I'd like to think I've searched the subject exhaustively, but if there's an in-depth write-up that I haven't found please feel free to flog me (after you point me in the right direction).

What I've found on the subject is mostly geared towards setups with tube headers - I'm thinking it's safe to assume no-one running log manifolds will be concerned with tuning this, but all discussion is welcome. From what I've read the drill is: measure exit temp with IR sensor, add fuel to the hotter holes assuming they're burning lean but don't discount the possibility of nearly dead cylinders reading cold from weak combustion (need more fuel). Rinse and repeat until you brick your ecm.

A little background - my current (and probably only efi) project is a freshly rebuilt '95 LT1 with a fairly sloppy cam and headers. My understanding of the subject is that the individual cylinder trims are used to balance low-flow AFR to accommodate any cylinder to cylinder airflow differences.

So here are my observations, questions, discussion points, whatever.

1) Reading primary tube surface temps on headers in a car with the engine running seems a far cry from controlled experiment. Assuming you have a clean line of sight to some point on each tube, through experience I've observed reading different points at different distances from the flange or points around the primary bend produce vastly different results on the same cylinder.

2) eehack's cylinder balance test - disables one injector at a time and averages RPM and MAP changes to extrapolate relative cylinder strength. Obviously this is extremely platform specific, but can this be factored into the equation?

3) Is there any value in aiming for a summed balance between the banks - i.e. once a temperature balance is reached should one then strive to scale the right and left bank trim values so their sum is each roughly 4.00?

4) Short of fitting your heads / headers with individual EGT thermocouples, is there a better way? The IR thermometer seems a bit haphazard and my experiments with a single portable contact thermocouple have been problematic due to the amount of time needed to bring the sensor to equilibrium with the tube temperature.

I'm more than willing to find answers through trial and error, but if anyone would like to save my PCM a few dozen flash cycles your contributions will be greatly appreciated!