BMW 745i conversion

[v8gas]

I am doing my first DIY conversion on a BMW 745i.
I have a few questions but will begin with a description of my install. I fitted a 82litre donut tank in the spare wheel well-slight problem is that it will have to be removed when its time to replace the battery. The only alternative was a LPG cylinder at the back of the boot but this covers the parking brake mechanism which could be a real problem if there was a fault with the electric park brake- it does happen. Also the boot floor is covered with plastic mouldings about 2 cms thick.

Next headache was where to fit the injector nozzles. The inlet manifold is magnesium and contains a plastic/fibre rotating variable intake system so nozzles cant be fitted anywhere thats accessable. The manifold has an upper and lower section which are factory sealed together, they do come apart but you will risk breaking something. The only solution is to remove the whole manifold - took me 3 hours- not easy, then drill the lower part of the manifold- see picture. I drilled by eye - just got to have a good aim so the nozzle entry is clear of the manifold fixing holes. I drilled from the manifold port outwards to get precise positioning of the nozzle outlets.




I didnt try to remove the petrol injectors I removed the spring clips which fix the supply rail to the individual injectors and removed the rail leaving the injectors in the manifold, they are a push fit but get jammed in place with grit because of their verticle orientation. With the manifold off the individual injector mounts can be flushed clean and then the injectors pull out but are very tight.



One good thing is you get to see the condition of the inlet valves with the manifold off.



I used Hana 2001 green single injectors to get each one as close as possible to the manifold and mounted vertically, there is just enough room for them under the petrol injector rail. They are not fixed but the short length of rubber pipe supports them in position. In line with the regs I used crimp clips for the pipe. This all has to be done before refitting the manifold otherwise there is no room for access.



For the water supply I cut into the return from the transmission cooler to the water pump, this flows hot continuously. Its the bigger of the 2 pipes towards the rear. The other pipe flow is under thermostat control and the flow will stop. As it happens the only space in the engine compartment for mounting the reducer is right next to the transmission cooler.

I used 8mm supply pipe from the tank to the reducer because the KME gold reducer has an 8mm inlet and others appear to have problems with 6mm pipe. Luckily there is an empty slot in the fuel pipe mounts which run under the left side of the car. It is possible to do it in 1 run, just a bit fiddly over the rear subframe when you are working on a car on stands with 6 metres of pipe.

The gas on off button is fitted above the boot release button which is easy to see but discreet.

At this point I refit everything to run the car to check all was good and it is, the car runs normally.

I cant complete the job yet because I bought a stag 300 premium controller but the harness supplied is for a 300 isa2 which is missing the permanent connection to the OBD, not even the spade connector in the plug. I am waiting for the supplier or stag to offer a replacement, up to now Stag have not been helpful.

Now my questions
Is the manifold pressure used simply to cut off the gas supply or does it have other functions? As you may know due to valvetronic the manifold pressure on the BMW is held at about 50 millibar only.

Which wire is used to connect the Bosch LSU 4.23 oxygen Sensor (BOSCH 0 258 007 146) am I correct in thinking that I need to break into the red wire? See picture.


The petrol pump is under ecu control, when you drive faster the pump runs faster, can this be altered to save the fuel pump when on gas?

Thanks in advance to all the experts

[Gilbertd]

No pictures but I assume you tried to upload them directly which doesn't work. You need to upload then to something like photobucket and link to them. There is a sticky at the top of the General Discussion forum with instructions but keep the sizes below about 1200 pixels wide or it makes them difficult to view.

[LPGC]

Valvetronic models are not the easiest install to get right, particularly for a first timer. They don't always run with manifold near atmospheric pressure - Problems start when the installer manages to get the vehicle to run well on LPG in either Valvetronic or throttled mode but not managing to find a combination of reducer pressure, nozzle size and software calibration that works well for both modes - partly due to the differential between LPG and manifold pressure changing at a different rate to the differential between petrol and manifold pressure when the mode changes (you know about the petrol fuel pressure), problem made more complicated due to different response times between petrol and LPG injectors. Earlier Valvetronics seem simpler to get right than later ones. For most vehicle installs it is possible to achieve proper fueling with either pressure a bit on the high side and nozzles a bit on the small size or vice/versa but the fussiest Valvetronics (not all) will demand you get the combination of nozzle size and pressure that works for both modes. I think a lot of installers play the issues down, thinking they can get good results following manufacturer advice not to connect vacuum pipes etc.. after all, old advice (AEB systems) used to be simply enter 127 across the entire map and don't connect the vac pipes, calibration sorted.. Works on some (usually older) Valvetronics but not on all. When I see some Valvetronic installs the installer has seemingly set changeover temp higher and changeover delay longer than they usually would - so the vehicle only gets to run on LPG when the engine is nicely warmed up and in fully Valvetronic mode, hopefully... I get them to work properly in both modes, sometimes connecting vac pipes only to the reducer or pressure sensor, always expecting to take some time to achieve proper operation in both modes. Only ever had issues with one Valvetronic - drove perfectly on petrol or LPG but when booting it on LPG the canbus was seemingly effected even without the OBD connected, even when power feeds etc where moved to different power pickup points. Was honest about it on the forum but nobody else had a solution either!

