I have a question for you all to discuss. Why is it that crossed gantry printers are not more common than they are? I realize that Ultimaker has made them, as well as a few DIY examples such as the Hypercube Overkill project( HyperCube OverKill Build Log) , but they really aren’t very common, even though, at least by first appearances, they have more or less the same advantages of the corexy configuration, plus one or two of their own.
While the crossed gantry setup has 3 more linear motion joints(additional gantry, plus the fact that the hotend attaches to both gantries, not one), it does away with the long complicated timing belt setup of the corexy, and all the extra idlers entailed. I realize that having two gantries means you have more total flying mass than a CoreXY setup, but the peak mass for a given axis should be no higher, which is what limits total speed and acceleration anyway.
I assume there must be a reason why crossed gantry(also known as CroXY) has never really taken off, but I can’t see it. It doesn’t seem to be any harder a technical problem to solve than the CoreXY configuration. I am looking forward to hearing what the rest of you have to say.
Hey there,
All in all there aren’t really too many corexy machines on the market right now if you really think about it. Sure there has been an influx of them recently, but a large amount of printers still rely on the outdated cartesian style mechanics.
It also might have to do with the buzzword that corexy is, with all the vorons and the bambulabs printers and the prusaXL there is a lot of hype around using corexy. So many companies are going to try to satisfy the demands of their customers. It could possibly come down to costs, if there are extra parts required to produce the croxy systems then it could dissuade manufacturers from choosing it over corexy.
I’m not really sure though, as you are correct in that this motion system can work well if implemented properly. I hope it is something that could be implemented in the future, but I guess time will tell for that one.
Anyhow, thanks for the interesting read, I have never stumbled across that specific thread before.
Matthew
You will see them in higher end printers (my Zortrax M200 uses the architecture) but it’s unlikely you’ll see them in consumer and low end printers as they require very precise assembly.
They really don’t have the same advantages as a CoreXY as they end up having a more massive gantry assembly and they restrict the size and operation of the toolhead. For example, my Zortrax doesn’t have PLA or any other material that requires part level cooling because the fan and ductwork can’t be integrated with the cross pieces. Instead of having two long belts, you’re probably going to have four or more because each side of an axis has to be driven - along with double the number of belts, your going to have the pulleys and idlers as well as the driving blocks (and figuring out how to tension everything).
If you have CNC equipment that can handle 300mm pieces of aluminum, you can probably put something together reasonably well, but you’ll probably have difficulty if you’re working with 2020 rails, M5 bolts and Tee Nuts.
Well, the primary advantage of a corexy design is the lower flying mass, which translates to more easily achievable high acceleration and speed values. Doesn’t a CroXY configuration have that advantage? I can’t think of any other significant advantages of either system over an H-bot.
It would be possible to configure a crossed gantry design so that hotend space wouldn’t be quite so limited. Four simple belts would still be simpler in my mind than the two long and complicated belts of a corexy. Might even still have the same number of idlers or less even with double the number of belts.
I hear what you say about 2020 extrusions, that makes sense.
I wish I could get a good picture of the inside of my Voron. It’s quite efficiently designed but an absolute bitch to maintain - when a belt goes, it’s multiple hours to replace and a lot of care has to be taken to make sure that everything goes back together perfectly, with nothing getting stuck between pieces.
I don’t know where you get the idea that a “CroXY” (or any kind of Cartesian architecture) will have less flying mass than a CoreXY. In the CorXY, you’re going to have a minimum of four gantry pieces whereas with a CoreXY you’ll have one or two. The flying structural mass (ie the rods) of a CroXY will be roughly twice that of a CoreXY (with a linear rail).
In the CoreXY printers I’ve designed, I use a single 2020 V-Slot rail for the gantry and 2020 V-Slot has a mass of 500g/m - as an 8mm optical rod is a bearing surface you should include something like a 12mm linear rail which has a mass of 380g/m (for a total of 880g/m for the CoreXY). 8mm optical rod (which my Zortrax uses) is 425g/m but for the CroXY you’ll need four of them so the flying structural mass will be twice that of a single 2020 V-Slot rail. If you were going to go apples to apples and use optical rods in your CoreXY design, the flying structural mass will still be half of what a CroXY has.
There are a lot of reasons why CoreXY is so popular.
I don’t think I said that CroXY or any other cartesian system could be lighter than a CoreXY system. At least that wasn’t what I was trying to say.
What I was saying, is with a CroXY architecture, both the x and y axis weigh about the same, whereas with a CoreXY architecture, one axis obviously weighs more than the other, although about the same as either the x or y axis on a CroXY printer.
What I mean, is that (at least as I understand it) the acceleration values are limited by the weight of the heavier axis. What the lighter axis weighs is largely irrelevant, as it isn’t the limiting factor. So the CroXY printer, while it has nearly twice the flying mass, still has its heaviest axis (which is either the x or y, as they are the same) about the same weight as the heaviest axis on the CoreXY printer, thus, at least in theory, meaning that the max acceleration possible for either printer is about the same.
Basically, the CroXY system could be considered somewhat more of a brute force solution, with the benefit of relative simplicity, at no real cost to performance, at least in theory.
You seemed to imply that a CroXY has the advantage of lower mass. No, you didn’t say it explicitly, but you seemed to feel that there was a mass advantage to the CroXY and I don’t believe that is true.
Again, I wish I could get a good picture of the XY mechanism of my Zortrax; the mechanism really is quite elegant but it’s not very simple with the optical rods at each end of the axis being used as the drive shafts for the other axis. The best I can find is this one:
which shows the two orthogonal drive/bearing shafts with the connecting belts going off from one axis to the shuttle on the other axis shaft.
It’s even hard to explain in words.
If you were only driving one side on each axis, I would be inclined to agree that there is something to it but for every printer with this architecture that I’ve seen, each axis is driven on opposite sides as I believe that the designers found that necessary to eliminate binding and the gantry becoming angled during movement. Below, I have highlighted the drives on each end of the axis from one of the images in the link your original post:
That’s quite a bit of extra complexity and, coupled with the need for absolute precision and the additional parts over other architectures, there really isn’t any kind of benefit to the CroXY. It’s definitely interesting but, having a printer that is built on this architecture, the superiority of CoreXY really stand out.
1 Like
What you are saying does make sense. Engineering is always a set of tradeoffs, and from what you are saying, I am beginning to see why the scales have tipped to CoreXY. Thank you for the interesting and informative discussion!
2 Likes
