EFI Discussion

[fj1289]

Thought I'd start an EFI discussion here to continue what was already started.  I think a some people are already thinking and few working towards EFI for the FJ.  I think it'll be handy to bring together some of that background work, knowledge, experience (especially from other motorcycles or cars), and questions.

Cheers   

[fj1289]

I've always been interested in EFI for the FJ, but never really could justify the expense or time or effort since I was very happy with the performance of my Kehein FCRs.  That changed when my trailer was stolen and I lost a lot of engine parts, tools, etc.  I started thinking seriously about a good EFI setup instead of another set of FCRs - especially since there had been a lot of growth in EFI on motorcycles and parts and expertise are a lot easier to find now.

I bought a couple boxes of parts from Marc Rittner (derbybrit1) when he was selling FJ mod stuff to help support his habit (aka getting an Aprilia v-twin over 200mph).  Marc has been very helpful talking through issues and passing on his knowledge.  Quick FJ group history: Marc did a lot of work with Steve Conklin who had his FJ1200 up and running on a set of modified Busa throttle bodies and a microsquirt ECU I think.  As far as I know, Steve never fully completed the install.  I think he was still learning the microsquirt and working on getting it tuned when he had his accident on the GSXR. 

My original plan was to mod the Busa TB's and use a customized microsquirt unit from Interface PNP.  I've already purchased the PNP microsquirt.  It will be set up to run batch mode for the injectors and wasted spark for the ignition so there will be no need for a camshaft position sensor.  The rest of this plan includes using the full Busa harness to replace my very ratty FJ harness and use Busa controls and as many stock sensors as possible.  One senor that will have to be changed is the coolant temp sensor.  I've found a cylinder head temp sensor from Trail Tech that works with their Vapor dash boards that is compatible with the standard coolant temp sensor (most CHT's operate differently from CTS sensors and are not compatible). 

WestOzXJR showed me a clean and easy answer for the crank position sensor off the late model XJRs.  That could save a lot of fabrication and even development work (crank trigger operation is often the source of problems on a project like this). 

Now, it has also been brought to my attention that the throttle bodies off an FJR might work on an FJ without needing to mod them.  I'll be exploring that option too...

Chris W.

[Dan Filetti]

Thought I'd start an EFI discussion here to continue what was already started.  I think a some people are already thinking and few working towards EFI for the FJ.  I think it'll be handy to bring together some of that background work, knowledge, experience (especially from other motorcycles or cars), and questions.

Cheers   

1st: let's start with the basics: has it EVER been sucessfully done?  I've read where more than a couple have tried it; has anyone suceeded in converting an FJ to FI?

Curious.

Dan

[fj1289]

Yamaha with the late model XJRs...but that's not what you're talking about.

Steve Conklin did have his running - but I'm not sure how far he got on his project.  I think he was having some lower end tuning and drivability issues.  But, the megasquirt/microsquirt community and code and hardware have advanced a lot since then. 

WestOzXJR has some time playing with an XJR with an  FI setup with an aftermarket ECU.  Sounds like he was mostly interested in some learning and sorting of the system in prepartion for some much more serious applications!  I'll let him fill in on his plans and experiences. 

I've got to think that there are some out there that aren't on the list - might be in some FJ powered buggy or sand rail!

Chris W.

[Pat Conlon]

I don't know about the FJR throttle bodies. I know room is tight between the FJ frame rails. Going by memory, according to Steve, the XJR throttle bodies would not work because the throttle angle sensor on the XJR bodies is on the left outside of the array, which is fine with the backbone frame of the XJR, but it would not clear the left side perimeter frame rail on the FJ.
That's why Steve and Marc went to the 'Busa throttle bodies with a custom manifold set for the 77-85-77 spacing for the FJ.

Chris, does the Microsquirt have autotune functions?
That's a cool feature...Set your fuel-air targets in the different rpm zones and in autotune, the ecu adjusts the pulse width on the injectors accordingly.

[fj1289]

Pat,

I'll have to see on FJR throttle bodies.  Looking at eBay pics they look pretty compact.  The FCRs I had also had a TPS - cleared with no issues, a little tight, but doable.  Part of the fun of doing something like this is discovering the "art of the possible"    Of course that usually means you spend more than just a bit of time in the WTF zone too!   

The more fabrication required and the more fiddly the project is, the less likely it is to get finished - like Dan mentioned.  I have the advantage of building on some other people's work and more options being available now - especially to facilitate tuning.   

