installing lpg injectors on intake manifold

[pibby]

Presently nutting out the best way to drill and tap the lpg brass fitting onto the intake manifold. Have had a trawl through some sites and in most instances (for arguments sake) it is considered better practice to keep the length of hose as short as possible ie have injector as close as possible to the intake manifold (say no more than 300mm). Though some say more importantly is to place the fittings on the manifold all eqi-distant from the point where the inlet manifold contacts the head, to put numbers around it, it is said no more than 50mm. both of these points I can see the sense in.

Where I am unsure (though others not so) is that it is recommended to tap the fitting in a matter which introduces the lpg perpendicular to the air flow. Though, on the contrary, I have come across the installation manual for Prins which depicts the fitting aligned so the lpg is introduced more in line with the direction in which the petrol injector discharges. The little I know about flow through the intake manifold would suggest you would want to direct the charge of lpg towards the inlet valve not at right angles to the airflow. (being generous to myself here, I have heard the term ‘flow bench’ but that’s the limit of my knowledge on matters such). I tried to take some photos of how the lpg was installed on the 3.9 intake manifold but batteries flat so no go. to my untrained eye the lpg will not ‘flow’ to the inlet valve when installed perpendicular to the air flow and on some of the outer intakes it discharges at a bend in the manifold and I would assume this would act like a cavity creating its own pocket of turbulence and be bypassed, or at least interrupt, the main flow of air being sucked into the motor.

So really what I am asking is why introduce the lpg at right angles to the airflow when with a bit of angling with the drill it will be possible to align it more closely with the orientation of the petrol injector? (and I assume get more of the injected pulse of lpg to the inlet valve in a timely manner). Or am I just splitting hairs?

[bee utey]

1. From experience, hose length in excess of the maximum recommended simply makes the change-over slightly rougher at idle. I regularly fit 300mm hoses where there is no room for the injectors close to the ports. Just means you set the changeover revs a little higher in the software. I had a Mitsubishi 3 litre V6 in for checking today and the changeover on the road with long hoses is undetectable, at idle it stumbles a bit.

The Impco/BRC kit I fitted to a 2.4 4 cyl Mitsi Triton came supplied with long hoses out of the box.

2. I was told that nozzles should be angled from 90 degrees to pointing towards the valve, but this wasn't critical. The important thing to note of course is that a measured time of vapour flow across a pressure difference is injected, so whatever angle is chosen you fine tune the fuelling accordingly. As the fuel is already vapour it doesn't have wall wetting issues like petrol or liquid LPG.

I was at the wholesaler's today, they were mounting nozzles in a FG plastic manifold. They had the nozzles at 90 degrees, I pointed out it was just as easy to drill at 45 degrees into the angle between the runners and the mounting bosses for the injectors. They said 90 degrees was the way they did it and it works.

[pibby]

hi bee utey

Re : point number 2 – that’s what I’m trying to determine. Installers put them in at 90 degrees and they obviously work but why is it not better to direct the charge towards the valve.

The response you got from the wholesaler reminds me of the monkeys in a cage story. There’s 5 monkeys in a cage with a ladder going up to a bunch of bananas. Whenever a monkey would climb the ladder all of them would be sprayed with a jet of water. Eventually they worked out it was best not to climb the ladder, hence if any monkey attempted to climb the ladder the others would stop it and beat up on it so they wouldn’t get sprayed with water. Eventually some of the monkeys were replaced with new ones and lo and behold one of the new ones attempts to climb the ladder to get the bananas. (Unbeknown to the monkeys there will be no more spraying of water.) So the original monkeys fearing a spray of water physically pull him off the ladder and beat up on him. This happens to each of the new monkeys till all monkeys will beat up on anyone going up the ladder. more monkeys are swapped over and fresh ones introduced. (now, there are no monkeys in the cage who were part of the original group which had water sprayed at them.) as expected one of the new monkeys goes to climb the ladder to get the bananas and gets beat up on by all but the newest monkeys. So he turns round to one of the monkeys beating up on him and says “why are you beating up on me when I climb the ladder to get the bananas?” to which the reply is “dunno, that’s just the way we’ve always done it round ‘ere”.

I have attached a couple of diagrams from the prins installation manual. (from what I have read prins are regarded as the first or second best gas systems in the world so as we say here ‘ they must know their s**t’. the first diagram shows the lpg injector in line with the petrol injector DEFINITELY not at 90 degrees to the airflow. I thought well maybe this is just how it was drawn so don’t take it too literally. However, the second diagram using the nylon hose shows a relaxation of this alignment to the petrol injector as the nylon hose protrudes into the intake chamber to get as near as possible to the inlet valve. So I would assume there is merit in having the lpg charge directed at the inlet valve (ie not at 90 degrees to the airflow) and to be as close as possible to the inlet vavle.

Anyone got any experience with this nylon hose setup. If prins have it as part of their installation manual then possibly it performs as good as or probably better than having the standard fitting. Anyone know who supplies these types of fittings in Australia? (some lpg suppliers are real PIA and won’t sell you anything if you don’t have a gas ticket. Notwithstanding that in the uk you can install your own system and then have it inspected.)

[bee utey]

Re (2) my wholesaler got one of the engineers out from Italy (for a fitting seminar I attended) who basically said the same thing. I guess the important thing to remember is that you are injecting vapour which just has to be deposited near the valves, so that the whole fuel pulse is sucked in by the next piston stroke. There is no "spray" pattern involved, the nozzle tip bore is much larger than a petrol injector nozzle.

The nylon tube type nozzles are available for installs where you can't drill the manifold close to the valves. I can get some if you are desparate. However you shouldn't have any problems getting standard nozzles next to the injectors angled inwards. I would rather have plain nozzles than bits of nylon tube rattling around inside the manifold.

[pibby]

Thanks bee utey. I will stick with the normal fittings for this run – I’ve already got these. I will probably have another play with the motor at another time and skim the heads and give the nylon hose a thought at that time. I have found another document which happens to be by the manufacturers of my lpg system. Here it clearly states to direct the gas charge towards the valve and to ensure it is aimed to cross the centre axis of the airflow, not to intersect it at right angles. I have no experience in this stuff so based on what I have seen I will be trying to aim my fittings towards the inlet valve though it’s not completely possible on these manifolds. Anyway, once again it’s a learning curve and good to have a little more understanding of how the lpg system/motor works. Thanks again for your contributions.

I had trawled the LPG Discussion Forum • Index page (http://www.lpgforum.co.uk) site and could only find the 90 degree theory but no explicit discussion of my question.

http://www.aulro.com/afvb/attachment.php?attachmentid=31503&d=1291412573

[peter51]

Hi bibby,
Here is some more food for thought.
I have an EMER SVi system fitted to a 2000 4.6. The car idles as smooth as silk and is very efficient - 17-18l gas per 100 km. Normal economy is 14l/100km - freeway only.
On replacing the valley cover gaskets etc. this week I was able to look at how the gas injectors were sited.
Because the upper plenum smothers the intake manifold, the installers (Carb and Gas at Woolloongabba, Brisbane) had to mount some injectors at 90 degrees and some at various angles to allow both angled and minimum distance routing of the downstream injector hoses. Hoses from the injectors are about 250mm.
I don't think that the efficiency between mounting at 90 degrees or otherwise would be measurable.
What is paramount prior to intake is homogeneous mixture of air and gas. Whether the gas enters at 90 degrees to the free stream or otherwise may have little measurable effect on the mixing of the two gases. All we can do is think intuitively about that, however what we believe will happen inside a closed system is often different to reality.
In any case, most intakes have some sort of vortex generating design between the fuel injector nozzle and the intake valve to make a homogeneous mixture, because they have known for the last 20 years that the low pressure fuel injection systems do not fully vapourise fuel - so vortex mixing is required at inlet to ensure the cylinder fills with the same homogeneous mixture throughout the swept volume of the cylinder. This allows for a well controlled chemical reaction.
You may return a bigger dividend by measuring the losses along the intake path and improve on any flow restrictions there and just tap the nozzles at 90 degrees. I would not put any tube into the intake as it will adversely affect the volumetric efficiency of your intake system.
I noticed the EMER gas nozzles are very simple. Several months ago I wrote to EMER to ask why the nozzle design is so simple, and why they had not incorporated some turbulator/ restrictor on the nozzle to change the LPG flow to turbulent rather than smooth laminar. A turbulent flow has “greater energy” and will penetrate further into the low velocity air stream – theory only, but highly probable. Of course I got no reply about improvements to the nozzles.
I have included some photos of my system.

