rangerovers.pub
The only place for a coil spring is up Zebedee's arse
Member
offline
1356 posts

OK I know this is a P38 forum.

Had a customer the other day with a 3.5 Disco, came in unable to run on LPG and didn;'t run great on petrol either. Sorted a few probs with the open loop mixer LPG system (needed new switch which wasn't seeing RPM so never turned gas on) and then fixed the bodged pipe between reducer and mixer and set it up a lot better.

But, he'd messed up ignition timing before he came. So, questions are - Anyone know static advance? Anything to know about timing marks on the front pulley? Likely to be any probs in the dizzy? He has his own timing light and wants to do it himself but I'd hazard a guess he'll be asking me the above questions at some point.

Simon

I don't think there's been mention yet of which vendor's C ECU it is, don't forget if it's a different make to your original one it will need that vendor's software (Zavoli / OMVL / Bigas / Etc sticker on ECU).

Yeh, would be good to read your results with the 7805.

When I read the stuff on calibration it always seems put over in a bit of an over-simplified fashion... will be because I know the iffs and buts and other stuff they've missed out or not explained properly ;-)

Nah, think more front garden with my mates, and going for midnight walks leaving the computer in the tent...But did used to carry a car battery when we went 'wild camping' (no tent) and slept by a fire in a field / bit of no mans land down next to a little river just a couple mile walk from home... Powered a CB, in case we could persuade anybody (females) to join us. Aerial was at least light and worked well, made a diipole using just wire and a loading coil bought from a ham specialist shop in Leeds, decent SWR too.

OB, I was planning on keeping the ECU in line with intention to always have a working old type ECU here but if your struggle continues I might sell to you just to help out; I am likely to get hold of another at some point over the next 6 months or so anyway. But if I were you, unless you can get one really cheap, I'd get a new version. Dunno where you got the switch, but if you go back to the same firm and say you just bought a switch, can they do you an electronics bundle but not include the switch it might work out well? All those crispy wires you'd be replacing, and you'd get a spare pressure sensor too... Or just buy a new ECU.

5v current draw won't be much, it only has to drive the few mA for the sensors.. unless it also powers the electronics in the ECU... But as memory serves there is nothing attached to a heat sink in the case, so would seem whatever combination of bits provide 5v they're not expected to need to sink much heat.. for a 7805 this implies not flowing much current. As a kid I used a 7805 and 7806 to power a BBC computer from a car battery when camping, this setup (overloading 7805 by some margin) needed a massive heat sink which I removed from some electric equipment I found down the quarry (come tip) when out shooting. I was always bringing stuff home to attempt to fix or dismantle for the components hehe.

Simon

The first one is an AEB2568 but with a slight deviation due to being made for Tartarini. It is pinout compatible but uses a different cabin fuel changeover switch and the spec of the temperature sensors it uses are a bit unusual. It you got the Tartarini switch it could be made to work, just cut the usual AEB119 type switch connector off and connect the connector for the Tartarini type switch up, colour of wires all match up. The only other slight difference is in the way Tartarini systems are calibrated, they effectively have a built in shape of map for the (limited) types of injectors that can be set, so the default figures in the user defined map are therefore all zeroes and you enter plus or minus figures to deviate from the built in map (actual map is underlying/unseen map plus figures you enter),.. unlike other systems where (if you have access to the full map at all, sometimes this needs V6 software or hardware key to access) you enter numbers between 0-255 and adjust the map directly. In cases where you can't use V6 software and don't have hardware key on other systems you cannot fine tune the full map, you can only adjust 8 numbers each of which effects a block comprising 1/8th of the actual map, so before V6 software better installers (who wanted fine control) had to buy hardware keys, a different key for each brand/sticker on the front of the AEB2568 ECU. You wouldn't really need the fine control for very good results on a P38.

I have 3 different types of switch connectors wired up to the LPG ECU on my own car to allow testing of ECU's and switches.. This is particularly useful for testing how modern ECUs / firmware fair on a 6 cylinder engine (more on this below) .

No need for me to check the other two ECU's since Gilbert has already done so.

Don't forget it's AEB2568 not AEB2856...

It is very unlikely you'd have any problems using a newer ECU, and even if there were problems they would point to shortfalls in other aspects of your LPG install that you might want to put right anyway, in which case it would run better after making the changes than it could have run with the old ECU anyway.

