Fuel Injection build.

Forum for Davrian topics and questions...
Charlie54043
Posts: 14
Joined: Tue Jun 24, 2014 5:57 am
Location: Lincolnshire Coast

Fuel Injection build.

Post by Charlie54043 »

Fuel Injecting a 1040cc Imp Engine

After a Rolling Road session it was clear that the 2 x 1 1/2” Stroms had very poor mixture control and were restricting peak power.

A manifold with four independent runners and Throttle Bodies would have better mixture control, increase power and improve economy.

The manifold design process has been described in a previous post.
viewtopic.php?f=4&t=716

This write up is to help anyone considering Injecting their Imp and will be kept brief by using photo's and parts description. Please study the photo's carefully for build details.

IMG_20170709_155156.jpg


So starting at the front of the Car....

Fuel pumps are 2 x Walbro 190LPH in-tank pumps, on a “mounting pipe” as shown. The assemblies are threaded down through the filler cap and the pump “mounting pipe” brought back out through a 10mm compression fitting in the roof of the tank.


DSC_0656.JPG

The original pump suction pipe is now the return from the fuel rail, the vent pipe continues as the vent, the “mounting pipe” can be seen between the two. There is no “swirl” pot as the tanks in the Davrian are quite deep, I haven't noticed any fuel starvation even with low fuel and hard acceleration!

DSC_0706.JPG

The return from fuel rail to left or right tank, is selected by a 3 way LPG valve, VSV12 from BioTuning.co.uk and is mounted on the Forward Footwell bulkhead.

DSCF0038.JPG

Engine sensors:
The crank sensor is from 1995 1.3i Ford Fiesta and mounted on the front of the Timing case. The 36 – 1 toothed wheel is from Trigger Wheels and can be seen sandwiched between the crank pulley and its mounting flange. The Distributor blank is an aluminium bar turned to the correct diameter, 50mm long with 2 O ring grooves and a threaded hole for the “strongback”.

DSCF0041.JPG

The Wasted spark coil pack is from a 1995 Mk3 1.4 Ford Fiesta and is mounted on the side of the crankcase above the Dip stick.

DSCF0020.JPG

The Coolant Temp sensor is mounted in the boss next to the dashboard temp sender and is a standard coolant sensor.

DSCF0024.JPG

Air Temp sensor is a standard “white” sensor and is glued into the lower plate of the air filter.

DSCF0035.JPG

The Throttle position sensor was included with the throttle bodies from a 2002 Triumph 600cc Speed Four.

DSCF0026.JPG

Injectors are from the above throttle bodies as is the Fuel pressure regulating Valve

The plastic fuel rail from the Triumph was not used as the Imp's exhaust is directly under the inlet manifold, instead a Stainless fuel rail from an Audi A4 2.0 petrol was shortened to fit.

DSC_0691.JPG

A wideband Lambda sensor is essential, it would be almost impossible to tune the injection system with a narrowband sensor and would be difficult to use the Megasquirt Autotune facility and Closed Loop System, an Innovate LC-1 wideband AFR sensor is used and connects directly to the Megasquirt ECU.

DSCF0031.JPG
DSCF0032.JPG

All the components are assembled as per photos....now for the difficult part....

Getting it Running:

Firstly, the use of a timing light is essential to determine which pair of coils are firing and when.

Maps for Fuel, ignition and basic settings are required, these are shown below. Remember these are for a 1040cc engine with an R20 cam and running 4 x 36mm throttle bodies with MegaSquirt.

The Fuel map controls the fuel at any given throttle opening and revs.



Fuel map.jpg

Note, the throttle opening scale is not linear because 4 x 36mm throttle bodies only need to open a fraction to supply the same amount of air as 2 x 1.25” Stroms!

The engine constants page tells Megasquirt the size of injectors and type of engine it is controlling.


Engine costants.jpg
Engine costants.jpg (150.49 KiB) Viewed 7798 times
When in closed loop and Autotune, the ECU will adjust the fuel to meet your target AFR Table. As can be seen below the the AFR at cruising speed can be 15:1 while at WOT it needs to be around 12.5:1 to avoid damage.

AFR map.jpg
AFR map.jpg (309.43 KiB) Viewed 7798 times



Acceleration Enrichment is quiet difficult to get right, correct fuel enrichment at High Revs Accel might be too much at Lo Revs Accel, so it is tuneable in the AE wizard

AE map.jpg
AE map.jpg (189.37 KiB) Viewed 7798 times

An advantage of Full engine management is 3D control of the ignition timing, which allows the timing to change with throttle position, unlike a Distributor.