[pibby]

Hi peter,

I agree they are a remarkably simply looking device the fitting on the manifold. I’m not surprised you had no response, I think a lot of this lpg game seems to be as bee utey commented from his conversations with other installers ie it works and that’s how we do it. it appears the detailed knowledge for lpg isn’t around as it is a specialized area whereas there is no problem getting opinions on petrol and what may or may not be better ways of doing things.

I am retrofitting this lpg onto a thor manifold the same as yours and also have the 4.6. as I’ve previously mentioned I have no experience in this area and as you say one’s assumptions may be wide of the mark when it comes to the behaviour of a medium in a restricted environment. Why I have the intial concern about the 90 degree was because my economy is not in the league of yours (though I have retrofitted lambda sensors) and was trying to identify any possible setup issues on the 3.9 manifold rather than replicate them on the 4.6 manifold. On the 3.9 plenum some of the outer runners have a sharpish bend near the face which abuts the head. The way the fitting was aligned was to fire across into the outer sweep of this curve. So in my simple way of looking at it I visualized how the outer sweep of a river moves and you find there is a dead pocket or water actually flows in the opposite direction to the main body of water hence any water ending up in this bend will be disconnected from the flow of the main body of water. So I took this completely baseless interpretation of a fluid medium and applied it across to the gaseous environment of lpg dynamics within the intake manifold. Worst case scenario I might at least get points for creative thinking!

Do you have any photos of you lpg setup specifically how the hoses weave past/through the upper manifold? I have seen various ways of doing this now and yours appears different again. Bee utey uses banks of 2 injectors and goes through the webbing and has the banks on the wheel sides of the motor above the rocker covers. I’ve seen photos of the way deacon auto in melb do it and have the banks of 4 mounted horizontally on the wheel sides of the rocker covers and run the hoses in across the top of the rocker cover which has a recess in it. this does look quite neat. I have seen a uk installer which does similar to bee utey going through the webbing but has the 2 banks of 4 on top of the plenum pointing across to the side it connects to. I reckon I will end up going with whatever method gives me the least bend in the line once I’ve got my fittings tapped into the manifold.

By the way, congratulations on your first post. Took you 6 months!

[peter51]

I can start you off with one picture of the forward part of the manifold.
It only shows 1, 3, 2 and 4.
I have a video boroscope, so I will get you video of the other four and some clearer pictures of the hose routing.
The fellow who installed was a perfectionist and spent considerable time doing the job.
Each lpg bank is fitted to the already tapped 6mm holes either side of the plenum. The filter fits to an already tapped hole on top of the plenum.
BTW this thor plenum is designed as “charge tuned” with increased volumetric efficiency and ideal for this application.

The vapouriser is on the passenger side mounted with a right angle bracket on an existing bolt.
The LPG ECU is sited at the coolest part of the bay forward LH
I'll get it all together and post it tonight.

You will notice I have put a low coolant sensor aft of the water temp sender. They are a life saver from Redarc - but where I put mine interferes with the fuel rail bracket - had to modify the bracket - however mounting forward of the ECT sensor may interfere with the hot coolant outlet pipe - can't recall.

