Megasquirt fuel injection

Competition engines and ancillaries - general discussion
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Megasquirt fuel injection

Post by Guy Croft »

This thread is solely for a detailed description of components, mode of operation, calibration of Megasquirt fuel injection by a member who is intimately familiar with it.

Read 'referring to vendors' at http://guy-croft.com/viewtopic.php?t=429
before posting and pay particular attention to copyright.



GC
Walezy
Posts: 167
Joined: June 25th, 2006, 10:56 pm
Location: Poland, Czyzowice
Contact:

Post by Walezy »

Hello,

I am using 2.0L Fiat 132 engine with 16v head from Lancia Thema. The intake manifold is modified Fiat 2.0 16v turbo with 4 x 44mm slide plate throttle body.
Camshafts are standard Lancia Thema non-turbo. Pistons are cast taken from Fiat Tipo 1,8L (it has a little dome). Conrods are lightened standard ones(690g), Flywheel is lightened(quite a lot but i have not checked the weight). Exhaust manifold is 4-2-1 with primary pipes 38mm and secondary 42mm.
Head has been ported by me but I am not able to check it on the flowbench.
I am using standard valves and valve seats are not modified.
Right now I am using standard wet sump oil pump with accusump(previously I used dry sump with Titan oil pump)

I am using Fiat Coupe turbo fuel injectors and fuel rail.

I have tuned the engine on MAHA LPS 3000 dyno. I have made it by setting the rev limit(on dyno) by 500rpm from 2000 to 7000rpm and I was using different load(light throttle to fully open throttle) on each rpm limit.
I have used speed density mapping with MAP connected to all 4 intakes with little fuel filter before the ECU to damp the pressure jumps on idle. I am thinking of Alpha-N right now but I will wait until I get somewhere good cams for this engine.

The engine has most torque at Lambda 0.86

I was able to tune the ECU to have smooth running engine and after i used mapped ignition(not on the dyno plot) the engine got a lot more torque than it was on standard bosch distributor but i have not checked it on dyno(I am thinking of at least 10Nm in almost whole power band as the ignition was way too early on distributor)

I am using 2 fuel maps(road/stage) and launch control. Everything works perfectly, engine idles smoothly at 700rpm, I has good throttle response and it never failed. I have done all the build and tune up with my friend(co-driver) and we never used any assist of any kind of professionals.

To tune up the ECU I needed to spend a full day at the dyno so you will need to count it as a cost (not too low for sure) but I think that it is the only way to make it good.

I was using MAHA exhaust gas analyzer to check the AFR (I hope that it was good to make it this way).

The engine was set up at intervals of 500rpm from 2000 to 7000rpm with at least 3 to 4 loads on each rpm point.
As the dyno is fully computer controlled so there was no mess with setting up any braking power (of the dyno brake) as I only need to set revs and the dyno will brake the engine not to go beyond them no matter how hard you push the engine.
I had BHP gauge on dyno display so I knew immediately which settings are the best.
After making injection map complete(12x12) I had to tune up the acceleration enrichments, this is done on the road and you need to make it by feeling as it is hard to simulate it on dyno. Basically if the car is chocking then it is too rich and if it backfires to the intake then it is too lean. You may need about 1-2 hours of driving with a friend who will change the parameters while you drive the car.

Setting up of ignition map can be made the same way as making fuel map although it is not crucial to make it on dyno(but for sure the best).
After such dyno tune up there is no need to use lambda sensor (narrow band) as it will not help too much and if it fails then you may get fault readings so I do not use it (but it is connected).

If you have any questions please ask.
By the way: Guy, maybe you have an idea of what to change in that engine to get about 200-210HP. I need that for rallycar.
Attachments
This is the dyno plot. I hope that it is clear as I have tried to translate it.
This is the dyno plot. I hope that it is clear as I have tried to translate it.
Power and Torque 44s vs MS inj.jpg (153.79 KiB) Viewed 50799 times
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Post by Guy Croft »

In the first instance I've checked out the power/torque curves and cannot see anything wrong with the data as such. You have done an excellent job of presenting the data in an open and honest way and I am sure from the way you have written that the back-to-back results of carbs vs injection are as valid as you can get so I accept them.