Ign map.jpg
Ign map.jpg (262.26 KiB) Viewed 7798 times

There should be enough info above to at least get the engine running and maybe out onto the road, but perfecting the maps will take many miles or huge cost on a rolling road.

When mapping there are two cardinal rules:

1.If in doubt run the mixture RICH, my engine will run at anything down to 10:1 AFR

2.If in doubt retard the ignition timing.

These are important, this engine MAY develop 3 x the original power and any detonation will finish the engine off before you can say “What hole in the piston!!!”

I haven't finished optimising the mapping yet, when I have, I'll put it on a rolling road and see if the results match that of the models in the “Length of manifold Post”

Finally:

Fuel Injection has updated the engine up to modern standards, it will toodle along at 1500 revs in town all day without baulking or chugging, then pull away at Wide Open Throttle without spluttering.

Floor the throttle and it's pure race engine, snorting and rasping it's way to 8000 revs as the acoustic waves travel up and down the manifold....

You just got to hear them!
davrian6140
Posts: 181
Joined: Thu Dec 23, 2010 9:32 pm
Location: Stedham,west sussex

Re: Fuel Injection build.

Post by davrian6140 »

Hi Charlie. Looking forward to comparing your fuel injection
with mine. Are you going to Castle Combe .
Dave.
Dave Cooper
davrian6140

Mk6a DAV 373H
User avatar
Nugget
Posts: 161
Joined: Wed Dec 18, 2013 8:26 am

Re: Fuel Injection build.

Post by Nugget »

Great post, good work
Charlie54043
Posts: 14
Joined: Tue Jun 24, 2014 5:57 am
Location: Lincolnshire Coast

Re: Fuel Injection build.

Post by Charlie54043 »

Hi Dave,

I'll be at Castle Combe, it'll be interesting to compare notes on fuel injection. Would you mind bringing your maps so I can have a look at them, I'll have my latest maps with me.

Hi Mark,

Thanks for the flowers and I look forward to seeing your Davian with the BMW K engine.

Cheers both, see you on Saturday, Charlie
User avatar
Nugget
Posts: 161
Joined: Wed Dec 18, 2013 8:26 am

Re: Fuel Injection build.

Post by Nugget »

Flowers??? Not from me, I think you might have a secret admirer ;)
User avatar
edgedj
Posts: 72
Joined: Sun Oct 31, 2010 4:10 pm

Re: Fuel Injection build.

Post by edgedj »

Looks really nice work there Charlie. Is there any way you can get MAP based in there at all with the ITBs? With megasquirt you can do blended. Have you got a reference pressure for you FPR? I guess not on the basis of the ITB part.

You mention using a fuel pump to return to the tanks, mine just returns straight from the FPR in the engine bay to the tanks with no pump. I imagine you might get some pulsing or interference with you rail pressure with your return pump and perhaps it just introduces another failure point, do you need it? I have a fuel pressure gauge in my fuel rail and looking at the data logs it is always tracking MAP +50psi.

What size injectors have you got? It's tricky getting small enough ones for an imp as they idle so efficiently that ordinary ones are barely opening down at the idle range and you get some very odd side effects where you tune the fuel down and down and it suddenly shuts off as the injectors just don't have time to open and flow. I attach a screenshot of a spreadsheet of the dead time testing and voltage correction for my dense 216cc/min injectors. Once these numbers are in megasquirt then idle tuning is very easy.
Denso 216cc/min characteristics
Denso 216cc/min characteristics
Cheers
Dave
Charlie54043
Posts: 14
Joined: Tue Jun 24, 2014 5:57 am
Location: Lincolnshire Coast

Re: Fuel Injection build.

Post by Charlie54043 »

Hi Dave

Nice to hear from you, I enjoy reading your posts on how you are developing your lower NVH Davrian

To deal with your questions:

The ECU mounted MAP sensor can be connected to the Vacuum line between the ITB's and fuel rail Pressure Relief Valve, the blue pipe to the left of the ITB's in the pic below. This is on the “To Do” list.
DSC_0849.JPG
I don't know what pressure the fuel rail runs at, I use the Triumph Pressure Relief Valve from the plastic fuel rail that came with the injectors so they are running at their designed pressure and know it is functioning correctly because at switch on, I can hear the Pressure Releif Valve “spill” and the AFR is correct as the head of fuel in the tank runs down.

The “spill” from the fuel rail returns to the tanks via the 3 way valve in the 4th pic, there's no pump, I wrote the “build” in note form and didn't made it clear, sorry.

Injector size? I was given the new Throttle bodies, fuel rail and injector assembly from a Triumph 600 Daytona/speed 4/TT600 whose power output is 110 bhp and the Modelled power output of this engine will be 90 BHP, so I used these parts.

