Hi all,

If you recall, I wrote last year about the conversion of my KLX to fuel injection (My Fuel Injected KLX). This project has been pursued this year with a custom throttle body replacing the "fuel injector in place of the carburator slide" configuration. Although the latter worked to my satisfaction all of last year, that solution was only intended as temporary, as I feeled it was not optimal for the following reasons:

- once you remove the slide, the carb main tract shows lot of grooves and discontinuities that can create air flow disturbances and cavities where fuel could condense;
- a carburetor uses a venturi which necessarily introduces a pressure drop, however small. Thus some hp is lost there, though I don't have any idea of its amount. With FI, there is no need for a venturi thus the intake tract can be made of a constant cross section until it enters the head so as to maximize airflow;
- a throttle plate after the venturi also introduces some further flow restriction. Actually, it's more the shaft that is causing the restriction, not the plate itself. That's the main reason why a slide carb is superior to one with a throttle plate. It was not possible for me to build a custom slide carb, but with a larger intake conduit the throttle plate restriction can be mitigated.

There's also an educational aspect that plays a role here.

Thus last spring we designed and built a custom throttle body with injector holder for the KLX. I planned to report about it earlier but decided to hold off until I could have the bike on a dyno, which finally occured last week.

The end result is quite good, I qualify it as 95% success. I've ridden over 3000km of mostly hard trails this summer, with near perfect behavior. Right after the installation and tuning of the new setup, I could feel an improved top end performance and the dyno results proved me right: 22.3hp peak @9200rpm, with just a slight rolloff to 10000rpm.

The only minor imperfection is that idle is a little inconsistent, varying from 1800 to 2000 rpm with the current adjustments. If I try to make it lower, it will vary too much and the engine nearly stalls in some circumstances. This is due to not having a separate idle fuel circuit like is done in commercial implementations. Fortunately this imperfection in no way impacts rideability of the bike. It's just cosmetically affected, not functionnally.

I think with my setup I get nearly all that can be obtained from the engine without internal modifications. That is, with the stock CDI ignition (yes, it's still the stock CDI that controls ignition on my bike). Replacing this CDI with a custom programmable one is the next step planned for this winter, so expect I'll come back next spring with updates on it.

I'll proceed in a subsequent post with some details of the custom throttle body, then the dyno results. I welcome your comments, impressions, questions and suggestions. Keep in mind that this project was for fun and for learning. You should take this report as entertaining, not see it as me claiming the final authority in EFI bikes or the KLX.

 

Here is a photograph of the custom throttle body. You can see the TPS sensor on the left, throttle actuation mechanism on the right, injector hole at the bottom.


As it is, it's a little bulky but it fits. If I were to do it again, it would be 1.5cm less wide. There would be a few other improvements too. But for the present purpose, it's all ok.

One interesting characteristic is the cross section: internally it starts on the airbox side with a diameter of 52 mm to match the airbox conduit. Going towards the engine, it tapers to a 42mm section that contains the throttle plate. Then another taper to 35mm to match with the head port.


Another view of it, from the airbox side, with throttle plate open:


Here, it is installed on the bike.

 

I've been missing dyno facilities since the start of this project. It would really be interesting to compare dyno curves with the stock bike, then with last year setup, and this year with the new throttle body. Different dynos might produce significantly different results in low hp ranges, thus it is best if comparisons are done with the same dyno. I have access to a dyno now, unfortunately I can't return the bike to stock condition for comparison purposes.

The dyno I have access to is in a shop nearby my house. It's a car wheel dyno, not a roller one: it is designed such that car wheel hubs are attached to two towers. This configuration tends to produce very accurate results as there is no uncertainty with tire to roller power loss. However it is not directly suited for a motorcycle. That shop did measure a few motorcycles in the past, with a custom made fixture based on a shaft that replaces the rear wheel. That fixture did not fit directly the KLX, we had to build something special for it.

