EFI Interest

Good day everyone.
I dont post often but I'm a reader of the forums. I have a Yamaha xj750 that I am planning to convert to EFI. I got the bike without an airbox and I'm interested in converting from carbs. They work but they are not the best.
I've seen other EFI conversions and I think I have the know-how to do it myself (I studied engineering). For those interested in EFI, what persuaded/dissuaded you from doing the conversion? Cost? Complexity? Inconvenience? Lack of electrical power? I'm trying to lookout for all the pitfalls.

 

All of the above...
Cost - you could do it with the myriad of diy efi systems out there if you have the skills. You will need some throttle bodies from a modern bike, and the skills to fit them. So maybe not so expensive?
Complexity - it’s very simple really, map it according to speed and throttle position with lambda correction?
Inconvenience - silly question.
Lack of electrical power - now this s potentially the killer blow - you have 20A or so to play with, the fuel pump will take half or more of that. Doesn’t leave much for lights and topping up the battery. It has been done though, one of the forum members on here has had much success with it.

 

Not an XJ but a short article about the 81 1100GPZ EFI.

FTA:
Almost all the ones I looked at on the internet had been converted to carbs, since no one could deal with the injection diagnostics. So to find a working fuel injection model was a rare find for sure. I had the impression that the fuel injection may have been what frustrated the previous owner, causing him to park the machine for years before selling it.

https://cdnbkr.ca/vintage-motorcycles/kawasaki-gpz1100-a-collectors-choice/

You may not be aware but Yamaha built an XJ750 EFI so you might want to do some research into it.

http://dave_jack.tripod.com/id27.htm

 

They were very early on in the electronic Injection era and quite a primitive system, converting them to carbs was very common.

 

Really it's that I got myself more projects than I should, and this idea has been on the backburner. I literally was looking at stuff related to this like two weeks ago, and then thought that I have to complete at least one more of the bikes I have disassembled right now before I get into it.

The real hurdles in this are all physical.

The first problem you run into is what to do for the intake.

The obvious solution is to use a set of throttle bodies from some fuel injected bike. The thing is, figuring out what fits is difficult. There is no database of throttle body and carb sizing that I found for bikes. Cars are much easier this way - there are common carbs, and there are FI kits made to fit in their place.

Less obvious is making something, but that presumes a fair amount of fabricating capability.

Second problem is mounting the oxygen sensor. These bikes have 4-2 exhausts, and FI brains like Microsquirt and Speeduino only do one lambda sensor, so you really want to change to a 4-1.

Third problem is timing. XJ is actually pretty well set for adapting to a timing wheel and sensor in place of the OE ignition rotor and pickup, but you have to make the wheel. There isn't one you can just get and bolt on. And getting cam timing is something that will need machine work, but you can run without it.

Everything beyond that is just time and some elbow grease.

 

I thought about these concerns as well. In particular, I was looking at using a colortune for tuning Volumetric Efficiency. However, the colortune seems to be getting rarer as time goes by.

For the throttle body, I was thinking of using plastics to 3D print/ resin cast. I'm still designing my own.

For timing, I was planning to use the Speeduino. Idk if you all heard about it. The Speeduino requires a conditioned signal though. That what makes it difficult. Either convert the crankshaft sensor or convert the signal itself. I'm hoping to convert the signal.

 

If the typical current supply is 20A, that ain't bad. From googling, fuel pumps pulls about 2-3A. Just a little revving hear and there to maintain power.

 

They aren't hard to find, but...

I have one. I never had much luck using it. Also, it seems a bit silly to involve it in FI. Swapping in a 4-1 exhaust and putting in a sensor bung is more expensive, but will get you a much better result in the end - you will have on the fly adjustment.

I wouldn't trust 3D printing in that application. You have everything that's bad for plastics in one place - heat, solvents, vibration, and mechanical stress.

Now, printing a model to use for casting an aluminum piece, that I can see.

Als0 - are you trying to make an entire throttle body rack, or some kind of adapter for an existing TB?

I actually mentioned it in my post.

Not what I meant by timing, though. I was talking about getting the timing signal, not which control box to use.

Yeah, this thing.

I have actually asked this question on the Speeduino forums, I think. I can't find the thread now.

What they told me was that the typical ignition pickup like an XJ has does not have the accuracy/resolution needed for an FI box. You have what is one big "tooth" on it that does not really tell you the position very well.



Speeduino and other FI controllers use a missing tooth wheel with 24-36 much smaller teeth and a different pickup sensor.

