Orangebean wrote:
super4 wrote:
But here is a question for you- if that vacuum lead is broken does it affect the running ? It has been broken for much of the cars life and running problems have only recently happened
No- cruise control has its own vacuum pump and is completely self contained. Won't affect running in the slightest..
Got me scratching my head now! I have no P38 here to check on at the moment but I was under the impression cruise control on them was plumbed to manifold vacuum? Sure I've heard and corrected vacuum hissing from broken cruise control vac pipe at the firewall end of passenger side engine bay in the past, correcting cruise control function in some cases, and the other end of that pipe went to the manifold..
Orangebean wrote:
Simon- if I can just jump into super's thread and ask a quick question re the 2 banks thing?
You say:
"On an older system setting 2 cylinder banks can affect a bit more than just allowing the twiddle factor for bank 2"
What else can it affect?
I've just replumbed the LP side on mine (Zavoli Zeta S, AEB2856C, Matrix injectors) to achieve an equal pipe run to each of my Matrix injectors- previously had 1 pipe at 600mm long and 1 at 300mm long from the splitter.
Pinj numbers are now out of balance, as you'd expect, so need to rebalance, but what else will be affected?
Ta
Now back to super!
Early AEB systems are not truly sequential, they base fuelling for all LPG injectors on only the blue petrol injector wires (normally front cylinder / same channel as gas injector A and E for banks 1 and 2). Really early systems didn't even have bank trim facility. Later AEB systems are truly sequential (just about anyway!). Some not truly sequential systems can still see firing order, earlier ones can't. Where they can't see firing order firing order is presumed but changing number of banks can change presumed firing order... But since gas injector ABCD still equate to petrol cut wires blu red grn yel, if teh system supports sequential cylinder changeover (which isn't the same as the system being truly sequential) even if gas injector firing order is changed it might still switch between fuels without engine hesitation.
Customer here now, I'll recheck I've replied properly later!
You definitely need to get it running right on petrol before you do anything with the LPG.
If the lambda wires are connected (and they by no means have to be connected) the readings should fluctuate regardless of what fuel you're running on if the engine is running as it should.
This line doesn't over-ride 'running right on petrol first' - I didn't realise you'd adjusted any settings yet. If it previously ran OK on LPG (and we assume LPG calibration really was OK) and the only setting you changed was reference pressure from 1.1 to 1.5 bar (and you haven't run autocal) it might be a good idea to change ref pressure back to 1.1, or calibration on LPG will be richer than before. On an older system setting 2 cylinder banks can affect a bit more than just allowing the twiddle factor for bank 2 but since the thread is probably confusing enough already I'll just say it is necessary to set 2 banks, even if there is the chance of doing so effecting how it runs on gas a little.
Simon
Nice pics Super4.
A bit of clarification on my last post...
I wasn't implying the vehicle will be running 25% rich or lean if actual pressure is different to reference pressure.
Calibration can still be perfect it software (reference) pressure is set to a different pressure than actual pressure, just that behind the scenes the ECU will be applying a more positive or more negative pressure compensation than it would if actual and reference pressures were exactly the same. I added more info on this too , explaining why installers sometimes have good reason for setting reference pressure different to actual pressure.
The info/pic more recently brought to light on Zavoli reducers (MUST having to be set at 1.2 / 1.5bat respectively for N / S reducers) agrees with what I said about Zavoli reducers, unlike the majority of reducers, being far less pressure stable if the allen key adjuster is adjusted to outside of 1.2 / 1.5bar.
The LPG ECU doesn't look directly to lambda readings at all during autocal. Autocal asks you to run the engine on petrol at a set rpm and keep your foot steady on the accelerator even if rpm changes over the whole process of autocal. Under these conditions it first runs the engine on petrol and monitors pinj, then switches to gas and monitors pinj while it adjusts ALL of the numbers in the map (every box in the map) by the same percentage until pinjs when running on petrol and when running on LPG in the ONE box that autocal used as the basis are same.
During autocal, the only reason rpm will fall/rise and/or engine load fall/rise will be due to if there is a different mixture when switching between fuels, however, autocal repeats the process a few times and as it nears correct calibration for the one map box it refers to, engine rpm and load will stay nearer to the same when the switch of fuels occurs, so by the time it reaches it's last comparison rpm and load won't shift when it switches fuels at all - So even though the box on the map that is applied during the early stages of autocal may shift, towards the end of autocalas long as the operator keeps his foot steady on the accelerator, the box that is the basis for the process won't change.
