Historians might look back on this time and call it the time when I finally flipped but I’ve decided to have a rant about what I call cargo cult tuning.
What is a cargo cult?
I’m not a historian, but my understanding is that during World War 2, both the Japanese and the Americans set up bases on Melanesian islands so that they could attack the other with limited range aircraft. To get the co-operation of the locals, they gave a lot of goods from their developed lands to the relatively undeveloped islanders. When the war was over, and the Americans and Japanese abandoned the bases, the goods for the locals stopped being airdropped.
The locals wanted these goodies to continue of course, but they didn’t understand the back-end infrastructure that goes into airdropping supplies to foreign bases. All they saw was the communications guys talking on the radios, and then the planes landed and brought supplies. So some people started to make “radios” and “control towers” out of what they had around, such as bamboo, coconuts and so on.
Richard Feynman, the famous physicist from the Manhattan project, popularised the phrase “cargo cult science”, where people dress up as scientists and do sciency things that they see scientists doing. But they don’t publish their results for peer review; they don’t encourage others to replicate their work and so on. Classical examples would be the Cold Fusion scandal from the late 1980s, and more modern examples would be the scam diets, fitness fads, fake cancer cures and so on that you see being pushed to vulnerable people.
I want to talk about cargo cult tuning, because I seem to see a lot of it. People not understanding what’s going on, or the theory at all, but going through the motions and to someone that doesn’t know, from the outside it looks as though they’re tuning. We know that engines generally need more ignition advance as RPM increase, and less as load increases, so if there’s a large flat area on the timing map, or a big step in the timing map, then that’s probably not what the engine really wants.
But it leads to memes like this one being posted on Facebook.
Where to begin… Firstly, it’s so self-righteous. As though there are two types of tuners, good and bad, and the creator of the meme must be a good one and isn’t he special. I could suggest that you try driving a car which has an 11 degree timing advance over a 400 RPM range from 1600 to 2000 RPM and see how smooth it is, but that’s not the point. The point is that the suggestion is that because the map is kind of smoothish, it must be good, unlike the steppy map from the bad tuner. But it doesn’t take any skills to interpolate a map from a few data points; and doing so just because other maps look smooth is a bit like talking into a coconut radio and expecting it to talk back.
I’ve also seen a lot of misinformation being posted. For example I’ve seen articles say that richer than stoich, at 12.5:1 produces maximum torque for a given amount of air, and that’s why engines are tuned to run rich on boost. Well, it may be true that 12.5:1 produces maximum torque but it’s not the only reason engines are tuned for richer than stoich on boost, and it’s not even the main reason.
I’ve also seen a lot of copying without understanding, and justifications like “my mate is running the same configuration”. For a start, your mate doesn’t have the same configuration. There are always differences. Secondly, even if your mate is running the same configuration, it doesn’t mean it’s right. Unless you’ve logged the fuel pressure on your mate’s car, you don’t know that the system is actually working. The pressure could be dropping off, and the cargo cult tuner just compensated for this in the fuel map. Or he didn’t, and it just runs lean in the top end. If you don’t have the data then it’s just a feel-pinion.
Now, even if your mate’s system does maintain fuel pressure and correct AFR, that doesn’t tell you how much margin it has. It could be just on the limit of being adequate. We can calculate the fuel requirements of an engine; if you saw my fuel model talk from PRI then you know how. From that you can work out what injector size you need, and how much fuel supply you need from the pump system. There are performance graphs available for pumps, so you can predict the flow rate at different pressures and supply voltages. This only requires high school level mathematics. You can use that to ensure that you have adequate margin, eg 20% more fuel supply than you predict you’ll need. But of course the engineer would then measure the physical system to verify that it performs as it was designed.
Another factor that makes me say always go back to basics if you’re confused is that minor differences may have unexpected or large impacts. As you may know, the Modular ECUs have a very sophisticated fuel model, including air mass calculation, fuel film size estimation and dynamic injector modelling where the flow rate and offset used in the model change live with changes in fuel pressure. I recently tuned a GTR on stock injectors, and then upgraded to ID1300s. I changed the injector settings in the ECU and the map was very close. However the transient corrections were way off. I was trying to work out why; without a good fuel model you’d just have accepted that you’d need to redo the transient conditions but I wanted to understand the physics of why would the transient correction need to change when injectors are the only thing that’s been changed. The reason was that the spray pattern on the new injectors was completely different from the factory injectors; as a result more fuel ended up on the runner walls and less went straight into the engine, and more correction was needed. I didn’t expect that to make such a difference but I’ve verified this on multiple cars now, and the only way I was able to correctly diagnose was by going back to basics.
Another problem with cargo cult tuning is fixing the problem in the wrong way because you don’t understand what’s happening. An example is a stall on return to idle. If the ECU is going into closed loop too early, as the engine speed is coming back down to idle, then it will start to bring down the duty cycle because the engine speed is too high. If you have a high enough integral gain this can be enough to cause the engine to dip or perhaps even stall. On our ECUs we have a minimum idle duty cycle which can be set, so that on non-monotonic idle valves (for example where 0% duty cycle is the same amount of air as at 50% duty cycle) you can keep them within a monotonic range. The tuner asked if he could just use that to fix the problem. I said that you couldn’t, because it doesn’t solve the problem of going into closed loop too early; it might help with the stall as he’s testing, but then when he tries again under different conditions (eg the air conditioner is on), the minimum value will need to be different. Really he’s applying a band-aid because he didn’t want to diagnose the problem correctly (or ignored my advice).
We also see this with diagnosing misfires; we can see a horribly bumpy MAP signal causing the calculated injector value to jump all over the place, but the customer still upgrades to CDI to fix a misfire on a car at 8 PSI of boost; I assume because he’s seen other cars running CDI and doesn’t understand that 8 PSI of boost is less than factory boost and it’s well within the capabilities of the factory ignition system.
If you’re watching this video or reading this article then I know you want to improve yourself and learn, so I trust you won’t fall into this method of thinking. The key things are: