16 months on.....
Turbo Mini Forums

Most of the wiring installed in the car but real problems with the TechEdge software talking to the MegaSquirt stuff (thanks to Ms Tobin....) I can get it to work on the desktop upstairs with the JimStim but not yet with the laptop outside.... so ignore evrything except the AFR figures.


About 0.6 - 0.8 apart so not great yet, BUT.......

the difference between Paul, Graham etc, is I did this with identical injection timing on inners and outers but with skewed VE tables.

If my theory follows through (long way to go yet...) I'm going to try and hit the best possible valve open time depending on RPM/load for individual pulses so there is no overlap between cylinders and dial out the charge robbing with the VE tables.

This can only work if I can get individual pulses short enough and accurate enough (timing) to send the required pulse width (detemined by the VE table calculation) into the correct cylinders.

Hopefully by tommorow I should have the rest of the wiring and the logging software conflicts resolved.

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

I look forward to seeing your results. I must say that this is the method I thought would work and would be used when I first started with the code implementation.

However, there is one thing that may become a problem and require more advance on the inners: the response speed of the injectors to the closing and opening sequence. You may need to shorten the injection window simply for that reason when you could instead advance the inners to make full use of the window available from the mechanical point of view.

But as I said, I look forward to see how that goes.

Jean

http://www.jbperf.com/

Although I think I know what you mean, the term charge robbing is more to do with air than fuel. With a turbo and a moderate cam, charge robbing should not be a major issue.

What is a problem is the carry over of fuel intended for the inner cylinders to the outers. I think this is what you refer to.

I tried this, albeit on an early version of the code and found that if I reduced either VE table, the AFRs on both cylinders would change.

In my opinion, there is so much interaction between cylinders caused by wall wetting, that it is near impossible to separate them in what would appear to be a logical manner like that.

Best of luck, all the same.

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 agree Jean, to make my theory work it relies on the actual injection pulse, including injector response time, fitting within the window at the actual load/RPM (ie, the actual transit time for any given flow rate of the mixture).

I don't know if it will work - Paul has made good progress with the overlapping hybrid single pulse mode, Graham has achieved good results (until he started playing with angled injectors) with the single pulse mode but I think we all accept that both methods rely on some of the first (or early part of the combined) pulse carrying over to the outer cylinder to prevent it running lean.

My logic is keep the pulse width well within the valve open (including transit time) window - which will obviously require the variable timing tables you have provided us with - but do the mixture variation with the VE tables.

Only time will tell.

Once your second batch of I/O-x boards are ready I can put the third wideband on the display too, at the moment the Inovative, which I have on the common exhaust, is giving crap results fed in through the MS2 as a 0-5V signal.

EDIT - out of order with Paul's post above...

Edited by Rod S on 7th Oct, 2010.

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

Paul, what I hope is that with very short runners the wall wetting should not be such an issue and with very large injectors and hopefully fast response times using low z, I can hit the window with enough margin to use VE tables to set the AFRs.

But it's just a theory at the moment.

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

Maybe its just the weather but I think I actually understood some of that !

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

One step at a time, I think !

Link between TechEdge and MS data sorted (a couple of files missing on the laptop.....) but before I do any serious testing, I thought I noticed yesterday that the AFR balance changed as the engine warmed up.

I saw this last year when I was testing the twin widebands on a carb so I tried to log it today....

Best representation is dumping the data into excel



Now obviously the AFRs are too high anyway at the moment and there is a lot of noise (or the injectors I'm using are rubbish at low pulse widths, or I haven't got their PWM settings right yet) but the change of AFRs with temperature is very obvious.

I can't put the RPM on in excel as it only allows two "y" axes, but it was from 1200 down to 1000 over the timeframe.

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

That's to be expected Rod. A lot more wall wetting when the engine is cold which means rich inners.

It's the same with mine, but I don't let it idle for that long.

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 assumed it was wall wetting from everything else discussed over the last two years, I'm just surprised at the difference when cold. Once I bring the AFRs back together at 85C (whether by timing or VE tables), it will be even worse when cold.

And I don't remember your deviation being that bad on your twin AEM gauges when you showed me it at MITP. I think I've probably got something else not yet set right for the warm-up phase.

I have no choice about how long it idles at the moment :)
I haven't finished building the brakes yet, let alone getting an MOT :)

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

It's probably exzagerated by the fact that you are running far too lean.

Ignore stoich and set warm idle at an AFR that gives a nice steady engine. Probably around 13.5:1.

You will see them close up even when icy cold.

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."

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

With almost everything working.....



Notice the difference between the TechEdge AFRs and the Innovative in the common exhaust after the turbo.

All have been FAC'd recently.

The Innovative goes in as a 0-5V signal to the MS2 whereas the TechEdges are true digital signals.

I still wasn't expecting such a difference.

I've checked the Innovative using their own LogWorks software (which reads its digital output) and the difference was the same.

