Fuel injected 860 GTS
-
paulh
- Cucciolo - the Lil Pup
- Posts: 11
- Joined: Tue Apr 17, 2007 8:23 am
- Location: Central Queensland, Australia
Fuel injected 860 GTS
This is my 1976 860 GTS which I have owned since 1980. In this time it has undergone many changes, the most recent being the addition of an engine management system that controls both ignition and fuel. This has been a completely DIY project involving a very steep learning curve and the sourcing of many parts off ebay and elsewhere around the world. I undertook this purely for the challenge and could not be happpier with the outcome.
Last edited by paulh on Tue Dec 22, 2009 6:12 am, edited 6 times in total.
-
cafe*racer
- SSD900 Darmah
- Posts: 364
- Joined: Tue Nov 16, 2004 1:49 am
- Location: Castro Valley, CA
-
cafe*racer
- SSD900 Darmah
- Posts: 364
- Joined: Tue Nov 16, 2004 1:49 am
- Location: Castro Valley, CA
-
paulh
- Cucciolo - the Lil Pup
- Posts: 11
- Joined: Tue Apr 17, 2007 8:23 am
- Location: Central Queensland, Australia
Macdesmo and austduke
Yes, you are right, it was a lot of work but probably more mentally than physically. The secret was in the planning and understanding how the ECU worked, what it was capable of and how it could be applied to the Ducati twin. It was an iterative process to find the right setup and configuration for my engine. The fuel part of it was easy but the ignition was much more difficult. I could write a book about what had to be considered to make this happen but basically it involved things like what size throttle bodies and what injector flow rates to use (throttle bodies, injectors and fuel pump regulator transplanted from 2003 Ducati Monster 800). These suited the tune of my heads perfectly. Another big issue was fuel supply because I didn't want to modify the tank to fit an intank pump. As a result I was forced to use an automotive inline pump which usually draw too much current an flow much more than is needed for motorcycle engines. The answer was to use a PWM motor speed controller to slow the pump down therefore decreasing the flow rate from 160 to 45 LPH.
Sensors used include a VR magnetic speed sensor on the crankshaft for ECU tach input and ignition timing, a coolant temperature sensor installed in the rear cylinder rocker cover to measure engine temp, an intake air temperature sensor installed under top triple clamp so it doesn't heat sink from the engine, a throttle position sensor operating on the rear cylinder throttle body, a manifold absolute pressure sensor measuring manifold vacuum from both cylinders via a small plenum chamber for smoothing, and a wideband O2 sensor installed in front cylinder header pipe to measure AFR for closed loop EGO control. Mostly this is used for tuning though.
Ok I've probably bored you already with this stuff but believe me this is only scratching the surface. Ausduke, I had to upgrade the alternator and regulator to meet the extra load imposed by the fuel pump, four high energy smart coils, high impedence injectors and the heating element in the O2 sensor. I transplanted a 350 watt alternator from a 1998 Ducati 916. The stator fitted without drama but the rotor was too wide and had to be machined carefully to avoid weakening the structure. This combined with the regulator off a similarly late model Ducati has worked perfectly to overcome a potentially big problem.
So, how does it go and what differences do I notice? Starting is effortless (electric start, though I have kick started with this setup). I control the intake air on cold starting manually by using the throttle stop and after hitting the button it fires straight up and holds the engine speed as determined by the throttle position. Even before getting any where near full temp it will hold a steady 1000 rpm idle and this does not falter no matter how long it idles for. Considering that I haven't yet tuned the acceleration enrichments, the throttle response on quick opening is excellent (compared to EFI Monsters that I have sampled which can be vague and fluffy in this area). Throttle response across the board is impressive and I'm comparing this to the 38mm flatslide Mikunis that this engine was using prior to EFI. The response of the Mikunis was very very good. In general the breathing of this engine is noticeably improved with the 45mm chokes but keep in mind that these heads are a long way from standard. Given the tune of this engine, the most noticeable difference for me is how the EFI has smoothed it out.
