The Smartrap Mini was his entry level design before he started the CoreXY machines. Its all listed on the Smartrap mini page of the RepRap.org website and also on Thingiverse.
The manual lists the print size from 150 x 150 x 150 to around 250 x 250 x 200. The first prototype is 200 x 200 x 150. When I built this it was closer to 160x160x125ish, but would all depend on your rod lengths.
I've been experimenting with the 300x300x360 CoreXY based on the Smartrapcore ALU design, and it is possible to add 100mm to all the alumiunum dimensions and rod dimensions, but the cantilever bed design doesn't work well at all. Also, I've noticed going 300x300 with the XY, 8mm rods really aren't that strong enough for that distance without flexing meaning to go 300x300 it would be better to redesign it with 10mm or even 12mm rods. I also ran into the issue of finding a 300x300 heatbed at the time and discovered aluminum plate with heat pad was the only way at the time to get to 300x300, so I settled on the 300x200 varipower heatbed from E3D.
I believe he did a Smartrapcore XL300 design that was intended to go 300x300, but not sure how well that actually worked out. I know the user WZ9V had purchased it and had some issues with the kit being incomplete or mismatched, but not sure if he ever got it working properly or as intended without modifications.
I got it mostly assembled but eventually gave up on it. I feel it needs screws instead of belts on Z and like you said better guide rods. I decided to go D-Bot instead.
The plastic corners of the D-Bot put me off on building it. I don't see how it would every really be solid unless you go metal on metal on the corners, and even with the Smartrapcore Alu that I have, I've seen the parts that hold the X axis crack near the linear bearings so I can only imagine what potential issues you get with the D-Bot plastic corners.
You can replace quite a few of the structural D-Bot parts with aluminum plates and corners. I do like the Smartrap drilled holes approach because it's simple and strong. If my D-Bot corners cause me issues I may do a hybrid of the two approaches but I'd probably do a self designed Delta before I attempt that.
I think I'm beyond a delta. I've done a ton of research about one that I've convinced myself they're neat to watch, but not worth the time and money building one. Also, as far as I understand, if someone can correct me if I'm wrong, the further your print is from the center of the bed the less accurate it becomes due to software rounding issues.
They are neat to watch, but have too many issue to make them what I'd call reliable or repeatable and with effector weight being critical your almost exclusively looking at being Bowden setup which I've already experienced isn't as accurate as a direct drive due to the flexing that happens inside the PTFE tube. That being said, I do believe if your goal is to print a lot of really tall objects a delta might be better that the Cartesian or CoreXY machines but to me that's just a niche.
300mm in x on smartrapcore works pretty well with a linear rail, no sag in the middle. Make it two linear rails at 90 degrees to each other and you could go pretty big. I imagine it would work just fine being extended in y as the rail is fully supported so in theory it could be any length as long as the 2020 extrusion could cope.
As for deltas, having built a no expense spared delta, which prints really well, it's definitely not the way to go unless you are willing to spend money and obsess about the precision to the point where you do solve the problems and its perfectly calibrated. True the edges of the bed might be less useful than the middle. But I'd say with small pulleys, 0.9 deg motors whilst resolution is lower due to carriages having to move more when the effector is at the edges than in the middle, you can print fairly close to the edge. They do print tall objects well. You can use flying extruders for modest size deltas or cable driven extruders on larger machines to achieve direct drive extrusion.
It's hard not to conclude that if you take out the desire for high speed, the desire for a flashy printer, the ability to print massive things which hardly anyone ever does, you're left with the conclusion that the best printer for accuracy is a good old cartesian with direct extruder, and given how many i3 mk2 prusa are selling I'd say the market agrees.