PiJohn & TurboTune
Hey, I've been puzzling over how the torque curve of a V6 can be described with a clean mathematical model—something like a cubic or quartic that captures the sweet spot and the climb. Have you ever tried fitting your timing or cam lift data to a curve and tweaking it to squeeze out that extra horsepower? I'd love to hear how you approach the shape of the torque map and what equations you think best represent the engine's performance.
You know the shape of a V6’s torque map, it’s a straight‑up “hill” that starts low, peaks around mid‑rpm, and tapers off near the redline. I usually grab a quick scan from the dyno, pull the rpm and torque data into a spreadsheet and run a polynomial fit. A cubic is the sweet spot most of the time – it gives that clear rise, a single peak, and a realistic drop. If the cam profile is super aggressive or you’re using a twin‑stage turbo, I’ll bump up to a quartic so the curve can handle a little bump or dip in the middle without getting all wonky.
The equation I lean on looks like:
T(r) = a·r³ + b·r² + c·r + d
where r is rpm and the coefficients a, b, c, d come straight from the fit. Once I’ve nailed the curve, I tweak timing, MAP, and boost to push that cubic’s apex higher. If the apex moves up, the engine’s breathing is cleaner – more air, more fuel, more power. I’ll throw in a little “lean” at low rpm to keep the torque smooth, and push the cam lift just enough to fill that mid‑range. It’s all about that clean, predictable climb in the torque map, not letting the engine stumble over a bump in the data. If you’ve got the numbers, I can help you pick the right degree, but keep in mind the real test is on the dyno, not just the spreadsheet.
Sounds solid – a cubic is usually the sweet spot for a single‑hump torque curve, and a quartic gives you that extra wiggle room for turbo spikes. One thing I always double‑check is the boundary conditions: make sure the derivative at the lowest rpm isn’t too steep, otherwise you’ll get a knee that feels artificial when you drive. Also, if you’re adjusting cam lift to push the apex, keep an eye on the cam’s lift profile – a sudden rise can throw off the fit and cause a small bump you won’t see until the dyno. Do you have the raw data handy? I can run a quick fit and see how the residuals look – sometimes a tiny fifth‑order term can clean up a nasty outlier without over‑fitting. Let me know what you get, and we can tweak the model together.
Yeah, I’ll drop the data into my rig and let the math do its thing, but I’ll still eyeball the curve. I love that you’re thinking about the low‑rpm slope—if it’s too steep it feels like a jerky dog that’s pulling too hard on the clutch. When I tweak cam lift, I make sure the lift curve is smooth, no sudden spikes, otherwise the torque map gets a nasty bump. If you can share the numbers, I’ll run a cubic first, check the residuals, and only add a fifth‑order term if the outlier is real and not just noise. Keep the fit simple, keep the engine happy, and let me know what you see.