HIF44E closed loop AFR anyone?
Turbo Mini Forums

Has anyone used a HIF44E with a wideband O2 sensor to control AFR?

I have an e carb that I'm looking at modding with a servo in lieu of stepper. With some simple feed back I should be able to regulate mixture. Given complexity of sequential injection systems and Siamese port timing it seems like a nice compromise.

Just wondering if it's been done before?

Sounds like a nice idea in principal.

If something is worth doing, it's worth doing half of.

I was not running closed loop but i could use my ecu to control the effect of the lean off pipe at part throttle
You can add or remove 10 afr quite easily across just about any load requirement
It did involved thinking out side the box
But what i did find once i learnt and understood the mechanism of the whole carburetta a properly sett up needle does almost everything you could ever need fueling wise.... and the lean off pipe works wonders at part throttle being able to easily tweek that at cruise is more important if mpg is a major concern
I ended up with a ( even if i do say so my self, a perfect needle) and a bleed valve as an adjustable lean off

Edited by Brett on 12th Oct, 2014.

Yes i moved to the darkside

Instagram @jdm_brett

Data logging rpm, manifold pressure, afr and the needle position is all the info you should need to profile the perfect needle
Rpm will help you work out the way the needle moves to its 'stations' then you can graph needle station vs afr it soon becomes apparent what the needle dimentions should be

Yes i moved to the darkside

Instagram @jdm_brett

So you going to drive the electronic choke to adjust the mixture?

That'll need a microcontroller with a PID function driving a PWM output to the servo.

Is that really less complex than sequential injection with a black box ECU that just needs wiring up to the sensors etc?

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

I think I've only ever seen one of these carbs many years ago and the stepper motor was to move the jet up and down for cold start enrichment (the "choke").

My first comment would be a stepper would be far to slow and probably not continuously rated but I guess you've considered that and hence said replace it with a servo.

Still not convinced a servo would be much quicker (small ones are normally just stepper motors with an H bridge drive).

So what had you in mind for closed loop control off say a 0-5V AFR signal ?

Surely you are going to need some sort of microcontroller with a PID programme written for it plus the hardware to connect the inputs to the outputs ?

And when I say inputs it's because, if you only have AFR as an input, you can only be looking for one fixed target for that chosen AFR value.

In practice an engine needs many different AFR targets from all the other parameters especially during acceleration (ie, TPS dot, MAP dot and so on).
And when it comes to acceleration, your movable jet servo would be fighting the dashpot that is trying to richen the mixture, not keep it constant.

Basically you would need a lot more inputs than AFR, a fair bit of hardware, some complex code, pretty much everything an ECU designed, and code written for the purpose, does with injectors.

But it is an interesting idea.

Schrödinger's cat - so which one am I ???

The choke in the HIF E functions in the same way as the HIF, but the choke spindle is turned by the stepper motor rather than by a cable.

agree with Metro turbo I have at least one down the garage from a K series 1.1 metro

You can use a wideband for 2 distinctly different things: tuning, and closed loop feedback. With EFI, it's much much better to get the tune right, than rely heavily on feedback and the same principle applies to your carb idea.

As Brett said, you really need to use the o2 sensor to help you profile the needle. Give it a perfectly profiled needle and this is the best your SU will be (and make sure it's in good health too).

If you want simple, then try this:

1)wideband in the exhaust, gauge on the dash.
2)Some kind of indicator that shows the current height of the piston in the dashpot.
3)Video 1 and 2, then playback later and work out which needle stations need attention.

Point 2 to is the trickiest. Driving around at wide open throttle with the bonnet off will probably draw a lot of attention. It would be better to somehow rig up a potentiometer to the piston and read the voltage from inside the car.

Edited by PhilR on 16th Oct, 2014.

Dazed and Confused....

Haha. I Knew it had been done before; no surprises it's on here! Nice one Brett.

Edited by PhilR on 16th Oct, 2014.

yeah it works well,
a couple screen shots from when i was profiling johns needle
this is my sensor logging on ms2 piggybacking over mjlj lol
id post pics from my mini but there on my mapping laptop

Yes i moved to the darkside

Instagram @jdm_brett