under "Fuel system tank to carbs". While I had it out of the tank I did check its behaviour and was reasonably happy... although I now realise that this was largely because I didn't understand how it works! Its a combined switchđ/level sensor in that the fuel level is indicated using a variable resistor operated by the float (max resistance at full, lowest resistance when empty- i.e. the opposite way around to most other cars!). Resistance is infinite when the sender is pulled right up. Low fuel is indicated by a separate switch which turns on at low fuel levels to activate the low fuel/choke warning light. I didn't know I even had a low fuel light!

My tank has been empty for years... you can't get a fuel level lower than that so given that I now know this switch function exists, I had expected the low fuel light to come on when I reconnected the sender... it didn't! I can find power to both variable resistor (GR wire) and low level switch (GB wire) at the sender cover, and if the latter is shorted to earth than the low fuel light does illuminate. Further when the ign is turned on the gauge does gradually move to "E". This means a good deal of the circuitry is OK, but in all likelihood the sender isn't? A new sender from SJS is £48 (and the rest), aftermarket unbranded senders are around £20 new but I don't know if they would fit. It seemed sensible therefore to have a go at fixing it but I was unsure how to do so.

I did find this account relating to an E type Jaguar which uses a similar  sender... in this account the W terminal is the green orange wire (low fuel warning light switch) and the T terminal the green black (gauge). On the bright side this does seem to be a pretty simple concept and I suspect, its nothing that can't be fixed by cleaning or maybe re-bending relevant contacts!- Hah!!!

"... Fuel Sender – stumped but fixed
For some reason the low fuel light on the dash wasn’t working, yet the fuel gauge was fine. The fault was traced back to the fuel sender unit, which has a removable cover plate. So it was easy to gain an understanding of how it worked to control both the fuel gauge and warning light.

W & T terminal mechanisms W terminal- low fuel warning light switch. T terminal- gauge

Low fuel light contact strip

As would be expected, the unit uses a rheostat to vary the voltage drop across the fuel gauge and the warning light is simply a contact switch. However I hadn’t realised they were two completely separate circuits, sharing a common earth – the sender unit housing.

As the float arm rises and falls with changes in fuel levels, its pivot rotates through approximately 80 degrees. Two slider contact arms are attached to the pivot within the unit and therefore follow the same arc. They are also in contact with the sender housing and so are the electrical contact to earth.

Fuel Gauge
One of the sliders runs along the edge of tapered coil of resistance wire which is connected to the exterior T terminal. When the tank is full and the float is raised to its maximum, the full length of resistance wire lies between the slider and the T terminal – a total resistance of 196Ω.

When the tank is nearly empty and the float is at its lowest, the slider will have moved shortening the length of resistance wire between the two. At empty, the rheostat resistance is 18Ω. The fuel gauge is calibrated to display Full and Empty for these two resistance values.

Low Fuel Light
There’s a copper contact strip on the inside of the cover plate which has a small diagonal break in the copper so the two ends are electrically isolated from each other. The W terminal, connected to the gauge, makes permanent contact with one end.

When the tank is full the second slider arm is in contact with the other end of the strip and moves towards the W terminal contact as fuel is consumed. The slider eventually moves across the gap making electrical contact with the W-terminal, completing the path to earth and switching on the warning light.

I couldn’t work out why it wasn’t working. The multi-meter confirmed the internal connections were working correctly. Yet the switching wasn’t evident at the external spade connector. It didn’t make sense as a metal rivet connects the internals with the external spade terminal.

Checking with the multimeter confirmed that somehow the rivet and the external spade terminal were electrically isolated from each other. A dab of solder solved the problem but I still can’t fathom how they could not be in contact with each other."

Thanks to ChrisVine for this text and pics

So I re-extracted my own sender... not too difficult, the fuel tank doesn't have to come right out- you can undo the two side retaining bolts and pull it forward enough to unscrew the sender retaining ring.

Having extricated the sender and tested continuity, it was clear that whilst I do have a variable resistance between the GB and earth as the float arm moves, there is no position where the low fuel contacts are activated and there was always an open circuit between the GR wire and earth. I did try to fix this- but to cut a long story short it wasn't successful! If you want to learn from my experience press on... otherwise just do as I ended up doing anyway and buy a new sender!

The Lotus sender is not quite the same as the Jag model described above but it clearly works the same way.  I haven't been able to find out what other cars share this unit which would I think provide a cheaper swap, but it seems to be a Smiths TBS 1214/013. Its similar to those used on the Landrover series III diesel and the Triumph 1300 (type TBS1214/000), but not I think identical- which is a shame as aftermarket replacement for the Landrover senders are quite cheap. Anyway perhaps the most significant difference from the jag unit is that its held together by 3 bent tags rather than screws.

Two side views of sender

three retaining tags visible.Top left, right and bottom behind float arm.

Other side of unit note screw on float arm.

Its easy to tap the tags upright to separate the sides of the sender unit.

Tags bent opening sender...

 The sender was very stiff to separate and it seemed to me that it would be necessary to undo the screw on the float arm...

Screw on float arm spindle bottom right
DO NOT UNDO THIS SCREW!!!

... but this was a disaster- firstly I now believe that you do not need to remove this screw and secondly turning it can rotate and seriously damage the contacts inside. What is worse the screw simply broke off anyway!

I had to cut it off and grind it flush.However I could still not open the sender because of the two piece contact strip shown here at the bottom and in side view below.

Note contact strip at the bottom- to separate the halves the case has to move down but its not free to do so until these contacts are released.