The graphs are in metric. For the benefit of non metric readers:
To convert PS to bhp - multiply PS by 0.986.
To convert Nm to lbf ft - multiply Nm by 0.74

To derive power from rolling road torque I¢ž¢m using rule of thumb formula below:

BHP = (torque (lbf ft) x rpm)/5250

My observations:

For comparative purposes a standard 2 liter Fiat Tipo (5 door code 160A8.046) engine with OE semi-sequential fuel injection and mapped ignition, common intake plenum develops:

148 bhp (DIN) at 6250 rpm (DIN is with all ancillaries fitted, equivalent to running with engine in car)
131 lbf ft @ 5000 rpm


1. On carbs
Power and torque results are always useful for comparison, so thanks for including them. They provide a really vital benchmark because you should always expect more from injection if throttle bodies are used. A top-end gain of 10% power is a reasonable target. More because injection doesn¢ž¢t generate intake pumping loss (no chokes), the fuel delivery is more precise and better homogenised, you can tune the ignition precisely, moreover if you don¢ž¢t get more power and torque from injection ¢‚¬Å“ there is no point at all in fitting it.

The torque on 44IDF carbs (obviously on split intakes) is 209Nm (154lbf ft). This is not bad for your 16v conversion engine ( 10/1 CR or over and normally aspirated OE cams). It¢ž¢s not far off what I¢ž¢d expect - you have gained torque over standard with the ram effect from the superior inlet setup and maybe from your head work.

I don¢ž¢t know what the ignition setup or settings were for carbs, maybe not optimised.

The top-end power on carbs is higher than standard at about 152PS (150bhp) but it¢ž¢s peaking at very low 5400rpm whereas the OE unit peaks at 6250. It should be peaking at about 250 rpm (6500rpm) higher and giving about 160-170 bhp bearing in mind the race carbs and head work.
Your projected torque (following the trend of the curve) is 160Nm at 6250rpm. This gives 140bhp, 8bhp less than standard. It should be higher than standard.
Typically I¢ž¢d call this ‹Å“no top end¢ž¢ can be because of cam timing, choke size, jetting or IDF inlet manifold losses, exhaust back pressure.

2. On injection
The power is 160PS (157bhp) and it¢ž¢s now in the right place and 9bhp over standard and 7 hihger than carbs which isn¢ž¢t a disaster by any means, I don¢ž¢t want to make a detailed comparison with the carb result as I don¢ž¢t think the carb result was anywhere as good as it could be.
Whereas the torque on carbs is dropping fast over 5600 the Megasquirt unit is holding its top-end torque better, suggesting to me that the fuelling is better.

Yes the torque at 5000rpm is higher than the standard Tipo unit (190Nm/140lbf ft compared with 131lbf ft) a good gain but how much is attributable to being fuel injected per se? Below 5600 the torque is very poor compared with carbs and it should be miles better.

I wonder actually if your injectors are too big for a start; you¢ž¢ve got Coupe Turbo injectors ¢‚¬Å“ well that¢ž¢s an engine developing 290Nm torque at 3400rpm ¢‚¬Å“ 214lbf ft torque. If your injectors are too big your duty cycles and pulse widths at lower ¢‚¬Å“ mid rpm are going to be all over the place. My advice would be to go back to standard 16v n/a ones, after all, your engine is not far off standard, and then an increase in fuel rail pressure might be all it needs.
I also say that you need to concentrate more on the steps you use for the load sites, you should be looking at 250-300 rpm steps over 4000 rpm, 500 is too wide to expect accurate interpolation from the ecu. I assume that you have done inlet air temperature correction in the map?

