Yup, there is no substitute for just building it and measuring.
To allow varying the magnet gap from 2.5 - 4mm when testing I'm thinking of elongating the holes slightly in the curved steel part. Then I should be able to swap the thickness of the plastic yellow edge part to change the gap.
I did some napkin math and the difference in hole placement of 2 vs 4 mm magnet gap is 0.5 mm on the outermost holes so I added 1 mm on all holes to be safe.
To allow varying the magnet gap from 2.5 - 4mm when testing I'm thinking of elongating the holes slightly in the curved steel part. Then I should be able to swap the thickness of the plastic yellow edge part to change the gap.
I did some napkin math and the difference in hole placement of 2 vs 4 mm magnet gap is 0.5 mm on the outermost holes so I added 1 mm on all holes to be safe.
I wanted to test if the cheap rolling mill (SJ 300) I bought is strong enough to roll my 2 mm thick plates so I made a test plate:
The real plates would have more holes so the idea is that if I can curve this plate, then I should also be able to curve the actual plates with more holes.
And as it turns out, it worked great!
It worked so well I'm thinking of redesigning the speaker slightly. The plan was to make the final speaker in 8 x 20 cm segments. But one problem is how to join the segments without creating a massive weak point.
And then I realized that I don't need to segment the driver. The segmentation was just a limitation of my 3d printer since it has a 25cm3 build volume but the main strength now is from the steel and I can order 160 cm tall laser cut parts no problem. A benefit of not segmenting is that I can also add an offset to the magnets such that the gaps don't all align.
For the next tests where I want to try varying the row holes and xmax I will still order 20 cm segments but then for the final driver I would then scale upp to 160 cm tall without segments.
Next step is to determine how big the lengthwise gap between the magnets should be. I'm leaning towards 0.25 - 0.5 mm. Smaller is better of course but 0.5 mm would be easier to glue and probably good enough.
The real plates would have more holes so the idea is that if I can curve this plate, then I should also be able to curve the actual plates with more holes.
And as it turns out, it worked great!
It worked so well I'm thinking of redesigning the speaker slightly. The plan was to make the final speaker in 8 x 20 cm segments. But one problem is how to join the segments without creating a massive weak point.
And then I realized that I don't need to segment the driver. The segmentation was just a limitation of my 3d printer since it has a 25cm3 build volume but the main strength now is from the steel and I can order 160 cm tall laser cut parts no problem. A benefit of not segmenting is that I can also add an offset to the magnets such that the gaps don't all align.
For the next tests where I want to try varying the row holes and xmax I will still order 20 cm segments but then for the final driver I would then scale upp to 160 cm tall without segments.
Next step is to determine how big the lengthwise gap between the magnets should be. I'm leaning towards 0.25 - 0.5 mm. Smaller is better of course but 0.5 mm would be easier to glue and probably good enough.