Some systems try to get around the differential pressure question by calculating a fake manifold pressure, this doesn't seem to work either.

I've never tried fitting the nozzles in the position you have, hope that works well and think it might, your choice of injectors and manifold nozzle location might rule out putting all the engine covers back on though? It is generally good to have nozzles close to the ports but a possible downfall is nozzles are closer to fluctuating pressure when the inlet valve opens (so more critical that each LPG injector fires at the same point relative to it's inlet valve as other cylinders to avoid some cylinders getting more gas than others) and it is probably just as important on Valvetronics (higher manifold pressure meaning lower airflow speeds at low loads) to have injectors firing into an area of the manifold with good airflow speed - directional manifold nozzles can help, Dai might remember a certain supplier pushing those many years ago but all suppliers have sold them for a few years now.

Don't connect the lambda probe wires.

Never messed with the petrol pump operation, would expect problems if you try.

Lots of previous discussions on the forum about Valvetronic setups.. I personally charge a bit more for converting Valvetronics than 'normal' engine'd BMWs because I know to do a neat install on them and get them setup properly takes more time - but I do get them right.

Simon

[v8gas]

Thanks for the reply Simon

A little bit of background for you, I am a qualified car technician and worked for various manufacturers but not for many years now. I find I have to do a lot of reading to keep up with technology.

About the nozzles, I did buy directional but looking at the size of the intake feeding 2 ports I decided against them as being too directional. I would have preffered to use the adapters which fit under the petrol injectors but was unsure if it would still be possible to fit the petrol rail.

I have refitted the engine covers by the way.

The petrol supply line is at 50psi adjusted down by manifold pressure.

At initial start the secondary air injection system operates for a 90 seconds to heat up the cats, so I would delay changeover until everything is fully warmed up, giving the petrol system a chance to operate and give time for any lubrication effect from petrol to do its work. I would not switch over to LPG until the Reducer is at 30C, so its easier on the LPG injectors which I've read dont like temperatures below 5C.

The throttle valve is only in control at initial startup so valvetronic is always operational.

The manifold pressure is maintained at minimum 50 millibars below atmospheric mainly for the operation of the crankcase venting and petrol tank venting.The petrol filter relies on a manifold vacuum supply to adjust petrol supply pressure.

The throttle valve is always fully opened (apart from tickover) on the valvetronic controlled engine. The load is controlled by the closing down the inlet valves to 0.18mm lift.So manifold pressure will fluctuate only slightly with engine speed and will be the opposite of a throttle valve system. At idle with valvetronic there will be minimal vacuum (50 millibars).

The petrol ecu adapts continously to combustion, so if your short term fuel trim is slightly weak on LPG it wil richen the mixture to compensaete and so long as its within 10% will just weaken again on petrol without altering the long term trim.

For the interest of others fixing nozzles to this manifold here is a diagram. The blue shaded components are plastic/fibre within the outer magnesium housing, so if you tried to drill above the joint between the upper and lower castings you will find a double skin which will be difficult to seal.
Items marked 2 and 3 are the rotating parts which alter intake length above 3500 rpm.



Peter

[LPGC]

Hi Peter,

Can tell from the pics you do neat work and have thought things through.

About the nozzles, I did buy directional but

Yes, I nearly said with nozzles where you've mounted them directional ones wouldn't be an advantage. I brought them up in case you looked up some of those other threads on Valvetronics, they have been mentioned on those threads.

I dont understand about fuel pressure differential. If The petrol supply line is at constant 50psi then petrol injection will be little affected by manifold pressure, does the lpg reducer at 15psi have a problem? will pushing the reducer pressure up help to counteract this?

By differential pressure I just mean relative fuel pressure - as you know, petrol pressure is relative to manifold pressure on most vehicles (most having a mechanical pressure regulator linked to manifold vacuum) but like you say the Valvetronic has 50psi relative to atmosphere. So, when in Valvetronic mode, the relative pressure is 50psi because there is no manifold vacuum, but in part Valvetronic mode manifold pressure might be .5 bar (7psi below atmosphere) so relative pressure increases to 57psi at low loads, when the engine isn't running Valvetronic mode (instead uses the conventional throttle) relative petrol pressure might increase to 61psi because the engine will idle with about .3bar manifold pressure (.7 bar / 11psi vacuum) like a conventional throttle controlled engine.

The petrol system knows which mode the engine is running in and has a different map and set of fuel trims for each mode, if it didn't then when petrol differential pressure was at 61psi above manifold then open loop either fueling would be 20% rich or closed loop fueling would see fuel trims go to -20%.