At this point autocal has hopefully set your map so that at least the box on the map that applied to rpm and pinj during the autocal conditions (fast idle maybe 2500rpm maybe 3.5ms pinj) is set properly. This is where it gets a bit more involved and refers to the shortfall in autocal's abilities... Before you ran autocal you will have entered type of injectors in settings, the type of injectors you selected set the default numbers in the map before you even ran autocal and the transition of those numbers across (rather up and down) the boxes in the map is what we refer to as the map 'shape'. Since autocal raises/lowers every figure in the map by the same percentage, autocal doesn't change the shape of the default map. The shortfall with the default map revolves around the fact that at best it's shape could only ever be an average map, where average means best attempt at suiting the average vehicle...but this default shape probably won't suit even one specific model of vehicle. So ultimately having ran autocal the user must then still change the map manually to make the shape of the map suit the specific vehicle... After running autocal, you should still do manual tuning.
TBH I don't even run autocal's anymore except on systems that demand that it is run before allowing normal operation (BRC and ESGI two such examples). It can be quicker to manually select all the numbers in the map and adjust all of the numbers in the map by a percentage (or even not a percentage, I prefer to work with linear changes because in my experience default maps have numbers too low with too great a map slope). 2500Rpm with lower than idle pinj isn't really a much seen real world operating condition, so if I change all numbers in the map by a percentage or figure at all... I might do my first ever switch to gas on a new install under similar conditions to prevent the engine stalling due to a system that at first contains only air but I'm unlikely to get as much insight into how much of the map will look from these conditions - which pretty much says it all as to the value of autocal on a lot of vehicles. P38's / injected Rover V8's are such low demand that you'd probably get away with just running an autocal, but still the results of just adjusting pressure reference and running autocal without doing any further adjustments might mean trims/mixture is less correct than when you started out.
SImon
Gov might have a point if they say 'Youre running your cars a bit cheaper, you can afford to pay more for domestic electricity now, and transport is good for the economy, especially now it's green, etc etc'.. But I largely agree with your point on this, as in people won't stomach raised domestic or industrial leccy prices even if transport costs are offset by it. Still, you won;t be charging a future electric car with 300 mile range from a 240v 13 amp socket and any special socket wired to your house to charge an electric car could very easily be fitted with it's own meter... sorted as easily as that. Domestic and industrial leccy the same price as now, car charging leccy to include 70% duty?
Yes a 2.0tdi has more bottom end torque, but I don't think it's fully fair to say this is 'accessible power' because the power band of diesels is generally much shorter than that on a petrol and inside/outside that power band equates to the engine making all/nothing power. The petrol has a wider power band and at least will make power at 5000rpm, furthermore that power band comes in and goes out more progressively. To really push both cars you'd need to do a similar number of gear changes, it is perhaps no coincidence that 6 speed gearboxes became more sought after as diesel engine power bands got more all/nothing... but you could probably get away with fewer gear changes if instead of the 2L TD you had a 3L petrol n/a which might cost a very similar amount to run. It's also not fair to compare what's wanted for road cars to what's wanted for race cars but the point about an n/a petrol of same size and same bhp as a td diesel generally being capable of going faster around a track probably does say something about what type of engine is more flexible in terms of power bands, and throttle response is surely another aspect of nice drive-ability?
Depends on what we call very recently. It doesn't seem long ago that diesels didn't have turbos, now we have lots of run of the mill factory petrol turbos such as 1.0 Ford Ecoboost engines fitted in Fiesta's etc and not designed to be high performance... but would certainly show up a 1.0 turbo diesel.
Agreed on short trips no good for diesels. Ignoring for the moment the other points on stink and noise, lack of long journeys damages diesel dpf's and egr's. That very low rpm torque diesels are so good also implies serious engine vibration on diesels with less than 5 cylinders, which is why they need dmf's, but is also the reason they break dmf's much more regularly than petrols, and dmf's are rather expensive... In the comparison between the 2L TD and the 3L petrol above, not much guessing on which would be more likely to need a dpf, dmf, egr. Like you say, stop start driving in diesels can be a false economy. Until recently I though Adblue was something necessary to brake down the crp in chemical toilets. Just reminded myself of my mate who has some unusual greetings such as (first ever sentence to a stranger) 'Have you had a sht this morning?' and (to old mates) 'Long time no see, are you still driving that old sh*tty toilet around?' I can imagine what he'd say to anyone pouring adblue into a car he didn't like lol.
