E30 with SR20DET

[Jon_Bmw]

Tempted to come up to that. Could bring the Wideband and metal pipe to do some AFR checking...

[Theo]

Sounds good Jon, I think you should come along. What's this metal pipe?

[appletree]

Great work looks ace sat in the bay.
Just a thought though with the dump valve after the maf could that make it run lean on part load when the dump valves releaving and putting unmeasured air in to the engine? Dont know if it would be a major prob, although i cant see the maf adding much restriction to the inlet tract and throttle response.

[Theo]

I don't think so (although I stand to be corrected). The air has already been measured by the MAF as it passed through it in the first place.

[gareth]

a dump valve only releases on the over-run at zero throttle doesn't it? boost control being done via the wastegate/exhaust side of things...?

[appletree]

It does open at zero throttle but at part load it can be part open to as its not an on/off valve the spring is working against the vacume from between the engine an the throttle plate.

See what your saying theo, i was thinking it might be like the VAG thing were there needs to be sum air resirculating.

If thats how nissans run them it must be right for that engine, probly should have thought abit more before i posted. Lol

[GeoffBob]

Just a thought though with the dump valve after the maf could that make it run lean on part load when the dump valves releaving and putting unmeasured air in to the engine? Dont know if it would be a major prob, although i cant see the maf adding much restriction to the inlet tract and throttle response.

Actually just the opposite Matt. If the BOV was releasing to the atmosphere then the engine would run rich since it would not receive air that it has been told (by the MAF) it is going to receive. However, Theo's BOV recirculates back into the mouth of the turbo after the MAF, so the total quantity of air is accounted for.

Where a complication can, however, arise is in the fact that the air now arrives in the combustion chamber some substantial time after it was accounted for by the MAF (where it is the signal from the MAF that controls the fuel injectors via the ECU). The engine could, therefore, run rich as the throttle is closed and the BOV begins recirculating (resulting in some bang bang type noises) and then momentarily run lean as the throttle is again cracked open.

A further complication can be as a result of the type of MAF used and whether or not air released by the BOV can leak through the MAF back to the atmosphere.

Sorry Theo, I don't mean to pick holes. These are just the facts with the system Nissan runs. However, on the original, was the BOV before or after the intercooler?

[Theo]

Where a complication can, however, arise is in the fact that the air now arrives in the combustion chamber some substantial time after it was accounted for by the MAF (where it is the signal from the MAF that controls the fuel injectors via the ECU). The engine could, therefore, run rich as the throttle is closed and the BOV begins recirculating (resulting in some bang bang type noises) and then momentarily run lean as the throttle is again cracked open.

A further complication can be as a result of the type of MAF used and whether or not air released by the BOV can leak through the MAF back to the atmosphere.

Sorry Theo, I don't mean to pick holes. These are just the facts with the system Nissan runs. However, on the original, was the BOV before or after the intercooler?

Would a wideband AFR gauge be sensitive enough/have sufficient resolution to show a very brief 'lean spell'? On the standard Nissan setup the recirculation valve is mounted after the intercooler just like mine.

[GeoffBob]

Theo, to the best of my knowledge (and others may say differently) a Lambda sensor cannot be used to respond to rapid or step changes in the ratio of air to fuel at the engine inlet. Since the Lambda probe responds retrospectively (it measures what has happened post the event) to tell the ECU to correct the mix, there is an inherent delay in the system. Technically speaking, this is described as the response time of the closed loop.

Knowing very little about the algorithms employed in the Nissan ECU I would suggest that this problem might be taken care of with a certain degree of ECU control coming from the TPS, to form a "load" input to the fuel and spark maps derived from a combination of the MAF and TPS inputs.

[Jon_Bmw]

Theo it is all very council, so perfect for Ace.

Basically I have a wide band sensor welded into some 1" metal tube. Poke that up its chuffer and the controller and display you hold in the dashboard area.

The wideband would pick up a brief lean spell, but you need a logger to look back at it. Unfortunately my brother hasn't completed the electronics board for that.

This is what it looks like:



Only an old picture, imagine a wideband and some wires feeding into the car.

Using the L shaped brackets, we can hold it very tight with bungies.

[GeoffBob]

I think I see where this is going. You want to use Jons Lambda to pick up sudden changes in the AFR? The answer is yes, a Lambda sensor will do this for you (assuming the display updates fast enough), but just keep in mind that it does it post the event. In other words you will be looking at data arriving on the display a split-second after the lean event has already happened. To use the data from a Lambda sensor to correct the AFR by feeding back to the ECU is standard procedure with many ECU's during steady state conditions, but notoriously difficult to implement in order to combat step changes in the AFR (such as when the BOV opens).