Yes, the microsquirt board that Tony sells does autotune - I don't remember if that is functionality that is build into his board or is part of the microsquirt.  Marc Rittner was  very impressed with the auto tune capability.  He used a full GSXR1000 setup on the supercharged N2O injected v-twin Ape.  How quickly and how well he was able to tweak the maps was impressive -- even more so considering how different his install was compared to a "typical" GSXR1000!

Couple comments on auto tune - it's not a tuning god in a box.  It requires you to datalog your ride, download the info and let the software calculate the changes needed for the map, then make the changes to the maps and reload the new maps.  Also, and this is a biggie - it only works towards the AFR map you ask for -- so you still have to optimize your target values.  So if you are asking for values that are too rich or too lean, you'll get them! 

As far as tuning EFI goes, it is a LOT like tuning carbs.  It is an iterative process - and what is "perfect" on the dyno is rarely the best setup for the real world.  Testing in the real world is the only way to really find out what the engine wants.

I'm hopeful this will give a lot more possibilities for tuning the engine too - with carbs it is always a balancing act on a streetbike between having enough fuel for acceleration and performance while also trying to keep the cruising range on the leaner side for mileage.  With the auto tune you should be able to optimize both regimes (like having two sets of jets in the carbs at once) -- set the large throttle opening settings on the richer side for performance, but able to keep the smaller throttle settings leaner for cruise.  In a sense I can have the "pilot circuit" optimized for both cruise and acceleration - always a compromise setting with carbs - acceleration pumps help, but have their drawbacks too. 

Chris W.

[WestOzXJR]

In the context of development of the Yamaha EFI system for the 2007 XJR, what came immediately before was a rather snatchy to ride in traffic FJR (which they later cured with re-mapping).
So I believe that the overall design of the XJR system was, at least partially, a knee-jerk (over) reaction to curing this matter. By this I mean that XJR had some interesting design features probably aimed towards making the bike very pleasant and civilized to ride in traffic.

Notably a secondary set of butterflies in front of the first that are opened by a stepper motor arrangement according to a table programmed into the ECU. I've tested the bike with this disconnected and it shows no difference on the dyno, at least at WOT, although the engine is noticably more responsive (not so soft) in transition from low throttle blade angles to WOT with the link to the secondaries disconnected and wiring them wide open. So I'm not going to emulate this idea in my home brew injection design.

Also, unlike a lot of EFI bikes that essentially position the injectors at the throttle blades, XJR's run the injectors in alloy manifold that replace the 36Y carb mount boots that we have in FJ's. (At the moment I don't have any pics with me as I'm at my second home in Thailand for a couple more weeks yet, but when I get back to Perth soon I'll post some up some photos as I have a head/manifold/Suzuki K1 throttle body combination on the bench I've been working on).

My assumption here (possibly incorrect) is that the Yamaha engineers probably had in mind the fact that FJ/XJR intake tracts are curved and therefore they set the injectors up at a rather odd angle (and at different angle for cyls 1 to 2 and for 3 to 4 to compensate for the different curvature of inner and outer cylinder intake tracts and to squirt the fuel out into centre of the turn radius of the intake tract in a probable attempt to improve low speed fuel atomization - probably a good idea, enough so that I plan to keep this part of set-up as my primary injectors.

The XJR ECU is an absolute COW to work with. I stay in contact with a guy in the UK who successfully got a piggy back controller to work when previously even Dynojet had given up and Dobek were doing some work with a bike on the dyno in Holland too but that also came to nothing... So although I wanted to keep it Yamaha, in the end I bit the bullet and decided to go after-market for a multitude of reasons including the difficulty getting an accurate rpm signal from the factory ECU for my air/fuel ratio data-logger and other electronics... It was just easier and neater to go after-market... And so there went a year of testing...

Another interesting part of the design is that throttle blade diameter is 34.5mm which probably contributes to these engines now make maximum torque at only 6000rpm. What I don't like about them is that it puts a large constriction in the intake tract at a point where either side of this is about 36mm and it make the shape of the tract less than ideal. I chose early K1 throttle bodies not only because they're the size I've calculated I want, but also they're very easy to change the centre to centre spec to suit the FJ/XJR. FJR bodies are less that ideal in the respect of this changeability.