The above is relevant regards how the LPG system is installed and calibrated. Picture the scenarios...

1. On most vehicles the LPG reducer would be connected to engine vacuum to facilitate LPG pressure being kept at a relative constant above manifold pressure, just like petrol pressure is usually kept at a constant relative to manifold pressure by the mechanical petrol pressure regulator which connects to manifold pressure. On most vehicles the LPG system's combined LPG pressure and manifold pressure sensor would be connected to both to allow the LPG ECU to react to fluctuations in relative LPG pressure and adjust LPG injector duration to keep fueling correct (with changing relative LPG pressure, as LPG reducers are not as pressure stable as petrol fuel pressure regulators).. If you connect manifold vacuum to the LPG reducer and the pressure sensor on a Valvetronic, LPG pressure then remains constant relative to manifold pressure (relative to atmosphere the pressure goes up and down with manifold pressure) while petrol pressure remains constant relative to atmosphere at 50psi. So, the LPG ECU sees that relative pressure (LPG versus manifold) remains constant and does not adjust LPG injector pulse duration on account of pressure but the petrol ECU decreases injector pulse duration to account for the increased difference in petrol pressure versus manifold pressure (the 20%). As the LPG ECU works as a slave to the petrol ECU's injector duration signals it responds by decreasing LPG injector pulse time by around 20% depending on the system calibration. Because relative LPG pressure has remained constant above manifold pressure while LPG injector pulse duration might have decreased 20% in copying the petrol injector pulse duration, the fueling can be expected to be lean by 20%...

2. If you don't connect manifold vacuum pipes to the LPG reducer or pressure sensor on a Valvetronic, the LPG pressure will be constant relative to atmosphere, much like petrol fuel pressure. So at first consideration this might seem a complete fix, but... Petrol pressure is 50psi over manifold in Valvetronic mode, rising to 61psi in throttled mode, a difference in the range of 20% which while running on petrol is fully expected and accounted for by the petrol ECU (as said above). LPG pressure might be 1.2bar (18psi) over manifold pressure in Valvetronic mode, rising to 1.9bar (28.5psi) over manifold pressure in throttled none Valvetronic mode. If you calibrated the LPG system in Valvetronic mode, in throttled mode the pressure would become 60% greater without the LPG ECU noting any difference in pressure, so the LPG ECU would not decrease LPG injector duration to account for the higher pressure. So now running on LPG in throttled mode the fueling might be 40% rich because the petrol system only decreases injector duration by 20% while the LPG system sticks in 60% more fuel without realising. The effect is of course due to the higher constant petrol pressure (50psi) compared to constant LPG pressure (18psi) - any change in manifold pressure represents a greater % relative difference in relative pressure for the LPG.

3. If you don't connect manifold reference to the reducer but do connect manifold reference to the pressure sensor, LPG pressure remains constant relative to atmosphere, so pressure changes relatively to manifold pressure and the LPG ECU sees the actual relative pressure. So now in Valvetronic mode the LPG ECU sees the relative 1.2 bar and in throttled mode sees the relative 1.9 bar. In throttled mode it decreases LPG injector pulse duration by 60% to compensate for the higher pressure, plus it decreases LPG injector pulse duration copying the petrol injector pulse duration. Result might be 80% lean.

4. If you don't connect manifold reference to the pressure sensor but do connect to the reducer - LPG pressure stays relative to manifold pressure but the LPG ECU sees LPG pressure relative to atmosphere, so in throttled mode the LPG ECU sees LPG pressure fall to from 1.2 bar to 0.5 bar and increases LPG duration by 60% while decreasing LPG pulse duration by 20% following the petrol pulse duration. Result is 40% rich.

LPG injectors are never totally linear, they take longer to open and close than petrol injectors, some are vastly better than others in this respect. If a nozzle size for a given LPG injector at a given pressure would flow the equivalent amount of fuel (when open constantly) as a petrol injector on a given vehicle, then since injectors pulse, for an LPG injector to flow the same during a pulse as the petrol injector pulse it would need to pulse for as long as the petrol injector plus the time it takes the LPG injector to open minus the time it takes the LPG injector to close. If the LPG injector takes 1ms longer than the petrol injector then for a petrol injector pulse of 3ms the LPG injector needs to pulse for 4ms, for a petrol injector pule of 16ms the LPG injector needs to pulse for 17ms.. a mutliple of 1.33 for 3ms but only 1.06 for 16ms. If we configured the multiplier in calibration to these points and all points between this would be the ideal multiplier map (K line) for this engine.

With a very linear LPG injector, using case 1 we could calibrate the system with a K line that would look similar to the K line of a none Valvetronic engine but could expect to compromise on fuel trims with +10% trim at idle in throttled mode, -10% trims in Valvetronic mode (thus steering away from too excessive extremes of fuel trims that would put the Mil light on by aiming for the average trim of zero).

