The test of our coreXY continue, now with the final motherboard and the DRV8825 stepper drivers. We start make simple test with manual control and pronterface to see if the mechanics is reliable and I’m satisfied. Using the prusa calculator we are trying different acceleration and top speed parameters. In the video the settings are 300 mm/s top speed and 9000 mm/s2 acceleration
http://www.youtube.com/watch?v=BDOjbzbCIzw&feature=share
Can it do the same speeds diagonally as well? Correct me if i’m wrong, but the way i remember it is that the firmware accelerates and limits the top speed for each of the cartesian axes individually - so when you’re moving diagonally with 300mm/s on each axis, the head would actually move at around 420mm/s, driven by a single motor on a coreXY setup.
I think you are right because the belt moves at 300 but the same speed is achieved for both axis, so the real speed is the diagonal resulting vector. I’m using the Marlin firmware and I’m studying how this is implemented, our previous printer design (a cumbersome one made with CNC linear guides) is Cartesian so the coreXY is new also to us, but I’ll keep you informed about our testing. Next step is to load some testing g-code and see the movement in all direction. I have to understand well if all the endstops and motor enabling works correctly. I read some commented line in the firmware that say that max endstops are not managed in coreXY kinematics, but it is not clear why to me. Anyway I was a little bit worried initially to use this type of kinematic, but now I’m very satisfied, I think I have also margins for faster movement.
While i’m still trying to wrap my head around its kinematics, i think coreXY is a sound concept that could drastically simplify Ultimaker-like setups - once the kinks are worked out, that is.
I wonder if with stronger belts you could build a CNC platform using coreXY.
Before start our printer we evaluate some different kinematics and I think that the ultimaker one is very clever and efficient, but there are a lot of parts moving and rotating. This H-Belt (one of the possible coreXY) it is more simple and keep all the heavy parts on the frame… We decided to use this after viewing an impressive video of Frsto pick and place running at 5 m/s…
@Nick_Parker I agree with you, the belt I’m using here are gt2 with 2mm pitch, only 6mm width, but they are very strong. On the old machine I use 12mm T2.5mm belts and they are awesome.
@Nick_Parker i’d say that depends on what kind of CNC application you have in mind (i mean, printing is CNC as well 
I’m using 15mm HTD-3M belts on my printer and i’m still holding back from mounting the Proxxon rotary tool that i had planned on using with it because there’s just too much flex in the mechanics. What also becomes a great factor in millability is the design of the printed parts - when milling, they will be loaded in completely different ways compared to printing with them. And the rods, which are self-supporting and will bend a lot when applying force to them.
It might work for tools up to maybe a Dremel if the rest of the mechanics hold up, but anything larger requires spindles to work properly in my opinion.
I agree, I’m not a mechanic, but also a simple manual drill press or manual mill have a very solid body construction to avoid bending and play. Anyway, in future, I’m curious to mount my dremel flex shaft on the carriage and try to mill some soft material, balsa wood or make PCB milling and drilling 
Just watched the video (Youtube is blocked at work), and there it is again - that high-pitched whine that made me throw out my DRV8825 drivers. Whosa whatsis suggested that using a higher voltage supply would help, but I haven’t gotten around to try it yet.
Do you guys not hear that noise?
Yes I hear the hissing sound but it become louder if I increase the current with the potentiometer too much, so I reduce the potentiometer until the sound was lower. I think the noise was normal considering that the motor are always engaged. I prefer to disengage motor when not used, but the coreXY need this to work because the two motors work on the same belt, so when one push or pull, in some cases the other have to be stand firm. My experience came from a Chinese 3Axis tb6560 that is quite terrible and cause me a lot of headaches, so these drivers seems like heaven to me the same for the voltage, initially, the old printer uses 48V to run with a lot of motor heat problem, a dual supply for arduino and cartridge, so we prefer to switch all to 12V. Anyway do you suggest a better stepper driver? What do you use in lieu of DVR8825?
I went back to the the standard Pololu-style A4983 drivers (e.g. Geeetech’s $2,99 clones). They heat up a lot more at the same output current and overall make more noise, but at least not with that annoying high-pitched frequency which drives me crazy.
I glued heatsinks to them (using Arctiv Silver) and have a fan blowing over them - this way i can crank them up until the motors get uncomfortably warm. That being said, the DRV8825s still manage to deliver a lot more current with no cooling - enough that my x-end warped under the motor’s heat.
Thank you thomas for the advice! We need to test a lot more the machine, but my intention is at least to print a fan duct and send fresh air to the drivers so I can take advantage of the higher current.
@Brigham_Valdez , we use Marlin that has support for corexy (that works also for h-bot). I think that if your machine is a pure H-bot (single belt without belt crossing) you can have racking problems. Means that if you have play on the bearings, or structure not rigid enough when you combine and change direction in X and Y the carriage is pulled with different direction by the belt and can become out of 90° between x and y causing circle to be like almonds or elliptical. This is from our experience. anyway our machine makes good circles, we try to reduce as possible the racking and in the next version we try to improve this more.