Fuel Sender Troubleshooting

Fuel Sender Troubleshooting

Author
Discussion

Swiss_Toni

Original Poster:

412 posts

182 months

Wednesday 17th September 2014
quotequote all
Before I go and buy a new fuel sender(s) I was wonder if anyone has any tips in troubleshooting the fuel level system?

I've got the original VDO tube type sender and stack gauge (which can be calibrated to sender).

I calibrate the gauge (I've done it a number of times) but other than full it really doesn't show the correct fuel level, especially on part tank.

It's the same on both tanks.

Any thoughts, do the senders sometime "go"?

anonymous-user

53 months

Wednesday 17th September 2014
quotequote all
I've never yet found a broken sender, but that doesn't mean it's not possible!

Two issues i do find:

1) the output voltage range of the senders is small, so errors can be large due to electrical noise or changing system voltages

2) The senders are capacitive, and under "sloshing" conditions seem to get "hung up" a bit and report more fuel than is actually present.


If you can remove the senders in situ (without taking tank(s) out!) it's pretty easy to check they work with a basic multimeter and a nice tall jar or glass full of fuel to dip them into! (usual caveats regarding highly flammable liquids, ignition sources/sparks and safety apply here!! )

Swiss_Toni

Original Poster:

412 posts

182 months

Wednesday 17th September 2014
quotequote all
Interesting....

As both side are reacting similar and I don't believe at this stage it's a sticking sender (sloshing).

That leaves wiring...

There are two senders, 1 gauge and a switch. I switch between the tanks which operates the low pressure fuel pumps (1 for each tank) and selects which sender is used.

Steve_D

13,737 posts

257 months

Wednesday 17th September 2014
quotequote all
Have they ever worked? With the follow on question you did remove the transit pins when they were installed.

That aside there have been posts in the past where owners have commented on the foam infill degrading so it is possible the float is jamming inside the tube. It is possible to remove the senders without draining the tank provided the tanks are not full to the max.

Next possibility is the electrical connections. As Max has said there is not much current so a high resistance along the way will cause problems. The tank change over switch could be the culprit as it is rarely moved. Try flicking it back and forth many time to help clean the contacts and see if that makes a difference.

Steve

anonymous-user

53 months

Wednesday 17th September 2014
quotequote all
Sorry, just noticed. No idea why i typed "capacitive" up there^^ in my earlier post, i mean't resistive obviously, hence the fact they can get "jammed" on their tracks a bit. I was thinking that you could fit the capacitive type to avoid this......


Might be worth breaking the signal wires out to a multimeter zip tied somewhere in view in the cabin and go for a drive and see if the output voltages are sensible. (the actual level gauge has a very long term averaging input circuit (or digital algorithm) so it doesn't show slosh and has a steady display, but a direct measurement of the voltage on the gauge might be a good idea as a first step)

MarkWebb

983 posts

216 months

Wednesday 17th September 2014
quotequote all
I have connected both of mine to an accurate ohmmeter and filled the tank litre by litre from empty charting the resistance as I go. The readings have been input into my Motec dash and work very well except when aircon is on one of them reads v low. As compressor cycles off and on with cabin temp it jumps up and down. This is obviously some kind of interference or volt drop which I have yet to get to the bottom of but clearly shows that they are veer susceptible to this. As an aside the ohm readings only begin to change once there is approx 1 gall in the tanks so I have this set as zero. Comments as to where the fault might be in the wiring are very welcome!
Check yours and if they don't change much then transport pins may be the culprit.
I will try to find my readings so I can give you an idea of what they were full and empty.

MarkWebb

983 posts

216 months

Wednesday 17th September 2014
quotequote all
2 ohms full and 72 when empty as an approximation. I can give you them all if you are interested.

Swiss_Toni

Original Poster:

412 posts

182 months

Thursday 18th September 2014
quotequote all
Thanks for the tips.

1st step, breakout the multimeter

I'll take the switch out of the equation and maybe "slave" in some wires to eliminate that as an issue if there still is a problem.

F.C.