3. Tyre growth correction
I don¢ž¢t know your rig but it is just possible that if you don¢ž¢t have ‹Å“tyre growth correction¢ž¢ on the rolling road, the engine rpm the dyno is reading from the tyres isn¢ž¢t true crank rpm, and it CAN be out by 400 rpm, so the actual power calculated from the RR torque is higher than you are seeing.
For example:
The torque reading of 183Nm at 5800rpm (135lbf ft) on inj could easily be 183 Nm at 6200 = 160bhp. If that is the case here then the numbers begin to look a lot more more promising.


4. For the future

Apart from a bit of top end all you have succeeded in doing is producing a smooth engine with reasonable top end power but not much 'below'. Don¢ž¢t be offended, I am making an objective observation here not a criticism. That kind of characteristic is not a valid 'raison d¢ž¢etre' for fitting mappable injection but sadly typical of what happens when people do it. One learns by experience and I know plenty of other engines in the same ‹Å“boat¢ž¢ including several on this site. I am sure it can be made to perform way better but it will be hard work.
Still, nothing easy is worth having.

Since you are fortunate to have a pair of IDF carbs around I would take off the Megasquirt right now and spend time re-developing the carb setup and only when you are getting the kind of power and torque/rpm I¢ž¢ve mentioned go back to the FI and try and match - and better it. You need a benchmark ¢‚¬Å“ only carbs will give you this.

Then you can start thinking seriously about cams and race pistons etc., the kind of things that will take the engine way over the 200bhp mark. Since I first tackled the 16v conversion in 1991 I have done and supplied parts for quite a few 16v 2 liter conversions running 45 DCOE carbs even with just reprofiled cams (Kent BD3) generating 190bhp PLUS at the wheels.

Bear in mind I have to do a lot of guesswork when working long-distance like this, I can get things wrong, sure, and I am not fully au-fair with the whole engine package, some good photos of exhaust, intake and other engine things would help. If there are any errors in my maths let me know, numbers are not my ‹Å“strong point¢ž¢.

These things are major issues with any conversion to FI and lately I have been having vital and intensive discussions with my new friend Phil Lanes of Lincolnshire based GG Motorsport lately about it all. I believe that joint high-level technical exchanges will enable to derive benchmark MAX results from engines like these.

GC
Walezy
Posts: 167
Joined: June 25th, 2006, 10:56 pm
Location: Poland, Czyzowice
Contact:

Post by Walezy »

On both carbs and EFI the ignition was standard, worn, never reconditioned distributor. It was set at about 15deg advance at idle but I am not sure exactly right now as it was some time ago. Right now I am using mapped ignition controlled by MS and it works way lot better but I think that this map could be improved for sure. I have not checked the engine on dyno with mapped ignition but it could be easily seen that the mid range was a lot better(something near or better than carbs).
The cams in both setups are set to standard marks(with adjustable camwheels).
The revs on dyno are taken from the clamp on ignition wire but it is also possible to take it from wheel revs(helps with diesel engines).

I will think about diffrent injectors but it is crucial to use high impedance injectors with MS in order to avoid problems with overheating electronics(and failure). Although maybe it will be good to leave the injectors that I have is I use wilder cams.
Next step will be using forged conrods and forged pistons that is for sure but i will make it on croma 2.0l block as I have it in garage(I do not want to use auxillary shaft). I will make it this year.
I have sold this set of IDF carbs and I do not think that it is good to leave that EFI as it works a lot better(no need to clean jets, no mess with setting up). As for the results, I will work on it and I am sure that I will get better results because when I have done a headwork on Opel Astra GSi(c20xe) I could get 206Nm and 163PS from it with standard engine. I understand that my modifications are not perfect but I the only way to get any improvements is to learn on my engines :) For now, every next engine i have tried to modify was better than previous one. I know that there is a lot of work that can be done to my Fiat head and this will be made when replacing internals to forged ones.
I wonder if my throttle bodies are right size. It is 44mm with no choke, maybe this makes the torque so weak in mid range(apart from ignition).

Guy, please look at my ignition table to see if there can be something changed. Also if you could look at the power and torque curves of that Astra GSi, maybe this will give you any idea of my head work if it is correct or it may be improved. I have shaved this head 0,5mm from standard and ported inlets and exhaust. Also inlet manifold was polished(80 grit)and matched as far as I could reach with the grinder. Nothing else was modified as it is a road car. Simmilar work was done to Fiat head.

