ESE's works engine tuner

[Michael Moore]

one that isn't integral to the carb body like the SU float chamber with this AMAL GP5:

cheers,
Michael

[oldjohnno]

Or just run electronic powerjets feeded directly from fueltank.

Good idea, bypass all the damn floatbowl issues altogether.

SU floatbowls are pretty awful, they can't take much fuel pressure and don't like vibration. I've had very good results using a remote Holley bowl; they have a high flow capacity, can take at least 6psi fuel pressure and will withstand being shaken a little. I used soft rubber mounts and an additional weight to keep it steady. But if you can do without a remote bowl altogether that's ideal.

[husaberg]

Good idea, bypass all the damn floatbowl issues altogether.

SU floatbowls are pretty awful, they can't take much fuel pressure and don't like vibration. I've had very good results using a remote Holley bowl; they have a high flow capacity, can take at least 6psi fuel pressure and will withstand being shaken a little. I used soft rubber mounts and an additional weight to keep it steady. But if you can do without a remote bowl altogether that's ideal.

I was more meaning just yo have it there as a extra reserve of fuel but still plugged onto the std float bowl.
Never seen a holley remote bowl.
But they all have to be mounted so vibes dont make em worse.
Some of the Amal ones are worse than others they used to hang them off a top mount.

okay here is a pic of someone using a bodged up Holley one Yeah you are right though Su's are low pressure, probably not as available as holleys in the USA either.

[peewee]

you saying use a float chamber from a race car holley carb ?

[speedpro]

Or just run electronic powerjets feeded directly from fueltank.

A tiny FI fuel pump could feed it on demand with a simple controller to turn it off and on, Or just leave it on depending on the power supply of the bike. I used a similar idea on my methanol turbo race bike except the "powerjet" was controlled by a cast valve that could be adjusted for start by spring preload, progression by spring rate and needle taper and initial fuel flow by a standard brass jet. The fuel tank was pressurised by boost through another valve. Maybe even a small pressurised reservoir filled by the FI pump. 3bar would make a nice little jet of fuel through a fuel injector. nitro might be hard on the seals though

[Reckless]

you saying use a float chamber from a race car holley carb ?

If you want to keep it very simple?
On our meth karts we just used a car fuel filter
like this


We cut them open at the top join - removed the filter then Re araldited them together.
Mounted them vertically above the carb inlet so they sat just clear of the top of the carb.
Very simple added enough reserve for the pump to catch up after a long straight.

Just adding an Idea.

[Carel H]

Thanks for the heads up.

Map sensor tech sheet says <2ms. other data sheets I have seen say 1ms so I expect it is somewhere between 1-2ms for the sensor to settle to 90% value of the change in the difference of pressure.

So not fast enough. It could easily miss the high and low point of a cycle. Certainly can't reliably make 8-10 measurements per cycle. If I can't time the measurements to be taken at predetermined points in the cycle I will have to think of something else.

Attachment 337433

When I attach a large volume (50cc) to the crankcase with a small diameter hose and measure the MAP value I get a clean average reading.

I am thinking of using the average MAP reading and adding the difference between the high and low pressure each cycle as measured by a fast Piezo sensor.

That way I could maybe make 20-30 measurements per cycle even at 12,000 rpm. The Piezo would only see the relative difference between high and low but with some math it could be dialed in to show a suitable maximum amplitude.

Adding them together, average map and the piezo measured difference between the high and low pressure each cycle could give a clean pressure curve that represents mass air flow through the motor.

I am not sure how fast the Arduino Nano is, but hopefully fast enough to manage 10-20 or so readings per cycle at 12,000 rpm.

It looks like all pressure sensors (Infineon, NPX) are created equal. Full scale response 1 ms. Piezo sensors I don't know. They will have like problems as the MAP sensors: scale, size, pressure, temperature compensation etc. And they call an MAP sensor "automotive", which certainly is an advantage, in contact with fuel.

The Arduino is fast, but how fast is the program? And what are the possibilities with the compiler? I work with the exact same processor, only here it is called AVRMega and I program in assembly.

My approach for 2 measurements per revolution would be (I actually made an ignition program like this):

Have a trigger (I used a Hall sensor as ignition trigger) and use the Time Capture Pin.

If the pin is triggered then:

Calculate rpm (time now minus last revolution time).
Calculate time to TDC at this rpm and set the timer.
If the timer goes off, start ADC conversion and add half rpm now time.
If conversion is ready, read it.
Wait till the timer goes off at BDC, start conversion.
If conversion is ready, read it.