The other big difference is tunability. With access to a 12 X 12 fuel map, a 12 x 12 spark advance map, a 12 X 12 AFR map and a sh**tload of other configuration stuff, the sky is the limit for tuning. The real gains to be made with this setup are in spark advance tuning and that will probably mean some dyno time. To say that I'm like a dog with two dicks at the moment is an understatement.
Anyway you have probably been bored sensless by now so I'll leave it at that.
Paul
Yes, you are right, it was a lot of work but probably more mentally than physically. The secret was in the planning and understanding how the ECU worked, what it was capable of and how it could be applied to the Ducati twin. It was an iterative process to find the right setup and configuration for my engine. The fuel part of it was easy but the ignition was much more difficult. I could write a book about what had to be considered to make this happen but basically it involved things like what size throttle bodies and what injector flow rates to use (throttle bodies, injectors and fuel pump regulator transplanted from 2003 Ducati Monster 800). These suited the tune of my heads perfectly. Another big issue was fuel supply because I didn't want to modify the tank to fit an intank pump. As a result I was forced to use an automotive inline pump which usually draw too much current an flow much more than is needed for motorcycle engines. The answer was to use a PWM motor speed controller to slow the pump down therefore decreasing the flow rate from 160 to 45 LPH.
Sensors used include a VR magnetic speed sensor on the crankshaft for ECU tach input and ignition timing, a coolant temperature sensor installed in the rear cylinder rocker cover to measure engine temp, an intake air temperature sensor installed under top triple clamp so it doesn't heat sink from the engine, a throttle position sensor operating on the rear cylinder throttle body, a manifold absolute pressure sensor measuring manifold vacuum from both cylinders via a small plenum chamber for smoothing, and a wideband O2 sensor installed in front cylinder header pipe to measure AFR for closed loop EGO control. Mostly this is used for tuning though.
Ok I've probably bored you already with this stuff but believe me this is only scratching the surface. Ausduke, I had to upgrade the alternator and regulator to meet the extra load imposed by the fuel pump, four high energy smart coils, high impedence injectors and the heating element in the O2 sensor. I transplanted a 350 watt alternator from a 1998 Ducati 916. The stator fitted without drama but the rotor was too wide and had to be machined carefully to avoid weakening the structure. This combined with the regulator off a similarly late model Ducati has worked perfectly to overcome a potentially big problem.
So, how does it go and what differences do I notice? Starting is effortless (electric start, though I have kick started with this setup). I control the intake air on cold starting manually by using the throttle stop and after hitting the button it fires straight up and holds the engine speed as determined by the throttle position. Even before getting any where near full temp it will hold a steady 1000 rpm idle and this does not falter no matter how long it idles for. Considering that I haven't yet tuned the acceleration enrichments, the throttle response on quick opening is excellent (compared to EFI Monsters that I have sampled which can be vague and fluffy in this area). Throttle response across the board is impressive and I'm comparing this to the 38mm flatslide Mikunis that this engine was using prior to EFI. The response of the Mikunis was very very good. In general the breathing of this engine is noticeably improved with the 45mm chokes but keep in mind that these heads are a long way from standard. Given the tune of this engine, the most noticeable difference for me is how the EFI has smoothed it out.
The other big difference is tunability. With access to a 12 X 12 fuel map, a 12 x 12 spark advance map, a 12 X 12 AFR map and a sh**tload of other configuration stuff, the sky is the limit for tuning. The real gains to be made with this setup are in spark advance tuning and that will probably mean some dyno time. To say that I'm like a dog with two dicks at the moment is an understatement.
Anyway you have probably been bored sensless by now so I'll leave it at that.
Paul
Last edited by paulh on Tue Dec 22, 2009 5:43 am, edited 1 time in total.