The main ECT sensor will not read correctly if you lose coolant as it hangs from the top of that gallery - see second photo.
Also notice from this photo that the water outlet here to the throttle body heater - which you will use for the vapouriser - has its outlet flush with the top of the gallery - this allows air locks to form in the circuit to the vapouriser - this is bad news when you refill and first start. In fact this circuit contains aerated coolant at all times- except at idle. I know because I have installed a clear tube in this line to observe the coolant flow for signs of cavitation.
Regards
Peter
PS: I also just posted on diagnostics as I use the MSV2.

[bee utey]

Thanks peter51 for the manifold pic, I never got around to taking one. You will see how easy it is to drill at 45 degrees to the runner in the angle between the runner and the flange. That is what I normally do these days. It makes the hoses quite straight up to the manifold top. I do this with the fuel rail installed to gauge clearances. The nozzles don't have a hex on the base (tightened by allen key) and are sealed with loctite.

The pic of the injectors I installed is here: http://www.aulro.com/afvb/attachments/technical-chatter/18247d1251366055-lpg-pics-injection1.jpg

In my opinion this was the easiest path to take. The holes in the webs between the runners are marked with the relevant cylinder number (stamped in to the adjacent metal) to ease diagnosis.

The attatched pic shows the same angle on a Toyota 2.7 Hilux I am currently converting.

[peter51]

Hi Bee Uty,
Your plenum looks different to mine. Compare your picture to mine about 2 posts back.
My plenum has no gaps between the runners , but it looks like you have routed your hoses between gaps in the runners. The upper plenums look very different.
Am I missing something there? . Is that a Disco engine?

In any case it's a very neat job.

What are your thoughts about controlling the LPG temp at the nozzle to a constant of about 5-10 degrees. This would require a temperature controlled flow control valve at the inlet to the vapouriser to control its LPG output temperature. This would cool the charge considerably and allow greater advance or higher compression. The ECU would continue to advance until knock is achieved.
This could be achieved with a stepper motor flow controller with feed back from LPG output temperature.
I would estimate at least a 5% increase in efficiency.
With my system the LPG vapouriser outlet temperature at the nozzle is random - depends on coolant flow rate(RPM dependant) and under bonnet temperature and ambient temperature.


For Pibby - I have attached a few more photos of my install. Hope it helps you.

[bee utey]

Hi peter51, if you look carefully at your first pic, the gap is between the runners close to the bolts that hold the top section on. There is enough room in the webs to drill two 1/2 inch holes between the front two runners and also the rear two runners. My pic is of one of 3 Thor engines I have converted this way, both P38A and D2.

As far as temperature control of the gas flow is concerned, it sounds like a good idea in principle. However my software has a minimum converter temperature for LPG operation of 30C. No reason why a PID controller from Auber Instruments couldn't control a solenoid valve to allow coolant flow within a narrow temp range. It would keep the vapour temp lower but not at 5C. The software keeps track of lpg and converter temps and allows for it.

I think when aftermarket liquid injection kits come more mainstream it will make this mod quite irrelevant. The specific heat of lpg vapour would be nothing compared to the latent heat of vapourisation.

[pibby]

thanks for the photos peter. the plumbing for the vapouriser on mine will come from the heater as it currently does. the ecu is on the firewall on that side of the car and so to the vapouriser. the pre heater line you are talking about i will just be blanking off.

am keen to get one of those temp alarms (not just low coolant alert) but have to offload my tdi first.

have you any diagnostics which identify the timing you are running when using lpg? my motor is a high compression too and i found it has to be waaay too advanced to get a ping sound from the engine and the car seems to go better without needing to be near the ping point. i once put electronic ignition on which i controlled from my laptop whilst driving (highly not recommended) where it was quite easy to repeat the ping test. i am curious to see what corrections the ecu is making to the timing on lpg as it is a very different curve to petrol and whether it is able to fully advance under low load to where lpg likes it (i've had no problems winding it up to 16 and 18 degrees at idle) or is it banging up against a configuration parameter which only lets it move x degrees from the standard timing curve?