About the following statement: "....It requires you to datalog your ride, download the info and let the software calculate the changes needed for the map, then make the changes to the maps and reload the new maps." Actually that statement is not correct... Auto-tune enables the ECU to alter the air/fuel ratio on the fly in real time within milliseconds according to the feedback from the oxygen sensor which is compared to the target a/f ratio programmed into the ECU and then the injector pulse widths are immediately changed. The ECU then awaits further response from the oxygen sensor to report if the target has been met and further changes made as necessary and so the loop continues. That's closed loop tuning. Open loop tuning is running on the fuel table "map" that is user programmable and whether the map is right or wrong, ya run what ya bring. Concensus advice would be to get it as close as possible first in open loop mode and then grab the last few percentage point with the addition of closed loop.

If you set up the hardware for auto tune, ie: closed loop tunable ECU + oxygen sensor + can data log or at least display the a/fuel ratio then tuning is easy once you've got the hang of how to adjust the fuel tables, no dyno time necessary, just do it in the real world. Good part is not losing 1/2 a day pulling things apart, just hookup with pc and make the changes. With the Edge ECU and optional wireless link then it's all just too easy.

Target air/fuel ratios for EFI systems are typically 14.7:1 because that is the stoichiometric ratio (most complete combustion): where oxides of nitrogen and carbon monoxide emissions are at the minimum. Of course more power at a richer target and better fuel consumption at a leaner target is possible. Achieving stoichiometric is WHY XJR went EFI, to meet tough Euro-spec emission requirement and Europe is where XJR's are still an incredibly popular motorcycle.

About TPS. In actual fact you do not need a TPS. It can be done with Manifold Air Pressure sensor alone. Only TWO things need to be measured/read for an injection system to function... They are rpm and load. TPS is a convenient way to measure load. Mark Dobek, one of the founders of Dynojet (and later sold the business) now have a system that indeed does not read TPS and in fact many systems both past and present do not use TPS, they measure manifold air pressure which is actually a more accurate representation of load. throttle position is merely a convenient guestimate of load.

About the statement: 1st: let's start with the basics: has it EVER been sucessfully done?  ...... Who cares if it's never been done before? Good thing NASA never asked that before designing a vehicle to go to the moon. Can you imagine that, "ohh nahh, fuggit, no-ones ever done it before so we won't even try..."

And I'm going to conclude with the controversial statement, that a well designed and set up fuel injection system will out perform carburettors every time, although the key phrase here is "well designed and well set-up". Easier said than done.

[Pat Conlon]

Nice write up Oz. Thank you.

I agree, closed loop autotuning is where it's at.... it will get you real close until open loop occurs.
I suspect the dyno operators have lost $$ due to autotuning features on ECU's.

I should say, my EFI experience is third hand. I hang out with the Miata crowd and, over the years, they have had some painful experiences with futzing around with aftermarket ECU's. By the time you put in hot cams, individual throttle bodies and or turbos, the stock Mazda ECU will not cut it.
What the Miata dudes have learned is that batch injection is fine for dedicated track cars, but if you want any semblance of a idle or smooth low rpm behavior for around town driving, you need sequential injection.
So....
What's the deal with the cam angle sensor on the XJR? Would you need a XJR head with the appropriate castings for the CAS?
I have not seen a picture for a XJR head with the CAS feature so I don't know if there would be any clearance issues in the FJ.
 

[fj1289]

Pat, it would be interesting to know what type of maps they were running on and what kinds of mods.  I think a lot of the older ECUs ran on either MAP/RPM or TPS/RPM.  Each have their benefits and drawbacks.  MAP/RPM is a better sense of the actual load the engine is under, especially once it is underway.  TPS/RPM is better at  reacting to small rider inputs, especially at slower speeds and if there are large cams or individual intake runners where the MAP tends to pulse more.  Most of the current ECUs use a blend of the two - biased toward the TPS/RPM down low and MAP/RPM for midrange and up.  And both can be used to determine acceleration fueling (or blended once again). 

I wonder if the problems were coming from the engines being extensively modified and only having the MAP/RPM tables to work off of -- maybe sequential was a way to improve the engine response relative to the manifold pressure changes; where using a TPS/RPM input might have given better driveability still at that point regardless of batch or sequential injection?

Anyone know of either a posted service manual for the 07+ XJRs?  Or a parts fiche that covers the european models?  Id like to see how they did the cam sensor.  Could be as simple as a boss on the cam cover with a sensor picking up a cam lobe...or a completely different design requiring different head, cams, etc... I wonder.