3,896 posts

207 months

Thursday 18th September 2014
quotequote all
MarkWebb said:
work very well except when aircon is on one of them reads v low. As compressor cycles off and on with cabin temp it jumps up and down. This is obviously some kind of interference or volt drop which I have yet to get to the bottom of but clearly shows that they are veer susceptible to this.
I have also experienced the same thing, LHS tank I think,
Your Ohm readings would be very useful.
Cheers.

MarkWebb

983 posts

216 months

Friday 19th September 2014
quotequote all
LH tank is correct. Interesting this may be a fault inherent in the loom. What gauge/dash are you using?
L=Ohms
0=72
2=68
4=65
6=62
8=59
10=65
12=53
14=50
16=47
18=44
20=40
22=37
24=32
26=28
28=23
30=17
32=11
34=2
There are other settings within the dash(my Motec anyway) that may effect the way that the gauges read. Such as pull-up resistor size and measuring method used by the dash.
I have 1000ohm pull up and variable resistance as method.

anonymous-user

53 months

Friday 19th September 2014
quotequote all
MarkWebb said:
LH tank is correct. Interesting this may be a fault inherent in the loom. What gauge/dash are you using?
L=Ohms
0=72
2=68
4=65
6=62
8=59
10=65
12=53
14=50
16=47
18=44
20=40
22=37
24=32
26=28
28=23
30=17
32=11
34=2
There are other settings within the dash(my Motec anyway) that may effect the way that the gauges read. Such as pull-up resistor size and measuring method used by the dash.
I have 1000ohm pull up and variable resistance as method.
No wonder the reading are all over the place!


Using a 1k pull up means you are only using 7% of the possible range of the fuel measurement input!

For example, if we assume the fuel gauge is using a 10bit ADC and a 5v reference voltage:

AT EMPTY (72ohms gauge resistance)
Signal voltage = 336mV
ADC counts = 69 (out of 1024 max)


AT FULL (2ohms gauge resistance)
Signal voltage = 10mV
ADC counts = 2 (out of 1024 max)


And because the signal is biased towards zero volts due to the large impedance mismatch any "ground bounce" or ground voltage level offset between the fuel gauge/dash and the tank sender will MASSIVELY effect the level reading (ie, for full tanks, your signal is only 2mV, or 0.002 volts! Considering the fuel sender is probably over 2m away from the gauge, and earthed at different points, no wonder the reading is all over the place!!


Hopefully you dash has a much lower output resistance pull up that is selectable??

MarkWebb

983 posts

216 months

Friday 19th September 2014
quotequote all
OK Max as an electrician(not an electronics engineer) I kind of get what you are saying.
I have never changed any of the pull up settings etc in my dash due to not really understanding what they do etc.
I always assumed that the pull up resistor was something physically put into the circuit or dash at the time of assembly. So I left these settings as they were in the dash software as I did not assemble it.
Do you have any experience of the Motec SDL (I think that is the model) dash to help iron out these problems?
It is the left hand sender which varies with the aircon off/on and this would be the longest set of wiring.

F.C.

3,896 posts

207 months

Saturday 20th September 2014
quotequote all
MarkWebb said:
LH tank is correct. Interesting this may be a fault inherent in the loom. What gauge/dash are you using?
I have a digital readout 52mm spa gauge at the moment, this only started happening after I fitted power steering and only with air con switched on.
I will investigate what the pull up is with this gauge.

anonymous-user

53 months

Saturday 20th September 2014
quotequote all
MarkWebb said:
OK Max as an electrician(not an electronics engineer) I kind of get what you are saying.
I have never changed any of the pull up settings etc in my dash due to not really understanding what they do etc.
I always assumed that the pull up resistor was something physically put into the circuit or dash at the time of assembly. So I left these settings as they were in the dash software as I did not assemble it.
Do you have any experience of the Motec SDL (I think that is the model) dash to help iron out these problems?
It is the left hand sender which varies with the aircon off/on and this would be the longest set of wiring.
Ok, quick summary of how these things work:


The "level sender" in the tank is a variable resistor in effect, whose series resistance changes linearly with the position (height) of the float in the device.