[peter51]

I'll get the figures as soon as I can.
I am currently overhauling the EAS air block and a few other projects.
I recall seeing 40 - 50 degrees advance on my old GEMS powered P38 at highway speed - petrol powered only.
I think all P38 brought into Australia are low compression - mine is anyway.
Regards
Peter

[peter51]

Well I had quite a bit of trouble getting an airtight valve block, but having finally suceeded I have grabbed some video files of the advance figures at various states for my Bosch Motronic 4.6.

I used the Blackbox solutions MSV2 extreme with its sister windows software.

You may need VLC media player to play the files - but should work on windows media player.
I have attached the following:


Idle on petrol
Idle on switching to gas - notice the misfires recorded - there is no human detectable stumble by the way -smooth as silk changeover.
50kph on gas
Accel from lights to 60kph on gas
Hard accel in sport mode from 40kph to 100kph on gas- reaching 4000RPM. At this stage the gas system beeps - assume insufficient gas pressure- bee-uty would know the answer. Notice the advance stabilises at 30 degrees when reaching 100kph.

At a constant 100kph the advance on all 8 cylinders remains at 30 degrees and stayed for the duration of our short observation period. However the ECU should continue to advance the timing in 2 degree increments until knock is reached.

I did have the battery disconnected for several days so I dont know whether these are learned or default values
Hope this helps

[bee utey]

Hard accel in sport mode from 40kph to 100kph on gas- reaching 4000RPM. At this stage the gas system beeps - assume insufficient gas pressure- bee-uty would know the answer.
If it drops out into petrol mode and stays that way then you are probably right about pressure dropping. If it swaps straight back to petrol on decel it is just programmed to run petrol under high RPM. Quite a common strategy, may be related to max injector speed.

[pibby]

hi peter,

thank you for capturing the data. seems like quite a powerful diagnostics tool. i had to download VLC to get it to play, WMP wouldn't play ball. the under under load concurs with what i found when modifying my advance tables when i had the electronic ignition. i would go no more than 25 degrees. i am a bit surprised it is in the low 30's under steady state driving as i found it was better in the 20's . i assume the knock control is pulling it back from the knock point as expected though i found lpg behaved different to petrol, it was better further away from the knock point whereas petrol is better close to the knock point.

these results also go to once again show the difference in timing curves between petrol and lpg. often people will advance their distributor to say 10 degrees at idle but this is really near the maximum to stop the knocking on petrol. it doesn't achieve the retardation that lpg needs when under load and is probably making it worse as it is advancing the timing curve more (though gives better initial acceleration until the advance goes past the magical 25 degrees). the knock control does not appear to be active at idle based on your results as it shows petrol and lpg at the same advance.

once again, thank you for collecting the data. it is very informative.

if i had the tool (and time) i would like to build a graphical representation of advance as determined by knock versus load (say MAP). even better would be a third set of data being torque, these are 4wd's after all. i have read of the iphone? having an app (possibly on aulro) which was able to calculate torque quite accurately when compared against a dyno.

brett.

[uzz32soarer]

Hi guys,

I have an older style Elco system on my 1995 4.6 GEMS Rangie and I'm in the process of converting it over to liquid to allow me to return the air intake system to standard and remove the restrictive in line mixer.

I'm struggling to find anyone that has done a GEMS 4.6 to try and understand where is the best place to mount the gas injector manifolds. My system has 2 x 4 injector manifolds, one for each side, but with the shape of the upper GEMS intake manifold, trying to mount anything on the Pass side of the engine is proving impossible.

I hear all these stories about how the downstream hoses have to be as short as possible and all be equal lengths etc, but how critical is this really?

I may have to mount the injector rails both at the front of the engine and sweep the tubes back to the manifolds which will make the rear most hoses around 400mm long, and curling up 400mm of hose in a pig tail to get to the closest front points seems silly if not totally necessary?

Has anyone done direct injection on a GEMS that could feed back some input here and even post up a picture of your manifold mounting?