Chris W.

[WestOzXJR]

Injected XJR's are batch fired/wasted spark. FJR's (and R1's - even the earlier ones) are sequential. Hence even the current XJR does not run cam sensor. Which in my mind is proof enough (from the way these engines perform) that batch fired/wasted spark is plenty fine enough. Sequential is MUCH harder to tune and requires dyno time to get the injection timing correct.

If you guys could ride one of these late XJR's you'd quickly become convinced that batch fired injection can give you the rideability you (and indeed I) want in a road motorcycle that gets ridden in traffic and therefore needs to be civilized around town at low rpm.

One area where injection systems have advanced substantially is nozzle design. Even only ten years ago, 4 cylinder bikes that made more than 150hp usually had either; serious driveability issues OR EIGHT nozzles. The problem was selecting a nozzle that atomized well at low fuel rates yet could provide enough fuel for the higher hp levels. Even 200hp bikes that used to run eight nozzles now have reverted to FOUR such is the advancement in nozzle design. If you're going to do this, chose late model nozzle.

Of equal importance, from what I've learned, is the placement of the nozzle has a noticeable impact on low end throttle response and driveability. Placing the nozzle closer to the inlet valve improves low end engine flexibility/response and further away equates to improvement in upper rpm performance. Those two facts together with my goal of a higher than 150hp target have led me to plan a system based around eight nozzles. If my target was less than 150 I'd stay with just the four in the inlet manifolds. With the curved inlet tracts I see the advantage of placing the nozzle in the curvature. Since one of the engines I'm building is a turbocharged development engine to test the idea of TIMED direct injection of water straight into the combustion chamber (kinda like a modern day steam engine hybrid), I'm planning on a LOT more that 150hp so much of what I'm playing with is geared to that end.

I heard before here on this very forum, of FJ's being referred to as, "car passing mother f**kers", which indeed they are. But believe me when I say what they have done with this engine now in terms of bottom end torque is truly astounding by comparison to even our ubiquitous FJ's. These engines now pull like a train in top gear from 30mph and it's a challenge to keep the front wheel on the ground in first and second gear even when slowly winding the throttle on from low rpm. Snap the throttle open in first from low rpm and it'll go over the top and buck you off quicker than blink and that'll be well before the 6000rpm where they achieve maximum torque. Yes it's true too that XJR's a fair bit lighter on the front end but these engines have pleasantly evolved.

Sure, they are now a little anaemic in the top end, my standard FJ outruns my standard XJR's in the 9 to 10,000 rpm range, not helped by a very restricted catalytic converter exhaust (albeit 4:1), but nonetheless I remain convinced from the dyno sheets I have seen from Yoshimura and the testing I have done thus far that the current engine with some minor exhaust mods can be brought alive in the upper range with no loss of bottom end torque and driveability.

Service manuals... I have factory service and spare parts manuals for '07 and up XJR in PDF format. I think the files are a little too large to email however, when I get back to Perth in a couple of weeks (to a postal service I can trust - don't yet know whether to trust the Thai postal service), I'll gladly snail mail manuals on a USB key to a few of you at no charge IF you'll pay it forward to others in the group who may also wish to have.

[fj11.5]

so,a few frame mods around the carb area, would an xjr motor slot in with efi still installed 

[WestOzXJR]

Will measure when I get home in a couple of weeks and report findings.

[candieandy]

This has my full attention. i would love to build my FJ to EFI. Oh man, i could run a wideband sensor and have the best of both worlds.. Power and good MPG. haha.

[Dan Filetti]

<snip> Oh man, i could run a wideband sensor and have the best of both worlds.. Power and good MPG. haha.

'haha' is right.  The Gixxer Seven-fiddy is fuel injected, it has power 133HP with a power commander and a full Ti Yosh pipe and a TRE.  However it decidedly does NOT get very good gas mileage ~35MPG when ridden 'normally' -I can make it get less than 30...  Hell, I typically get that same 30+MPG with my Accord...

Seems to me it's one or the other, good MPG or Power, but not both...

Guessing that's what you meant by 'haha'

Dan

[WestOzXJR]

Well maybe if you add some back pressure back into that exhaust system and get a good tuner to have a play with the cam timing you would see some real improvement.

But of course no-one's ever done it before (that you know of) so it's no doubt not possible right?

Or perhaps it's just that kind of engine with such a gutless bottom end it's never going to be efficient at anything but ludicrous engine speeds.