Terrific and detailed responses here, thanks!
If the delivery is non-real-time text, you will probably want a multiband ARQ system.
Do-able, but, the free software available now is really challenging. Are you a programmer of any type?
I can tell you this is in the top-three problems you are going to encounter.
Next is ERP. Your tak-tenna and similar designs are SO far below a dipole it's not funny.
Anecdotal reports of "gee I made contacts galore on this thing" have no scientific basis.
You need ERP to overcome path loss, matter of physics, and the dynamic range of your typical desired path is on the order of 80 dB.
As N6BT says: "everything works!" You need to read this article:
http://www.ok1rr.com/index.php/antennas/94-everything-worksI am thinking of a good way to phrase this next part, I know you have some HF background but your comments are showing an exuberance / expectation set that may not be reflected in a reduction to practice ... my goal is to help you not make newbie mistakes that will cost you $$ ... I haven't come up with one yet, so, bluntly, here's what you should do first...
Get a conventional used 100W rig for $400 and put up two antennas, a 40 meter dipole flat-top at 30 feet oriented east-west, and, one other compromise antenna like the tak-tenna or similar you suggested. Get a switch to flip between the two, and then spend a couple days a week for 6 weeks (preferably over a change in seasons) sitting on the SouthCARS and/or eCARS nets.
Just that and only that. Doing so will give you the hands-on, on-air experience to determine what really will get the job done versus what might be aesthetically acceptable but not functional.
The bands you're interested in will primarily be 60-40-30. The antenna you need for this will be a remotely-auto-tuned doublet fed with open wire line, but, I will let you arrive at this conclusion by yourself.
Back to the radio. Most software now available doesn't have the needed rig control integrated with the AxSK mode control on the PB, so, the fiddling you have to do with com port sharing and virtual audio channels is a huge pain in the neck, not to mention the gross lack of robustness. That's one big reason the Peaberry won't really catch on, besides the fact it's no longer available anyway. Until the software out there gets real, it's a big problem. Also, unless you are a real builder, the PB is not a real choice, even if it
was available. Since a built and tested unit sells for $250 or so (I know, since I just sold two at that price), you can do way better than that for not much more.
The Elad solves part of the software problem by exposing each radio component (control plane and signal plane) separately. It also is 100% digital, which is good for what you're doing, and has a far more robust RF architecture than the PB. And then you need an amplifier (all this ain't working well unless you can crank up to 50W or so). Yes, I know, the cost is prohibitive. Instead, you can also do what you want with some old IC735 or IC728 or similar at a fraction of the Elad price, using an embedded Linux/windows controller and some programming on your part.
Steve KZ1X