-
cafe*racer
- SSD900 Darmah
- Posts: 364
- Joined: Tue Nov 16, 2004 1:49 am
- Location: Castro Valley, CA
-
paulh
- Cucciolo - the Lil Pup
- Posts: 11
- Joined: Tue Apr 17, 2007 8:23 am
- Location: Central Queensland, Australia
ECU - Microsquirt (http://www.microsquirt.info/)
Throttle bodies - 2003 Ducati Monster 800 (complete with 3 bar Magneti Marelli FPR and 330cc/min Marelli pico injectors and also the throttle position sensor)
Engine temperature sensor - GM style closed element
Intake air temperature sensor - GM style open element
MAP sensor - GM 3 bar (generally not used on Ducati engines because valve timing with big overlap and the use of individual throttle bodies do not allow good enough resolution of MAP to be usefull as an input for fuel and spark control, but I'm an optimist and will eventually try to get it to work)
Wideband O2 controller - Tech Edge 2J1
Wideband O2 sensor - Bosch 4.2 LSU
Fuel pump - Walbro GSL393 (160 LPH)
Coils - Delphi LS2 style high energy with built in ignitors and dwell limiting capability
Crankshaft - ai-tek high sensitivity magnetic speed sensor (7085-1010-120)
High pressure fuel filter - Napa Gold
Trigger wheel - made my own 24-1 wheel from 1/8" mild steel mounted on a boss which is also a split collar clamp for fixing to crankshaft
Lumpy, I have probably gone over the top but I hope this answers your question.
Throttle bodies - 2003 Ducati Monster 800 (complete with 3 bar Magneti Marelli FPR and 330cc/min Marelli pico injectors and also the throttle position sensor)
Engine temperature sensor - GM style closed element
Intake air temperature sensor - GM style open element
MAP sensor - GM 3 bar (generally not used on Ducati engines because valve timing with big overlap and the use of individual throttle bodies do not allow good enough resolution of MAP to be usefull as an input for fuel and spark control, but I'm an optimist and will eventually try to get it to work)
Wideband O2 controller - Tech Edge 2J1
Wideband O2 sensor - Bosch 4.2 LSU
Fuel pump - Walbro GSL393 (160 LPH)
Coils - Delphi LS2 style high energy with built in ignitors and dwell limiting capability
Crankshaft - ai-tek high sensitivity magnetic speed sensor (7085-1010-120)
High pressure fuel filter - Napa Gold
Trigger wheel - made my own 24-1 wheel from 1/8" mild steel mounted on a boss which is also a split collar clamp for fixing to crankshaft
Lumpy, I have probably gone over the top but I hope this answers your question.
Injected GTS
Truely a unique and very interesting GTS!! A monument to your engineering abilities. Just don`t tell my old GTS or she`ll be very jealous. The VR speed sensor, is that VR as in VR Holden Commodore??? Also the ECU, I gather this is a programable unit. Please forgive my ignorance as electronic fuel injection is a complete mystery to me. Just trying to get my head around how you got the ignition timing to work in conjunction with fuel injection system. I`m very interested in the ignition system as I am currently saving my bickies to lash out on one of those DMC systems for my 76 GTS as I have a real dislike for the old Ducati Electronica systems and their advance step.
-
paulh
- Cucciolo - the Lil Pup
- Posts: 11
- Joined: Tue Apr 17, 2007 8:23 am
- Location: Central Queensland, Australia
Lumpy,
VR stands for 'variable reluctance'. It consists of a magnet and a coil so therefore it does not require an external power source. Essentiallly a magnetic flux is created and as a ferrous metal target approaches and passes the pole of the sensor an AC voltage is generated which becomes the input pulse for the ECU. I don't know about the engineering ability as I don't have an engineering background. Basically it came about because I read an article in a magazine about a project that involved the conversion of a later model belt drive 900SS from carb to EFI by grafting on a complete system from a later model Ducati. I wanted to go further than this and implement a fully user programable system so I got into some pretty extensive research to learn about engine management and to see what was out there. This led me to 'Microsquirt' because it was completely DIY and affordable.
