lambda sensor placement
Discussion
Hi all.
Im currently (slowly) building a custom made exhaust manifold to a different design layout than the original however now im a little worried about lambda sensor location. I can never remember how to post up pictures so I will try to explain. Where the original manifold had a sensor both sides before the exhaust joins in the Y piece separating each bank of cylinders. One sensor reading cylinder 1,2,3,4 other reading 5,6,7,8. My design doesnt join like that, its more like 1 and 2 join 5 and 6 while 3 and 4 join 7 and 8. So is this going to screw up the way the ecu runs the car? Or is it only going to matter when you plug a rover gauge in and the left bank reading is really 2 from each side.
Im happy to post up pictures if needed. I hope to finish it today so hopefully better pictures posted then
Im currently (slowly) building a custom made exhaust manifold to a different design layout than the original however now im a little worried about lambda sensor location. I can never remember how to post up pictures so I will try to explain. Where the original manifold had a sensor both sides before the exhaust joins in the Y piece separating each bank of cylinders. One sensor reading cylinder 1,2,3,4 other reading 5,6,7,8. My design doesnt join like that, its more like 1 and 2 join 5 and 6 while 3 and 4 join 7 and 8. So is this going to screw up the way the ecu runs the car? Or is it only going to matter when you plug a rover gauge in and the left bank reading is really 2 from each side.
Im happy to post up pictures if needed. I hope to finish it today so hopefully better pictures posted then
You building something like this? 
Should look awesome, something I would like to try myself. Would be interesting to see what hp gains you can get. Have you had your car dyno,d recently? Re. The O2 sensors that's an interesting one. The system in not very sophisticated ei. A misfire on one cylinder will show up lean on that bank and so the whole bank gets given more fuel. If you put one sensor on each bank now only two cylinders monitored by that sensor will be getting fuel added but the other two cylinders on that bank will go to the other sensor which will show rich on that bank so it will be leaned off which will prob result in one bank ending up full lean and one full rich if you see want I mean. I wonder if you can just pick up two cylinders on each bank and (obviously by putting the sensors at a two into one junction on each bank) they aren't measuring volume of gasses or unburned fuel but oxygen left in the exhaust gasses so maybe it could work. Hopefully one of the clued up tuners on here will tell us if that will fly or not. Good luck.
Just had another look at the picture I posted and even two from one bank doesn't look possible. I,m stumped might have to go to go to an after market ecu with single lambda sensor.
Should look awesome, something I would like to try myself. Would be interesting to see what hp gains you can get. Have you had your car dyno,d recently? Re. The O2 sensors that's an interesting one. The system in not very sophisticated ei. A misfire on one cylinder will show up lean on that bank and so the whole bank gets given more fuel. If you put one sensor on each bank now only two cylinders monitored by that sensor will be getting fuel added but the other two cylinders on that bank will go to the other sensor which will show rich on that bank so it will be leaned off which will prob result in one bank ending up full lean and one full rich if you see want I mean. I wonder if you can just pick up two cylinders on each bank and (obviously by putting the sensors at a two into one junction on each bank) they aren't measuring volume of gasses or unburned fuel but oxygen left in the exhaust gasses so maybe it could work. Hopefully one of the clued up tuners on here will tell us if that will fly or not. Good luck.
Just had another look at the picture I posted and even two from one bank doesn't look possible. I,m stumped might have to go to go to an after market ecu with single lambda sensor.
Edited by NZ fan on Sunday 3rd August 09:49
I wouldnt be claiming i was building anything like that, mine looks well rough round the edges compared to that. That was kind of another question I was going to ask. Would it be better to put it in just where two cylinders join rather than where 2 from each side join. I haven't had it dynoed. Once I finish it I will post up some pictures but please bear in mind I am no pro welder im just a lad who bought a hobby welder and likes little projects
The sensors were over looked when designing and building if im honest. So now the choice is- one senor where cylinder 1,2,5,6 join and one sensor where cylinder 3,4,7,8 join.
Or one sensor where only cylinders 1,2 join and one sensor where only cylinders 5,6 join.
these are pretty much my only choices so please someone tell me the best option out of the two. Opinions quickly please guys
Or one sensor where only cylinders 1,2 join and one sensor where only cylinders 5,6 join.
these are pretty much my only choices so please someone tell me the best option out of the two. Opinions quickly please guys
rburnham11 said:
The sensors were over looked when designing and building if im honest. So now the choice is- one senor where cylinder 1,2,5,6 join and one sensor where cylinder 3,4,7,8 join.