I like to keep at least one good working example of an old type AEB ECU in stock but at the moment I only have an old OMVL one that is broken because it doesn't connect channel 3 on bank B petrol injector wires together when switched to petrol... works great on gas, but doesn't allow running properly on petrol! They prove useful for confirming install issues are due to modern AEB2568s, which seem to have issues particularly when fitted to some 6 cylinder engines possible example here, leading to another disagreement with Tinley.... I am just about to buy a complete old Zavoli (AEB2568 based) front end, trusting the seller on this that all parts are in good working order.

Simon

ECU in 'such as this' (Orangebean above) should work, provided the initial install was done to spec. Due to the none truly sequential nature of early ECUs (not reading any petrol injectors except front on each bank), on an older ECU you might get away with incorrect wiring on rear 6 cylinders where you wouldn't with later truly sequential ECUs. Then, when it works, there is scope (in certain limited situations) for the old ECU to deliver better results than the new one - SInce old ECUs only read Pinj from front cylinders, enrichment during acceleration for other cylinders could be adjusted/mixture enriched or leaned. On newer ECUs the adjustment only allows mixture leaning during acceleration, so, if you have very long pipe runs the older ECU could compensate for delayed gas delivery to ports by pulsing gas injectors for a bit longer, new ECU cannot.
Taught Tinleytech about this a few times, this was probably the most recent

Agreed with all points on 7805 trick, tbh I don't remember now where I connected 5v from 7805 but since it worked when I tried it I suppose must have been to serial connection. I only have one reservation on this... was it a fluke in my case, due to something unrecognised going on... e.g., an actual (communicative) connection on serial might wake some part of the main chip in the ECU, which might be the real reason for ECU working when serial connection is connected.

SImon

Replied to the LPG 5v issue on your other thread..

Been down this road many times before! Old ECUs sometimes fail in this way. Usually before the system fails to operate completely the 5v line goes wrong, which leads to incorrect readings such as temperatures, pressures and tank levels, all of which barring tank level throwing calibration off.
While you're connected with the laptop things can seem better because the 5v is held correct by the serial connection - Installs effected by the latter symptom can seem OK while setting up with the laptop but as soon as laptop is disconnected calibration goes wrong. If the serial connection can make all the difference, so too can a 7805 wired to the 5v wire at the serial port (or attaching 5v from elsewhere on the vehicle, since many vehicles use 5v sensors).
Gilbert might disagree... He took the 5v advice on LPGforum (working on someone else's car) a while ago and found it didn't work in that case.
I suppose if the 5v falls to miilivolts, could be that the main power feed voltage is pulled down when the system tries to power solenoids / injectors, i.e. solenoid coil or wiring short. Or 5v could be pulled down by failed pressure sensor. The temp and level signals can't really pull the 5v down even if shorted because of the ECU internal resistor on each sensor between the 5v rail (each sensor forms half of a voltage dividing circuit, meter of which is what respective AD converter gets).

The bit of track coming out of the switch seems ominous too.

Simon

You're welcome. I also meant to include... I would never fit 6mm on a P38, even for a mixer system, but wouldn't replace 6mm with 8mm if I was, say, replacing a mixer system with a sequential system on one. I've driven loads of P38s using 6mm with sequential systems and none of had a problem.

Simon

I just noticed this thread...

I generally go with 8mm for just about everything I convert these days, but that's just because I stock loads of 8mm Faro pipe and many various types of Faro fittings for this diameter, including fittings that (for 8mm Faro) emulate 6mm copper pipe ends.

Like Gilbert said, 6mm is considered OK to around 250hp. Very unlikely there'll be problems due to 6mm on a P38, even running a sequential system (instead of a mixer system).

Even on most stuff that needs 8mm pipe, 6mm fittings are usually ok unless you're dealing with something a fair bit over 300bhp. Even on a 310hp vehicle, which might use a reducer with 8mm inlet, it would be unusual to ask a supplier for a tank valve with 8mm outlet. For reasons that I don't fully understand but have a bit of an idea on, flow of liquid LPG through pipes isn't totally related to minimum diameter within that feed, an 8mm pipe with 6mm fittings will flow more than 6mm pipe with 6mm fittings.

Because tank pressure fluctuates with temperature and pressure versus diameter dictates flow capability, if you don't have problems in the depths of winter you're unlikely to have problems all year round. The first point here also explains 'even running a sequential system' - A mixer system supplies gas at near atmospheric pressure while a sequential system might supply gas at 1.5 bar. Tank pressure minus supply pressure is the pressure that pushes gas down the pipe, so it's possible to get away with smaller diameters on mixer systems for any given bhp.