Like I said - it really is not a big problem to fit a trigger wheel in place of the ignition rotor. There is lots of room. And not at all a problem to replace the pickup coil with a VR sensor. The problem is only that you have to make the wheel.

 

Well you could halve your problem by leaving the ignition system as it is for now - use both existing triggers to feed the fueling ecu. Position is absolutely clear and accurate, but what is not is the time in between the two tdc signals - so effectively you lose accuracy when the engine speed is changing. Having built from scratch a system that only used the equivalent of points to drive it I would say it's accurate enough until all the other issues are sorted out - sort of eating an elephant - do it in small bites.
I'm with @JPaganel on the printed manifolds, go aluminium every day. Using a single throttle body would be a backwards step imho, you'd lose the sharpness and some power - why would you want to?

 

I can see wanting to use a manifold with a single TB for a couple of reasons.

One is simplicity - you don't have to keep them in sync. The other is the ability to use a widely available and common TB with common sensors.

I think if you used a TBI setup where the injectors are in/near the TB you would definitely have some slop in response.

However, most cars today use a multiport setup where injectors are at the intake ports and the TB is separate. This may be less of a problem. Really, how bad can it be in comparison to the delay you get with vacuum carbs between the butterflies opening and the vacuum pistons raising the needles?

As far as generic TB...

You can get them really cheap - this one is on Aliexpress for 20 bucks. Says it's for some Honda.




What I don't know is how you'd calculate the TB size. I am pretty sure you need the overall area of the opening to be less than what the total area of carb openings is, but not really much beyond that.

Finding a motorcycle TB rack that fits is still a pretty good option.

One thing I noticed about those is that some are cast in pairs, and some are completely separate, so, in theory, you could adjust the spacing. Still need to find something that fits in the manifolds, though.

 

Ok, so you will lose response ( maybe not much compared to cv carbs, but the delay in the slides lifting is what gives you tha rich kick, so it is needed). You will lose power. But you gain a whole load of simplicity. Pays yer money takes your choice..
This is the guy on here who'se done it and got the t shirt.. Happy reading
https://www.xjbikes.com/forums/threads/look-ma-no-carbs.10574/
Edit to add this from @SQLGuy himself
Personally, even as things stand now, I don't see any issue with the single throttle body. The bike pulls fine and has pretty good throttle response. I went with a single TB for a couple of reasons: one was easy availability from a local scrapyard, the other was that the TB includes an idle air valve, which I think will be important in achieving the kind of drivability I want; hooking one in with a quad TB bank would have been pretty complicated.

 

Does a fuel pump really use that much power once up to pressure?

 

Just replying to one of your responses.

- 3D printing on its own may not be sufficient with the equipment I have. I could use a higher temperature print material (like polycarbonate) with FDM, supplemental materials like CPVC for plumbing, adhesion (paint remover, gorilla glue) with cross-layered print-path parts (one orientation print over a different orientation print), and resin mold casting. From googling, some manufacturers actually created some hostile-condition parts like fuel rails and intake manifolds so I think with engineering, simulation, design and testing could build some parts. You would be surprised what plastic could handle when designed right. The main issue is the anisotropic nature of 3D prints and the material's lower melting point.

 

You'll still need to machine it to fit a spindle and butterfly, and throttle potentiometer. I can post you a Rover 200 plastic one ready made though if you want, big enough for 130bhp, with a potentiometer stuck on the side ready?

 

Im thinking it's time someone designed a throttle body by the people, for the people. For too long, we have been depending on the car/motorcycle manufacturers for our throttle body supplies. It's time we take the power that belongs to us and use it for our good.

In all seriousness, I like Aluminium but I don't have easy means of manufacturing right now. If I could make a reliable throttle body by 'plastic' means, I think that would be a big step for the hobby community.

I'm planning to create a rail of individual throttle bodies. I never had an intake air box so I cannot convey that experience but I like the ITB look and feel. Pod filters FTW (don't stone me). I heard tuning the ITBs would be difficult but I'm up for the challenge. I could also design-in IAC, TPS, and MAP. I even thought about using MAF but Speeduino doesn't currently support it and I don't want to make a new firmware. As the other guy said, one bite at a time.

 

I thought of a potentiometer but their problem is their limited lifecycle and reliability. I'll use a standard TPS.

I'm thinking of using a barrel-design throttle body. A plastic butterfly valve is too thin and might break apart under the vacuum. Even if I use a metal plate, I'll need to design a metal rod to control the plates as plastic might deflect too much.