Here is a photo of the KLX attached to the two towers


Here viewed from the back


And here is the much awaited result: blue is power, green is torque


Notice the absence of sharp power rolloff above 8500 rpm we typically see in other published power curves. Power stays above 21 hp from 7300 up to 10000 rpm. I think this is indicative that the CVK34 is the main limiting factor for the high end of the engine range. It was also my greatest reward, although I sincerely hoped for something closer to 25hp (recall that there is no wheel to roller loss in this setup).

Actually, one might question whether that 22.3hp figure is accurate: we got this on a car wheel dyno having a range of 800 hp, so it does not appear really well matched to the task of measuring the KLX. Here are a few facts that help giving credits to the results we got:
- In terms of torque, we use approximately 1/6 of the range of the equipment, which is not that bad compared to the hp figures. We do the runs at wheel speeds lower than typically done with high power cars (the runs were in 3rd gear), and the dyno works fairly well even at these low speeds.
- successive runs show very repeatable results, within 0.1hp;
- even if we are uncertain about absolute accuracy in the 20hp range, we believe the shape of the power curve can be relied upon.
- for comparison purposes, they did a few CRF250 in the past, showing around 32 hp for stock ones, up to 37 hp for highly modded ones. Apparently these figures are about right for these bikes.

At least I can tell you that the high end improvement in power is really apparent when driving the bike: on the street, I tend to upshift at about 10K rpm when I want the greatest acceleration, vs 8500-9000 rpm before.

 

Very sweet project! Are you gonna sell your throttle bodies when you get them ready? I'd love to have fuel injection. Wonder how much it would cost to buy the parts from a D-Tracker?

 

There is no definite plan yet, but I doubt we will go commercial with this within the next year, if ever. There remains a lot of development work to have a commercial ready kit or whatever. Our main business is more in the indutrial sector. Anyway, we will have more precise plans within a few months.

Add to that the arrival of those Ecotrons kits that TNC pointed us in another thread. If it works as they say, it makes for a very affordable FI project! Honestly, I wonder how they do it for the price shown on their web site. If we were to offer a similar kit it would hardly be under $1000...

 

Big standing ovation to Ray!! <where's the clap-clap emoticon..??>

Great work, and good on you for taking such a challenging project all the way thrugh!

Couple questions.
- Do you have measured A/F curves?
- What is your air box configuration? Lid on/off, snorkel on/off....

--
Mikko

 

I started a similar project a fews years ago but stopped when I couldn't find some needed information on timing of the fuel injector.

1. Are you triggering the injector on every stroke or have you found a way to time it to the open intake valves?

2. Do you vary the start of the injector pulse in relation to RPM?

Here's a couple shots of my injector body and injector mount>

Ride on
Brewster

 

Yes, the A/F curve is measured while on the dyno. The fueling rate is adjusted such that the A/F is nominally 14.5 for 0-60% throttle, down to 13 at 100%. In practice it varies a little over the operating range, but stays within 14.5 to 15 at 0-60% and 13 to 13.5 at 100%.

My airbox configuration is lid on, but no snorkel. I know from my experience that even with the snorkel, there is very little air flow restriction from the airbox. What people are seeing when they change their airbox configuration is the pattern of acoustic waves that changes how much fuel is drawn from the jets. An FI system is not influenced by air waves like a jet, thus I expect that I could run with the snorkel without any difference in performance.

 

Hey Brewster, what's your take on the discussion over on ADV about that FI issue currently running that eakins brought up?

 

Your throttle body looks much cleaner than mine That project was for which bike?

1. Yes, the injector is triggered at every revolution, there is nothing in my setup that detects intake stroke. We know from experience that performance issues are negligible with this scheme, compared to one injection pulse every cycle. There is a significant benefit though in regards to emissions, we experimented cases where there were 20-30% more HC when operating "dual pulse" (i.e. every turn), vs properly timed "single pulse". But I guess that I already have much lower emissions with my FI bike compared to a non FI.

2. The injection pulse is phased to start at 0 degrees, indepedant of rpm and load. We can program the ecu for a phase that varies with rpm and throttle position, but there is no significant benefit when operating in "dual pulse". Thus I did not bother to program it with a varying phase.