With less linear injectors we can use injector latency to advantage in case 2. In case 2 with linear injectors we'd expect the system to be rich in throttled mode but when we do the maths it can work out OK, at least for idle... Just doing the maths for fairly linear injectors that take only 1ms to open we can see the effect - multiplier of 1.33 at 3ms petrol duration, LPG injector pulse of 4ms (actually flowing 3ms worth of gas) for correct mixture in Valvetronic mode. With vac pipes disconnected actual pressure is 1.9 bar but ECU still reads 1.2 bar, Petrol injector duration goes from 3ms to 2.4ms so with same 1.33 multiplier the LPG injector pulses for 3.19ms of which it is only flowing gas for 2.19ms, this is 27% less than the 3ms so fuel trims would only be out by 13% (from the ideal injectors 40%). With this setup, the situation improves if we use less linear injectors and improves if injectors become less linear as gas pressure versus manifold pressure increases (can take longer for some injectors to open as pressure difference across them increases)... No wonder in the old days with less linear injectors manufacturer advice was generally to use method 2. But in this case a flatter K line should be used...

Maybe needing to read a bit between the lines but you'll see how choice of injectors, nozzle size, working pressure, piping arrangements, calibration are all very much inter-related and how sometimes, with some piping arrangements, the best spec components are not the best components for the job.

Simon

[v8gas]

Hi Simon thanks for A very thorough answer.
aaaaaah now I see what you mean. I would always go for a rich rather than lean mixture with gas since the cat temperatures can rise several hundred degrees above normal when lean.

The fact that the N62 engine uses a vacuum pump for the brake servo does suggest that BMW dont expect much vacuum to be present.

I have never checked the actual pressure within the intake manifold but this is a quote from BMW

The throttle valve Ensures a constant vacuum of 50 mbar in the intake manifold This vacuum is needed to exhaust the blow-by gases from the crankcase and the fuel vapors from the activated charcoal filter.

50 mbars = 0.05bar = 0.725 psi very slight vacuum -its virtually atmospheric.

So maybe using atmospheric to control gas injection should work well especially if you avoid the complicated start procedure.

Also just read about the KME gold pressure limit valve venting down the vacuum tube. I would prefer to vent to atmosphere (is this compliant?), god knows what a variable intake full of gas may do.

Another quote
During the starting procedure at a temperature between 20 ºC and 60 ºC, airflow is controlled by the throttle valve. If the engine is at operating temperature, it will be switched to non-throttle mode approximately 60 seconds after start up. In cold conditions, however, the engine is started with the throttle valve fully opened, which has a positive effect on the starting characteristics.

The easy way I see to get around this is to delay switchover until the engine is warmed up.You can hear the valvetronic operate about 5 minutes after starting from cold.



Some other info which may affect LPG injector location If it favours 1 port or the other. The petrol injector is central, so maybe another reason to use LPG injectors which sit below the petrol injector. Myself I wouldnt worry about it -but you never know.


VALVETRONIC engine control system. So-called phasing is implemented to assist adjustment in the lower valve lift range. The intake valves of a single cylinder are opened synchronously up to a lift of 0.2 mm. Valve 1 then begins to lead (advance). Therefore, valve 2 opens with a slight delay behind valve 1 and catches up to valve 1 again at a lift of approximately 6 mm. From here on they open synchronously again.Phasing results in a lift difference between both intake valves of up to 1.8 mm in the lower partial load range. Consequently, the flow of fresh air is distributed asymmetrically.


Manifold Pressure Sensor on rear of manifold

Barometric Pressure Sensor within the ECM



Peter

[LPGC]

Peter, you're welcome.

I haven't previously read the info from BMW you mentioned but have noted much the same things as you've said while working on these engines. Would also add that there is the other mode I mentioned where manifold vacuum is held at constant .5 bar and that I've seen them switch to throttled mode for longer periods following engine start-up and following periods on over-run.

Some installers actually do just set them up so the LPG system can only give good results when the engine is in full Valvetronic mode. This can work well for maybe 99% of the time, the other 1% might be in some of the circumstances I mentioned above.

As an installer you have very little control over the mixture the engine gets with a sequential setup, particularly one with wide band O2 sensors that runs closed loop almost all the time even near wide open throttle at high rpm. The LPG system works as a slave to the petrol system so if the LPG system is calibrated rich, the petrol fuel trims will simply go negative to compensate and if that goes too far the Mil light will come on and the engine likely switch to throttled mode anyway, same on the vice/versa. The problem regards mixture is that if you don't set the LPG system up in such a way that the petrol system can realise the mixture it requires / dictates, then the Mil will come on and likely with all the open loop / restricted performance implications this brings. Minor issues that the petrol ECU notices, not always mixture related, can also see the engine switch to throttled mode on what might usually be a temporary basis but if the LPG system cannot provide proper fueling under throttled mode this will lead to mixture fault codes also being created and these faults might then prevent the engine switching back to Valvetronic mode at least until the LPG system has been switched back to petrol for a while.