I'll check-out the Rover's throttle body, if it could work, be my guess.

 

The Rover is a single body, would be too big for multiple setup. When I said potentiometer I meant tps - it is after all just a potentiometer.

 

@Minimutly correctly points out the hard parts of making a TB from scratch.

Shafts and butterflies are precision parts (have to be, to avoid air leaks), and also the main mechanically wearing parts. If you want them to last any length of time, you are probably going to wind up with some metal inserts, or something.

You could still go down the path of using common parts by making pieces to adapt them. For example, make a manifold and some parts to adapt these into a rack:

https://www.aliexpress.us/item/3256...ePPC__xxxxxx__1687049309&gatewayAdapt=glo2usa


Or these



This one is actually really not a bad idea. They are very, very common - they are made to replace Weber carburetors on things like old VW Beetles and BMW. There is a whole bunch of people making them and they are made in several different sizes. They aren't 8 bucks, but they will save you a pile of work and manufacturing cost.

This is what a manifold for four of them looks like



I can see that being a 3D printed piece - no moving parts would make it easier.

There is nothing wrong with pod filters on an FI system, so long as you use something quality, like UNI foam pods.

They have a bad rap here because people often use the cheap cloth cone filters from eBay which let in too much dirt and because they cause problems with the vacuum carburetors.

 

Thank you for your knowledge. I do agree that Aluminium throttle bodies could be manufactured. I just want to try and see what happens. If proven impossible, so be it but I think it's a great opportunity to try and make a plastic throttle body. I've actually 3D printed a plug valve before. I didn't install it or use it for anything and yes, there was an air leak but I think there are ways to rectify the problem. O-rings for example. Again, TBD but I'll try it.

 

Ok. Update so far:
Btw, sit down, this would be a long post.

I thought of calling the project:
- MC Carb Substitute
- Motorcycle All-In-One Individual Throttle Body Carburetor Replacement
- Or some combination of the above.
All suggestions will be considered.

Next, I've changed the project scope for simplicity/design cost sake.
The project aims to be a platform for creating EFI All-In-One Throttle Body for motorcycles. This includes other models of Yamaha and other brands. This does not mean that one throttle body could fit on all motorcycles. It means that with consistent parts/layout/configurations/dimension relations, the throttle body design could be adapted to other vehicles and similiar/proportional tuning could be used across motorcycles. Think of the Holley EFI TBI kits for cars and trucks.
Also, the kit should be easy to install. Few modifications. An amateur should be able to put it in. Just take off the carb, mount the kit in the same manner, connect some wires, and be good to go. Pre-tuned and all. The only modification I think is required is a Wideband O2 sensor for detailed tuning. I found a kit on Amazon (https://a.co/d/5J0DNMi) that only requires drilling and clamping the bung in place for the O2 sensor to mount. It would be intimidating to drill into your exhaust but simple enough. No welding required.
Btw. The scope is subject to change.

Next, I've gone through so many design iterations.
Initially, I wanted to make a peristaltic pump (https://en.wikipedia.org/wiki/Peristaltic_pump, Google It) driven throttle body. I searched the internet and did not find such a device. I thought I was on to something, something new. I didn't want to tell anyone. But as you guessed, I gave up that plan. After weeks of research, I found out that peristatic pumps could have fuel metering issues with wear. It just would not be constant between strokes or with operation time (the time it spent running). Also, the failure of the device due to lack of maintenance could lead to catastrophic consequences. A ruptured hose could lead to a fuel leak and a fire. I didn't want to risk that, especially if other people used the design. Also, I would need to design my own firmware to run this device. I realized that it's better to piggyback off of other people's work (no, I'm not stealing).



This introduces the Speeduino (again, google it). An open-source DIY ECU that uses an Arduino Mega 2560 for processing power and control. This would be my platform for controlling the electronics and running the engine. With this, however, I would need to do modifications to the bike electronics, especially the ignitor pickup. I need to design a new Variable Reluctance (VR) circuit to send the proper signal to the Speeduino to give it a sense of engine speed. That's the most critical work to be done. I don't need to do this now but it is vital for the future if I want the throttle body to be easily swappable and non-intrusive.

