The AM Forum

I saw this idea, used for 40m. I’d like to do this for 17m. I have a few questions:

1 How do you calculate the length of the delay line?
2 Is the delay line made from the same coax?
3 I assume the cable length from each dipole needs to be the same?
4 How do you calculate the spacing between the two dipoles?

50 ohm coax to each antenna should be 1/2 wavelength
phasing line should be 72 ohms 1/4 wavelength

spacing between dipoles typically could be 1/2 wavelength
you could set up a simple ratio using existing data to get into the ballpark for spacing:
33 ft/40M =Xft/17M
X=~ 14 feet

If you have trouble calculating dipole dimensions, use this:
https://www.omnicalculator.com/physics/dipole

The problem you might face is that the dipoles are balance antennas and you're feeding them with unbalanced lines so it's conceivable that you could get antenna currents flowing back over the coax shield. This might require tweaking of the dimensions and proximity to other objects.

COAX CABLE VELOCITY FACTOR: The velocity factor of a coax cable is important in applications where phasing of signals is important – phased antennas, matching stubs and many other applications need a knowledge of the velocity factor.

If you can use open wire line, there's a more versatile solution.  Put up two dipoles, spaced a quarter wave apart.  From each dipole, drop a quarter wave of real open wire line, not the brown window line stuff.  Parallel the lines, but first flip one over to get a 180 degree phase shift.  Then run open wire back to the station from this common point, to a tuner.  This gives a bidirectional pattern, figure eight, broadside to the dipoles as you'd expect.
If you model this array, you'll find that this 180 degree bidirectional configuration gives the lowest angle of radiation, lower than the unidirectional phased setups.  On 17 meters, low angles are where it's at.  I use one of these on 80 meters, with a Phillystran "boom" stretched between two pine trees.  It's the cat's meow for DX and the pattern holds up on 40 meters  with more  gain and even up to 30 meters.  Above that it goes into funny cloverleafy modes, but even those are usable if the lobe falls in a useful direction.  Built for 20 meters, with a 33 foot width,  this antenna would take you up through 10 meters in fine style, provided you are willing to change tuner settings between bands.  This is a modification of the old Kraus 8JK antenna.  Don't cheap out and use the crappy brown window line as it changes impedance when wet or snowy.  I was holed up on Cyprus years ago and spent a few happy evening modeling this antenna and it works just as predicted.  Mine's been up for 15 years now . If you can manage it, balanced line is the way to go, and with a tuner you get frequency agility.
73 de Norm W1ITT

In my situation open wire is not an option. Looks like this will be my next spring project.
Thanks for the suggestions.