You should never see the KME Gold PRV open. If it is connected to the manifold and does open the likely result would be the engine would stall and then the gas would escape from the air filter.. Some installers would pipe it to the manifold, another school of thought is to route a pipe to somewhere away from heat/ignition source under the vehicle - after all we don't run piping form tank PRVs to the inlet manifold.

Some other info which may affect LPG injector location If it favours 1 port or the other. The petrol injector is central, so maybe another reason to use LPG injectors which sit below the petrol injector. Myself I wouldnt worry about it -but you never know.

I can't see it being a problem either but it might be something to think about in case of problems later. Injectors do seem to be in the opportune position but if, say, the LPG ECU starts LPG injection on cylinder 1 immediately on receiving the start of the petrol injection pulse for cylinder 1, perhaps with cylinder 1's inlet valve just starting top open... then for cylinder 2 there is a slight delay between the LPG ECU receiving the start of cyl2's petrol injection pulse and beginning cyl2's LPG injection pulse (cyl2's inlet valve has been open longer when the pulse starts), there might be a difference between the manifold pressures in the position of the nozzles between cyl1 and cyl2 which could result in one flowing more gas than the other for the same pulse duration.. this can occur when nozzles are too close to the valves but would expect this issue more relevant with older spec sequential ECUs.

50 mbars = 0.05bar = 0.725 psi very slight vacuum -its virtually atmospheric.

So maybe using atmospheric to control gas injection should work well especially if you avoid the complicated start procedure.

.... In that case it wouldn't matter if you connected the reducer and pressure sensor to manifold vacuum or not.. Yet if you look on this forum there has been a fair bit of discussion on the subject..

Most of my customers wouldn't be happy with an install that takes a minute to switch to LPG with a warm engine. Picture the scene - Start the car in the morning, customer wants it to switch to LPG asap i.e. in about a minute from flat cold. Get to work, do a bit, time gets to 10:30 and customer decides to nip out for a bacon butty , engine is still hot, customer wonders why they get nearly to the sarnie shop before it switches to LPG, same on the way back to work, same any time after they've switched the engine off. Perhaps the car develops one of those intermittent and latent BMW cam sensor issues and this sometimes causes a switch back to throttled mode.. Bad news if the LPG system cannot provide proper fueling in throttled mode, what would be a minor issue that the customer might not ever be aware of results in the customer becoming very aware of a problem with the engine running very poorly on LPG but miraculously running fine when he switches it back to petrol - your fault.

Simon

[LPGC]

Everything I said above was relevant in this case.

I' saw the vehicle shortly after the LPG system was fitted. I could sort it out if the owner left it with me overnight (give me chance to see it warm up the next day) but I don't think he wants to do that, problem with lifts etc... and I think he would prefer to try and sort it himself.

A fantastic job of install though, probably the best install workmanship I have ever seen. Top marks to Pete, done no wrong at all.

Been a long time since I saw the vehicle now, so Pete might have sorted calibration himself. But I doubt it.

Simon

[v8gas]

Thanks Simon for your assistance with the car, really helpfull.
Yes it still has a problem at tickover- mixture going weak. Code 272C/D Additive Adaptation - Control Limit Reached-System Too Lean at Idle.

Have now driven 1200 miles and get 20 mpg on gas. It can go all day on the motorway without engine management light but in stop start traffic at tickover eventually the light comes on. I beleive what happens is fuel trim adjusts for cruising but this is then too weak at tickover.

MAP figures seen on 745i Minimum seen 0.3 (-10 psi) maximum seen 0.95 (-1 psi)

If reducer is connected to MAP
MAP= 0.35 and Gas = 1.0 effective pressure on gas injector will be -10 + 14 = 24psi
this is tickover pressure so is used for autocalibration. Problem is Petrol ECU adjusts injection time down to weaken mixture at higher vacuum.

MAP= 0.9 and Gas = 1.5 Resulting pressure on gas injector will be -1.5 + 21 = 23psi
so gas is good on valvetronic. This will be altered by petrol ECU which will increase injector time 20% to allow for changing pressure on petrol supply rail of constant 50psi. Result richer gas mixture. Gas ECU sees no difference in effective pressure so makes no alteration.
petrol pressure is 50 psi so with manifold 10 psi below, effective pressure is 60psi an increase of 20%

If disconnected Stag ecu will see MAP 1.0 and will not alter fuel much. But gas Ecu uses map to guess engine load so needs to be disconnected from map for valvetronic mode. MAP pressures on valvetronic do not follow a curve but switch from one state to another.

This is where I have a problem-

with manifold vacuum disconnected (and plugged) from reducer, stag software shows 0.92 bar MAP.I believe this is correct for valvetronic mode.
with manifold connected to reducer this changes to 0.35 bar instantly which does not make sense.

