Yup, the throwbar is the weak point in all N-scale turnouts, but it's not just figuring out how to increase the surface area of the solder joint that will solve the problem.
As some of you have noticed, there is less tension on the solder joint at the throwbar using Code40 rails than using Code55 rails, which are quite a bit stiffer than the smaller rail. This is because Code40 rail is much more flexible than Code55 rail, and bends with much less effort when the switch is thrown.
When using the no-hinges one-piece-closure-rail protocol for your closure/point rails, the stiffer the rails, the more torque is applied to the throwbar. This can be alleviated somewhat by leaving the combination closure/point rails unsoldered for a long distance on the throwbar side of the rail gaps at the frog, but then gauge and radius problems can occur.
Ideally, the point rails should be hinged at the point rail heels with a sliding hinge that allows lateral movement as well as radial movement...like using a shortened rail joiner like an ME #6 turnout. This gets rid of the torque problem, but introduces a potential electrical problem because the sliding hinge isn't reliable all the time to carry electrical current because of oxidation, dirt, paint, ballast glue, cat hairs etc. A fairly easy way to make sure of electrical continuity at the hinge is to solder a fine, flexible wire across the hinge from the toes of the closure rails to the heels of the point rails. This should be done so that the wire isn't tight, but has slack in it so the point rails can slide back and forth in the hinge.
This will usually solve the torquing of the solder joints at the throwbar.
I prefer to use Proto87Stores' etched "heel block hinges" instead of shortened rail joiners because of their appearance and utility.
Photo (1) - Proto87Stores "Heel Block Hinges" on one of my turnouts... 
From an appearance aspect, since if we are building our own turnouts we are building models of real turnouts, most prototype turnouts have point rails that are standard lengths, and are hinged. This is pretty obvious when looking at real turnouts from a human eyeball height because of the foreshortening. You can see the "kink" at the hinge which connects the closure rail and point rail very easily, but, if you build your turnouts with single-piece combination closure/point rails with no hinge, this "kink" is missing, and IMO it detracts from a realistic model turnout appearance.
Photo (2) - Prototype "kink" at the point rail heel hinge... 
One way to get the benefits of a monolithic closure/point rail from an electrical/alignment aspect but still have a hinge, is to file a "V" on either side of the long closure/point rail where prototypically the hinge would be. I file mine using a triangular jeweler's file, filing away the rail foot and rail head at the tips of the "V" on both sides of the rail, just kissing the rail web. This makes an excellent, strong hinge which will always be aligned and always carry electrical current.
When using the "notch" hinges on my turnouts, I solder in one or two PCB ties at the closure rail toes, just before the notch hinges to keep the closure rail properly aligned with their respective stock rails.
Photo (3) - A photo of "Notch" hinges on one of my older turnouts...
It'd be great if all you had to do next was to solder a PCB throwbar to the point rail toes, BUT...the shortened point rails and the non-sliding hinges you've just created will drastically magnify the torque at the point rail toes' solder attachment joints at the throwbar and will chronically break, even using silver-bearing solder, which is 5 to 6 times stronger than plain ol' lead-tin solder....especially with Code55 rail.
The solution for Code55 point rails is to hinge them at the throwbar, which will get rid of that damn solder-joint-breaking-torque that non-sliding point rail heel hinges induce.
Here's how I do it...and get a somewhat prototypical look at the same time...
Photo (4) - Here's a drawing and instructions on how to fashion Code55 point rail toe hinges...
Photo (5) - Here's a photo of my point rail throwbar hinges in the flesh...
So, because of the thickness of the wire I use to attach point rail ties to the throwbar, this will make flange clearance tight for Code40 rails, so I use my point rail toe hinges at the throwbar only on my Code55 turnouts, and I use my "notch" hinges at the point rail heels instead of the Proto87Stores' heel block hinges (actually, I use them, but deactivate them by soldering one side of them to rail heels & toes which makes them take the place of a wire because I like their appearance, but a "notch" hinge will still work just fine). The hinges at the point rail toes at the throwbar eliminate any torque problems.
I also use Proto87Stores' Tri-Planed Point Rails, which eliminate filing away the stock rail feet on either side of the point rails for a prototype appearance. Unfortunately, this also eliminates having single closure/point rails in my turnouts but, making use of P87Stores' Heel Block Hinges is a good way to look good and retain electrical continuity at the same time if you "deactivate" them as I've described above.
For my Code40 turnouts, I use the Proto87Stores' heel block hinges, solder a fine wire between the rail heels & toes at the hinge, and reinforce the solder joints between the point rail toes and the throwbar with 0.003" brass shim stock plates bent to a very shallow "Z" shape. The sliding/rotating P87Stores' heel block hinges eliminate any torque problems on the solid solder joints at the throwbar.
Photo (6) - Here's a good photo showing P87Stores' heel block hinges and their tri-planed point rails and my first iteration of my hinged point rails at the throwbar. Notice the stock rails at the switch aren't filed away on the inside... 
I always use PCB ties for both headblocks on either side of the throwbar, and on either side of any gaps because reinforcement at these positions is necessary IMO. I probably use more than needed elsewhere, but I have a large stock of PCB tie material, so the cost is irrelevant for me, since I won't be running out.
Although this may seem complicated (and it is) and take more time than the Fast Tracks methods, my turnouts don't break any more, are smooth as silk, and look better IMO than my old, simpler ones. Following my methodology will virtually eliminate throwbar problems from normal use.
Cheerio!
Bob Gilmore