[Jon_Bmw]

Agreed Geoff. It is a fraction late the data you are receiving, but it is as good as you are going to get without mega ££££s. Some data should hopefully be better than no data!

Interestingly cars fitted with wideband lambda sensors are pretty bloody good. We rolling roaded a tuned R35 GTR at work about a week ago. It was on 'cobbs' acess port.

I kid you not, all of the power runs, the AFR was totally straight across 13:1. Amazing really. The AFR was independant of the car btw.

It made 482bhp ATW and an amazing AFR graph, happy days.

[Ziggy]

Great stuff

Jon, I've ordered the lead I need to talk to my ECU, so I'll have to pop over to yours soon!

[Jon_Bmw]

Wideband party.

Whose chuffer can we stick it up.

[Ziggy]

[GeoffBob]

FYI Theo, if you're interested, this is the controller I use with the Bosch LSU4 wideband Lambda sensor on my car. It's a little bigger than a matchbox and sells for around £80. The AFR is displayed on a row of LED's that run up and down the side of the box. The unit includes a 0-5V analogue output (representing AFR 10:1 to 20:1) which can be used to drive a dash mounted gauge and/or an external AFR input to an ECU. It's made in SA, see here, but I think they export. I'm using mine to drive both a guage on my dash as well as my Adaptronic ECU.

[GeoffBob]

Agreed Geoff. It is a fraction late the data you are receiving, but it is as good as you are going to get without mega ££££s. Some data should hopefully be better than no data!

It's not so much that it's late Jon, its the fact that it happens post the event. If you measure a lean mixture, the mixture is already lean. Likewise, if you measure a rich mixture, the mixture is already rich. AFR control using a Lambda sensor is reactive (rather than proactive), because it responds to the air and fuel that has already been combusted. A MAF, on the other hand, is proactive because it tells the ECU how much air is coming in so that the ECU can set the injector duration in order to achieve the correct AFR from the get go.

Don't get me wrong, closed loop AFR control is an important part of engine management and I use a wideband Lamba to achieve exactly this on my own car. However, because of the speed of the control loop, a closed loop AFR control is good only for steady state and slow varying control, and cannot cope with rapid (read that as near instantaneous) changes in the actual AFR.

I believe what you say with regard to the R35 GTR, but I have yet to see an ECU that does away with MAF, MAP, TPS and other sensors and simply controls the injector duration by measuring the RPM and AFR. All ECU’s still have to measure (in some way or another) the mass of the inducted air in order to proactively set the injector duration, and then, using closed loop AFR control from the Lambda, to reactively adjust out any inaccuracies in the proactive setting of the fueling.

[Gunni]

Ah the nice people at Perfect Power, I´ve done business with them for ages.

Anyway from my findings in tuning cars where the WBO2 is in the downpipe then you can assume that there is a slight mismatch to injector pulsewidth and the response time of the wbo2 sensor to pick it up, this shows as a slight delay in logs. Nothing to bad. But needs to be known when reviewing logs.

When the sensor is in the end of the tailpipe the delay is more but the readings to show up.

I also agree with GeoffBob that I do not know of a ecu that can run the car on rpm and WBO2 alone.
It could under cruise but any other then that it would be slow to react. Let alone under rpm rise

[Theo]

I think I see where this is going. You want to use Jons Lambda to pick up sudden changes in the AFR? The answer is yes, a Lambda sensor will do this for you (assuming the display updates fast enough), but just keep in mind that it does it post the event. In other words you will be looking at data arriving on the display a split-second after the lean event has already happened. To use the data from a Lambda sensor to correct the AFR by feeding back to the ECU is standard procedure with many ECU's during steady state conditions, but notoriously difficult to implement in order to combat step changes in the AFR (such as when the BOV opens).

I see, I thought it would be nice to have a fiddle with the wideband to check out the AFRs and also see whether the recirc causes a momentary lean spell. I’m not too concerned about it to be honest, a lot of people recirc their charge without issue. Before I had the Walbro fitted I really wanted to check the AFRs to see if the standard Bosch external pump was running out of puff, but that’s not a concern anymore.

That perfect power device looks neat and could easily be accommodated in my car. A friend of mine suffered engine-killing detonation on his Silvia last month and even something as simple as that row of LEDs could have given him sufficient warning to avoid it.