The Fuel gauge, dash, or other unit used to "read" the fuel level must be able to measure the resistance of the level sender, so it can report the fuel level accurately.

Unfortunately, measuring resistance directly is difficult, so the technique used is one where the dash and the level sender are arranged to form a "voltage divider", see:

Voltage divider Wiki

The resultant voltage can be easily measured as the output is now a voltage, that varies in proportion to fuel level.
In the "old days" that voltage was used to directly drive an old analogue gauge, with a needle, where more volts = more deflection etc. These days, we simply measure the voltage with a microcontroller (using it's internal Analgue to Digital Convertor (ADC)) and that chip then drives a digital display, stepper motor or whatever.

To form the voltage divider the measuring device needs to source current to the level sender. It does this by having a "pull up" resistor inside. This is used to pull up the signal line to the sender, and the resistance of the sender "Pulls in down" to ground, depending on it's resistance.

The important point is that to maximise the resolution of the system, and minimise the effects of any noise or voltage variations, we need to arrange the voltage divider to output the highest possible voltage "swing" for any given input change.

In the case of the Motec dash mentioned above^^^, using a 1000 ohm "pull up" setting with only a 72ohm sensor results in hardly any voltage change over the output range. This is because the 1000ohm resistor dominates, being many times bigger than the 72ohm sensor (ie, max resistance is 1072ohm, min is 1002ohm, which is not much change considering the 1000ohms!)

Using a lower resistance pull up, say 100ohms, will massively improve the systems resolution and noise rejection (by 10 times! as maxR = 172ohm, minR = 102ohms). The downside is that more power (and hence more heat) is used by the voltage divider. Generally, with devices that have "selectable" pull ups, the lowest possible setting is chosen to still be high enough to avoid damge to the unit in all cases.

anonymous-user

53 months

Saturday 20th September 2014
quotequote all
Sorry, should have added:

My assumption is that changing the pull up setting in the calibration for your dash changes the physical setting in the device. But i could be wrong, and you just enter the pull up value that is physically present in the dash to ensure you get the right output numbers etc?

I don't know how these dash's work in that respect. ie. is the fuel level pull up "software selectable" or not?

MarkWebb

983 posts

216 months

Saturday 20th September 2014
quotequote all
Time to download and read the manual I guess.

Steve_D

13,737 posts

257 months

Saturday 20th September 2014
quotequote all
As there has been comment on the readings changing when the A/C is on it sounds like the problem is voltage drop. A number of owners have found the ignition switch getting warm due to the current it is taking and (like myself) have fitted a relay beside the fusebox to take some of the load.

In this case this mod may help or failing that bring a supply to the gauge direct from the battery.

Steve

anonymous-user

53 months

Saturday 20th September 2014
quotequote all
Steve_D said:
As there has been comment on the readings changing when the A/C is on it sounds like the problem is voltage drop. A number of owners have found the ignition switch getting warm due to the current it is taking and (like myself) have fitted a relay beside the fusebox to take some of the load.

In this case this mod may help or failing that bring a supply to the gauge direct from the battery.

Steve
I'd suggest its more likely to be a ground offset between the earth of the dash and the earth of the sender units, especially if people are running with very low sender voltage ranges due to using large value pull ups etc!

MarkWebb

983 posts

216 months

Saturday 20th September 2014
quotequote all
It seems that my fuel gauges are connected to an input that has the 1k resistor built in. Other inputs require you to fit a resistor. Getting to the back of my dash dammed near impossible!

Swiss_Toni

Original Poster:

412 posts

182 months

Saturday 20th September 2014
quotequote all
Richard at the factory has been helpful as alway.

Among other things he's provided the figures for what the sender should read.

Full is 2.1 ohms
Empty is 71.9 ohms

I did VERY quick check directly on the sender and I got 3.9 ohms at full and 54 ohms with about 10l in the tank. The resistance moved smoothly as I filled the tank.

More troubleshooting is needed.