toyotec

If you can confidently do all the work yourself, understand engine technology, know where to source components and know how to target drive feel then go for it. The drive feel and experience allows you to target the level of torque and ultimately power the engine will have to deliver in all conditions. This dicates which components are required or not and of course a rough budget. If you will be paying out get a 225PS spec 20vT. When I built mine, there where few who had engines like this and none who had similar objectives. I made and broke a few custom components before the car ran "right". Now that the chassis has been significantly upset by 315PS, attention, time and money are now shifted to suspension components to effectively harness the grunt to the tarmac. The current engine has done over 10,000 miles and the car still runs as a demo. I can be pm'd for more details but only do so if you are sure this route is feasible.

The adaptation ( long term) occurs over a certain amount of cycles and operate in addtional mode for idle, muliplicative for part load. You also have general lambda regualtion that would wobble the fuel mixture between 0.97 to 1.05 lambda for best cat efficiency. If a disruption to lambda 1 occurs, like an air leak or dodgy injector, the lambda controller will add or remove fuel. You car seems to be running lean as the lambda controller is adding fuel by up to 20% in the long term. 25% is usually the clip for both the short term and long term items hence the fault. The 02 reading relates to general lambda control. Nominal values are +/- 3%. I would look at items such as the purge valve as with age these can stop sealing or seating leading to a throttle bypass and messing up the aircharge calculation.

Forget all that "OBD2" malarky. The US "OBD2" engines with M5.9.X controllers have 2 lambda sensors + two under floor cats. It is an American requirement. The European engines from August 95 with M3.8.1 controller have a single feedgas lambda sensor as do previous engines running M2.9.1. The lambda sensor you require is # 021906256AK as fitted to Golf 3 VR6 from August 95. I am assuming the engine controller you use would be 021906256.

A lambda value of -15% suggests the "fr" value or regulation factor from the lambda controller is removing 15% of fuel at that time to allow the lambda sensor to correct to 0.455v or lambda 1. If there are no air leaks ACF ( purge) or diagnostics going on your fr value should be +/- 5%. The upper and lower limts for fr are calibrated to +/-25% and the upper and lower limits of both idle and parlt load adaptive lambda are +/- 30%. However if you change the ECU for another and have the same issue then the problem is not the ECU it is the hardware.

If it suffers from the occasional juddd....dddd....ddder at low revs, this could be a plug lead or spark plug. This assuning all other things that can shut a cylinder down are fine.

Same here. My leads are less than a year old but the vehicle developed a misfire at idle while warming up after a soak. It was a very wet day and the car had been parked for 3 days. Closer inspection displayed arcing at the coil pack, suggesting a build up of resistance between spark plug and lead. Removing the leads and plug at cylinder 4 cured the issue after passing a dry cloth on the offenders. If I have ignored and driven, a path would have burned to earth and there would be permanent damage to coil pack or driver. Perhaps as the car has taken a few hits like this, the damage may have already started but we will see. I would imagine the values in the calibrated tables are the match for IGBT amplifier and thus the coil pack's forward voltage as well as flashback. There is a lot of effort that goes into populating a table like that and it would not be recommend to tweak. Not for NASP engines anyway and once with definite hardware issues are described. For standalone's I would tend to use the same infomation seen in the bosch settings to set dwell.

Kev, The humidity of recent may be causing the plug leads and or the metal fitment over the spark plug to arc to ground causing the misfire as experienced by members. I have had this twice after my vehicle was parked over a weekend, when it was very wet. Yes in the M3.8.1 and 2.9.1 software, there is a Kenn Feld for Coil dwell time, but I would imagine that this is the last place to optimise as the problem is really mechanical. Perhaps a change of coil pack to different chracteristics may warrent such a cal change to dwell time, but you will have to take into account the rating of the coil pack drivers if this change will be effective. My coil packs are on 150K miles, but I do keep and eye for issue like these as they will only cause more expensive issues if not quickly rectified.

Yes. If you have an issue with packaging and the standard unit is not suitable to be used or modified, then the option you spoke about, once it can flow and does not induce heat soak into the charge, would be a good option.

It does not have to be an OE unit if packaging would prohibit. Large surface area cone will also work provided it is properly shielded from engine bay heat. It is the enthusiasts who bin their std filteration thinking that brand x filter kit is "better" without understading why should take note.

I did a series dyno tests on an ABF 16v engine where the intake system i.e. from filter box to throttle duct was investigated to see if there could be any improvements to minimising pressure drop. In that exercise there was a stablising baseline then the factory airbox was stepped from std system with and w/o panel filter, to pod filter , to modified airbox with and w/o OE/aftermarket filter. In the end the single one improvment on that airbox was the removal of the lower and upper airhorn + the creation of another hole towards the fender side of the car. This modified airbox produced slightly more torque post 4500rpm to 6800 than the rest of combinations. It may just been very slightly better than the best oiled cone filter( not that you would notice). The aftermarket panel filter had no performace benfit over the new factory paper element. There was also no benfit or loss to torque running without. The airbox was free, the mods took time but were free and the filter kit was 80Quid. So 80Quid saved. I use a similar airbox to kevhaywire in my 315PS 16v turbo, with subtle modifications to tweak sound. My VR6 also uses a factory airbox modified for some noise and similar to the 16v AIS study. We must also understand that testing a vehicle on a dyno with a fan blowing at the engine bay is one thing, but drop the bonnet over a hot engine and the temps do shoot up to Stick to your the black plastic air boxes. With the right modifications they can be tuned for the best response.

This is the forth lambda sensor. First was an after market ebay job which had the heater element short to ground code your talking about. The car actually run better with this one fitted. The second was a geniune Bosch sensor. The car run like a pig and was giving the lambda sensor short to ground and no other code. I took it back off and noticed I had put a big dent in it, so I thought that must be the problem. I am now on the third new Bosch sensor, same code! I dont think its the sensor or the wiring maybe? I was starting to think maybe a slight leak in the downpipe? You cant really see anything because of all the heat shield covering. I'm really annoyed as I have had to give up now as its driving me mad! I usually do everything on my car and can't really afford to give it to someone else at the moment. At least its drivable and not drinking lots of fuel. My main concern is the MOT is due in Jan. Thanks for any advice you can give. I just re-read this thread and saw that you had changed the sensor several times so apologies and you also tested the sensor wires. I am assuming here that the current ECU is remapped? You said you have another ECU have you tried this? Also when did you first notice the problem?

Was gonna ask earlier but are you the guy that did the ABF turbo megasquirt. Good to see you on here if so! Yep. I do mess around with aftermarket stuff quite a lot in my spare time. Of recent though I have been investigating and mapping older VW controls.

Replace lambda sensor. It would seem the heater on the sensor is faulty, causing high current draw from the heater circuit causing the fault to be triggered. The sensor should be replaced with an OE bosch unit. 0.5v to 0.0v to 0.5v is abnormal. The std settings in that cal suggest the system must wobble through 0.45v from ~0.1v to 0.7v then back to 0.1v or lean, rich, lean...

You can also use Moats Ostrich 2.0 and calibrate the ECU live, once you have confidence in the XDF. To do so you would need an EEPROM. The ones I use are multi-time programmable SST 27SF512.