Talk Morgan Forums Morgan Sports Cars Morgan Three Wheeler & Super 3 Fuel gauge sender/rack

Thanks for posting this Dan,

It is interesting to see a product like this on the market. I did not realise you could buy one because it is very similar to what i made and have been testing/ developing.

It is quite expensive and there MAY be a problem with using it on the morgan 3 wheeler.
It depends upon how quickly it starts working when the ignition is switched on:

The the gauge takes the signal from the fuel sender and then digitally filters it.
This is becuase you do not want the sloshing of the fuel in the tank to make the needle (or number!) jump around excessively.
This type of filter is a low-pass filter. When digitally implemented, it works like a weighted moving average. The most recent input voltage figures have a higher weighting than the older ones. Really old readings (say over 2 minutes old) have little relevance to the calculation at all.

People seem to think that the M3W fuel gauge still jumps around too much.
VDO/Morgan kind of 'shot themselves in the foot' by having a resolution of 1%, becuase the fuel sender jumps around by more than 1% in normal operation. In order to cure the jumping, they should have made the recent reading less weighted. But hey, I think the fuel gauge needle in my audi probably jumps around just as much, however the gauge only has a resolution of 12.5% (an 8 bar gauge).


There is no need to filter when the car is first turned on. THis is becuase you know that the fuel is not sloshing around.
The only time that the gauge does not filter is when the ignition is first turned on after the key has been newly inserted. This is so that you dont have to wait for the filtering lag to pass after filling up with fuel AND so that you get an accurate reading of the tank when you first turn the car on to go somewhere.


This means that if you turn on the ignition with a disconnected fuel sender and then attach it after a few seconds, most of the values in the gauge calculation will be zeroes. If the gauge has switched into filtering, it will take a few seconds for the gauge values to start to rise high enough to register above 0%. It will take a further couple of minutes to get near to the accurate reading.

Similarly, if you turn on the ignition with the fuel sender attached (and a full tank) and then disconnect it, you will see that the fuel gauge slowly falls in a lagged manner taking time to go to zero.

A problem with using a digital transposing deveice like the SN74 is that it does require power in order to send the correct value to the gauge.
As explained above, it is necessary for the correcting device to send the correct value to the gauge when the ignition is first turned on. Otherwise you will have to wait for the filter lag to pass before you get an accurate reading. If the correcting device takes a few milliseconds to turn on itself, then it wont be able to send its value over quick enough before the filtering starts.

There is a way round this.
You could install a battery within the correcting device. Perhaps that is why the SN74 is so large, i dont know if the SN74 has an internal battery? I cannot figure out another way of getting power to it before the gauge.
Or,
you could connect the correcting device direct to the 12v vehicle battery and bypass the ignition loom. However this last suggestion may drain the battery and would not be fused.

So i imagine that it would work, (definitely would be accurate after 2 mins), however would have to check its switch on behaviour.

Regards

ps you would have to buy option A and manually enter the resistance range

Thanks for the insight Charles. Your explanation makes perfect sense. Unfortunately with my car and seemingly others, the logic that performs this filtering to dampen display is somewhat awry. I get wild fluctuations, premature low readings and often find a particular percentage indication pegged for as many as 20 miles, whilst at other times it drops like a stone.

Whilst the resistor mod has improved the situation to the point I only get a consistent 0% indication above 200 miles from full tank, I have suffered these anomalies before and after fitting.

For these reasons I don't believe the issue can completely be solved by simple sender resistance correction. I believe you alluded to this in an earlier post.

Following a cold start the other day I noted an initial 0% reading (tank 2/3rds full) until system volts rose along with display, suggesting this may be another variable. Normally the indication drops after start consistent with your explanation of filter function.

I think I've found a car with a worse fuel guage behaviour than my M3w - and it's a 2016 Hybrid Merc!

At 50 miles range an amber reserve fuel warning comes on and the range keeps counting down - so far so good.

Stop the car and switch off, now when you restart a red reserve warning appears but no range display (even though it should have been 47 miles)

Drive 12 miles to work and switch off expecting around 35 miles left.

Switch on get large red warninng that only E mode is avaible at low power - fill up now.

Set off on elec motor for petrol 5 miles away and after 1 mile every light in the display comes on and the motor cut out - really impressed (not!)

Now the good bit - switch off, switch on and normal hybrid service is resumed so petrol/ electric to the petrol station and based on fuel put in still about 4 litres in the tank and 10% left in the drive battery.

And to think a team of (probably) German engineers worked long and hard to come up with this good idea.....

My fuel gauge ( with the aftermarket longer sender) has been steadily improving! I'll get 180-200 miles now before she shows "0". My guess is that somehow the float is becoming "better". Maybe it's settling somehow? No idea, just know she's been improving.

On my Piaggio MP3 scooter the LCD trip meter would switch to miles to empty when the low fuel warning came on. The capacity was only 12L but it would seem more accurate & 'cleverer' than a M3W or Merc hybrid costing over 10 times more.

My thoughts are that the damping mechanism in the electronics has a degree of self learning of the current, before and after readings from the float sensor, possibly similar to the ECU adjusting the mapping to suit individual engine's. If so the problem seems to be it can learn in the wrong direction!

I don't think it is clever enough to learn anything, in any direction.

Fully agree, it's only accurate in the middle...:

Tim

I am no engineer, so some of the discussion on this thread is above my pay grade; but I have just spoken to Morgan about the problem, which it does at least acknowledge. One of its electrical boffins told me:

1. That this is a problem more prevalent in 2011-14 M3Ws. Since then, an improved sender has been fitted. He claimed that the older ones will show zero when there is about a third of a tank left; the newer ones when there are c.two gallons left.

2. That fitting a longer sender is unlikely to help by itself as the fuel gauge is calibrated to the factory-fitted device.

His advice (and this is very Morgan) was to carry a two gallon can of petrol for a trip, run the car till empty and see how far it goes once it reads 0%.