Warning: Final Editions fuel indicator

[Inversed]

I would offer a word of advice for those who plan to take their Final Editions on track. I don't know if MX2E dash or analog dash work the same way, but FE fuel indicator is lagging on track A LOT. Nowhere close to any car I've ever tracked. It showed 20% after stint, and I was thinking to do another short one before going to petrol station. Thankfully I decided to go to petrol station, before I even left the track the indicator was on 0%.

I think the other FE Exige on track ran out of fuel on track while still showing 20%.

[Paul_D]

That's really strange. There should be no real way for the system to lag...unless AIM have built in some damping and massively messed up the numbers. The fuel level is reported constantly to the dash, and on the aftermarket AIM dash seems to change in increments of around 0.8% at a time. 

Here's the AIM trace from my last day at Snetterton. 1 lap on the main part, and total session at the bottom.

 

One thing that can appear as a lag on all versions is the fact that when you brim the tank, it goes a fair bit higher than the top of the level sensor. So you can do a fair few miles before the reading starts to drop. From that point on though the indicator should be accurate. 

I'd be very tempted to try filling the car up with some fuel while the ignition is on and you can see the fuel level change. That might clearly show you any lag in the system and allow you to highlight it to Lotus if it's an issue. 

[Inversed]

I will cross reference AiM dump to what I know was visible state to see if display is heavily damped.

[Paul_D]

Yeah, that would be interesting. 

Were you looking at the standard bar graph indicator? If so, you could also set the dash display to show the numerical value for ECU FUEL LEV and see if there is any difference between them. (Edit - Can you actually do that on the Final Editions once you've accessed them? Or is the display completely fixed?)

Edited March 14 by Paul_D

[Inversed]

So I did look into the last stint + drive to petrol station and data is quite worrying.

Here is a snippet of 7 laps, track lenght is 4.1km (time around 1:45). And then last long part is very chill drive to petrol station, also around 4km away. ECU FUEL LEV before lap1 was 37%, exiting track 23% and at gas station 4%. This prettymuch corellates with what I described in previous post.

To answer Paul_D-s comment, there is no % view option on FE dash.

[Paul_D]

That's bizarre.

I think that it points to a measurement issue itself rather than any sort of lag. Your data shows a drop of 14% in 7 laps. I had a drop of 15% in 7 laps of a slightly longer circuit, so that seems to tally up quite nicely. You wouldn't genuinely use 37% fuel in just 7 laps. 

I'm trying to think what could make it exhibit that sort of behaviour, and I'm struggling to be honest. 

[Inversed]

I think other driver needs at the track day needs to add his input here, he is on forums. I drove 40 laps in total, went to gas station after 30ish. He was out of fuel after 35.

35 laps approximates to 140 race kilometers, which I think is about normal. If you say 15% drop is normal in 7 laps, that would mean you can do 46 laps * 4.4km = ~200 race kilometers. Ofcourse it depends on track type aswell, this one is quite fast flowing, 2 times over 220kmh, never under 100.

[Paul_D]

Sorry, what I meant was that our fuel readings seem similar in the area where it seemed to be working ok. I'm not sure how accurate the absolute values are.

I've been looking at more of my data, and the results do seem to be a little strange. I checked the readings at the end of one session, compared to the beginning of the next and there is a definite change after the car has sat for a while.

The change between turning the car off at the end of one session, and turning it back at the start of the next session varies with tank level too. 

Examples.

Turned car off with 85%. Turned back on at 90% Increase of 5%

Turned car off with 69%. Turned back on at 71% Increase of 2%

Turned car off with 60%. Turned back on at 52% Decrease of 8%

Turned car off with 31%. Turned back on at 17% Decrease of 14%

I checked this for two completely different days, and the results were similar. 

This very much ties in with what you're saying, in that the car over-reads by quite a lot at low levels until it 'catches up'. I probably haven't been caught out by this yet as I try not to run the tank too low in case I get fuel starvation.

I still have no explanation, but there definitely is an issue to be aware of, so thank you for highlighting it. 

[TonyKL]

FWIW on my old Exige V6 it would "lose" fuel whilst parked up - come in after a session with it showing 1 bar and then start the car 20 minutes later and it's almost empty.  I'd say it's more of a Lotus issue rather than AIM.

On my 250FE the fuel shows empty on the dash but still have about 8 litres in the tank !

[jimichanga]

Fuel readings are always problematic on track because of surge due to cornering g, depends where fuel sensor is but Exige V6 drinks heavily on track in any event. Easy to get caught out if you push your luck.

[Paul_D]

Obviously you’ll get surge on track, but most cars are still relatively accurate in my experience. The reading isn’t live direct from the sender because it would be very erratic. It’s fed into the ECU and the output is calculated, I assume using a fairly clever (normally) algorithm to account for surges and such.   

Indeed, the data traces above clearly show no particular change in reading around the bends, and the sender is fully over one side of the tank in the Exige attached to the fuel pump  

What’s unusual here is that it appears at higher tank levels (above 70%)  the ECU is reporting a lower than ‘real’ tank level when out on track. When everything settles the reading goes up.