Simon.

Gilbertd wrote:

Lpgc wrote:

Post contains a couple of minor errors, unlike my posts ;-)

What's that then?

Only messing mate ;-)

Gilbertd wrote:

So if the air filter gets a bit clogged, the mixture will be permanently rich (not enough air) so the short term trims will be going constantly negative to reduce the amount of fuel and get the fuel/air ratio correct again. This will cause...

Would be true of a petrol carb, LPG mixer system or AlphaN based fuel system but isn't true where fuel injection system meters air flow and/or manifold pressure..

Lpgc wrote:

Morat wrote:

Steady on Richard, that's almost as long as one of Simon's posts ;)

Hehe... Yes Gilbert, everything in 'mod'eration mate! Post contains a couple of minor errors, unlike my posts ;-)

Simon

Piped in series on my own old Chrysler Grand Voyager because it has a matrix in the front and another in the back (matrix's flow in parallel of course)..The reducer is in series with the joint return pipe from both... Thinking was along Gilbert's lines, I didn't want half the flow through the reducer and one quarter flow through each matrix...

There are a lot of iffs and buts regards temp correction. For starters, it seems AEB don't presume gas temp correction works entirely correctly anyway, because otherwise they wouldn't have thought necessary to have compensation for reducer temperature too. AEB default temp compensation figures are a bit exaggerated and reducer temp shouldn't even enter the equation. On some installs (and where it won't effect neatness etc) I try to fit the gas temp sensor in position where it's least likely to pick up heat from the engine (or radiator) or blown cold by external air while driving. On other installs position doesn't make a great deal of difference... most V8 powered stuff gets hot everywhere under the bonnet anyway. The old advice was to fit the temp sensor as close to injectors as possible, thinking was it is at the injectors where the temperature and pressure of gas matters and gas temp could be effected by pipe temps etc between the reducer and injectors ... But I've come to think that in reality when driving gas is shifting so quickly through the pipes it will have little time to be effected much by pipe temperature / At idle gas is shifting slower but still quickly enough for pipe temp not to have much effect on actual gas temp / Though pipe, injectors and temp sensor are more likely to be effected by under bonnet conditions at idle and with slower moving gas the sensor might be effected more by the temperature of whatever it's mounted in than by the gas. One of the iffs and buts concerns compensation particularly at low pulse duration, because unless the ECU firmware correctly accounts for injector latency and temperature related response, no single set of (default) compensation figures will work properly over all possible configurations of pressure, nozzle size and gas injector pulse duration. I.e. you cannot accurately apply Boyle's law without accurately accounting for how injector response is affected by temperature and accurately accounting for injector latency, but almost all firmware seems to get these factors wrong, so temp compensation is best adjusted on individual installs. But again, not much of an issue on a sequential Rover V8 where pulse duration at idle should be higher than it is on most other vehicles and where the petrol system is very forgiving in terms of fuel trims etc - default compensation seems to work fine on them.

Simon

I can see justification for all above points but it's still true that over 90% of P38s and most other LPG converted vehicles have reducer plumbed in parallel without issues.

Both the parallel method and the series method could mean the heater matrix and reducer each only get half the flow - for obvious reason with the parallel method, the series method can mean twice as much restriction in the same pipe so the total flow can only be as much as the part with the most restriction. Series wouldn't be great with a fairly clogged heater matrix, in fact I have seen certain factory converted Vauxhalls fail to run on LPG due to this problem and during summers have re-piped the reducer in parallel to allow running ob LPG until the owner had the matrix problem sorted (even following matrix repair, parallel can be left in place with no ill effect)..

Which method works best (keeps both reducer and matrix hot) will depend on aspects of the vehicle. Some heater circuits will be capable of flowing a large volume but without much of a head of pressure, others will have limited volume but with a high head of pressure while flow is limited. So, with parallel method and only half as much total restriction, some heater circuits will flow twice as much water in total, meaning a no compromise reducer or matrix flow situation. Other vehicles like some Mercs, intentionally restrict water flow under certain conditions such as engine above idle rpm, seemingly so heater water flow is consistent between idle and off idle, and where you can't tap into the un-metered side on such vehicles series is best. In all honesty as long as the reducer is low enough on a P38 it doesn't seem to make any difference.