If I bypass stag ECU map sensor then switch to gas, stag software shows map pressure at 0.97 bar (atmospheric-correct when disconnected) but gas pressure quickly drops to 0.4 which sounds alarm and switches back to petrol.
Reconnect map sensor and everything operates normally and pressure is stable at 1.0 bar with map showing 0.35 bar

Questions
How can MAP pressure sensor affect gas pressure?
Why should reducer connection affect map pressure?
It seems to me that the only answer can be that the pressure sensors are interconnected either physically or more likely by software, is this normal?


N62 Power output at 3000 rpm = 150bhp = 50% load so 90% of driving is done under this rpm and load

Peter

[rossko]

Reducer maintains a fixed pressure relative to whatever is at it's vacuum nipple. In a normal application, that is MAP, so the output gas absolute pressure varies up and down whilst driving.

While differential pressure sensors have been used in LPG systems, they are rare. Nearly all are absolute sensors.
The Catch-22 is that the reading from the absolute sensor is not is what is reported on screen. Because it's not usually helpful. As you realise, it's the pressure across the injectors that counts, which is absolute supply minus absolute MAP.
So, laptop software generally displays this calculated relative pressure (and you don't get to see the absolute gas pressure reading at all).

In an ordinary application, displayed MAP goes up and down with driving load, and gas output pressure tracks that, but displayed pressure appears constant.

If you start fiddling with the vac pipes, you need to take account of effects on both real and displayed (calculated) pressures.

One way to minimise the Valvetronic effects is to start out with highest gas pressure possible - variations in MAP are then a smaller proportion in percentage terms. The 1.0 bar you have seems a bit pansy to me. If you can lift it significantly, you may need to reduce injector sizing to suit.

[LPGC]

Rosko bet me to saying much the same things (nice on the odd occasion we agree mate!). I have said as much on this thread and in person to Peter.

From memory, I think Pete fitted Hana injectors of the none Gold variety, so nozzles not interchangeable without fitting different injectors. Nevertheless, it is possible to get the system setup properly, I've done a few at around this pressure with reducer connected to manifold pressure reference.

As you say, the closer you get to petrol pressure the less pronounced the discrepancies will be - though if petrol pressure is 50psi he's not going to get near that with gas pressure so there will always be some discrepancy.

I beleive what happens is fuel trim adjusts for cruising but this is then too weak at tickover.

Pete, It will have many fuel trim ranges, each covering a different load and rpm range, then in cases for Vtronic, part Vtronic and throttled modes... A lot of different fuel trims.

But gas Ecu uses map to guess engine load

What Stag uses to calculate engine load doesn't make any difference except for the auto-adaption (novice user self tuning) functions. Like any other system, Stag uses Pinj calculated gas pressure (as Rosko said this is just GasP - ManifoldP), gas temp and in some cases injector spec to arrive at Ginj. Stag engine load reading isn't implemented on most systems and isn't at all useful, possible the only reason for it's existence is Stag's 'switch back to petro beyond X engine load' setting. Duty cycle readings would be much more useful.

How can MAP pressure sensor affect gas pressure?

As discussed - GasP shown on screen is relative (to manifold pressure).

Why should reducer connection affect map pressure?

As on the phone - If the reducer diaghpram was leaking and reducer & sensor were T'd together, this would introduce gas into the vac line that would negate the sensors vacuum reading. But I rather think in this case the fueling may drift far enough to cause the petrol ECU to determine a problem situation and revert to throttled mode due to the fueling error - If you fitted injectors that will fuel the engine proper;u at 1 bar Gp versus Mp relative pressure with a fairly low multiplier, then you double relative pressure ny disconnecting reducer manifold vac reference, then the gas injectors would need to wprk at mearly half the pulse time and probably won't be able to do so properly...

Simon

[v8gas]

Thanks but I find the answers confusing.

Are you saying that the Manifold pressure shown by the stag software is wrong?
I understand that the gas pressure shown is effective pressure and not absolute and any figures I refer to are effective pressure as shown in software--

MAP 0.35 gas 1.1--so if you do the calculation this gives an absolute gas pressure of 1.75 bar -- already at the limit for the KME gold reducer.

MAP 0.95 gas 1.7

I did check for leaks from the overpressure valve on the KME reducer but there was none.


I got to say at this point that I wouldnt recomend the Stag system. I have tried contacting them directly for advice but they either cant or wont give me any useful info. I chose the stag system because it is advertised as compatible with valvetronic but they give no info how to adapt it. I also paid for the premium ecu with permanent OBD connection. Another problem- when correctly wired it tries to connect (yellow dot) but after acouple of minutes it says no connection. Even worse when I connected BMW INPA I got messages saying that the bus was disrupted-not continuously. End result I was not prepared to risk messing up the whole car so I removed it.

At the moment the car runs well on gas from tickover to full throttle and at 20 MPG on gas (26MPG petrol) is within what I expected- if a little low.

