Set up Lambda target

newbee here I was looking too know how to properly set up your lambda target for when the engine goes into load (full power run). In road tuning how do u determine your target lambda when u can’t see how the numbers effect power or going too rich?

General guidelines are provided in many places on this website. But you can choose 0.88 - 0.90 as a staring lambda for naturally aspirated engines are you will be be safe and very close to optimum. Often you can't measure much of a difference anywhere in that range.

So Say if Im tuning a 1jz gte with single turbo and 1000id injectors. I would roughly try a lambda target And then once going through a full power run if I have a lambda set too 0.85 and see engine going too lean. then I would simply change the lambda target too a richer mixture? Would this this be best way to get performance out of road tuning? Thanks for your reply

If the engine is going too lean (ie, Lambda greater than your target of 0.85), you change the fuel table to make it hit the target.

After the fuel table is tuned so it can hit the target at all RPMs, then you try a test where the target is a bit richer (say 0.83) and see if it makes more power (or the same power, but the coolant temp doesn't rise as much). You might try a bit leaner (say 0.87) and check the results -- make sure you don't see any knock. You might repeat this test with the coolant temp at the maximum end of the "normal" range -- to make sure you don't have any indications of knock as knock is more common at higher combustion temperatures.

That is how you pick a target lambda that makes the best power and keeps the engine safe.

Cheers lad I appreciate it all makes sense to me now

David you stated that "General guidelines are provided in many places on this website.". While that may be true, More specific (than general) would be great. I have been unable to find guidance on how to formulate an accurate Target Lambda Table. Since pretty much all the tuning processes depend on it, it seems like having a proper Target Lambda Table is step #1, Unless I'm misunderstanding it, adjusting VE (AFR) is relative to target Lambda. If the Lambda table is off, it's garbage in / garbage out.

A common target lambda table for a normally aspirated race engine, is 0.90 Lambda at all RPMs and loads. That is the specific table that I would start with.

Thanks, but I think you and I have a different definition of the word "specific".

Why,? It's perfectly adequate Lambda Target, and will allow you to fully tune the VE / fuel table. The ECU uses the Lambda target, the mass of air (Manifold Pressure and Temperature), and the VE value to determine how much fuel to inject at a given RPM. You tune the VE table value until the result (Lambda/AFR as measured by a wideband sensor) matches the target.

Then you change the target to find out what the engine wants. You do not have garbage output, due to the initial target table input being garbage.

It's not possible to start with much of a better Lambda Target table unless you have experience with that exact engine setup. I just tuned an engine I've never seen starting with a table of Lambda .9, and good power (above the owners expectations). Spark plugs look great.

After the VE (fuel table) is tuned, you can change the Lambda Target and your VE table should not need to be changed -- the ECU will still hit the target.

This is how you would approach a standalone ECU, when starting from scratch. If you are flash tuning an existing ECU, then the OEM Lambda Target would be a great place to start. OEM engineers spend a lot of time finding the ideal targets for the combination of fuel economy, power and emissions.

You might find these webinars have what you are looking for:

Normally Aspirated

https://www.hpacademy.com/previous-webinars/065-planning-your-afr-targets-naturally-aspirated-engines/

Forced Induction:

https://www.hpacademy.com/previous-webinars/066-planning-your-afr-targets-forced-induction-engines/

David,

Thank you for talking the time to respond. Maybe I'm over analyzing it, but my frustration with this is that everything in a refined tune hinges on an accurate (optimal) Target Lamda Table, yet there is no practical way to achieve that for the novice (self) tuner.

I can rent Dyno time and spend hours doing steady state/Ramp Runs MBT timing, or VE tuning, but neither will truely be "optimal" if the Target Lambda is "less than optimal".

So lets say you determine VE at a certain point in a specific segment of the table at steady state. If you modify the Target Lambda, you throw that VE number (and optimal Ignition timing) off to some degree, so you need to run it again to establish the new corrected values. And you need to do that every time the Target Lambda is modified.

Moreover, how do you know when you have truely achieved optimal Target Lambda?

PLEASE ask Aundre to do a Webinar on "Finding Target Lambda in a Stand Alone ECU"

Preferably using Haltech NSP.

Thanks again,

Cheers!

What specific ECU are you tuning with the Haltech NSP software? Perhaps that's where the disconnect is. I'm not sure all the ECU models offer the same tuning capabilities.

Looking at the help for "Base Fuel Tuning", you find that only the 'Volumetric Efficiency" method would use the Target Lambda to calculate the amount of fuel injected. You tune the Base Fuel Tuning table so that the measured Lambda from your wideband matches the Target Lamba -- this is often the easiest, if you use the same Target Lambda or all cells. On cars with Catalytic Converts, you might chose to Target Lambda 1.0 (handy because the measured Lambda value is exactly the correction multiplier for the VE value to reach the Target Lambda).

