Ideal print bed: Glass or Aluminium?
Glass is always level, easy to clean, easy to work with.
Aluminium allows for the addition of automatic bed leveling with an inductive sensor and distributes heat a little more evenly.
When printing mostly ABS and PLA, which one is better?
Interesting... Do you have a link handy to the exact kind of foil tape you are referencing?
I can't recommend an exact one, but anything that's conductive should work. I got the recommendation from here. Some people just use aluminum foil. Most likely, it would go on the corners or under the bed so adhesion shouldn't matter. Alternatively, you can use a contact sensor, such as the BLTouch, which is material independent.
The guys who make the DeltaPrintr have found a pretty neat solution to auto calibration where they put pressure sensors at each foot of the bed. Although not as straightforward to use with "normal" printer firmware, it means that you no longer need any specific mechanism on the end effector itself. http://deltaprintr.com/shop/3d-printers/deltaprintr-kit/
How to setup up auto bed tramming leveling tilt compensation using inductive sensor and aluminium bed: https://youtu.be/EcGFLwj0pnA?t=106
Al is a material commonly used in heatsinks, so it's NOT going to distribute heat more evenly, it will cool down rapidly as one moves away from the heat source. Glass is very hesitant to thermal changes, so while it takes longer to warm up, it takes longer to cool down as well, which means the extremities will be closer in temp to the heat source once a thermal equilibrium is achieved.
I believe printing directly on aluminium is unwise, simply because it will expand when heated, typically giving the bed a concave or convex shape. Glass, on the other hand, does not (at least not significantly).
As pointed out in the comments below, the heat expansion of aluminium could potentially be mitigated by increasing the thickness of the bed, as well as heating it evenly. Also, a common solution is to place a glass plate on top of an aluminium bed, at the cost of a slightly longer heat-up time.
In my experience, printing directly on heated glass can be very practical and give a nice surface finish for some PLA variants and other materials that support it. I don't know if printing directly on aluminium can give similar benefits.
It seems that this can be mostly mitigated by getting quality aluminium, 3+mm thickness, and ensuring that the entire plate is heated evenly. Thoughts?
On the other hand, many 3d printers have an aluminum bed then a glass on top of the aluminum.
hairsrpay or abs slurry is the way to go with bare glass and ABS. Or PET tape or Kapton tape on it. The latters also work with bare aluminum but not sure about the former.
Also some say with glass you have less uniform heat distribution. I don't think that's a problem myself and maybe it has to do with not having a ful size or uniform heating element in the first place.
I print directly on my heated aluminium bed with glue stick on it. It works well for PLA (not tried with ABS). The bed shape don't change (3 mm thick but it's a Delta with a small diameter). I've got a nice surface finish.
Basically having an homogeneous temperature on a aluminium bed is impossible, or possible only at ONE temperature: the sides are losing heat more than the center, it's inevitable. You can have a non-homogeneous heating below it, tuned to heat more on the sides, but it would be optimal only for a specific temperature, the one for which you designed the heating path below. One compromise would be to heat only the circumference, maybe a sort of ring 3mm wide, so that the center is always slightly colder and it is pulled apart/stretched by the sides. At least it will not bend up/down.