Simon
The option to start on gas will be shown in modern software even if your ECU's firmware is too old to support it.. The option to start on gas may or may not actually work.
You've got to press enter for any changes you make to stick.
As you were told by Jaymic, if you make any changes to the pressure setting you will have to do a full re-calibration. The pressure setting is the pressure at which the system won't factor in any compensation for pressure when calculating gas injector pulse duration, i.e. if actual pressure reading is the same as the reference pressure you set in software it doesn't need to factor in any compensation for pressure being actually higher or lower than your reference pressure. A higher actual pressure than reference pressure will result in the LPG ECU pulsing gas injectors for less time and vice/versa. So, if you make changes to the reference pressure you also have to do a full remap. If you simply change reference pressure from 1.1 to 1.4 bar without doing a remap, the system will compensate for what it now sees is an under pressure situation by pulsing injectors for longer duration (1.4/1.1 = 1.27), so we might expect your mixture to be richer by around 25% unless you remap.
AEB systems are particularly good at compensating for pressure differences accurately. The following isn't to say that it would be a bad idea to change your reference pressure from 1.1 to 1.4 and re-calibrate but is something to consider especially if your system is older... Some of the older AEB systems don't allow you to set the pressure at which the system will switch back to petrol, on such systems (and behind the scenes) the change-back pressure is around 50% of the reference pressure, so that might be 0.55 bar with reference of 1.1 bar and 0.7 bar with reference of 1.4 bar. Your reducer is just about up to the job for fuelling a Rover V8 flat out but might expect pressure to drop a bit when flat out. If you set ref pressure to 1.4 bar, and if other factors are also at play, you may find it more likely to switch back to petrol when you put your foot down even though it might have been perfectly capable of running your engine flat out without switching to petrol. While (again) if you change that figure you'll have to do a full recal anyway. Save your config before making any changes anyway!
Simon
Imagine what the world would look like if 100% electricity came from renewables. Imagine how much extra battery capacity would be needed to store power from renewables... or if we're talking grid storage in terms of batteries in electric vehicles, going to be bed at night thinking you've got X range in your car and waking up to find you have half X range in your car and it would take 4 hours to charge back to X range.
Imagine comparing a 1400cc diesel turbo to a 1400cc petrol or LPG turbo, or a 3000cc diesel turbo to a 3000cc petrol or LPG turbo. Comparing a turbo to a none turbo isn't fair but I do remember a track test (by the likes of Top or 5th Gear) that compared a BMW 3L turbo diesel to the 3L normally aspirated petrol version of the same car and the petrol version bet the lap time of the diesel even though it had about 20% less bhp. For a fair test you'd have to be comparing the 330 turbo diesel to something like a modern M3, then come back and tell us which car was the nicest to drive ;-)
Easy to tax electricity in the same way as petrol really - If 1kwh costs X, force the electricity suppliers to charge gov's 70% duty on top of X.
My bit about China and India wasn't intended to be taken completely literally but raises a concept...They missed the boat in terms of early industrialisation, it may be cheaper for them to proceed in the same way the West did and go through the smog stage etc but they'd be doing it on a bigger scale and consuming resources that the Western way of life has depended on for decades, so perhaps it's just as well that they should be at the forefront of developing new clean technologies in order to progress even if it that slows the progress.
I would've clicked 'Like' on the last 4 posts but the facility isn't there ;-)
Had a similar discussion/argument with the mod of a UK Subaru Legacy forum (http://www.uklegacy.com/forums/index.php/topic/144446-tomorrows-world-of-motoring/) before getting myself banned for telling it how it is on a different subject.(explained below). If you read the thread bare in mind that the mod DAZRSK had previously taken every opportunity to criticise LPG conversion over various threads and it seemed to me he started this thread as yet another way of taking a pop.