With the 'Volumetric Efficiency' method, you can change the Target Lambda later (by using a dyno and/or a 5-gas analyzer to determine the best Lambda for maximum power or Emissions / Fuel Economy). YOU DO NOT NEED TO CHANGE THE VE TABLE if the modeled air mass is correct.

But I believe the other methods (specifically "injection time", and "fuel flow rate") would require you to re-tune the fuel table if you wish to reach different Lambda Targets.

Your assertions that VE tuning requires an optimal Target Lambda Table is just not true in my experience.

Haltech Nexus R3. I'm building a Base Map for an application that doesn't have an included Map. Datsun (Nissan) L28ET, 2800 cc, 6 cyl, Turbocharged.

So maybe I'm not understanding the connection between Timing, VE, and AFR.

For AFR, if I use a base Target Lambda of 0.9 across the board, the wideband (in open loop) is going to indicate how far off from the target the AFR is. If the VE table values are adjusted to match the target lambda table value, 0.9 is going to be very rich at idle and low load conditions, and probably too lean at high loads, especially under Boost. Sure, you can back the timing way off to compensate, but that leaves a LOT of performance on the table.

For Timing, if the AFR is too rich, it's wasting fuel (at best), and pushing knock if too lean, so Timing needs to be altered to compensate, resulting in sub-optimal MBT.

Either way, a proper Target Lambda Table is the key to an "Optimal" tune as I see it. Please explain how that's wrong.

I have watched every Webinar on the subject.

Please understand, I'm not trying to argue, I just need it to make sense.

As David has mentioned, 'optimal' lambda for power (and pump gasoline/most of its derivatives) is almost always .89-.9 lambda - this doesn't change with any engine configuration, it's chemistry at this point.

Generally everything to either side is a compromise -

Leaner AFRs at idle and cruise on a street car? Your lambda (and AFR) targets can - and should - be leaning pretty close to stoich unless the configuration is telling you otherwise by how it behaves. Thankfully this is a pretty safe sandbox and you can get away with just pulling fuel out until it feels 'fluffy'. ALSO, if you've used .89 lambda as your target, and your injector/flow/engine configuration is accurate Haltechs fuel model SHOULD be close enough to get you to the point where STFT and LTFT does almost all the work of dealing with any 'fudge' factor. If it doesn't, I'd investigate how you've configured your map.

On boost the compromise in the target lambda world is the cooling affect of enrichment allowing for more boost - this starts to be platform dependent depending on a myriad of things as well as this sandbox is a little bit less forgiving so bias towards being rich - there are plenty of good examples floating around but what I tend to start with is just taper down to .78 at 1 bar of boost and see how it behaves before trying to make more. Again this is NOT optimal for making power in and of itself - it's optimal for SAFETY which a dyno wont tell you.

Thank you for the input DG.

Isn't an AFR of 14.7-1 = to Lambda 1.0 (Stoich)? Is there any reason a Street/Street performance tuned, basically stock Engine shouldn't idle (low load) well at .98/1.0 at normal operating temp? .89/.9 seems a bit fat for Idle and low load ranges. I'm not going to be drifting or doing long max power pulls here, but I do want a well tuned and efficient Engine, that will stay together for a long time.

Here is what I have so far:

[img="blob:https://www.hpacademy.com/332b45da-d246-4f08-a1e4-8337f9d0f1ae" alt="" alt="" ]

One side note, I want to avoid any "Burble" or "Popcorn" on decel. That's for Race Cars, no purpose for it in a Street Car, plus, it's just annoying.

Yes, a basically street car should be at 1 lambda/stoich at idle and cruise - I very clearly stated that here: "Your lambda (and AFR) targets can - and should - be leaning pretty close to stoich unless the configuration is telling you otherwise by how it behaves."

I think what you're concerning yourself with/forgetting is that to tune a VE table you DO NOT NEED a refined lambda target table so long as its safe. Tuning for 13:1 at idle and cruise is not your be all end all however it's a VERY easy number to just punch in and tune to outside of boost whilst checking other things so long as its safe, read: atmosphere and below. Your job as the operator/tuner is to tell the ECU how much air the engine can process - the ECUs job is to then do the math to get to the lambda target with the injection data you've given it and I find (and it looks like David Ferguson does as well) that taking guess work out of the equation and finding the shape of the VE table at atmo and below is easiest/fastest/causes less headache at 13-ish:1.

Thank you David (both of you) for taking the time to respond. You have given me a lot to consider.

Cheers!