Probably an old question - 4 cyl honda cb900?

Good day.

I hate carbs. I mean I really, really hate them and I have 4 of the bloody things on my 81 Honda CB900c.

I can never get this thing to run right for more than a few minutes. As soon as it gets a bit hotter or colder, it runs like crap. Most of the noticeable problems are with idle. I probably also have issues at speed but they don’t manifest in running quality, just crappy mileage, which I would live with if the idle issues weren’t so maddening.

I miss the fuel injected reliability, smoothness, economy and power of my injected BMW bikes, but I don’t feel like giving up on the old girl. Besides, now it’s kind of a quest, just to see what it would take.

I would prefer a throttle body setup for the simplicity and cost effectiveness, tho I’d consider multi port. I wouldn’t have an issue fabbing an intake.

It’s a daily driver and I’m not looking for max performance. I have no doubt even the least optimized system would be better than what I have now.



Welcome! :wave:

You’re in the right place then! :beers:

So although I always say NanoEFI is for 2 cylinders maximum, it can still run some 4-cylinder engines in some arrangements. It lies a bit outside of my realm of experience so far, but if you’re willing to dig into the technical details with me I’m sure we can figure out a way to send those carbs to the recycler :bomb:

After checking out '81 CB900 crank and cam pictures online, it looks like you have a normal inline-4 setup. You do have a DOHC, right? Just want to make sure I’m looking at the right thing.

Wow. That was quick. Thanks for getting back to me.

Yes it is a DOHCam setup, with pistons 1&4 and 2&3 paired on the crank. There are 4 - 32mm Keihin constant velocity carbs, 1 carb per cylinder. It idles at about 900-1000rpm and mostly operates in the 3-6000 rpm range.

It has electronic ignition triggered by a pair of hall effect sensors on the crank. 2 coils and 2 plugs per coil. It uses the wasted spark system where each plug fires once per crank revolution, (as opposed to the normal 4 stroke thing of firing once per 2 crank revolutions); once at the top of the compression stroke and a ‘wasted’ spark at the top of the exhaust stroke.

I think that’s to allow for only using 2 coils (with 2 leads each) in total.

Just thinking about it, maybe you need a CV type carb to work with the wasted spark system, but probably not.

The way I figure it, each carb is only working 1 out of 4 piston strokes, so in theory (if these weren’t the stupid CV type carbs with a vacuum operated sliding fuel needle), a single 32mm carb could fuel all 4 cylinders, which means single throttle body injections should theoretically work.

I’ve been mulling over this for a while. Originally, I was going to try doing something with auto carbs, but then, even if it was better, that would still be a lot of work and I still have carbs. EFI would be a much more elegant solution, if it could be done cost effectively.

I have been thinking about going to the pick n’ pull to see if I could salvage enough efi bits to make a go of it, then I stumbled upon this site. This is EXACTLY the kind of project I have been looking for.

I have been working on vehicles (including professionally) for over 4 decades. I’m a big fan of efi conversions, except I think they are more expensive and complex than they need to be. They can never be true plug and play, but I think they can be close.

I have a spare bike that would be a perfect guinea pig. Let me know if you need any more info.


1 Like

joekanuck…I feel your passion and desire. Whilst I do think the the Nano setup can be made to work on any multi, I really believe that you should keep involved and show patience.

The reason for that is age and dimension, in your case. The variation in your cam w-o-b-b-l-e is enormous. Yes, correctable.

I have two potential apps. A single 250, which has FI already, and more important…a 50cc 4T that could double it’s HP or boost torque.

The 900 looks fun! But you should think about all the variables…there is a lot!

Hey Thai

As I mentioned I came into a spare bike and my current one has served me well for over 3 years. I’m not in any rush to get anything done now.

My main interest is in the potential for a simple, affordable efi kit. There are thousands of old carb bikes out there and the number one problem, is carbs. Getting just the carbs rebuilt is at least 2-300 bucks, not much less than what I believe is possible with EFI and some creative minds working together.

The goal is to turn the bike (or any similar bike) into something modern running without losing that old school feeling and to have fun along the way. My engines are almost 40 years old. You’d have to be nuts to want to really wring them out, but they can go indefinitely if they’re not pushed to breaking.

What do you mean by cam wobble and what effect might it have that could be cause for concern?

Luckily still awake! My issue…not the Nano issue, is that all alpha and beta be done with known and precise data sets. The wider you make your cam, the more variation you have from simple loading to heat expansion.

My example would be: Gather data from a single/dual cam single cylinder and go from there. You can see that your initial data set need be tiny, in order to extract accuracy.

You have a great project purpose, but could easily chase your tail.

Hey Joe.

The CV carbs are to keep the intake air velocity up, as a normal carb would choke if you twisted the throttle too hard off idle or low rpm.

I wouldn’t go from quad 32mm carbs to a single 32mm.
It would have good airspeed, but be a massive bottleneck and make for a disappointing ride.
Possibly a simple efi solution in the fuel area would be to have the injection batch fire, although not as elegant as sequential, it would be a huge step up from carbs, and still allow quad throttlebodies without too much trouble.

@joekanuck You might be able to repurpose your original CV carbs in order to use them as throttle bodies for EFI. I’m thinking by removing the large internal springs, locking the vacuum slide in the full “up” position, and capping off the fuel inlet and overflow outlet.

Adding a TPS sensor to the shaft will be a headache, but likely less than fabricating a combined manifold. Then it should be a matter of working out an appropriate location for MAP and IAT sensors. Looks doable with the CB900 carb set from what I can see :+1:

This is good. NanoEFI’s injector drivers can handle up to 6a each. So you actually can run four injectors (driven in pairs). As long as they’re Hi-Z (10 - 12ohm) and set to run in saturation mode, you should be fine pulling around half of max rating on each driver.

I’m interested to see what the grouped timing would look like. I’d need to put together a graph to get a visual.

Should be no problem to use the hall (or perhaps even the high voltage output) as our trigger source. Engine phase may be an issue, but that can be worked around as well.

With a bit of work, could be a stealth install, keeping the old school look.


I was just using the ‘single cv carb’ as an example of the concept. If I went with single carb, the idea was to go with a small displacement auto carb or maybe a Harley carb.

The goal (at first at least) is to be better than carbs, but not necessarily the best it can be. Batch injection would more than meet that criteria, but with simplicity also as a goal, I’d like to try reduce 4 injectors to 1, and try a throttle body unit.

That being said, it would be easy to rig the 4 carb setup into a dumb throttle body. There is about 2 inches of rubber boot connecting each carb to the head. It’s enough room and I would just have to work on a way to mount the injectors. There is a huge airbox behind the carbs which I could eliminate giving me lots of room to move the carbs away from the head a bit if I needed longer runners to mount injectors, and run filter pods on the carbs.

That part shouldn’t be a big deal.

There are vacuum ports on the head used balance the carbs that could be used for a MAP sensor. I’ll have to do some scavenging for a TPS. Mounting it would be a matter of farmyard fabrication but I can’t see it being a problem.

I’m liking the ‘carbs as throttle body’ concept the more I think about it. The grand, overall, plan would be to have a system that would be as plug and play as possible for a DIY mechanic.

I’m glad I found you guys and dropped by. I’m enjoying this.


For fun, and so people know what I’m working on, here are my bikes. The one with the fairing is an 81 Honda CB900c and the bare bike is my latest acquisition and 82 CB900c.