Your entire printer will need to be custom.Quote
georgi
Thank you for the elaborate answer. I find the direct drive delta the most fascinating design, that's the reason I was asking.
My intention is to build and program a controller and motor drivers myself anyway (I got attracted to 3d printers while looking for some embedded programming project to do). If it starts to look that I'm successfull with these I think I'm going to try building it.
Deltas like that cost a fortune for a reason. You need high speed, high precision, industrial motors and a system to run them (along with custom programming). This is all even more important if you expect a lot of speed. High speed alone can overload an Arduino board buffer, and stepper drivers are nowhere near strong enough for the motors you will need. Which means a full cnc style system ($$$$). I would expect to have 5k into it before you even see your motors move.
What many also forget, is the filament.
You will need to design and build an elaborate extruder system and possibly a nozzle. The pressures at high speed can be extreme, and while common stepper could possibly keep up, the hobbed pulley probably won't and will simply strip the filament. A larger nozzle diameter will help, but not solve it. You will also need a way to cool the previous layer fast enough, before the next layer is layed on top of it. This becomes a very delicate balancing act because higher speeds, often require higher nozzle temps to facilitate faster flow, but you need to cool the plastic faster and dissipate even more heat, and it goes up exponentially.
Once you cross 150mm per second on anything, it starts getting tough, and that speed is within the boundaries of many well built printers. By the time you crest 200, you start having severe issues with extruders skipping, low end pneumatic fittings can blow out, and vibration becomes an issue. On an Arduino, you also start hitting the buffer limits if you don't tune it. By 300, if your machine isn't perfect, forget it. In the last year, in a quest for better, faster printers, I have built 3 printers, used a dozen extruder designs, 4 extruder motors, 4 hobbs, 3 heads, destroyed half a dozen bowden fittings and frayed a set of belts. Most of that before I even broke 250mm per second. I can hit 300, but not very reliably, and from what I have seen on Youtube and in person, pretty much no one is even getting that. Speeds are highly over-estimated.
I would recomend you build a normal Delta first, you will learn a lot regarding the how speed effects things for only a small fraction of what the machine you want will cost. It will save you a lot of frustration and if you do decide to press on, you can sell that printer for what you have in it (if it runs, a working printer is worth more than the sum of it's parts).