Wouldn't want to pipe some reducers in series.. some such as old Bigas only have 10mm water pipe connections...

Simon

The next step should probably be check main live feed and switched live voltages at the ECU when the switch fails to light.

Gilbertd wrote:

Maybe you had an iffy connection on the common supply to the petrol injectors? That would explain the LPG system getting offended and the popping and banging when running on petrol. If you had a high resistance in it anywhere, the LPG system would be seeing something less than 12V so wouldn't wake up and the petrol injectors wouldn't be getting a full supply.

I reckon that's a good call.

Off subject slightly but it' remains an unwritten rule of thumb for installers to avoid connecting LPG ECU's switched live or rpm detection wire to the front cylinder of any bank of cylinders. Not relevant to P38's though!

On ECU's such as Stag and maybe including AEB2568 B suffix the switched live simply throws an internal relay, may be solid state in other ECUs. Main power for ECU outputs (injectors and solenoids) always comes through the fused main battery connection, wouldn't want to draw potential peak of maybe 10 additional amps from the petrol injector live as that could cause problems for some vehicles.

Good that the switch stays lit now, if all other wiring was OK but brown rpm wire was broken would expect the switch to light only momentarily when switched live was 12v.

Measuring voltage between chassis and an earth point can give insight into the condition of the ECU, sometimes voltage here increases when the vehicle is running on gas (when it is powering injectors and solenoids), any such voltage can throw off readings such as pressure / temperatures and can lead to more problematic connection with the laptop.

Sometimes the first sign of ECU trouble is difficulty connecting. Connection can mprove more difficult if the laptop is powered from the car (some earth loop type thing going on?). In some cases ECU's only work properly when they are connected to a laptop, dodgy ECU might benefit from the power connection over the serial link. Again, earthing the ECU chassis can help.

Simon

Orangebean wrote:

Simin- can I assume that any solid black cable is (or is meant to be) an earth?

Lpgc wrote:

While you're under the cover, check the earth wire connections in the loom under the cover are all good.
Simon

Yes but watch out for the rpm wire which is brown inside a black sheath.

Looks like you've found the connection problem with the broken red/black in the Molex connector anyway.. Well, former problem with connection before the switch wouldn't even light. The later fitted Supaseal connectors are much better than Molex though you do see the odd bent pin.

Simon

Later on I'll see if I can dig out the pinout info I made myself years ago, I've never seen an official or online pinout for AEB systems though.

Probably no need, as only difference between 2568 systems is switch and on later looms the additional wires that can be connected to the vehicle's OBD port. Without re-reading the thread I seem to remember Gilbert saying Vogel fitted OMVL systems, though it is Tartarini 2568s that I'd expect to be maybe different regards the switch.

While you're under the cover, check the earth wire connections in the loom under the cover are all good. Check the pins on the ECU itself aren't tarnished particularly on main earth and main power feed.

Simon

I edited since you quoted me OB, bit more info.

As Jackk and Ferryman said, also there should be a tie-wrap holding the cover to the loom where the loom first enters, probably need to cut the tie-wrap.

Did you get my bit that implied it may have worked OK before if the chassis was earthed before (but now with ECU not bolted to chassis you lose this other earth point). Chassis shouldn't need to be earthed but I've found on older ECUs earthing the chassis can help. When I see a problem like yours I measure voltage between chassis earth and battery earth, any voltage above say 0.5 when all solenoids and injectors are working may imply a problem (which can cause analogues signals such as pressures and temperatures to be read incorrectly), sometimes just earthing the chassis can help.

Simon

Morat's point was worth mentioning when he did, the programs do only show 3/4cyl options unless talking to an 8/10cyl ECU.

As there's power on the red/white, are battery live and earth good? And if so, are those earth connections underneath the ECU plug cover good and no corrosion on ECU pins? Was the ECU chassis previously earthed / is it now?

I'm sure you'll be right on the error codes... Just that a common definition of P1000 is 'check cycle not complete', a constant code on Jags after engine start until cycle is completed during current run cycle even without the battery having been disconnected, occurs on other vehicles when battery has been disconnected. Some older vehicles lose IACV calibration when battery is off, especially pre2000 Jeeps. Could also see potential for airflow versus throttle angle related codes occurring until IACV calibration is relearned.. But is TPS working properly?

Simon