I'll recheck the FACs tommorow on all three just to be certain and once I've got one of Jean's I/O-x boards all ground variations etc on the 0-5V signal are discounted as all three will be logged digitally.

But I'm surprised at the difference unless something is wrong with the FAC.

Oh, and radiator fan while I'm experimenting...

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

FAC?

I'm guessing Fucking Analogue Curve,

Free Air Calibration :)

Sorry Joe, I'm too used to TLAs (Three Letter Acronyms) from where I used to work....

It's what you are supposed to do regularly to the ones that use the built in calibration resistor from the factory (usually installed in the connector).

The AEM new stuff and also the TechEdge new stuff get around this somehow - I don't quite understand how - but the original Innovative and TechEdge still require calibration to be accurate.

Well that's what they say and I have a clear difference in readings between the Innovative and average of the TechEdges after calibrating them both as specified......

Edited by Rod S on 8th Oct, 2010.

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

Ah yes, make sense now,

I run 11:1 AFR @1200rpm when dead cold and 14.3:1 @880rpm above 75C. It's a bit rough when warm, becomes much more regular at 13:1 AFR but too close to 2.5% CO level limit. 13.5:1 at 80C gives around 1.5% CO.

std 998 A+, g295, MD266, RHF4, 109hp @0.8bar/5400rpm

How about fitting the Innovate sensor in one of the sample chambers for a cross reference?

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."

http://www.jbperf.com/

Nice idea, hadn't thought of that one.

But, assuming you mean the Innovate sensor and its controller, unfortunately not very practicable as the Innovate controller is down near the gear change and the TechEdges are up on the dash tray next to the MS2.....

The three LSUs are identical Bosch part numbers so logic says it has to be either FAC or the fact I'm just sending the Innovate one in via the MS2 as a crude 0-5V signal (so there could be a small ground offset) whereas the TechEdges are fully digital through to the laptop.

Reading the Innovate digitally through it's own LogWorks programme though showed the same offset suggesting it's FAC, so I'll re-do the FACs tomorrow (too damp out today) but as soon as Jean has his next batch of I/O-x boards available I'll set one up to read the Innovate digitally.

EDIT - out of order with Paul/Jean's posts.....

Edited by Rod S on 9th Oct, 2010.

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

Just to avoid confusion, at the moment it's two identical TechEdge controllers on two seperate wideband LSUs before the turbo - sample chambers are on siamesed exhaust 2/3 and the second one on single exhaust 4.

These are the ones that need reasonable confidence they are reading the same for setting up the code.

The third one is a single wideband using a standard Innovate LC-1 on a standard tapping (bung) about 800mm after the turbo down near the gearchange.

I really only installed it because I had it, and to add a "confidence" factor and to be able to put a simple single gauge on the dash as a "warning" of problems.

It's not quite the average of the other two yet (as I was expecting) but, on the "bananas" arguement, if a Free Air Calibration doesn't sort it, I could just subtely alter the calibration curve in the Innovate firmware to bring it in line with the TechEdges...... :) :) :)

All three LSUs are the same Bosch part.

With the benefit of hindsight, I would be using three TechEdges.... unfortunately I bought the Innovate one before I knew what I was doing :)

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

Rod,

I knew you were using the 2 TechEdge controllers in the "right" configuration. I just wanted to emphasise it.

And I suspect the issue is not really calibration but more fundamental than that. It's basically how each company interprets the Bosch specs for the sensor and how they control it. I don't know how they perform their own tests to come up with what they have but it seems that each manufacturer is systematically different for the other ones.

It might be that the differences are more due to the analog outputs and what I have seen is more due to errors on that than actually measurement differences. There definitely seems to be a trend where the value seen in MS is different from the value seen on the PC or on a gauge. Whether this is due to wiring issues, unreliable analog specs or other issues is debatable but there are definitely different behaviours with different controllers.

Jean

http://www.jbperf.com/

I took this snapshot a couple of days ago but only just transfered the file off the laptop (I was setting up an old laptop just for data logging).

The third AFR is still analogue (I haven't finished wiring up Jean's new board yet.....) so we know there is a yet un-explained offset but it's pretty constant at +1 when the engine is idleing at ~900 RPM

Red is inners, green outer (No4 only) and orange common exhaust about 700mm after the turbo.



Ignore the numbers at the top (they match the end of the plot only) and ignore the large spikes when I raise and lower the RPM (AE is turned off at the moment).....

Inners and outer deviate - no surprise as I've only optimised the injection timing and VE tables for idle so far - but the common drops to (average -0.2) ie, much closer to what I would expect it to read relative to the other two.

Although I've only got the one snapshot, this is 100% repeatable.

I'm wondering if there are some pressure related effects on the LSUs in the sample chambers (as Paul suggested in another thread) or whether taking the "outer" off just cylinder 4 is a bad idea.

Note, the overall irregularity - probably down to my VE ables being very rough in the bottom left corner which creates a poor idle - is absolutely identical on all three plots which suggests all three LSUs and controllers work perfectly.

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

Interesting........

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."