The only problem is the idle mixture code putting on the MIL light. I intend to play with it but the only adjustment I can see is to increase the multiplier at tickover but this is obviously affecting the fuel trim.

I also still cant see how this happens
"with manifold vacuum disconnected (and plugged) from reducer, stag software shows 0.92 bar MAP.I believe this is correct for valvetronic mode.
with manifold connected to reducer this changes to 0.35 bar MAP instantly which does not make sense"

I supose the end result is that I should have chosen another system.

peter

[rossko]

MAP 0.35 gas 1.1--so if you do the calculation this gives an absolute gas pressure of 1.75 bar -- already at the limit for the KME gold reducer.

Don't follow the 'rithmatic here? The displayed MAP is in the range 0 (perfect vacuum) to 1 (nominal atmospheric). It won't be terribly accurate, vary with weather, etc., but good enough for this purpose.
It's easy enough to perform a sense-check, does live MAP display show around 1 when the vac pipe is pulled off the sensor?

The displayed gas pressure is relative to MAP - so in this case absolute gas pressure will be 1.45 bar or about half a bar above atmospheric.
It's easy enough to perform a sense-check, does live pressure display around 0 when both vac and gas pipes are pulled off sensor?

Those seem reasonable numbers for an idling conventional engine.

This reducer will run 1.75 bar relative (i.e. up to 2.75 bar absolute, or more on a turbo with positive boost, but we're not really interested in that figure).

MAP 0.95 gas 1.7

Under what conditions? You can't rely on accurate regulation from the reducer unless gas is actually being consumed - trapped pressure in the supply hoses will give false high readings from time to time e.g. on the overrun, or when engine stopped.
If you're seeing that say under full throttle but 1.1 at idle, there is summat wrong.

Really sure the MAP/pressure sensor type selected in software matches the physical type connected? A number of different sensors have been used over the years, with different outputs; the software needs to be told the correct type to interpret its output properly.

[v8gas]

Thanks for the suggestion about pulling pipes off the sensor, I've tried each in turn but not both i,ll try that.

I was working on the KME gold reducer having a maximum working pressure of 1.7 bar, which is what they quote.

This is doing my head in! I find it easier to visualise in Psi because I,m old.

Figures shown are rounded and taken from stag software with stag sensor all new with latest software upgrades.

MAP 0.35 bar = (below atmospheric) minus -10Psi gas pressure 1.1 bar (above atmospheric) = plus 16 Psi resulting effective pressure = 26Psi or 1.79 bar--pressure in reducer = 1.1 bar

MAP 0.95 (valvetronic mode) = (below atmospheric) minus -1Psi gas pressure 1.7bar (above atmospheric) = plus 24 Psi resulting effective pressure = 25Psi or 1.72 bar--pressure in reducer = 1.7 bar


These MAP values can and do occur at any stage of "throttle" with valvetronic ultimately controlled by the engine ECU.
They are all whilst running on gas so reducer should be stable. Maybe someone with valvetronic expertise could tell us why it is not continuously 0.95 bar MAP.

Please correct me if I have the wrong end of the stick which is possible.

Thanks Peter

[rossko]

MAP display of 0.35 bar means 0.65 bar below atmospheric, it's P for "pressure" as opposed to "vacuum" being measured.

Working pressure quoted for the reducer is the useful pressure available across the injectors, developed against the reference point, the "vacuum" nipple.
Absolute pressure is irrelevant and is not displayed.

The important thing is that the displayed (relative) gas pressure should appear constant in service, regardless of MAP. (with the proviso that during overrun cut-off it may not be "in service").
Yours does not appear to be that way? Though it's getting difficult to tell if you're telling us displayed readings or what you calculated?

[v8gas]

My gas pressure definately changes as MAP changes, I thought this was normal.

My reducer is set at 1.65bar gas pressure at atmospheric.

As the MAP shown changes then the gas pressure shown changes to always add up to 1.65bar .

If I pull the vacuum pipe off the reducer when running the gas pressure shows 1.65bar instantly.

I am looking at testing my valvetronic since I believe it should be nearer 0.95bar 90% of the time, but it seems to be 0.35 on tickover most times I connect the software.


Peter

[LPGC]

Pete,

Sorry it's been a while. I have gotten a good way into a lengthy reply several times only to have to break off due to work, and then suffered laptop problems which have meant I lost my draught... ready for a new laptop!

Thought we had already discussed how the pressure readings work, anyway have been described again by Rosko above.

Case a: If vac pipes are connected to both reducer and pressure sensor, pressure reading in software should remain almost constant regardless of map reading, only map reading should change as (like Rosko said) the reducer may have a max pressure of 1.65bar but this doesn't mean absolute pressure, this means in respect to the manifold pressure connection. i.e, it could at least in theory deliver 4.65 bar absolute gas pressure on a turbo'd engine that runs 2 bar manifold boost (atmosphere being 1 bar absolute, + 2bar boost = 3bar, + gas pressure = 4.65 bar). In summer with good tank pressure anyway!

