Engine build progress
- Thread starter TTM
- Start date
TTM
Well-known member
Nick, what puzzles me is that I get no knock when running lean and on the wastegate spring. When I add fuel to run lower EGTs I then get knock. I get away with the lean spike/knock when running lean and on the wastegate spring, but as soon as as I plug the piggyback to both add fuel and run a flat boost ramp (in order to get as possible to natural boost ramp of the wastegate spring) the lean spike/knock are both back. This is what I noticed yesterday evening. I have one more test to do, that is with the Piggyback with same fuel level, the same flat boost ramp and a lower peak boost pressure. If the lean spike goes away and there is no knock then the turbo probably spools too fast for the EMS to keep up.
DivineE
New member
I got chatting (in a pub) to one of Jaguars engineers who deals purely with mapping the engines every day. I tried to explain this fault that seems to affect all of the bigger 944t engines and he said he would be certain its all relating to the boost control not controlling the spike as it comes on boost. He reckons if you can get a boost control system in place that's good enough to keep the boost curve smooth it would go away. I got wondering about the waste gate design as it exits the crossover pipe at more than 90* from the crossover pipe. It was pointed out to be by one of the guys at Hartech and looking at the book of theory it should never be like that. You want the air flowing to the waste gate to be as close to the direction of flow into the turbo as possible so that the pressure is the same.
Would knock caused by over boosting cause a lean spike to read on a wideband even if the AFR was correct. Misfires etc can increase the AFR indicated on a wideband because of the way it detects AFR. Maybe high AFR's are not causing the knock but the other way round. With regards to the boost spike issue, it is a typical control systems issue commonly known as overshoot. In turbo terms the wastegate isn't opening soon enough to stop the turbo driving boost above the desired level. With a DPW type boost control with either a MBC or EBC the wastegate can never open the until boost pressure at least equals the spring pressure in the wastegate (maybe you can use engine vacuum to get round this). If this spring pressure is too high then the boost control cannot open the wastegate early enough to stop the boost spiking too high. With a weaker spring you can get the wastegate open earlier and prevent a boost spike. You might need a closed loop boost control system to take advantage of this. On a side note I started building a new EMS solution for the 944's DME last night, should have MAP/MAF and on board closed loop boost control. Hopefully I can deal with this issue.
Reading back a bit it seems that you don't get the spike on the natural wastegate spring. So it sounds then like the piggyback is bleeding too much boost through to the spring side of the wastegate keeping it shut for too long and causing the boost to spike. Does a boost spike register on any of your sensors Thom? Even if it doesn't register it may be damped out by the sensor to give a smooth reading.
TTM
Well-known member
I cannot spot any boost spike on the logs. I don't think I have ever had an issue with boost spikes. In fact, some of the logs I posted in older posts show that the manifold pressure rises sharply right at the moment I open the throttle and is followed by the lean spike. If I open the throttle progressively to avoid this "step" in the way the manifold pressure rises, no lean spike.
That's what I was beginning to think. As detonation/preignition is caused by some unburnt fuel exploding after/before the ignition phase then perhaps if I feed the cylinders with "too much" fuel I'm likely to get "more" detonation/preignition? In that case the base fuel map works well with the ignition map even though the engine runs lean, as I don't get any knock running this way. The ignition map may not work with more fuel. The knock I am seeing when adding fuel may be preignition, not detonation. Because the system suddenly adds too much fuel when I open the throttle, and this quantity of unburnt fuel doesn't make it in the AFR logs. -> Maybe I should fit smaller injectors or upgrade the ignition?
Looking at the logs, the knock/lean spike happens well before the car is building boost. Just after you floor the throttle. Is this correct? If so then we can eliminate boost/boost spikes as the issue. Ignition would be my next thought, at low throttle and plenty of vacuum the ignition is quite advanced as the mixture burns slower, when you floor it you get good cylinder fill but does the ignition remain advanced? It would only take a delay of a couple of revolutions where the car is at WOT but the DME is still running very advanced ignition. How does the Vitesse system work out what the advance should be? It can't see the increase in vacuum and there may be a slight delay in seeing a reponse from the MAF and in switching to the WOT maps as the throttle opens. This would cause knock as the cylinder pressures would jump very high. This in turn may read as a lean spike on the wideband. Without the piggyback the ignition timings may be slighty less advanced to begin with which is why you don't get the same symptoms despite running leaner.
