Why are the gauges so inaccurate?

[allante]

I wonder why the gauges are so inaccurate, It looks like a float the same as a petrol tank. Is it the nature of the gas, does it move about more, and why don't they make a gauge that uses pressure?

[Gilbertd]

LPG tanks don't have baffles to stop the liquid from sloshing about and, unlike a petrol gauge sender that is inside the tank, the gauge part has a magnetic linkage between the float and the sender. If a gauge used pressure it would read 3/4 full all the time in the winter and full all the time in the summer until you actually run out. Pressure in the tank is temperature dependant not level dependant until it is empty.

[e301988325i]

Mine's accurate and repeatable, but I've calibrated it myself, it does take lots of fine tuning to get there over several tank fulls, not really viable for an installer unless you take it back regularly.

[Morat]

Lordy - long time no see! How's tricks?

[LPGC]

Also, LPG level floats are at the end of an arm which is fairly horizontal, so there is a bit of weight (from the arm) on the float. LPG is only around half as dense as water, less dense than petrol, so it takes LPG to be a good way up the height of the float before the float will actually float (if you will) and lift the arm to allow the sender unit to see a change in level above empty. That said, it is possible to calibrate LPG gauges better than the 'standard gauge reading' but they will never be as accurate as most petrol gauges and there will always be a bit of gas in the tank before the sender unit can notice any difference from empty.

One of the things I do on a new install, on the way to a pump for a first fill, is to turn on tank and reducer solenoids (to ensure earth reference is the same as will be in operation), check the reading from the gauge and set the 'not quite empty' first light on the switch to come on when the reading from the sender is just above the empty reading. With a D/A convertor that has 255 steps for reading level sender position (like in AEB and other ECUs), the default position for empty might be around 17 steps but this can usually be adjusted down to maybe 7 steps.. For systems that show voltage, the 255 steps would represent 0 to 5v one way or the other... On a 5 state gauge I like to set the gauge so each light represents 1/5th tank volume, so the gauge works linear but it isn't always possible - depends on how well near empty level is represented by the float. For any of this to work in practice it is essential to use linear senders, no good using the 5 step variety.

Simon

[scottyf]

I can't be bothered. I reverted back to my old ways like I did when I had a motorbike with not gauge on it.
Used the trip. Worked out how much distance I got out of a tank full and go from that.

Reset the trip each time I fill up.

You could always work out a rough mpg that you'd do on lpg. Then work out how many liters you stick in to give a rough calc of how far you can go between fill ups.

[Morat]

Also, LPG level floats are at the end of an arm which is fairly horizontal, so there is a bit of weight (from the arm) on the float. LPG is only around half as dense as water, less dense than petrol, so it takes LPG to be a good way up the height of the float before the float will actually float (if you will) and lift the arm to allow the sender unit to see a change in level above empty. That said, it is possible to calibrate LPG gauges better than the 'standard gauge reading' but they will never be as accurate as most petrol gauges and there will always be a bit of gas in the tank before the sender unit can notice any difference from empty.

One of the things I do on a new install, on the way to a pump for a first fill, is to turn on tank and reducer solenoids (to ensure earth reference is the same as will be in operation), check the reading from the gauge and set the 'not quite empty' first light on the switch to come on when the reading from the sender is just above the empty reading. With a D/A convertor that has 255 steps for reading level sender position (like in AEB and other ECUs), the default position for empty might be around 17 steps but this can usually be adjusted down to maybe 7 steps.. For systems that show voltage, the 255 steps would represent 0 to 5v one way or the other... On a 5 state gauge I like to set the gauge so each light represents 1/5th tank volume, so the gauge works linear but it isn't always possible - depends on how well near empty level is represented by the float. For any of this to work in practice it is essential to use linear senders, no good using the 5 step variety.

Simon

That's a nice touch. The last step is the most important, and I'm sure I'd fill up less often if the last stage on the gauge didn't cover about 60 miles of travel. Then again, I do keep petrol in the tank...

[e301988325i]

Lordy - long time no see! How's tricks?

Very well thanks, you? My E30's still sat in the garage waiting in anger until these pesky nursery bills are out of the way!

The 330i lpg was sadly written off, although it was such a weapon I needed to slow down anyway. I had to make a court visit for a fine & 6 points,
. I didn't even stir the box!

[scottyf]

A nice 330i will probably be my next car

[joshnosh]

Mine's accurate and repeatable, but I've calibrated it myself, it does take lots of fine tuning to get there over several tank fulls, not really viable for an installer unless you take it back regularly.

same
took a while but its not to difficult

[joshnosh]

id say though, the main reason for the inaccuracy isnt sloping about. that gets averaged out
its that the gates set in the softwear for set resistances.
then,
the variable resistor has a tolrance, the holes in the resistor/gauge allow it to be rotated,
theres a little magnetic link in there which also dosent help and last but not least 1 multi valve can go in different tanks!

every install is going to be different and the scale might not also be linear. the float arm certainly inst linear
if you take a while to set the gates up there actually quite good
however it takes around 1 month of running to do it really so its not something you could ever inspect an installer to do

[LPGC]

As said, for most installs there will need to be a fair bit of gas in the tank before the float will even lift from the bottom. so the empty position is likely to be the light that would be on longest if each tankful was run to empty.

Another thing seems to be the resistors (or resistive strip) used in senders seem to be affected by temperature.

Installers could make better effort than they usually do, for instance, most installs I see have fuel gauge parameters set to default settings. with experience it is at least possible to guestimate more accurate parameters, the installer always has the opportunity to calibrate the empty, quarter and maybe half full lights.

Simon