There is a small pip holding the two parts of the contact together- I drilled this out with a Dremmel and small bit so that the blades could be separated and now at last I could separate the two halves of the sender.

Drill through retaining "pip"

Inside the upper half the contact was similar in nature to that seen in the jaguar- although clearly a slightly different shape. The action of the low fuel switch is however identical and its activated when the upper contact sweeps off the smaller arc of the copper plate and over the discontinuity to earth the other side.

Copper contact plate inside sender- top right arc-shaped  section bounded by the dark lines is not connected to the bulk of the plate. The arc-shaped region responds to variable resistance for the gauge action and the switch for low fuel is activated when the contact crosses the gap towards 3 O clock

The upper plate was very worn, and there was an obvious groove in the copper. However, this was nothing compared to the state of the sweeping contacts which were deformed and bent and were no where near making contact with the wire-wound variable resistor. I'm pretty sure that this damage resulted from trying to remove the float arm screw so again I repeat- don't undo it!

Contacts scrambled and no where near contacting the wire wound variable resistor. note corrosion on the upper contact. The contacts were damaged I believe by attempting to remove the float arm screw.

I was able to bend the contacts back into a reasonable position:

Contacts swivelled to correct position...

... and bend them down to contact the variable resistor strip

I then refitted the back for testing, hoping to see smooth and variable resistance through the GB wire (gauge) and activation of continuity between earth and GR wire when float fully down.

It took a few attempts but I was eventually able to find a  position where both these functions operated. However reassembly presented a little problem as I had removed and ground out that float arm screw. 

I therefore drilled the arm out at 3mm and tapped to M4 so that I could and use a new short M4 screw to hold the arm.

Float arm drilled and tapped

New screw inserted

... and tags turned down again.

This was all looking very promising but sadly this is where it all went wrong. Having reassembled the unit I found that both the smooth, variable resistance and the switch activation were no longer working. 

I took the thing apart again to discover two problems:

Firstly the contact arms had become (perhaps were always) loose on the float arm spindle- this means that they tended to jam in one position and let the arm move underneath them.

Contact arms have become loose

But could be refitted with a star washer.

I could reattach the arms firmly using a star washer pushed over the spindle- however although this gave a nice variable resistance function of gauge operation, the low fuel switch was erratic. 

Secondly, I had failed to appreciate that the connectivity of the copper arm-contact plate is complex. It is not all connected to earth (this is obvious in hindsight since the switch has to function separately from the gauge; if both were earthed then the switch would always be active!). Although I had physically secured the float arm spindle using a screw, this had also created a short circuit to earth meaning that I had now inactivated the switch function I had just restored! There are probably ways of covering the spindle with a plastic sleeve and using a plastic M4 screw to retain it, or for that matter machining a new spindle from PVC, but overall I think the law of diminishing returns now applies .... plus the last thing I want to risk is any potential for sparking in the tank, so at this point I retired, wounded, despirited and in need of a beer!


Well the new sender (used) duly arrived from Lotusbits (I ordered a flowlock as well which came already fitted with its unions so I was very pleased with that) and I will fit it with the new gasket and ring seal I have obtained.

However I also located a second non-functional sender and wanted to see if I could rescue  my original.

This time I took the new sender apart without unscrewing the float arm screw. I just drilled through the contact pip and straightened the 3 retaining clips. The halves then separated- the resistor half just slips along the arm.

Separating the halves of the second sender unit- float arm still in position. The front of the sender can be removed by sliding it along the float arm. The variable resistor in the case top has a shiny groove worn in it.

The second sender was in worse condition than my original since the copper contact plate was far more worn, and there was a deep groove along the variable resistor. The arm can be removed from the resistor half by removing the float and feeding the wire arm through.

Back (copper contact) side of the sender case. Note float arm in position with contacts as they should be. Copper contact is very badly worn in this float. I applied some penetrating oil around the base of the float arm spindle before ...

The contact arms are in much better condition in this unit- straight and unbent. You can see that they should be located one above the other.

... pushing down on the screw head and the spindle slides out of this half of the case.

The float arm and spindle can be removed from the contact half of the case by pushing the screw inwards through the casing so unscrewing it wasn't necessary.

Float arm spindle removed- screw head still in place and visible inside spring

It might be possible to swap the copper contacts or resistors between cases but as these are riveted in (and I don't have any like rivets) this would be a little bit tricky, its easier to swap the arm. I could insert the float arm into the resistor  half of the case from my first sender and feed it through until the spindle located in the case.

Float arm inserted to my other sender body 

... I could then fit the top cover passing the screw through the hole in the case and clipping it into position.

and the cover added allowing the screw to poke through

This restored function of both gauge and switch, but only if the unit was held tightly closed and compressed. It seemed that the arm must be held tight over against the screw side of the case to make good contact with the copper strips inside. This is probably the function of the screw which acts both to hold the contacts against the internal copper but also to dampen out any sudden movements  caused by fuel slosh. Sadly after my previous efforts to fix this sender, the hole in the cover was now a little too large for the arm screw; and if this was meant to act as a friction lock... then it doesn't do so now! Instead I fixed the screw in the pushed-through position by compressing the two halves of the case before tapping the fastener wings over. I then pushed on the float arm so that the screw poked through the unit as far as it could, and fitted a fibre washer as insulator and finally held the screw post in this position with a C clip.

Screw with washer and C clip

This had restored function- I think with hindsight I could have done a better job and repaired my unit without need for a second sender- however I made some mistakes and this wasn't possible. I wont fit this repaired unit because I have a replacement in better condition now but I am pleased to have worked out how this could be fixed.