TTM
Well-known member
Absolutely. The Vitesse software adjusts ignition timing in function of manifold pressure - this is what the MAP sensor is used for. Plugging or not the Piggyback doesn't seem to have any effects on how advance is handled. If ever there is an ignition map hiding in there, the end user cannot access it through the Piggyback. As there is no reason the ignition map should vary when using the Piggyback or not, I want to test if setting the EBC to the same boost charateristics (peak boost value + boost ramp) as the Wastegate spring has any impact to the lean spike/knock.
The piggyback is what allows the Vitesse system to change ignition isn't it? I thought it manipulated the crank triggers etc to get the ignition advance it wanted? Where does the MAP sensor connect too? How does the software on the vitesse board in the DME communicate with the piggyback? Can you explain how the vitesse system hook's up to everything? It will help me understand how it all works. It seems like the piggyback is totally separate from the MAF system. It just changes the position triggers and the MAF signal to change the fuel/ignition on the engine. So without the piggyback plugged in the ignition and the fuel maps will revert to the default supplied on the MAF chip. Does this make sense with your experience of the system?
TTM
Well-known member
Tom, base maps are on the chipboard in the DME box. Running without the Piggyback plugged to the modified wiring loom means running one of the maps on the chipboard, these are the maps that I call "base maps". Through the Piggyback the user can make alterations to the fuel map and control the EBC, but has no access to the ignition map. The MAP sensor connects to the manifold pressure line and the software on the chipboard adjusts ignition timing by itself in function of the voltage it gets from the MAP sensor. Again, the end user has no access to this, and it makes little sense that plugging in the Piggyback would mean switching to some ignition map other than the one already implemented on the chipboard. As for the MAF sensor it sends a (calibrated) voltage to the DME. All the knowledge, including the MAF function, seems to be in the chipboard.
Where does the MAP sensor connect too electrically? Is it wired into the existing wiring loom? Vitesse states you need the MAF and Piggyback for V-Flex, this suggests the MAP based ignition tuning is disabled when the piggyback is disconnected. My guess is that the piggyback is reading the MAP sensor and tricking the DME computer into changing the ignition by changing the crank trigger signals.
TTM
Well-known member
The MAP sensor connects to the plug for the Piggyback. When running without the Piggyback a short is plugged on the plug (!). There is no provision in the original wiring harness for a MAP sensor (though I saw a car in Germany recently that suggested otherwise, but this is another topic). Piggybacks being swappable suggest the ignition map is stored on the chipboard only and the Piggyback has no effect on it. It's the software on the chipboard that handles the ignition map in function of the MAP signal.
Are there any wires coming out of the DME from the new chip board? Or does it just keep the original DME connector? I'm wondering if the Piggyback can talk to the chip board so it knows what values are in the fuel maps. Or do you just add or subtract fuel rather than changing the actual value.
Got ya, the final piece of the puzzle is how the chip board reads the MAP sensor. Either it uses the existing analogue inputs on the microprocessor in the DME or it has its own analogue input which would require the MAP sensor to be wired up directly to the chip board. Are there any wires coming from the chip board out of the DME? Edit: John, says its all done on the DME side like you say...
Thom, my understanding is exactly the same, that Vitesse (John) has his own ignition maps that are much more agressive than factory. He does not want people to know what his ignition maps are or how much advance he runs, I think he also sells the fact that he doesnt want people adjusting the ignition to stop them grenading their engines too. I know that in order for Pauls 3.2 engine to be tuned properly there was some adjustment required to the timing on the piggyback, hence how I know it can be done. If you tell him you are having this problem he may well let you have the password to try and resolve it. I am not sure if the turbos he sells now spool as fast as Simons, I know my Stage V Vitesse turbo doesnt for sure!
]Fatherhood and moving house! Now the move is done I am focusing back on the car while the good weather is here. Not much to do really just lots of niggly time comsuming or money consuming items. Only trouble is I sacrificed a big garage for a big house and garden, the garage is there but its 30's style very narrow so the dream of having a drive in workshop still eludes me!
Thom, if you read the info on the Vflex software on Vitesses site it specifically states that it adjusts the ignition timing relative to boost pressure and promotes this feature as a Vitesse innovation. Therfore if the values he has are not working for your setup then that is most likely the cause of your knock. The time the turbo comes on boost is the most critical point to get the timing right and the most dangerous time for your engine as that is when it hits peak torque. Therefore this is the most likely time to get knock if the timing is too advanced. If this is the cause of your knock and if your a/f is good then only retarding the timing a touch will cure this.
Posts made and opinions expressed are those of the individual forum members
Use of the Forum is subject to the Terms and Conditions