A few owners of Subaru's all with twin scroll turbo's all fitted with BRC LPG equipment all had the same problems which their installers couldn't sort out even after many return visits. I predicted what the problem would be and drew a few conclusions about the systems and the installers. Everything I said was proven to be true but one guy with the problems took exception to my drawing conclusions (that were correct) because they painted his installer, who was his mate, in a negative light. He threatened the forum with legal action and implied his installer and BRC would also take legal action... He went from saying things like 'I thank Simon profusely for all his help on this' to threatening the forum resulting in my being banned and the whole thread deleted, but only after his problems were solved lol. Had any of the threats been directed at me I would have been happy to go to court with them.
I'm on good terms with the boss of the forum who banned me (not DAZRSK but I'm not on bad terms with him either), the boss is just scared of the hassle he's been threatened with.
Diesels sound crap, kick out crap from the exhaust and are crap to drive.
The electricity used to charge electric only vehicles (and run everything else that runs on electric) comes mostly from power generation plants that burn fossil fuels, this is likely to be the case until (and if) fusion reactors are invented. If fusion reactors were in place tomorrow the electrical supply infrastructure would still need upgrading before the majority of homes could have an electric vehicle.
Countries need revenue from vehicle fuel tax on the one hand but want everyone to be green on the other hand. If it currently costs a motorist something like 15p to drive a mile in a car and government get half the 15p in revenue, that's 7.5p gov gets for the average mile driven. As a nod to green gov allows reduced taxation on green fuels such as LPG and electric vehicles. If there comes a day when everyone runs an electric vehicle, the electric will have to cost at least 7.5p for every mile driven plus whatever extra needs to be factored in to cover the cost of upgrading infrastructure.
The savings enjoyed by running electric vehicles may only last until the electric vehicles get better and start to become more mainstream. While-ever anyone produces petrol, LPG will also be produced and governments that like to nod to green will tax it less heavily than petrol. We could even see a situation in future where it costs less to run a vehicle on LPG than electric... Electric cars are not totally clean, they shift the problem from vehicle exhausts to power station exhausts. Then we could talk about (ref Brian's post) what it takes to get enough lithium to produce batteries for all these electric vehicles and copper for electric cables and transformers.
In the future maybe we'll have to take our dead in our electric cars to recycling centres for rare earth element extraction lol. One of the biggest problems with regards oil use is the rising demand from developing countries like, say, China and India. What would happen if developed countries took a hard line on this - 'Sorry China / India, we developed first and have become accustomed to our way of life, but we're not going to let you do the same as we have/do with oil because there isn't enough to go round'. War? OK. So what if we say to oil producing countries 'We know you only have oil to offer and have based your economies and standards of living on selling oil to the West but we're not going to buy any oil in future, sorry'. War? Well maybe they'll start turning oil into more plastic products such as ships ,etc which become much cheaper than current metal ships but the new products bring problems of their own (plastic on the sea bed etc). Then we might say 'Sorry still no good, you're still causing pollution'. Is it war yet? If only we had nuclear fusion, and that Star Trek technology for producing food etc out of thin air (that runs on lectric).
Simon
Only thing is, at higher rpm in neutral the engine manifold pressure will be lower than at normal idle speed, and since gas pressure is relative to manifold pressure gas pressure should be lower at high idle speed so less gas should escape from any leak.
Yeh that's good!
Must have been quite a leak on the 4mm pipe. Even without a spring clip or even if you cut part way through that pipe would still expect the pressure sensor to read system pressure.
Simon
As said, a bit of a leak (hiss when you bend pipes etc) on the 4mm pipe between Matrix and pressure sensor is unlikely to be enough to cause a low pressure reading at the sensor. But you'll want to eventually fix this problem anyway and it would be a good idea to fix it now so you can at least rule this out.
Whilst on the subject of the errors page - The software may show an errors page but we don't yet know if your ECU/firmware are new enough to support error reporting. But it's worth checking to see if any faults register for gas vapour temp sensor... If the ECU picks up on such error it will switch back to petrol anyway.
Recap...
You have ruled out some potential causes such as blocked liquid filter and solenoid. It is very unlikely that the tank pickup pipe has come off, far more likely that there's a problem at the front end such as faulty reducer. The leak(s) as you described them seem to be inconsequential.
Simon
I think it's unlikely that this leak will be cause of the system reading low gas pressure - If the leak was bad enough to have sensor read low gas pressure you'd certainly have known about it before. Also, any leak that you can start and stop by moving pipes probably equates to only a small fraction of the gas that the engine could use if you put your foot down.... so although it might hiss and stink and generally be less than safe, such leak is probably negligible in terms of a proper functioning reducer's flow capability. The one exception would be if you pulled the pressure pipe to the pressure sensor just about all the way off but then you'd smell and hear it.