[p38arover]

I have an injected system on my 4.6 GEMS. I'll get some pics up later today.

Mine is a KME system.

To keep the hoses short, and of equal length, the installer, KLR Automotive, made up two mounting brackets for the injectors which are in two banks of four.

[uzz32soarer]

Thanks Ron.

Going to be quite difficult I think. Here's a picture of the gas rail that is in my kit.

[bee utey]

I assume you're talking sequential vapour, not liquid injection. The hoses can be as long as they need to be, all that suffers from long hoses is a slight leaning out during changeover as the hoses fill with gas. It's only noticeable at idle, you won't feel it under load.

As for 2 banks of 4 injectors, I've only done a GEMS with 4 lots of 2 injectors. I've used hoses up to 300mm long on other installations without any real problems.

http://www.aulro.com/afvb/attachment.php?attachmentid=61456&stc=1&d=1370562649

[bee utey]

Thanks Ron.

Going to be quite difficult I think. Here's a picture of the gas rail that is in my kit.

One at the front and one at the back perhaps, or one on top of or below the throttle body.

[uzz32soarer]

Thanks for the help guys.

I see two ways to accomplish this job, neither is great. To maintain a tidy engine bay with at least some degree of symmetry I have two mounting options.

The simple way with the most room to play is to mount both rail to the upper manifold on the drivers side of the engine.

An alternative, which will be a struggle with the alternator but may be possible, is to mount both rails vertically at the front of the manifold as shown below.

If I mount together on one side, I estimate the longest run would be 380mm.

If I mount together at the front, the longest run is 360mm.

So negligible difference and either way, longer that the 300mm often deemed to be the maximum distance allowed?

[uzz32soarer]

Given that everything is computer controlled, I'm sure I can control it to the point that I can split the rails. ie: take number three and four around to the other side of the engine, and use one and two to feed the opposite side. This would allow me the best balance of hose lengths and keep basically every hose to around 360 - 380mm

[bee utey]

Given that everything is computer controlled, I'm sure I can control it to the point that I can split the rails. ie: take number three and four around to the other side of the engine, and use one and two to feed the opposite side. This would allow me the best balance of hose lengths and keep basically every hose to around 360 - 380mm
The lengh of the hoses is not critical in that a longer lose just needs tuning for more injection time, but unequal length hoses are much more difficult to adjust for. The software won't allow you to adjust individual injectors. Much better have one rail feed the front 4 cylinders and one the back 4. Drill the nozzles into the runners from the gaps between the pairs of cylinders so the hoses end up at similar lengths. I'm quite sure there's enough room to do that.

[uzz32soarer]

One at the front and one at the back perhaps, or one on top of or below the throttle body.

I scrounged one picture off the web that I think Deacon did. It has one set on top of the intake and looks seriously ordinary. I couldn't live with that.

[p38arover]

I haven't had a chance to take some pics today but here are some I took some years ago.

Front view injector block 1
http://www.aulro.com/afvb/attachment.php?attachmentid=61464&stc=1&d=1370571267

Rear view injector block 1
http://www.aulro.com/afvb/attachment.php?attachmentid=61465&stc=1&d=1370571267

Front view injector block 2
http://www.aulro.com/afvb/attachment.php?attachmentid=61466&stc=1&d=1370571267

Rear view injector block 2
http://www.aulro.com/afvb/attachment.php?attachmentid=61467&stc=1&d=1370571267

LPG ECU
http://www.aulro.com/afvb/attachment.php?attachmentid=61468&stc=1&d=1370571595

[uzz32soarer]

One at the front and one at the back perhaps, or one on top of or below the throttle body.

Good thinking, I reckon this will work. I looked at the complete engine in the car and certainly one gas rail will fit attached to the upper manifold across the back and the other across the front with the primary gas feed line running down the clear drivers side.

As you say, feed the rear four cylinders and then the front four cylinders off a rail fixed across the front of the upper manifold.