Project X

[madforitkev]

Progress update.....

I have a fuel injected, fly by wire Kettle!!!

Initial hiccup was the speeduino fuel calculator.

My triumph injectors are 230cc approx so feeding that into the calculator I got....

This gave a required fuel load of 8.9.

It didnt start........I started to disbelieve my installation, thought something weird had happened with ignition timing, wiring, injectors failing etc.....

After lots of testing......Back to basics....compression check, easy start straight in the motor....ignition obviously OK.

That left fuel and air.

Butterflies just open and still no start and it p***ed fuel.....way too rich!!!

To make sure throttle position wasn't the issue (position sensor noise making it seem like I was asking for a higher number in the fuel table), I disabled and grounded the input.

Now the VE table can only work at zero TPS and RPM is the only driver.

Next step was to half the required fuel.... it popped and banged....still too rich, easy to identify because if I opened the throttle (more air) it ran smoother.


After a few iterations I am down to 0.8 fuel load!! 1/10 of the Speeduino estimate.

With a small throttle opening, it runs the cranking fuel load setting (adjustable, mines at about 120% of base fuel for a 20C ambient), starts easily and runs the After Start Fuel curve (basically like having choke open for a bit) for 20 seconds, then the Warm Up curve as the engine gets up to temp. Once I get to about 30c it sits there on tickover. Not to shabby!!!

Next step is programming the throttle to get the right air to fuel butterfly setting. My bodies are big (46mm) so I have to restrict the air flow or it will run too lean.

I have the AFR gauge all hooked up now and in the stinger and at idle it runs about 1:10/11 which is Ok. With that running I can start doing tuning runs for the VE table.

That feeds into the speeduino data logger so I can review the AFR post the tesst runs.

I managed to work out the oil pump linkage...

One issue with this was the standard oil pump return spring combined with the throttle body springs and the cable friction is a bit too much for the bodies motor so I have made a temporary spring setup for testing. Will make it more 'production' later.

The fuel regulator sits nicely on a bracket I made that attaches to the throttles...ignore the cable tie, version 1 photo.

I had to make a high pressure fuel input adapter for the fuel rail and because my tank hinges up, i needed it to rotate and have a QD fitted to ease tank removal.

Few bits to tidy up, then I can start test runs........and its only taken 2 years....

[badgereng]

A medal will be struck in your honour. Great work and perseverance. Top job sir.

Pete

[madforitkev]

The misses keeps asking me when will I finish it......

The honest answer is 'i probably never will', I am having too much fun trying new things ....just happy to be locked in the nerd-shop... and fettlin'.

I know I have to start making it look pretty soon, but i'm not in a rush (as you can tell)

Plenty left to do yet....I am just sooooo pleased the fuel injection has worked out OK so far!! And who knows I might even inspire some other fool to have a go! The forum definitely has the candidates!!

[Alan H]

I have every admiration for what you have achieved, but there is NO WAY I would ever attempt such a project - and I used to program PLC type computers.
Good luck with the (hopefully soon) dyno runs and injection system set up.

[madforitkev]

Made some progress this week.

Basically trying to get the base fuel load right so it will tick over.

Part of the problem I had was the fly by wire controller not being able to cope with a just open/start position, so I went back to the trusted mechanical adjustment screw to set the base air level.

With that sorted I was able to make the throttle response better from the mechanical hard stop.

Small tweaks on the fuel load and it start and runs nice.

Speeduino has the following options;

Cranking - add lots of fuel
After start enrichment - slowly taper off after a user set time on the extra fuel
Warm up enrichment - once up to temp (user set) then just use the fuel in the VE table with no additions

It now fires up lovely and the throttle response isnt too shabby so the tune (at least at low revs) must be close.

I've got the AFR gauge logging now and a bluetooth link sorted to the ECU so I can run MSDroid (a tuning app) on my phone while riding.

I have a bit of throttle position sensor noise probably due to an old sensor and wires being too long at the moment. Will fix later.

Oil pump spring is still a bit too strong or I have too much friction in the linking cable. Full throttle struggles to get there mechanically. It will be easier to fix that off the bike.

Sorted the high pressure fuel installation with a few QD's to help remove the tank quickly.

Time to ride it again now and collect some data This is where I find out if my throttle bodies are just too big....... Hopefully not as I have the limits reduced in the motor driver.

Getting closer to the strip down and start again!¬!

[madforitkev]

Silly question.

I dont have a standard Kettle at the moment......when the throttle is wound wide open, how open is the oil pump? All the way until the lever stops on the pump?

Just trying to calibrate my throttle bodies.

Cheers

[teazer]

Basically yes, but it depends on how the linkage or cable is adjusted.

[madforitkev]

Lots of progress since the last post here.

Bottom line is I have Alpha N (TPS) EFI working on my 2 stroke 750cc triple now.

As you know I am using Triumph Tiger throttle bodies, so have it running fly by wire as well which is very nice on the wrist compared to 3 x carb springs!! Today was the first time testing it and it is NICE......

I completed the first real road test runs today with full EFI. Below is a snippet of one run from Tunerstudio's Mega Log Viewer program (I use Tunerstudio to do the tuning setup).

I have been experimenting with a wide band AFR sensor in the end pipe (similar to the dyno setup I use).

I did one run and it clearly had a low fuel load at revs (over running on throttle close/lean) so I upped the base fuel a bit (so much easier than buggering about with jets!!).

The second run was smoother (still running in so kept the revs low) It seems to idle at about 13 AFR, drop just below 12.8 @ 4k revs and at 6k its still about 12.8 AFR so not much change really under load and it still over runs slightly so may still be a bit lean?

At this stage I dont really trust the AFR readings, mainly because they dont seem to change much (and its a challenge with a 2 stroke) although my readings are relatively low (indicating its rich), which based on the first run and then adding base fuel load to make it smoother, I dont think it is.

Everything else seems fairly normal and the running temp is good (it logs in degrees F)

As it seems to start and tickover OK, I think the next step is developing the VE table with the dyno runs.

I am toying with dropping to Tiger 800 throttle bodies to reduce the air flow which I currently control by just reducing the opening with the motor controller (WOT is about 1/2 open) but that has an affect on my oil pump link and the 800's might be better for that.

I use an Android App called MSDroid while doing the runs and its a great data logger that exports to Megalog viewer for full analysis. The ECU connects via a bluetooth link I added to the PCB which makes life easier.

Still fettling but basically I am ready to strip it down and do the pretties. One more dyno run me thinks

[Alan H]

I used to work on plc computers to run different kinds of process on a multitude of sites, and when you do it regularly, it's usually straightforward, but I've been out of that game for over 10 years now, and seeing your graphs, I'm quite happy with not doing it any more!
I get the gist, but I'd need the full mapping data to make more sense of it all.
Best of luck with the progress though.

[teazer]

The "static" AFR is a little odd. We took a couple of our Suzuki triples to the dyno recently and had really significant swings in the apparent AFR. At low revs, below say 4,000, AFR read lean even with the sniffer probe inserted all the way into the belly or divergent cone on the expansion chambers. We believe that is essentially the effect of reversion in the pipe that sucks air in as the pressure wave changes direction.

Can you plot the voltage from the WEGO sensor to see if it is responding from startup to hot running? Looking at the charts makes my head hurt. I find it easier if I use the cursor to move left to right so I can see the changes in say lambda and TPS % without the other data. And then I can layer in other variables to see how they correlate.

You are making great progress though. Congrats on the journey.