Do it again.

And of course you are free to decorate your flywheel with stickers, indicating to a dedicated sensor:

"This is where I want to start my conversion". Both versions still have the 1 ms grey area, so I would try this one first. Always nice to have a processor doing almost nothing.

I like your curiosity.

[SwePatrick]

Good idea, bypass all the damn floatbowl issues altogether.

SU floatbowls are pretty awful, they can't take much fuel pressure and don't like vibration. I've had very good results using a remote Holley bowl; they have a high flow capacity, can take at least 6psi fuel pressure and will withstand being shaken a little. I used soft rubber mounts and an additional weight to keep it steady. But if you can do without a remote bowl altogether that's ideal.

I´m going to run a electronic powerjet modded to be easily adjusted by a variable jet and electronically controlled by Ignitech.

As someone spun further on this idea with a fuelpump, i won´t try that as i recon the desired fuelflow will dynamicly adjust itself some, due to airspeed.

[TZ350]

The Arduino is fast, but how fast is the program? And what are the possibilities with the compiler? I work with the exact same processor, only here it is called AVRMega and I program in assembly.

My approach for 2 measurements per revolution would be (I actually made an ignition program like this):

Have a trigger (I used a Hall sensor as ignition trigger) and use the Time Capture Pin.

If the pin is triggered then:

Calculate rpm (time now minus last revolution time).
Calculate time to TDC at this rpm and set the timer.
If the timer goes off, start ADC conversion and add half rpm now time.
If conversion is ready, read it.
Wait till the timer goes off at BDC, start conversion.
If conversion is ready, read it.

Do it again.

"This is where I want to start my conversion". Both versions still have the 1 ms grey area, so I would try this one first. Always nice to have a processor doing almost nothing.

Great, thank you for the heads up.

It is going to be little while before the engine is backup and running so plenty of time for me to learn some more about programming. Assembly, well that will be a challenge. I hope I can get away with a high level language.

[Frits Overmars]

Assembly, well that will be a challenge. I hope I can get away with a high level language.

Must be your sense of humour, TeeZee; assembly is about as low as you can get. No 'high' operating systems getting in the way and deciding what's good for you and what isn't; with assembly you talk directly to the processor. I love it. OK, I'm definitely old-school; I grew up with Algol, Fortran, Pascal and of course Basic .

[diesel62]

Great, thank you for the heads up.

It is going to be little while before the engine is backup and running so plenty of time for me to learn some more about programming. Assembly, well that will be a challenge. I hope I can get away with a high level language.

Before it all gets to hard why don't you just put an oscilloscope on the map sensor. You will soon see if it is going to work or not.
No programming required.

Sent from my SM-P555 using Tapatalk

[lodgernz]

Must be your sense of humour, TeeZee; assembly is about as low as you can get. No 'high' operating systems getting in the way and deciding what's good for you and what isn't; with assembly you talk directly to the processor. I love it. OK, I'm definitely old-school; I grew up with Algol, Fortran, Pascal and of course Basic .

Actually, machine language is as low as you can get, unless you want to program the microcode in ROM. Assembler is, strictly speaking, a high-level language, but the instructions correspond pretty much one-to-one with the machine code, so it is very efficient.

C language is quite efficient though, for a HLL, so you might not need to learn assembler. It's fun of course, but a bit tedious.

TeeZee, if you do want to use assembler, PM me, I have some docs you'll need, if you don't already have them.

[TZ350]

Before it all gets to hard why don't you just put an oscilloscope on the map sensor. You will soon see if it is going to work or not.

Yes, good idea, I will try that.

Have scoped things before but the rapid changes in Frequency (read RPM) makes it hard to adjust the time base so I can see the trace properly.

It is very hard to hold the RPM steady and synch the scope. Maybe I need to work it the other way around by setting the time base for 200 Hz and then revving the bike around 12,000 RPM. That way the trace will steady up at the target point instead of continuously running to the left or right across the screen. Either way its hard work and the trace is only readable over a narrow RPM band. Synchronization becomes a continuous moving target and has me working like the proverbial one armed paper hanger.

Anyway your suggestion makes very good sense and I will try it. Maybe there are more suitable scopes ........

[diesel62]

I'm hearing you. Need a good quality laptop based scope not little screen old electronics scope.

Sent from my SM-P555 using Tapatalk

[TZ350]

Need a good quality laptop based scope

Tell me more. What should I look for?