There were many designs I made but I will show you a few:
Next design, PVC housing. I thought that PVC pipes were a good material to use for the substrate for mounting and modifying. I learned later that PVC is not considered safe with gasoline but that wouldn't hold me back. I could use HDPE pipe if I tried hard enough or CPVC for higher temperatures. However, the true stopper was that I could not create a reliable design to isolate the downstream from the upstream of the throttle body. I was basing my design on a barrel throttle body and the upstream is close to atmosphere while the downstream is close to the engine and is regulated by the throttle mechanism itself. I thought of o-rings to seal but they would wear-out and they are difficult to hold in position. Eventually, I abandoned the idea but if you see of a way to hold a seal/isolate the chambers, let me know. I would consider the design. Just look at the following picture, PVC looks so cooool.



Next design, Sliders.


I designed all kinda slider designs. A vertical slider as seen above, to cam profile-following slider seen below.


The main issue with the slider designs is limited space. I was creating a design to retrofit into the previous carburetor region and the sliders were limited by size constraints. Also, I found it difficult (again) to isolate the upstream from the downstream. Yes, the vacuum would help seal upon a face but what was difficult was the throttle cable. It needed to be isolated and most designs I developed required a rubber bellow (they are not standardized and I would need to make them myself) or I need to use an aftermarket throttle cable (the Yamaha XJ's hand throttle is rare enough). Abandoned. I could reduce the throttle body size but I don't want to hamper performance. I had alternative desings in my head anyhow.
The alternative, Cam-Profile. This looks cool as the cam profile moves the throttle plate from the closed position to an open position following an oblique path. Bearings would help in the follower. Again, problems as mentioned above but also, the design is susceptible to binding which is not good for when u are trying to get off throttle for the red light ahead.

While all of this was going on, I was learning my CAD and CAE software, Solidworks in greater detail. Better workflow, more features. Now I'm venturing into using surfaces. Might be able to create 3D files of the Yamaha XJ bikes soon (fingers crossed).

Next Design, Hinge Joint.




Ok. Most throttle bodies are butterfly flow control valves. Common alternates are barrel and slide. I introduce to you a new design. Hinge joint. Instead of having the butterfly valve shaft mid-stream, you could put the shaft at the top (or bottom) of the stream and let the 'paddle' cover or open and control the flow of air. Some folks might think, 'Wouldn't the hinge be affected by the vacuum inside the chamber in terms of forces required to open and close?'. In a sense, yes. However, these calculations I found to be minor to the other forces needed to open and close the throttle body, or maybe I'm doing the maths wrong. Anyhow. I'm going forward with this design.
Btw, if anyone is into engineering, I highly recommend the website McMaster-Carr. Holy moly. They have all kinda hardware and CAD files for them too. This helped me brainstorm my designs. From there, I got some of the components for this hinge throttle design.
The hinge design uses sheet metal (I finally gave in) as the forces on it are too high for plastic and it deforms too much (from simulations). It would require torsion retention springs like a carburetor and it would use a rubber-like seal to mate to when in a closed position. I might use a resin laminate with carbon fiber or fiberglass with a 3D print core or a cast core for the casing/body (if you know, you know). I would mount the electronics enclosure below and give it access to the manifold for the MAP sensor. I could also mount the fuel pump below (this is a whole nother story). Fuel injectors and rail on top at the intercept angle. The fuel relief valve on the fuel rail with lines running to the fuel pump enclosure. I'm considering designing a sheet metal plate to replace the petcock. This would allow for return fuel and intake. I might also just put a new in-tank fuel pump (like a car) inside and let people install the barb and such for it. Idk yet, TBD.
The fuel pump. OMG. This is scary. Either I build an enclosure for the fuel pump (who am I? the EPA?) or get some kinda compact fuel pump enclosure, OR I use an inline fuel pump. Small inline fuel pumps are expensive but I don't want my DIY fuel pump enclosure to explode. It's difficult to get material that is fuel compatable. Basically metals. Then how do I manufacture the enclosure? This is my biggest issue so far. I might give in and go with the Walbro $120 inline fuel pump. Get all the features, little headache, and move on with the project. There's a lot of work to do.


PS. 'Why didn't you use the butterfly design?'
- Well, it would require a shaft to mount to each flap and a hard cylindrical surface (pipe) to mate to when closed. I might go to it if all hope is lost but right now, I'm working with hinge throttle.

If you have any suggestions, AGAIN I'm all open ears. Read the above mess at your leisure.
TLR: I am creating an All-in-one carb replacement EFI throttle body kit for my Yamaha XJ 750 bike. It should be easy to install. I am currently in the design stage of the project and I would use a Speeduino for the ECU.