Going back to MS. I think that it is very good EFI for the money paid for it. It has very good support although it is available only online. It is not made for people that have no idea about such systems as it will require some time and knowledge to tune everything up but if you think that you are able to handle such problems then it will pay you back with quite good results.
Attachments
This is intake manifold with throttle bodies that I use in my engine.
This is intake manifold with throttle bodies that I use in my engine.
intake.JPG (77.31 KiB) Viewed 50623 times
This is my intake manifold with slide plate throttles.
This is my intake manifold with slide plate throttles.
Image00001.jpg (142.68 KiB) Viewed 50639 times
Opel Astra Gsi, after headwork.
Opel Astra Gsi, after headwork.
Image00002.jpg (158.95 KiB) Viewed 50650 times
Opel Astra GSi engine before headwork(200000km done in normal road use)
Opel Astra GSi engine before headwork(200000km done in normal road use)
Image00001.jpg (157.9 KiB) Viewed 50664 times
Ignition table for Fiat 16v. This was not used on dyno but the engine has lot more torque
Ignition table for Fiat 16v. This was not used on dyno but the engine has lot more torque
ignition- MS.JPG (72.06 KiB) Viewed 50666 times
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Post by Guy Croft »

I need to study this in detail and then I'll reply, OK?

GC
Walezy
Posts: 167
Joined: June 25th, 2006, 10:56 pm
Location: Poland, Czyzowice
Contact:

Post by Walezy »

Sorry for that late reply but I had a busy time on the university at the moment but I have succeed with getting Master of Science in transportation :), by the way my thesis was about modifications that make the cars go faster(not only engines).
No rush about this reply, If you need any additional informations please let me know and I will try to give it to you as soon as possible.
Merry christmass to all GCRE forum readers.
petert
Posts: 61
Joined: July 13th, 2006, 12:38 pm
Location: Sydney, Australia
Contact:

Post by petert »

I take it 100kPA is the full, wide open throttle load on the Ignition Map? If so, it seems strange to me that you'd need 44deg. of total advance. I have no experience tuning these engines, but I would have thought maximum advance would have been around 30-32 degs. in the 100 to 90 kPa lines.
vandor
Posts: 108
Joined: November 30th, 2006, 3:24 pm
Location: Texas, USA
Contact:

Slide throttle

Post by vandor »

Walezy,
Interesting slide throttle arrangement!
Did you make it yourself?
I'm loking for something just like that, could you give me any pointers on how to make them?
Thanks,

Csaba
Walezy
Posts: 167
Joined: June 25th, 2006, 10:56 pm
Location: Poland, Czyzowice
Contact:

Post by Walezy »

This throttle bodies are taken from ex-works FSO Polonez 2.0l 8v engine with IDF waffle intake manifold(I had that some time ago and sold the set but I left the throttles for this engine) .Those throttles ware used with Kugelfischer fuel injection. It uses needle bearings to avoid getting stuck open but it needs often to be checked and cleaned to work smoothly.
All you need to make such throttles is two piece of aluminium plate about 20mm thick, a plate 3mm thick made of brass(bronze?) and some needle bearings(about 5mm wide in plastic "stripes"). The aluminium plates need to be machined to so that you will be able to put that brass plate between them but it needs to be done very carepully so that it will not be loose or too tight. THe needle bearings need to be put on the "head side" of the throttle as the engine will be sucking the plate and if you do not use the bearings it may stay open even after using some grease. Two plates need to be matched so that no false air will go into the engine as the idle will be very bad. The plate aluminum plates and the brass plate need to be drilled together at one time to diameter that you need in the wide open position. This is basicaly it, nothing difficult if you know what you are doing or you have a qualified machinist with proper tools.
I have put trumpets from a set of IDF 44 carbs that I had. Now I will use kevlar airbox instead of the two air filters from those IDF carbs.
NickRP
Posts: 126
Joined: September 28th, 2006, 3:26 pm
Location: Nis, Serbia, Europe (A)
Contact:

Post by NickRP »

I agree, the ignition map seems odd to what I've seen on these engines. With Guy's permision, I could post stock Tipo 2.0 16v ignition map.