Jon, that looks like an incredible fabrication you’ve done with that bracket for the wide band. It would give me great pleasure if you could stick it up my chuff please.

[Jon_Bmw]

All true again Geoff, all I was saying was that I have yet to see a car NOT fitted with a closed loop wideband make a pretty much perfectly straight line across the graph.

I assume what it does is almost learn from its mistakes, if that makes sense. Once the car has been remapped, you assume the standard ECU sensor figures will be out. So unless the guys at Access port spent months getting it spot on, I don't know how it has done it. I guess it is a continuous learning process till it optimises the map completely. Perhaps I am wrong, who knows. Agree about the sensors, I just just suggesting that the wideband is perhaps there to help the learning process.

[Jon_Bmw]

I guess the other question is why so many MAF/MAP etc cars get the AFR so 'higgldy-pigldy' that are not fitted with widebands.

Theo i'd love to stick it up your chuff, do you have a battery in the boot for recharging?

[GeoffBob]

Great stuff then gents. Interesting reading as usual.

Jon, sounds to me like what you are describing is what Adaptronic call adaptive learning. Adaptronic in Oz were the crowd that pioneered both adaptive fuelling and adaptive ignition. It's a process whereby the ECU maps itself. It needs a decent base map to start off with (preferably created on a dyno) but from then on it adapts the fuel and ignition values using real-time RPM and AFR data. Its a very slow process, I know because I've watched my Adaptronic ECU doing it, but it is something totally different from closed loop AFR control.

Gunni, I know the people at Perfect Power very well. Have you ever used their ECU?

Theo, popping a wideband Lambda up the exhaust and looking for moments of high or low AFR is a good idea. I did the same with a mate of mine who has one of the top of the range units from Autronic. It compared fairly well with my dash mounted AFR from Perfect Power, so now I use that for reference. You'd do well to get something like that on your car as it could well save your engine in the event of a fuel pump or FPR going faulty.

[GeoffBob]

I guess the other question is why so many MAF/MAP etc cars get the AFR so 'higgldy-pigldy' that are not fitted with widebands.

Because they are mapped badly to start off with. No matter how accurate and uncorrupted the signals from the various sensors on the engine, the ignition angle and injector duration is only as good as the person who created the large 3D table of data (the map) that lives inside the ECU. As always, the weak-link is a human being. The more time and effort that is put into mapping a specific engine, the better the result. Mass produced engines where "one chip fits all" are seldom as refined as an ECU custom mapped for a particular vehicle. Another factor would be the age of the vehicle and whether the original map is applicable to an engine with worn components and dirty injectors etc.

[GeoffBob]

I thought it would be nice to have a fiddle with the wideband to check out the AFRs and also see whether the recirc causes a momentary lean spell.

I am going to put my head on a block here and predict that your mixture will briefly go rich when you snap the throttle closed under boost (causing the BOV to recirc) and then momentarily go lean if you go WOT immediately after. I'm basing this prediction upon the size of your intercooler. Not something to be worried about, but just for fun I am making this prediction.

Please be sure to post up that I got it completely wrong after yourself and Jon have tested Have fun testing - I hope your weather is better by then!

[appletree]

Another good AFR gauge is this one i have made buy AEM there alittle dearer but come with a 52mm dash gauge.

http://www.aempower.com/ViewCategory.aspx?CategoryID=67

here it is in action


When dave walker compared mine to his on the rollers there was .2 difference but his was an up the exhuast type as min was in the down pipe

[Gunni]

Gunni, I know the people at Perfect Power very well. Have you ever used their ECU?

Yes I have used the XMS3 on a few cars. Although I wish I hadn´t.
The non interpolating one. I was not impressed one bit.
I haven´t used any of the new ones. But wish I had on those other cars.

And I have used tons of SMT6´s and have just gotten a few SMT8t´s shipped. Not that I´m a fan of Piggybacks , it´s just that sometimes they are the cheapest and fastest solution when doing minor changes.

[Gunni]

I guess the other question is why so many MAF/MAP etc cars get the AFR so 'higgldy-pigldy' that are not fitted with widebands.

Because they are mapped badly to start off with. No matter how accurate and uncorrupted the signals from the various sensors on the engine, the ignition angle and injector duration is only as good as the person who created the large 3D table of data (the map) that lives inside the ECU. As always, the weak-link is a human being. The more time and effort that is put into mapping a specific engine, the better the result. Mass produced engines where "one chip fits all" are seldom as refined as an ECU custom mapped for a particular vehicle. Another factor would be the age of the vehicle and whether the original map is applicable to an engine with worn components and dirty injectors etc.