Or one sensor where only cylinders 1,2 join and one sensor where only cylinders 5,6 join.
these are pretty much my only choices so please someone tell me the best option out of the two. Opinions quickly please guys
Nether option is ideal- I have a similar problem in so much that I have a balance tube between the two sides of the V8 before the lambdas- so a certain amount of the exhaust gas from the alternate bank affects the Lambdas- It still runs OK like this, although the fuel control is not as fine as it should be but it probably good enough just to keep the catalysts sweet, although you may end up with shunting if the fuelling on each side is significantly different and the ECU is trying to make big corrections to the wrong bank. A far better option is to forget the lambda's and catalysts all together, and switch to a non cat fuel map now we now know how to map these things- Or one sensor where only cylinders 1,2 join and one sensor where only cylinders 5,6 join.
these are pretty much my only choices so please someone tell me the best option out of the two. Opinions quickly please guys
So how exactly do these sensors work? Do they measure the amount of gas or just the burn mixture. Surely if its just the burn mixture then reading it off two cylinders from each side shouldn't affect to much? You're right not ideal and if I had thought about it a bit more I would have changed my design. I want to install a megasquirt system in the future but thats a long way off yet.
Blitz what manifold are you running then?
Blitz what manifold are you running then?
Each side of the V8 has 4 injectors that are fired together at the same time with a single signal from the ECU- so you can only alter the overall mixture for each side of the V8 depending on how long the 4 injectors are fired for- so you have one lambda probe per side of the V8, and per 4 injectors on each bank to control one bank at a time. So lets say you use a mixture of the left and right banks exhaust for your lambda signal half the time it will read the correct bank, and half the time it will read the wrong bank, so half the time it will be applying the wrong fuel trim level to keep the mixture correct. The V8 does not fuel quite evenly between the left and right side, so the fuel trim values will be different even in a perfect world, but now you have a right old mix up for the ECU to deal with. It will do this, but the range of fuel correction applied will be wider as its getting it wrong half the time. I think you will get away with it MOT and emissions wise, but wide fuelling corrections can lead to shunting at low speed. Hence its best to junk the lambdas completely, junk the catalyst , run the correct non catalyst fuel map and find a TVR friendly MOT man. The engines are much smoother without lambda control anyway.
As for how it all works- the probes simply measure unused oxygen in the exhaust- lots of oxygen means a lean mixture as the fuel is burnt before all the oxygen is used, and a rich mixture leaves very little oxygen. The whole system is designed to cycle around a mid point of 14.7:1 AFR, but it cycles both too rich and too lean by design due to the complex way a catalysts reduces 3 main pollutants by running the exhaust mixture oxygen rich and oxygen lean to get different chemical reactions to work inside the cat'.
My manifolds are a combination of Range Rover cast headers (due to the tiny engine bay in the Ginetta, tubular ones wont fit), plus my own designed down tubes and side pipes. There's an intricate balance pipe that joins the two sides that runs under the sump. I added this post adding lambda probes- hence the issue with the probes getting the wrong banks mixture now- but Im running the non cat map these days anyway.
As for how it all works- the probes simply measure unused oxygen in the exhaust- lots of oxygen means a lean mixture as the fuel is burnt before all the oxygen is used, and a rich mixture leaves very little oxygen. The whole system is designed to cycle around a mid point of 14.7:1 AFR, but it cycles both too rich and too lean by design due to the complex way a catalysts reduces 3 main pollutants by running the exhaust mixture oxygen rich and oxygen lean to get different chemical reactions to work inside the cat'.
My manifolds are a combination of Range Rover cast headers (due to the tiny engine bay in the Ginetta, tubular ones wont fit), plus my own designed down tubes and side pipes. There's an intricate balance pipe that joins the two sides that runs under the sump. I added this post adding lambda probes- hence the issue with the probes getting the wrong banks mixture now- but Im running the non cat map these days anyway.