Walezy, could you put some photos of your slide throttle setup disassembled? I'd like to discuss some details, if possible.

By the way, 2 liter Tipo engine has fully sequential injection.

Regards,
Nikola
Walezy
Posts: 167
Joined: June 25th, 2006, 10:56 pm
Location: Poland, Czyzowice
Contact:

Post by Walezy »

If you can put such map then it would be great as I could make it as default to start tuning it on dyno. You may also send it on my email if it is not allowed to make it here. I will look for some photos or I will disassemble the throttle from the car to make some photos but I will be able to make it next week as now the car is kept outside workshop as I am making the building a bit warm as it was too cold inside to work on the car in winter.
Does it make that much difference(fuel consumption is not an issue) that injection system is sequential?
NickRP
Posts: 126
Joined: September 28th, 2006, 3:26 pm
Location: Nis, Serbia, Europe (A)
Contact:

Post by NickRP »

I will post the ignition map screenshot tonight, no worries.

For full power operation, it makes no difference whether you use fully sequential, semi sequential or batch injection configuration.

I am looking forward to seeing the slide throttle internals photo!

Best regards,
Nikola
NickRP
Posts: 126
Joined: September 28th, 2006, 3:26 pm
Location: Nis, Serbia, Europe (A)
Contact:

Post by NickRP »

As promised, stock maps.

If I were you, I'd start with those 1.8 16v ignition maps.

Regards,
Nikola
Attachments
Ignition map of an unmodified Tipo 2.0 16v ('91-'93) engine.
Ignition map of an unmodified Tipo 2.0 16v ('91-'93) engine.
stock000.png (74.16 KiB) Viewed 50559 times
Ignition map of an unmodified Tipo 1.8 16v engine.
Ignition map of an unmodified Tipo 1.8 16v engine.
stock001.png (73.92 KiB) Viewed 50550 times
christophe
Posts: 3
Joined: August 2nd, 2007, 10:53 pm
Location: Lincolnshire, UK (A)
Contact:

Post by christophe »

Hi Guy and all,

I¢ž¢ve been a megasquirt user for a little over 6 months now and so far have been very impressed with the project so though I would share some off my findings with you.

The car I run is a Ford Orion mk3 with a zetec engine used for road rallying and after a few mild upgrades I¢ž¢d experimented with different Ford ECU¢ž¢s as well as a ‹Å“superchip¢ž¢ and having found different gains from differing combinations I decided a fully mappable system could well yield good results.

The system I wanted would need to be cheap (student loans don¢ž¢t stretch far!) and I would like to do as much off the installation and mapping as possible. This left Megasquirt as virtually the only option with most similar systems Emerald, Omex etc. costing much more and needing most off the work done by authorised tuners. The DIY element of megasquirt also had me intrigued and the online help I¢ž¢d found via the forum and MegaManual looked excellent.

I therefore purchased the ECU along with a megasquirt stimulator from a dealer in the Uk for ‚£180, there are a few different chip and pcb options but these are tailored to the users needs and easily selectable.

The build process was completed in a few days and could be done by anyone with reasonable soldering skills, code writing skills are not required but you do need to know how to read circuit diagrams and work out component values.

The stimulator is a separate PCB, which pugs into the ecu to simulate the engine, allowing the ecu to be tested at each stage off construction, as well as establishing connecting with a laptop once the relevant software has been installed.

The MegaManual has excellent troubleshooting sections which should help you with any building or software issues, my construction I¢ž¢m glad to say was trouble free connection with the laptop was established first time and all off my out of practice seemed to have taken to the board well enough.

Once the ecu has been built the installation into the car is quite varied depending on the application, in most cases a new loom will need to be created to plug into the ecu or an OEM loom can be adapted to fit the megasquirt connector.

