You would have to use computer modeling for the precise dimensions.
Using four 90 degree elbows should be OK. I would make the spacing 3 inches, unless that creates a problem with sweating the joints together. The ends of the folded dipole are most sensitive to being touched; they are high impedance points. Use insulators for support of the tubing.
I have found that using the formula 5540 divided by the mid channel frequency in MHz gives the length in inches that is close enough for a test folded dipole.
CH 2, 54-60 MHz, mid channel 57 MHz, 5540/57 = 97 inches
CH 3, 60-66 MHz, mid channel 63 MHz, 5540/63 = 88 inches
CH 4, 66-72 MHz, mid channel 69 MHz, 5540/69 = 80 inches
CH 5, 76-82 MHz, mid channel 79 MHz, 5540/79 = 70 inches
CH 6, 82-88 MHz, mid channel 85 MHz, 5540/85 = 65 inches
CH 6 low end: 5540/82 = 68 inches
CH 6 high end: 5540/88 = 63 inches
If you are concerned about wasting expensive copper tubing, I suggest that you make a test folded dipole 65 inches long using 14 gauge solid copper THHN insulated building wire like I did for a CH 3 test antenna.
If your CH 6 signal is still marginal with the test folded dipole, it might not be because of insufficient antenna gain; it could be that the antenna isn't in the best location or there could be interference from electrical noise which is worse on VHF-Low.