Edited by blitzracing on Sunday 3rd August 17:50
If I'm thinking clearly about this, there is nothing to prevent you from rewiring the harness to fire two cylinders from each bank. There is no predictable timing relationship between an individual cylinder (piston and valve position) and the firing of the injector. Splicing in some extra wire so that four injector connectors can reach over to the other side should not be difficult.
That is just brilliant, its far better than a mixed exhaust set up. I think the only possible small gremlin is you get slightly different mixtures between left and right bank with the same injector period, probably down to plenum design(?). Id suspect this is the reason for the use of 2 lambda probes in the first place- I dont have any data to back this up, but I remember it being an issue when I had my car mapped many years ago on the non cat set up, but I doubt its enough to worry about.
First, I think you're using non-standard cylinder numbering which could cause some confusion. Normally it is referred to as odd bank (1, 3, 5, 7) and even bank (2, 4, 6, 8). But I think I understand what you are doing. The front half of the engine is plumbed together and the rear half is plumbed together. Is this correct?
Using standard cylinder numbering, the injector connectors for the odd bank should now control 1, 2, 3, 4 (the front half) and those for the even bank should now control 5, 6, 7, 8 (the rear half). Just make sure that the cylinders feeding into the odd side Lambda sensor are all being controlled by the odd side of the wiring harness.
Using standard cylinder numbering, the injector connectors for the odd bank should now control 1, 2, 3, 4 (the front half) and those for the even bank should now control 5, 6, 7, 8 (the rear half). Just make sure that the cylinders feeding into the odd side Lambda sensor are all being controlled by the odd side of the wiring harness.
Man this has given me much more of a headache than I thought it would! Im thinking I might put them in where the 2 pipes join not the 4 and hope for the best and correct later as rewiring sounds like more work than I planned. Would it pass emissions like this? Im interested in the non cat map though blitz
Im not sure your how your manifold is working- what cylinders connect to what down pipes? - a diagram of the V8 would be good to prevent confusion. Rewiring would not be bad- you only have to extend the injector wiring far enough to get all the same injector pulses in the same down tube. Currently you have 1,3,5,7 (left bank) feeding one and 2,4,6,8 (Right bank) feeding the other. Id put the probes where 4 cylinders join, in preference to an exhaust pulse from just one or two cylinders- this will be too pulsed, especially at low RPM to get a sensible reading. You can certainly try running the car on the non cat map you have already- it wont kill it, but I think there is a tendency to run lean at 3500 rpm plus, so you dont want to hold it at those rev's mile after mile, until you have had it on the rollers or at least measured the the lambda outputs under load. Im sure by the time you have done this there will be some more maps around anyway. Steve Sprint on the Griff forum has a map for a 4.6 hes worked on thats down loadable that may be a good point to start. I really dont know about the MOT, but the lambdas will keep cycling as best the ECU can manage, and that be good enough for the cats to work.
Edited by blitzracing on Monday 4th August 18:31
Im not following the rewiring idea. Surely changing the injector wiring will be changing the way the engine actually runs. I dont fully understand the concept. All I can imagine is the ecu telling the injector in cylinder 1 to spray but due to rewiring it actually sprays cylinder 6 or something. Like I say I dont think I followed it probably. And by rewiring do we mean just swap the connector leads to that plug on to the injectors.
You are making things over complex- the injectors simply fire once per four stoke cycle. It does not matter where the piston is, or what part of the cycle you are on, as the sprayed petrol simply hangs around the in the inlet manifold until its drawn in as the inlet valve opens. So we have two lots or banks of 4 injectors firing once per 4 stroke cycle, and you have just two circuits feeding 4 injectors each- All you have to do is match the lots of 4 injectors to each lambda probe so they have a the same injector circuit whatever side of the V8 they are on, it does not matter.
Sorry guys but I still haven't sorted this yet because of buying a house so obviously this had to be put on hold for a while. I hope to do more at the weekend, anyway I had another thought..... I will go for the idea of wiring the injectors differently unless somebody thinks the following idea is a better option- would placing both sensors in the downpipe after all cylinders have joined? I personally cant see why this wouldn't be a good idea only thing is that the ecu would make adjustments to all cylinders instead of left and right. So you would lose discrimination between the banks but is that necessary? I dont know.
Input please gents
Input please gents
Gassing Station | Chimaera | Top of Page | What's New | My Stuff