Many people have used these feeders on Ultimakers. Here is one expert user, below, who used just the Bondtech drive gears. The instructions look pretty good, the only concern might be price?
Neotko FatIRobertI Bondtech Feeder for UM2+ / UM3
Edited by GuestI persona don't have issues with my UM3 feeders so for me it would be a waste
But if your having issues then it maybe worth it.
But like @lrodriguez has shown. There are cheaper options that could give the same result.
Many people have used these feeders on Ultimakers. Here is one expert user, below, who used just the Bondtech drive gears. The instructions look pretty good, the only concern might be price?
Neotko FatIRobertI Bondtech Feeder for UM2+ / UM3
That's a fine feeder 
I made it exactly because bondtech isn't cheap. But seeing that they have change the design so it removes the bowden curvature I might change my design someday, is a nice idea to remove as much bowden pressure as possible, just as a general rule.
@Tomnagel about "the standard Cura printing profiles won't be optimal anymore."
What settings do behave worse on a Bondtech than a UM2+/UM3 feeder?
The only I can think of off the top of my head is the PVA retraction distance, since it actually doesn't retract the filament, but just move it the distance long enough to remove the pressure without moving the material on the core. Any other than that or is more a gut feeling? Any test done by UM on a um3 with bondtech? Any issues found?
I general if you invest in an upgrade you need to think about the problem you want to solve. In the old case of the Ultimaker2 it was very clear the feeder was a weak point, and the bondtech was/is a very good solution for that.
With the Ultimaker3 the feeders are IMHO not the weakest link of the extruder driver train anymore. The head, with its weight and weight distribution is clearly the limiting factor. When I see "be able to print much faster with sustained quality" on the bondtech site I'm pretty sure thats not the case for the UM3.
Besides this I have no idea what Tom means with his statement on Cura profiles not being optimal anymore. I don't see any issue there.
In short, would I spend the 400,- ? nah... putting a Ferrari engine into a Fiat500 will not make it a Ferrari.
2
There are certainly alternatives for certain parts on an Ultimaker, which behave different than the Ultimaker-part. A reason for considering an alternative could simply be price (if they would behave the same), or because they behave differently, and perhaps this would suit your workflow better. It can be. That is also why we have such an open system, so there is a level of freedom for our users to tweak their tools and workflows to optima forma.
What we (try to) say is that if you replace parts by a third party product, which is not just cheaper, but behaves differently, the settings will need to be adjusted to this new situation, to make it 'optimal' again. The current settings are tailored for Ultimaker-parts. So if you want a different set up, you have to know this influences your settings too.
1
Are there issues you're having? The only extruder I have upgraded on our Ultimakers is on the 2Go and that uses the old extruder design so it was a must. Our 2+ and 3 have no problems in our usage.
No I haven't had any issues. I was just wondering for the 3 if there was any great advantage to a set of $400 extruder motors.
What settings do behave worse on a Bondtech than a UM2+/UM3 feeder?
Not perse worse, just different. And different behaviour of a subsystem creates different behavior in the total system, which could be, in combination with the Cura default profiles, suboptimal.
Every feeder has something I call microslip, with which I mean that the amount of filament that is transported per revolution of the motor depends on the force on this same filament. In the UM2+/UM3 feeder, this microslip is probably more than in the Bondtech feeder, because that feeder grips the filament on both sides. Some people might say that therefore the Bondtech is superior. But that is not always true. A somewhat higher microslip yields a more 'forgiving' feeder, which could be beneficial in the first layer for example.
And there can be other differences, I don't know.
Anyway, the default Cura profiles are tuned towards the UM feeder, and not towards other feeders. So if you are a user who wants 'no hassle', stay away from non-UM 'upgrades'. If you have special needs, and are willing to invest a lot of time optimising your system, go ahead.
1
Interesting about the microslip. I call it missing steps.
And I mean, that because the extruder motors lack of precision under torque, they can, and do, miss steps when under torque specially. So when extruding and retracting the feeder doesn't actually move the real amount of mm asked, since it can retract more distance (no torque) than what extrude (sudden change of torque when reaching the extrusion point).
