Hello. I would like to share my view on the EG topic with regard to cnc design; I will also try to be as adequate as possible, although English is not my first language. We all know that EG is great in compression but not so great in traction or torsion. Both gantry beam and column mill design solution are primarily more subject to torsion than either to traction or compression. Especially in a gantry design, the vibration impact on the dimensional accuracy of a work piece lags far behind the Z axis overhang. I believe that having to deal with the Z overhang requires solutions that will also partly satisfy the requirements of a good vibration dampening system, and here I concur with Peteeng that the greater the mass the better. Also, as Pippin 88 pointed out, it is better to have an increased section of the beam than an increased wall thickness. Now, it is common knowledge that any Z axis that is off centre of a beam will impact the torsion angle of that beam, no matter how stiff the beam would be. So, a designer’s quest would be to find what angular deflection of the beam would be reasonably acceptable and at what distance from the center of the beam (overhang), so that the tool deflection caused by the beam’s torsion be less than a certain threshold. Let’s run some numbers, to make it more tangible: 1) steel, 1 m unsupported length beam, fixed ends; 2) square section 200mm/200mm/12,7mm to take Catahoul’s dimensions ; 3) cutting force= say 300 N; 4) Z axis weight = say 1000 N; 5) (horizontal) distance center beam to Z axis centroid = say 250 mm; 6) (vertical) distance center beam to the table= variable between 100 mm- 400 mm (300 mm the light under the beam) . Let’s calculate tool deflection from gantry torsion (A), tool deflection from gantry bending during cutting in the X axis (B) and Z axis displacement from gantry sag (C) for different Z axis overhangs. If the tip of the tool lowers to 101 mm from the centre of the beam (basically 1 mm bellow the bottom of the beam), than we will have: A= 0,00442 mm; B=0,00014 mm; C=0,00068 mm. If it lowers further to 200 mm from the center of the beam, we will have: A=0,00951 mm; B=0,00014 mm; C=0,00068 mm. When it lowers to 300 mm, we have: A=0,01555 mm; B=0,00014 mm; C=0,00068 mm. Finally, at 399 mm we have : A=0,02243 mm; B=0,00014 mm; C=0,00068 mm. Please bear in mind that A is the tip of the tool deflection from gantry’s angular deflection alone, and not from tool bending due to cutting forces . Therefore, apart from this deflection (A) there will be the well known sources of deflection from : tool holder runout, spindle runout, speeds&feeds cutting forces, milling bit runout, etc. Now , what we see is that max A is 0,022 mm and only the designer assesses if that is acceptable or not. For milling metals, most likely not acceptable over 0,01 mm, which basically would mean an overhang of max 220 mm from the center of the beam, a.i 120 mm bellow the bottom of the beam. It means that the user of the mill will either use risers of 180 mm in height, in order to mill plates, or the beam gets increased to: 300 mm/300 mm/8 mm (91 kg and 0,01 mm total deflection from combined gantry torsion and gantry bending during cut on X axis direction, at 450 mm from the center beam) . Now, out of 1 torsion and 2 bending forces, please take a look at the minimum value of torsion (A=0,00442 mm) as compared to the maximum bending (C=0,00068) and we see that the impact of torsion is about 6,5 times bigger than the impact of bending. At the other extreme, (A=0,02243 and C=0,00068), torsion impact is about 33 times bigger than the bending impact. Now let’s fill in the beam with EG; it means for the 200mm/200mm/12,7 mm an added weight of about 80 kg. The displacement from gantry sag C goes from 0,00068 mm to 0,00087 mm, basically peanuts. So adding EG is not a problem from a weight standpoint, but obviously the acceleration will suffer a bit. I wouldn’t worry though: first we calculate the inertia of the portal, ballscrews and stepper motors (assuming there will be steppers), altogether, as if there is no acceleration (the gantry+legs+railways +ballscrew+ Z axis with motor, should come in at about 300 kg with EG) and then we multiply it by the desired acceleration and see what torque would be able to deliver that. For that, assume 300 rpm is a speed where a stepper feels good, a.i where the trade-off between the drop in torque, increase in speed and the heat generated is in its right mind. We will have 300 rpm=31,4 rad/sec and this is the angular speed that will move the gantry , if the ballscrew pitch is 5 mm, then there will be a feed of 1500 mm/min (ok if you ask me). The acceleration time will be anything between 0,5 sec and 1 sec, therefore the angular acceleration in 1 sec is going to be Acc 1 sec = 31,4/1 = 31,4 rad/sec^2 and the acceleration in 0,5 sec is going to be Acc 0,5 = 31,4/0,5 = 62,8 rad/sec^2. Suppose the total inertia of the gantry would require 0,1 Nm to move the gantry at 0 acceleration; then the torque needed to bring the gantry to the desired speed would be anything between T= 0,1 x 31,4 = 3 Nm, and T=0,1 x 62,8 = 6, 28 Nm. Although I have not performed any calculations on the inertia of the portal , my guess is that a Nema 34 , 8 Nm stepper will do, plenty good. Now, we can see that filling EG in the beam doesn’t really screw things up either from a weight or from an acceleration point of view, assuming the conditions set out in the beginning of this calculus. The question is: does it have any measurable benefits as far as vibrations is concerned ? Well, it depends on measurements, something that a hobbyist has little means to do. Can we anticipate anything ? Maybe, but I doubt there is a recipe that can be carried over from one design to another. There are so many elements that can affect the overall dynamics of the system that only one minor element that changes will change response frequencies to various scenarios. The difficulty is that this is a trial and error thing, a.i you measure effects and then you work on possible causes. If the science were better at explaining what vibrations really are, we would have a different understanding of vibrations and probably a better way of tackling them. Frankly, I have never been a fan of imagining vibrations as oscillations in the strong force. Basically, more weight means more atomic metallic bonding, which means more strong force that should be able to dissipate the kinetic energy generated by vibrations, by either of the methods explained here as well. In conclusion, I wouldn’t worry trying to calculate to what extent some EG dampens vibrations and in what circumstances; I would try to include EG in my design anyway and I would try to calculate/assess where it screws up the design (acceleration, motor torque, etc) and how much that would bother me. If it would mess with my desirable parameter, I would include it in design. On the other hand, some reasonability in desires/expectations would not hurt anyone. I have seen some people trying to squeeze the life out of aluminium by reinforcing aluminium profiles with EG or steel plates, in a manner that made little sense to me (other than dissipating heat). Better steel than aluminium on steroids. The point is that heavy, agile and cheap don’t go well together. No affordable hobby cnc mill will do best in both steel and aluminium, at least that’s my understanding. Some choices have to be made from the outset and then it is advisable to stick to them. Sorry for the lengthy post.