The opposite is true as the level falls, and the error becomes progressively greater until you’re reading 15-20% high when you’ve actually down to less than 20% ‘real’ level  

This is not an affect I’ve noticed in any other car I’ve taken on track. (Although that’s not to say it hasn’t happened and I just haven’t noticed)

 

[Toqcars]

Hey, I am the other driver who run out of the gas on the middle of the track 😂

Few points:

- on the road, last half of gas runs out way faster than the first half. My friend with an Exige Cup430 FE also confirmed it. It's like last half is like 20% of the whole range.

- I was doing the first trackday on the same track just fine. Went on the track with the 20% of gas, finished it nicely. I was quite slow at that time, so less side accelerations

- Two days ago run out of gas when the car was showing 20% of gas. Was quite quick, so quite a few side accelerations which I guess affected the measurement as well. When I was waiting on the track for 10-15 minutes (there was a massive crash at the same moment, so they were cleaning it), fuel indicator dropped to the 0% and switched on the light for the gas. Only difference was time and standstill status of the car. No fuel consumed, because no fuel 😂

 

So my point of view is that the fuel indicator is affected heavily by the longitudinal accelerations. Equation integrated into the dash for calculating the gas capacity is not correct (if there is an equation). There could also be some lag reporting the level... some bug in the low-pass filter to be 10x longer? There could also be some offset on the sensor.

My solution was to take some spare gas on the track, never go on the track with less than 50% of gas (yeah that's the new point, 2 days old). When on road, refill when under 50%. 

 

Question: if I integrate the AIM dash... I could fix the equation for the gas indicator, correct?

[Paul_D]

There is definitely a difference between the first and second half of the tanks. A big part of this feeling may be because you can do several miles before the indicator will start to move. On the more powerful cars with the bigger tanks, the fuel pump / level sensor is mounted lower than the top of the tank, as per the diagram below:

On the S and the 350, it's mounted level. 

 

That doesn't explain most of the problems we're talking about though, when the indicator error level changes depending on how full the tank is.

Interesting that you run at at 20%, and yet I've been out on track at 17% with no issues. Maybe I was then about to run out? 

With regards to: Question: if I integrate the AIM dash... I could fix the equation for the gas indicator, correct?

I might be wrong, but I think that all of the calculation will be done within the ECU of the car and not in the AIM dash. It sounds like people have seen the same thing happed with the standard dials. I don't think there's anything you'd have access to in the AIM dash that would allow you to compensate for this. 

I guess the important thing is that now we're aware of it, we can try and learn from it. I will start trying to note what my fuel reading is when out on track and what it changes to when I re-start. I will also try and start noting how much fuel can be added from different levels. Eventually I should roughly know how much fuel I've got. 🤣

[Toqcars]

Thanks for clarifying that first half of the tank is different. I was also just fine on the first track day with less than 20% of gas... but I was a bit slower.

The best would be to totally empty the tank and check fuel level via OBD or something similar. Add some amount of fuel and check again. We could create a correction table. If possible, correction table would be integrated into the AIM later.

Or to weight the car with a different level of fuel (with the exactly the same luggage). 

[Paul_D]

Because this is potentially quite important, I decided to do a little more data analysis tonight. I basically put every single session I've done into an Excel chart. The key values I was interested in were the final fuel level reading when out on track (i.e. level on the last lap of a session) and the values once everything has settled / caught up. I took this as the first lap of the next session, and considered this to be the 'real' tank level. The data is from multiple sessions over 4 separate track days.

This data is obviously a bit hard to interpret at a glance, so I put every single entry into a scatter graph. The graph plots the error when out on track (in %) vs the 'real' tank level. Despite the data coming from multiple different days and times with refills in between, (and with some anomalies) the correlation is clear to see.

If the tank level is above 80%, the reading on track is likely to be showing less than this. 

From 60 - 80 % the reading is likely to be within 5% accuracy in either direction.

Below 60% the error becomes greater and greater with the level on track reading higher than actual tank level. The trend line of the graph appear to head towards a tank reading of roughly 22% when the tank is actually empty. This obviously matches very closely with what Inversed and Toqcars have found.

[Toqcars]

@Paul_D that is a proper scientific approach. Great data extraction (last and first lap) and nice correlation of the data. So now we know what is going on. Question left is whether we would see the same behaviour of the fuel level indicator on a road. Any idea how to test it? Because I am pretty sure that indicator is wrong when driven on roads/highway as well.

Thanks Paul to provide correction function for the fuel level indicator. Now we only need to find out how to implement it into our cars/AIM dash

[Paul_D]

I don't know the exact behaviour during 'fast' road driving, but what I can confirm is that during a journey home with 0% left on the indicator I was still able to drive for several miles. 

This (very strongly) suggests the the indicator is not exhibiting the same problem of over-reading at low levels.