About steppers behavior:
http://hackaday.com/2016/08/29/how-accurate-is-microstepping-really/
So Cura profiles take that into account. Nice. But also, let's not forget that except infill the max real speed of that profiles doesn't go beyond 30mm/s. And that not even taking into account the low accel/yerk that lowers the visible areas to a even lower number.
Wouldn't it be better to get as much precision as possible and build up from that? I mean, Cura already uses a way too big number of tricks, removing variables and adding precision should allow for better and specially faster prints.
Edited by GuestNo, missing steps is an entirely different mechanism.
What I'm talking about with microslip happens in the pinch.
The same phenomenon happens in cars: the distance covered with one wheel rotation of a car is dependent on the forces between the road and the wheel. So when accelerating, there is a high force between road and tire, and one wheel rotation gives a smaller traveled distance then when cruising at constant speed when the forces are (much) lower.
Back to filament transport: even though the knurled wheel "bites" into the filament, the same effect is there. You can even see it: look with a magnifying glass to the teeth marks on a piece of filament: the distance between the teeth marks decreases when the extruders forces go up.
A guy called "Nophead" reported on this effect already in 2009. He has photos of the teethmarks somewhere on his site, but unfortunately I can't find the exact page.
I thought that phenomenon was more related to the distance between the wheel and bearing and how much pressure the filament does to it and how the bite can go deeper after passing over again increasing the diameter of the rotation, making shifts on the esteps moved.
Micro slips are very common in the production environment. @tomnagel is correct. There are mechanical influences but there is always some sort of slip between the material and the drive wheel and lots of things effect this slippage.
Also it's not about the diameter as the circumference traveled is still the same.
Edited by GuestBlessed be the fruit
I think in cars they call it "partial slip", most noticable when braking very hard. On dry surface, you get the best grip of the tires when they slip about 15% (according to an article in an automotive magazine I read years ago). At which point they already start making a grinding noise. A 100% slip is when the wheels are locked up while braking. And (almost) 0% slip is when the car is just rolling down the road, without adding power, without braking. So most ABS systems are designed to prevent locking up, but allow the partial slip that gives the highest friction and best braking performance.
In 3D-printing filament, under a microscope this slip is visible in that the square pits that the feeder bites into the filament, get stretched out into diamonds, due to the mechanical force and rotation of the feeder wheel against the filament. So if the pits get stretched out by 20%, that would amount to a slip of 20%, and thus 20% underextrusion. So, less friction in the feeding system, should give less deformation of these pits, and thus more accurate extrusion. I guess that is how it works? 
1
Acceleration and deceleration will have different slippage percentage compared to constant feeding an this percentage will change depending on the pressure required to push through the nozzle. Temperature will change the viscosity and could minutely have an effect also.
Our production lines that use plastic film we had driven unwind and dancer rollers that control a potentiometer. You can then drive the film off the roll at a constant tension to give a constant micro-slip throughout the machine. This way a new roll of film gives the same results as a nearly used one.
so back to your original question 
yes i am using the bondtech extruders on 3 of my 5 printers.. all um3.
I certainly see advantages!
the biggest one for me is that because i mounted the bondtech feeders on the backside top, i can use the large spools on the UM3 extended.. i can also see that they provide much more pressure and i can work better with flexible materials. to me it also seems the material retract at the end of the print is a bit cleaner..
as far as, should you buy them for the UM3? I think in very specific scenarios they come in handy. I am happy wiuth them on the 3 printers that have them, but I won't be upgrading the other 2..
2
ArunC posted a topic in UltiMaker Cura,
.thumb.jpeg.0b7a05eafc09add17b8338efde5852e9.jpeg)
Dustin posted a topic in Firmware,
Recommended Posts
tomnagel 126
If it behaves differently than the UM3 feeder, the standard Cura printing profiles won't be optimal anymore.
If it behaves the same, then it's not worth $400.
It very much depends on what you are looking for. In general, the UM3 feeder performs well, but in certain cases the Bondtech feeder will be better. But you will have to optimise the slicer settings yourself.
Link to post
Share on other sites