With the front solenoid guts removed there's still the tank solenoid, so if you create a leak on the vapour side gas will leak for quite a while (until liquid gas between tank and reducer is used up) before dying down.
You can't use petrol hose for LPG.
5mm or 6mm internal diameter pipe between injectors and manifold, 4mm internal pipe between Matrix injectors and pressure sensor.
Simon
super4 wrote:
Have checked wiring and things I can get to, managed to get 12 volts to the solenoid on the reducer which sort of worked but did not seem as active as the one on the tank - it may be that the connection I was making was not very good. But what I want to ask is - does the pressure come from the simple fact that the gas in tank is under pressure and therefore makes its way to the intake manifold provided solenoids etc allow it or does the pressure get built by the reducer in addition. If the pressure is in the system all through then presumably I have a blockage or as suggested, a solenoid simply not opening ? Is it as simple as that ?
The tank solenoid will likely be a lot louder than the front solenoid because it is on the tank.. Tap the tank with a screwdriver and it will make a lot louder noise than if you tap the front solenoid with a screwdriver. Plungers in solenoid posts do get fouled with crud / heavy ends but if they open at all they usually flow enough to allow the engine to run OK and it's a similar story with filters. But you could rule out both the plunger and the filter by simply removing them temporarily to see if the same happens with them removed. Even if removing them seems to effect a cure (which might point to plunger post or filter problem) the real problem could still be voltage drop when the injectors pulse for longer duration (could still be main power feed issue).
Simon
Orangebean wrote:
Handy hint, not LPG related, but...
Wheels stuck on hubs- loosen wheel nuts (sufficient to allow wheel to wobble on hub but not fall off). Drive car forwards and backwards a bit while turning steering from lock to lock.
Most of the time just loosening nuts and turning steering from lock to lock will do it.
Much easier on the back!
In my boy racer days I followed a group of lads who were in a mk1 Escort down the A1 at 100mph, they pulled up due to steering wheel wobble, turned out they'd changed a wheel earlier but can't have done the nuts up properly because 3 nuts were missing and the one remaining was nearly at the end of its threads!
Sorry I missed this thread until now.
Unlikely to be a case of solenoid not opening enough to allow enough flow of liquid gas for idle but not enough flow of liquid gas for blipping the throttle or light acceleration... Unless the electrical feed to the solenoid is intermittent, in which case the solenoid might close when the throttle is blipped or in light acceleration - but unless this was due to a main power feed problem (to the LPG ECU) it would more likely be due to coincidence.
I have known plenty Zavoli reducers fail in a way where the maximum gas flow is severely restricted, in some cases the restricted gas flow has been enough to maintain vapour pressure for idle but for no greater load than idle... so blipping the throttle or light acceleration would cause pressure to fall away very quickly (can't maintain pressure of volume throughput is restricted, so pressure falls away, system senses lack of pressure and switches back to petrol). The situation described here would call for a new reducer but this may not be the case and it's worth doing other checks first.
The gas vapour temp sensor will need fixing, it's worth checking the solenoids and for issues such as bad main power feed / earth.
You could monitor voltage to the solenoid(s) while switching to gas, at idle on gas, and when you blip the throttle. The injectors pulse for longer with higher engine loads (such as when blipping the throttle) soif there's a bad main power feed to the LPG ECU the extra current drawn can cause voltage to all LPG components to lower which could cause solenoids to close or sensors (including pressure sensor) to read incorrectly.
After the system has beeped and switched itself back to petrol you could read vapour pressure on screen again, N/A (not applicable) means no pressure... If you do a pinch test on a vapour pipe (or disconnect a pipe) you'll know if this seems to be a true or false reading. False reading points to pressure sensor, voltage or ECU problem. True reading points to solenoid, voltage or reducer problem.
Another thing about Zavoli reducers is that they only really like to work at the pressure for which they were designed (1.2 or 1.5 bar depending on model). Like other reducers they have a pressure adjusting screw but on Zavoli reducers if you turn pressure down they tend not to be capable of flowing nearly as much gas (throughput) as at standard pressure.
