@N0ZYC: Where do I find the mod you're talking about?
I saw this get bumped and so wanted to just retype up my reply here with a bit more cohearance and detail. Yes HDU's link to JM's page is exactly what I used. DON'T bother with the original coax-phasing design, it's an ENORMOUS headache and is so easy to solve with the other design.
It works VERY well. I've started with the doppler in the rdf bible (transmitter hunting, radio direction finding simplified) and I currently have a unit that has interchangeable antenna arrays, is tested to work on vhf, uhf, AND hf, and has a ton of smart mods. I just finished my portable array. By "portable" I mean like hand-held. (bicycle mobile, the huntmaster here has taken to hiding off bicycle paths to thwart my doppler on my suv! I love a good tech escalation!)
It uses 1/2" pvc for the most part and 1/8" brass rod for the radials. (dipole) This is the prototype but works well and is mostly waterproof. The next one will substitute tape measure for the radials, and is meant for ON FOOT doppler. (they aren't going to like me very well methinks...)
As for your antenna questions, TBH I have read a LOT of "armchair quarterback" on dopplers, both under the hood and on the antenna. Before I understood the theory well, I was trying to follow instructions that I later determined were arbitrary and a waste of time. So watch what you read. Try to understand what they're telling you. Insist on an explanation for WHY they think it should be done that way. If they start doing a lot of hand waving and wildly tossing around buzzwords, just nod, smile, and ignore. According to TWO sources I was trying to go by, doppler on HF should have been impossible. But my 10m doppler is a killer.
For your array, try to get the antennas as well tuned as possible, for best sensitivity. One mod I did that I recommend for others is (A) a way to "step" the counter when in the stopped position, and (B) an indicator for which antenna is selected. This is useful for both testing and tuning. Nothing more annoying than a grumpy doppler that turns out to be one antenna with a problem. That will likely pack you up for the day if you have no way to test specific antennas during the event.
So tuned dipoles work best. The pin diodes and resistor are likely to mess with your tuning. I really haven't spent any time optimizing that part of mine, but I bet the right amount of L and C would really improve mine, if I can find the time to play with it. Vertical whips work well too, the ground plane is NOT as important as some will tell you. Your vehicle will work well enough, the ground radials won't provide much benefit. A dipole however does make a noticeable improvement in sensitivity. They're just a much larger pain to construct. Probably not worth it for most people.
Spacing. That is the question. "just under a quarter wavelenth between each antenna" is what I go with. Less spacing will yield lower FM deviation on the perceived tone. If the deviation can't beat the switching noise, you won't get a good bearing. My 10m array is only 6ft on a side, and that's enough to work with for me. But what about MORE deviation? I really ought to do more testing on it, see if I can hunt 2m using my 10m array. But several have said this will produce overlaps on the deviation, and essentially make you chase shadows. All the "two dipole" system space their dipoles at 1/2 wavelength, and THAT is optimal. Someone will have to do the math, I'm not THAT far into it yet to confirm whether or not you can go up to 1/2 wavelength on a side before the signal starts to distort. Going past 16 LEDs to 32 LEDs is a waste of time, unless you've done an amazing job constructing it, your hardware doesn't have much more reslution than your 16 LEDs support, so why bother.
Buy your pin dodes from this dude: http://www.ebay.com/itm/280703143876
Antenna placement must be in the correct rotation, and in order. Specific placement does not matter. My array rotates clockwise.
any of those would work. If your display is turning in the wrong direcction when you turn, put your antennas counterclockwise instead.
you should be able to correct for any of the above using the calibrate dial and the +180 swich. There are two gotchas: 1: remember that some radios add 180 deg and need the +180 switch in the other position. write it down on your box so you remember which one to use in case you need to swtch radios. 2: calibration may change if you switch bands. At any rate, I always zero myself on the local repeater or something as part of my prep for a hunt, so neither of those issues really affect me.
There's also no reason to have them square. diamond works the same.
A little out of date and somewhat incomplete, but here's my goodies: http://cable.vftp.net/n0zyc/doppler/
you really should read the list of improvements I made.
And if you, or anyone else reading this thread for that matter, has a ddf1 or any other variant on the Roanoke or Montreal Doppler you're not using, contact me, I'd be interested in it.