From my experience and how MAF systems are supposed to work, tuning them should be as easy as just putting in the AFR´s you want and that´s it.

But that´s not the case with them and would take tons of variables to get spot on under all sensor inputs.

One of the biggest problem is the incorrectness of the MAF scalar inside the ecu. If I had plenty of time I would put up a AFR chart/fuel table I like and then adjust the MAF scalar to make the table be right.

But next car with the same MAF but less dirty will be off.
And so it will never be right.

[GeoffBob]

One of the biggest problem is the incorrectness of the MAF scalar inside the ecu. If I had plenty of time I would put up a AFR chart/fuel table I like and then adjust the MAF scalar to make the table be right.

But next car with the same MAF but less dirty will be off.
And so it will never be right.

This is why I like the Adaptronic ECU so much. It includes the facility to calibrate all connected sensors (a small 2-column table) mapping actual sensor electrical output to the true value. It was quite a process with my MAP and MAT sensors I can tell you since it requires that you have a more accurate "reference" sensor against which to calibrate the one on the engine. The up side is that my fuel table started life as an "ideal" table, created using a spreadsheet and a little knowledge of my engines VE. After that it's a case of using the adaptive learning feature to correct the tablulated data, or more properly using a dyno.

On the issue of maintaining correct AFR, I have a similar issue with MAP based systems. Following the ideal gas law, it is necessary to tell the ECU both the manifold air pressure and manifold air temperature in order to correctly determine the mass of air inducted into the engine. Pressure alone is not enough ”“ inducted air temperature must be measured (common knowledge I think?).

The problem I have with a MAP based system, however, is that the input to one axis of both the fuel and ignition tables is manifold pressure, and not mass air flow, as it should be. The output of the table (injector duration or ignition angle) then has a correction factor applied to it as a function of the signal from the MAT (temperature sensor) in order to take temperature into account. Honestly, this is soooo wrong. It works, I know, but the correct way to do it would be to combine the measured manifold temperature with the measured manifold pressure to determine the mass air flow, and then input that value to the fuel and ignition maps. Not this nonsense of applying some not-very-well understood correction after the tables output the ignition angle and injector duration.

What makes it worse is this correction (A percentage trim as a function of measured manifold air temperature) is defined by the tuner, so if the tuner gets it wrong you may get correct AFR across the range of RPM and air flow on a cool day, but the wheels fall off the AFR on a hot day!

Even worse, you get stuck in traffic on a cool day long enough for your radiator to start heating up your intercooler and before you know it, your manifold air temp starts climbing and the engine battles to idle because some idiot got the correction factors wrong, and now the AFR is badly out!

My point is this (similar to yours Gunni) so long as the output of the MAP and MAT sensors are well defined (calibrated) all that should be required is an accurate fuel and ignition map that requires little or no fiddling and/or correction factors applied.

[Gunni]

This is why I take the time to set up those correction tables at various temperatures in random situations, often the air temp sensor is heatsoaking and not giving out accurate readings so as with everything Engine management it´s all a damn compromise. Luckily VEMS has another table that will allow you to adjust for the sensors heatsoaking.

[GeoffBob]

Here's trick for you Gunni. I reverse the wires to the electric fan/s (obviously assuming the car has electric fans) so that each time the fans come on they deliberately heat up the front mounted intercooler, raising the temperarture of the inducted air by a few degrees at a time. This provides the opportunity to perform the air temp corrections. I discovered this trick once when I accidentally wired the fan backwards and watched the signal from the MAT sensor climb slowly.

Apologies for scribbling all over your thread Theo .

[Theo]

Right. Car is now properly aligned and corner weighted and the results are:

Overall mass with 40 litres of fuel - 1094kg (not including driver)
Weight distribution: Front 52.6% Rear 47.4%
Front camber -1.75*
Rear camber -2.25*

Also set the castor and front and rear toe, now all that's needed it a trackday to see what tweaks are needed.

[appletree]

Thats a pritty dam good weight distribution!!! NICE

[Theo]

Yep, I'm pretty pleased with it. There's 1kg difference between front left and front right. There is however almost 60kg difference betweeen rear left and rear right, which is odd. Has anyone else encountered this when putting your e30 on corner scales?

[gareth]

that's quite a lot

have you got a missing rear spring pad or something?

[Theo]

have you got a missing rear spring pad or something?

No missing spring pad. It's the nearside which is heavier.