This is all assuming the feed from the tank is good - If the tank were outputting vapour instead of liquid (on a single hole tank, particularly 30 degree tanks) if the flexible pipe that picks up liquid from the bottom of the tank has come off inside the tank the tank will output vapour. The reducer needs to be fed liquid gas but might flow enough for idle if fed vapour.
Simon
Yesterday I removed a complete LPG system from a Disco3 with Jag engine.
System comprises 95L 4 hole full toroidal tank (was underslung), Agis electronics and injectors, KME reducer, BRC style forced lube system.
The owner will be selling the lot, probably on Ebay.
He said he doesn't much like LPG anyway, and besides, Guy Salmom (his preferred Landrover dealers) won't have the vehicle in their shop with the LPG system installed, not even just to service it. I did mention he could have taken it to one of the many other LR specialists but he wanted it removing and had me spell out very specifically on his receipt that all of the LPG system was removed to appease Guy Salmom!
The tank was a pita to remove, bolts all seized and difficult to access... angle grinder job! Cut the rear bolts easily but couldn't get to the front bolts until I'd slipped a jack in between the tank and Disco underframe to pry the tank down a bit, even then I had to put a 9" cutting wheel on a 6" grinder to reach a front bolt... the big disk spinning so close to my hand in a hard to reach spot wasn't something I enjoyed! With only 1 bolt remaining I fatigued it off rather than repeat the risky cutting.
The bits would fit and work on other LR models including P38s but the front end bits would be a bit difficult to get spare parts for, The 4 hole tank will be the most valuable and useful part here but will need a bit of work - new solenoid post, coil, will come without a filler or JIC filling pipe, I'd advise a new level sender, it'll need wire brushing and paint.
If anyone contacts the seller don't mention this post ;-)
Simon
As Gilbert said.
The 4 hole design can also be mounted internally or externally.
The advantages of the Icom and zero degree design over 4 holes are that they don't have that big box on the side so you might keep your jack storage etc,.. and the flat valve base on 4 holes can mean you lose a couple of litres capacity compared to a 0deg tank of same dimensions, though not usually in 720x270 tanks.
Not that I'd be concerned but all internal pipes should run inside external venting.
Simon
Bigas was a surprise to me, most often where I've seen Icom tanks fitted on aftermarket installs the front end components are a bit unusual too (e.g. AG / Voltran / etc), there's nothing unusual about Bigas.
You'll probably need to change or at least fix the tank valve, would start by looking under the tank's gas tight housing cover to see what type of valve(s) it has, Icom tanks with this type of gas tight housing may have one of a few different types of valve setups. Your type of tank has it's advantages over single hole and 4 hole tanks, I could think of a few applications where they'd be ideally suited but I'd like them more if parts were more easily available.
It is possible to safely fit a normal 95L zero degree (external) tank in the wheel well of a P38. The old way of doing this was to have the valve protrude through the side of the spare wheel well at about 315 degrees (too far North and the anti-roll bar will catch it, too far West and the valve won't reach the tank from the outside of the wheel well due to the squared off bit of well where the jack etc fit)but it's a lot of extra work for the extra few litres the full toroidal external tank offers over the 30 degree internal tank of same measurements because not only does the valve have to go through the side of the well in the correct position but also brackets have to be made to secure the tank from it's top (which is where an external tank's mounting lugs usually all are... there are exceptions, but suppliers don't tend to know which of their external tanks have lugs at both the top and bottom). Internal tanks are secured with a couple of bolts straight through the middle of the boot floor. In case you ever decided to change tanks rather than fix yours, these days it would be possible to fit an external tank upside down in the wheel well using a flexible valve cover, with the tank upside down it's mounting lugs could then be bolted through the well floor saving work making brackets and work fitting the valve, this type of setup would also allow you to keep with much of your existing piping. But an internal 30deg toroidal wlll hold just as much gas if it's valve is bent a bit, is much easier to fit and costs less.
This valving arrangement by Icom is unusual on aftermarket conversions but many are fitted on factory converted vehicles.
They're a pain when they go wrong and it seems yours might have gone wrong. If the problem is only at the tank end then it's likely the float is sometimes sticking or the fill shut off sometimes doesn't work. Or one (or both) of those issues could be constant, in which case the reason you don't always get 95L in it would be because you don't always run the tank empty, maybe because of a problem with the under bonnet components.
Obviously something went wrong at the tank end when you managed to get 95L in it, so, perhaps different subject - do you know what parts are fitted under the bonnet?
Simon