In my case I wanted the ecu to be completely interchangeable with the original ecu, this is because the car is in almost daily use and I decide to adapt the loom to megasquirt and it doesn¢ž¢t work I will have problems getting about the next day!

Therefore I utilised an adaptor board pcb from the USA which allowed the DB-37 connector fitted to megasquirt to be turned into a Ford 60-pin EEC-IV connector meaning I could plug the megasquirt ecu straight into where my original ecu went and because megasquirt could be utilised to work with my original sensors and injectors this meant in the initial stages I could swap between my Ford ecu and meagsquirt in a matter off seconds.

This has changed recently as I¢ž¢ve taken advantage off some off the extra features, (removed the MAF sensor, fitted a Wideband Lambda sensor and bigger injectors) but I can still swap in the original ecu and it will run the engine well enough to get me home which is often a concern if we are doing an event 150+ miles away in the middle off the night with no trailer to take us home!

The home tuning element of megasquirt is either feared by a lot of people or embraced, I think in my case I was lucky as the engine I use had been used by many other people and as so tuning maps could be downloaded from the megasquirt site. I did this for both fuel and spark maps and after going through the software setting up the engine parameters I was ready to run.

To my surprise the engine fired first time and with little tuning could be got to a very reasonable idle, initial I tuned via feedback from a standard fitment from a narrowband lambda sensor as fitted to many fuel injection vehicles. This provided feedback on whether the engine was running rich or lean and initially proved useful but infact only provides an accurate exhaust gas reading at a ratio of 14.7:1 therefore this could only be used for tuning low load areas off the map.

My next purchase was therefore I wideband lambda sensor and controller, this adds quite some expense to the project costing ‚£100 as not only is a new sensor needed but a control box is need to drive the complex device, although it is then very easy to interface into megasquirt inplace off the original sensor.

With this sensor in place I can accurately read AFR ratios from 10-20:1 allowing me to tune all areas off the fuelling maps, as It turned out the map I had downloaded was very good and only needed fine tuning in certain areas, I have since been running with this map for the last three months and find I can do very little more with it without the use of a rolling road.

The spark map I have also left virtually untouched as this would be even more difficult and potentially damaging to tune without a rolling road with knock sensors etc.

This is now the current stage off my project, the maps are tuned in as much as I can get them my self and will need to be done on a rolling road to further perfect. Another element of the project is to see what power and torque can be gained, the car originally made 130bhp and 125ft/lb on the rolling road and I have since only made minor modifications and intend to take it back to the same rollers to have the two sets of results compared.

From what I can tell now though the car feels much nicer to drive, low down torque seems to be much improved and even though not fine tuned the overall driveability of the car seems much better, no more hesitation and jerking and smoother power delivery on the whole.

I¢ž¢ve recently been utilising some off the extra features such as closed loop lambda fuelling control, shift lights and am currently implementing a clutch switch to allow me to set-up a launch control and flat shifting programs.

I will report back as soon as I can on what gains the system has shown on the rolling road but even before that I feel it has been well worth the investment and has allowed me to visualise how many other changes have affected engine operating conditions.

Hopefully my insight has been off some use, if a little basic and if anyone would like help or advise I¢ž¢d be more than happy to lend a hand (I¢ž¢m no expert though!) I¢ž¢m only up the road from you Guy so if you have any customers struggling with the system or want to have a poke around yourself let me know and I¢ž¢d be only to happy to pop by.


Chris
Walezy
Posts: 167
Joined: June 25th, 2006, 10:56 pm
Location: Poland, Czyzowice
Contact:

Re: Megasquirt fuel injection

Post by Walezy »

Hello,
I have recently upgraded my EFI to Meagasquirt 2 v3.0 with extra code(1.0.2) and i have changed the ignition to Coil on plug system based on VW/Audi coils. The whole system is running on Alpha N fueling which is essential for individual throttle bodies. The car has been roughly tuned with wideband lambda sensor and it works really nice now. Of course some small problem occured during install but i was able to make it all work.
Book#378
Post Reply

Who is online

Users browsing this forum: No registered users and 42 guests