If vac pipe is (case b) only connected to the reducer or (case c) only connected to the pressure sensor, then pressure reading should go up and down when the engine switches between valvetronic / none valvetronic modes. Pressure reading should change in opposite directions between case b and case c. Don't bother trying case c except for diagnostic purposes, it won't practically.

If case A and vacuum reading plus pressure reading always total around a specific figure, then you may have a problem with the pressure sensor (unlikely, as it is, after all, reading vacuum correctly and reading approximate correct gas pressure). More likely you have a problem with firmware - I have known this issue before with Stag. If you like, I have a Stag pressure sensor which we can swap in to try if you visit, I would also help you try different firmware.

As memory serves, though, I thought when I saw your car (and with vac pipe connected to sensor and reducer) that calculated pressure reading remained fairly constant in both valvetronic and none valvetronic modes? I would have noticed a problem otherwise...

Simon

[v8gas]

Thanks for the replies.

I think I found the main problem.

I found out that even after clearing the lean fuel codes that the fuel trims stay at their last positions until reset with BMW software--there is no other way of doing it. I have BMW INPA diagnostic but could never see the info on fuel trims. I had to alter the program files and got there in the end.

I guess due to the way out fuel trims the engine ecu ignores the trims and puts the engine under throttle control.


The problem is that there is nothing on the instrument display to tell you this and you cant tell from the way it drives it feels perfectly normal, no loss of power or rough running.

Valvetronic is mainly a fuel saving device to reduce pumping losses.


So even after clearing codes you are recalibrating the gas with the trims already at max lean as in my case.

So there you have it if you get a mil light you need to clear the fuel trims before doing anything else.


I am hoping that the petrol ecu doesnt decrease fuel by more than a few percent when it sees vacuum in the manifold.We will see.


I have never seen the MAP change much from 0.93 bar under any conditions since resetting the fuel trims.

Maybe this will mean more MPG for me.


Peter

[LPGC]

Pete,

Are the pressure readings making any more sense now?

I found out that even after clearing the lean fuel codes that the fuel trims stay at their last positions until reset with BMW software--there is no other way of doing it. I have BMW INPA diagnostic but could never see the info on fuel trims. I had to alter the program files and got there in the end.

I guess due to the way out fuel trims the engine ecu ignores the trims and puts the engine under throttle control.

If short term trims are maxed out, the petrol ECU can interpret this as a fault condition and therefore not adjust long term trims accordingly, this situation can arise on a wide range of vehicles and can certainly be caused by incorrect LPG system calibration. However, where this occurs, without using vehicle specific software it is possible to steer long term trims back to appropriate figures - Simply adjust LPG calibration temporarily so STFT's are not maxed out and the LTFTs will then shift.

As said above.. Any OBD error, even latent, can result in Valvetronics switching to throttled mode, you have confirmed the fuel trims getting out of usual working range as one example, there are also the 'phantom BMW' codes such as latent cam sensor errors that a driver may never usually be aware of because they would never develop into a hard code. But the engine will occasionally switch to throttled mode anyway, even without detecting a situation. E.g. - the petrol OBD system will occasionally check the fly by wire throttle (to ensure the vehicle will still be driveable in case, say, a Valvetronic actuator fails) and the only way for it to do that is to run in throttled mode. If you only get the LPG system setup so that it is capable of properly fueling the engine when in Valvetronic mode, then the first time it runs in throttled mode (maybe due to an issue, say a phantom cam sensor issue or as part of a self check) while running on LPG you will be back to needing to use the BMW software again to reset fuel trims.... but at least you'll know of the easy short term fix next time!

I have never seen the MAP change much from 0.93 bar under any conditions since resetting the fuel trims.

If you watch MAP long enough, say a month, I bet sometimes it will. This could be a case of one morning the engine starts in throttled mode and remains in throttled mode long enough for the LPG system to switch to gas... I know I've said as much already in this post, but, without having the LPG system setup so it can provide close enough trims while in throttled mode, this is when the fault will probably arise again.

As said early on - as it is now it should be right 99% of the time but not 100%... Unless trims are OK in both Valvetronic and throttled modes.

Simon

[v8gas]

Further information about my conversion. Now done about 25,000 miles on lpg and the only problem Ive had is that the petrol injectors started to occasionally misfire on startup from cold I guess from lack of use. I put some expensive grade petrol with cleaning additives in and it seems to have cured it. Still getting 21mpg on LPG and the fueling on lpg is set so that the short term trim only reduces by 1.0 (max is 10) so stays within the limits ok. I havent seen the engine management light for a long time now. The car actually runs smoother on gas than petrol. In the end I set the reducer at 1bar and can floor the throttle without any problem.


Peter