With regards to how the ignition part works, I am struggling to find a way to briefly summarize it for you. But I'll have a go anyway. Basically everything starts with the crank sensor and trigger wheel. The Microsquirt has a number of user selectable ways of dealing with this input to create the correctly timed ignition events. One is 'dual tach input' which I chose because of the odd firing angle of the Ducati 90 degree V twin engine. I use a multi tooth trigger wheel because it delivers a higher level of timing accuracy as a result of the software using a tooth counting and time prediction algorithm to determine when the next tach input should occur. The missing tooth on the wheel creates a reference point. The tach event occurs on the first tooth after the missing tooth, but this can be shifted to another tooth by using a delaying configuration just in case the wheel's physical mounting position cannot be altered. I have set mine up to provide a tach input on the first tooth after the missing tooth (declared) and this physically occurs at 25 degrees ATDC. After sync this event is assigned to ignition output #1 and injector output #1 which are tied to the rear (timing) cylinder. The tach event for the front cylinder is created by skipping 24 teeth (360 degrees crankshaft rotation) from where the rear cylinder event occurs. Because of the odd firing angle ( front cylinder fires 270 degrees after the rear cylinder) this event has to be offset by the appropriate amount to get correct timing. This event is then applied to ignition output #2 and injector output #2. The ATDC timing happens because the ignition events are setup in advance to allow adequate dwell time and the application of spark advance as determined spark advance table. The hardware and software of the ECU do the rest. The ignition outputs from the ECU are logic level signals containing the user configurable dwell periods and these are cabled directly to the coils which do the rest.
I think that I am finding it much more difficult to put into words a description of what I have done than it actually was to do it.
Here are some screenshots of the tuning software interface and tuning tables that I am using.
VR stands for 'variable reluctance'. It consists of a magnet and a coil so therefore it does not require an external power source. Essentiallly a magnetic flux is created and as a ferrous metal target approaches and passes the pole of the sensor an AC voltage is generated which becomes the input pulse for the ECU. I don't know about the engineering ability as I don't have an engineering background. Basically it came about because I read an article in a magazine about a project that involved the conversion of a later model belt drive 900SS from carb to EFI by grafting on a complete system from a later model Ducati. I wanted to go further than this and implement a fully user programable system so I got into some pretty extensive research to learn about engine management and to see what was out there. This led me to 'Microsquirt' because it was completely DIY and affordable.
With regards to how the ignition part works, I am struggling to find a way to briefly summarize it for you. But I'll have a go anyway. Basically everything starts with the crank sensor and trigger wheel. The Microsquirt has a number of user selectable ways of dealing with this input to create the correctly timed ignition events. One is 'dual tach input' which I chose because of the odd firing angle of the Ducati 90 degree V twin engine. I use a multi tooth trigger wheel because it delivers a higher level of timing accuracy as a result of the software using a tooth counting and time prediction algorithm to determine when the next tach input should occur. The missing tooth on the wheel creates a reference point. The tach event occurs on the first tooth after the missing tooth, but this can be shifted to another tooth by using a delaying configuration just in case the wheel's physical mounting position cannot be altered. I have set mine up to provide a tach input on the first tooth after the missing tooth (declared) and this physically occurs at 25 degrees ATDC. After sync this event is assigned to ignition output #1 and injector output #1 which are tied to the rear (timing) cylinder. The tach event for the front cylinder is created by skipping 24 teeth (360 degrees crankshaft rotation) from where the rear cylinder event occurs. Because of the odd firing angle ( front cylinder fires 270 degrees after the rear cylinder) this event has to be offset by the appropriate amount to get correct timing. This event is then applied to ignition output #2 and injector output #2. The ATDC timing happens because the ignition events are setup in advance to allow adequate dwell time and the application of spark advance as determined spark advance table. The hardware and software of the ECU do the rest. The ignition outputs from the ECU are logic level signals containing the user configurable dwell periods and these are cabled directly to the coils which do the rest.
I think that I am finding it much more difficult to put into words a description of what I have done than it actually was to do it.
Here are some screenshots of the tuning software interface and tuning tables that I am using.