Knowing where the injectors came from, I used the calculator on:
http://www.megamanual.com/v22manual/minj.htm
Which gave me an approx 15 lbs/hr (160 cc/min) flow rate, then entered this into the “calculate Required Fuel” box in the Standards Constant page 1 in Tuner Studio.

This value only needs to be an approximate to get the engine running because the VE Bin values can then be adjusted to give the correct AF ratio.

In hindsight, these injectors are too large. As you mentioned, very little fuel is required at idle. The present injectors are only opening for 2.0 mS at idle and as they take 1.0 mS to open/close and MegaSquirt contols in 0.1 mS steps the adjustment is getting close to it's minimum limit.

From your injector characteristic table, it looks as if 216ccmin injectors deliver 0.0055cc of fuel for a 2mS pulse.....phew, thats a very small drop!!

While on the subject of size, the 36mm throttle bodies are also too large! As mention in the "build", the butterflies only need to open a fraction and they have the same open cross section area as part opened twin Strombergs.

The problem comes in the resolution of the throttle position sensor, TPS. On this installation, idle to WOT is an ADC count of 140 , with the first 5% of opening being an ADC count of just 7. It can be seen on the Fuel map that the ADC scale is “expanded” at the lower end, but ideally it would be better to have a far greater range than just 7.

A much better solution would be to fit smaller throttle bodied as on Mark's BMW K1100 engined Davrian, 27mm TB's might be a be a good compromise for a road vehicle.

Even with all these "Non Optimised" parts the Fuel Injection works really well and gives an excuse to “go for a spin” and do a bit more mapping, 2500 miles of mapping has been done and the fuel map is nearly there. The closed loop system is "tuning" the AFR between -10 and +10% at cruise most of the time.

Best wishes Dave and I look forward to seeing how you get on, Charlie
matnrach
Posts: 51
Joined: Sun Nov 15, 2009 9:57 am

Re: Fuel Injection build.

Post by matnrach »

Do you use a damper between the manifold and the pressure sensor?
When I was naturally aspirated I did some data logging and found that without any form of plenum the pressure spikes were huge. Having too large a plenum then made the response laggy. I have always used one of those large plastic fuel filters to dampen the pressure spikes. Seems to work quite well
User avatar
edgedj
Posts: 72
Joined: Sun Oct 31, 2010 4:10 pm

Re: Fuel Injection build.

Post by edgedj »

I've not had to use damper myself, it seems ok with just the small chamber in the 28/36 manifold that I've got (which must be about 200cc chamber) and some MAP averaging in the Megasquirt code.

Charlie, your injectors sound pretty good if they are from a 600cc engine. Are you running 1 or 2 squirts per cycle, I seem to remember that MS defaults to 2 per cycle which really makes the resolution issue worse for small engines. I'm running sequential injection based on a second sensor on my dizzy drive with a single dog tooth wheel which forces me to run 1 squirt per cycle since the aim is to inject on the back of a closed valve for maximum evaporation. Idle torque is hugely increased with sequential injection, I can just lift the clutch with no throttle and it will pull away at 900 rpm with my sport cam.

If you are within 10% then happy days, perhaps MAP would be closer but frankly 10% has to be a lot better than carbs.

Sounds like TPS resolution is worth chasing down as like you say 5% throttle on those butterfly valves is going to be a big range for just 7 ADC counts.

I strongly recommend looking into sequential injection, it's very simple with a dog tooth in the place of the dizzy and really makes a big difference to low end torque.
Charlie54043
Posts: 14
Joined: Tue Jun 24, 2014 5:57 am
Location: Lincolnshire Coast

Re: Fuel Injection build.

Post by Charlie54043 »

Hi MattnRach,

A damper was fitted to the Manifold Air Pressure (MAP) sensor line after your question.

The sensor line was T'ed into the Vacuum line from the Throttle bodies (TB's) to Fuel Rail Pressure regulating valve but, as you found, the pressure spikes were huge. An in-line plastic Fuel filter about 30 dia x 50mm long with 1mm restrictor was fitted which reduced the spikes giving a smoother trace.
IMG_20171021_094838.jpg
The fuel control was then switched to Hybrid Alpha N to use the MAP's rate of change in pressure ( DeltaP) to trigger Acceleration Enrichment and it was found that the AE threshold had to be set to 70 Kpa at idle so as not to trigger AE.

At this level, the MAP DeltaP at small throttle changes on the road was not sufficient to trigger the AE threshold level so the engine was sluggish to pick up.

The fuel control was returned to Alpha N with the TPS deltaV controlling the Acceleration Enrichment and the MAP sensor was not used.

However in another post, I've shown how the MAP sensor can be put to good use.

Charlie
Post Reply