yeah i have the fillers on the cheap. sand and sand pebble from the building center. works well on the cheap side. sand blast media would get really expensive and with epoxy being the real cost of the process. i know what i have is at least as strong as concrete on the compressive strength side of things. the epoxy works well it is just harder to mix and takes more than 12% and ends up being like 15% any less it becomes to dry and to hard to work with. i seen conversations here on using far less than 12% even but it does not work with what i am using. what i have is pretty impact resistant as well so i was happy with what i came up with. probably not the most cost effective mix but it seems to work well.
Hi All - Finally I have been able to put some quality time into the dwgs again. Checking all the holes line up is a mission. I was check "fitting" the gantry to the outside rails and noticed the holes missed by 0.427mm. Decided to track this down as it "should" have been a round figure. Worked back and front through the assemblies and what a mystery!! Turns out the bearings on each side were slightly misaligned and the actual hole spacing was correct. Now to true the gantry up and that little delta will disappear. Onward drawing by drawing, hole by hole.... I've had kitting costs in for the table and they are really good. Seems cutting up wood is so much cheaper then metal. Makes we want to go back to my plywood router design, later....Plus a couple of the machined parts so the cost sheet is fleshing out as well...Peter
Hi All- This weeks job is to get the laser cut dwgs done and out to quote. Plus finalise the drive parts dwgs, out to quote. I have been mucking about with the graphics settings in my CAD. I use the illustration mode for manuals. Here's some snaps. I have finalised some of the technical details of the saddle soldered parts and now very happy. Have been playing with phosphoric acid and hydrochloric acid cleaning and both seem to be OK. Hydro cleans the mill scale faster.... Just have to get the dwgs over the finish line... Peter
Hi All and Sundry - Been checking through YaG and decided to look at the motor sizing. The dual axis can accelerate at an average of 250mm/s2 which is good for this level of machine. The cross gantry axis will do better as it's the same motor but only moving the Z axis and saddle which is about 22kg. The long axis is about 29kg per motor. Also found a cute photo of a lizard... The inertia calcs show the rotational inertia is dominant with 1.28Nm required vs 0.13Nm for the linear component.
So if we want to accelerate to 5m/min in 0.33sec, the machine will take off at 500mm/s2 as it has 3Nm of torque then as it gets to 500rpm will decelerate to 250mm/s2 as it only has 1.5Nm at this rpm.
For those wanting to use this calc in the Z axis remember the load has to include the weight of the Z axis as its being lifted. It also has to be subtracted when descending. In summary the sizing of the motors is Ok for the YaG machine intent. Peter
Hi Buoyen - Thanks for checking yes you are correct. When I converted [800gcm2] I didn't square the cm. That's good as the acceleration will be much higher. I'll redo the numbers. Lower inductance is always a good thing. I looked thru stepperonline and they do have 1.7mH versions with 8mm shafts. I like bigger shafts 10mm as the bearing is bigger but the 10mm shaft series has the higher inductance. The torque curve for this stepper is fine but I can review prior to purchase... Cheers Peter
Hi Bouyen - again thanks for the pickup. Here's the corrected calculation(removed to correct again) . Much betterer . Peter
Guys and gals out there please note that to acheive 0.1g accelerations is tough on maker grade machines even though the numbers indicate the motor can do it. Friction, stiction, clearances (chatter) racking, jerk control all of these things go against achieving high accelerations. Re "jerk" Jerk is the rate of change of acceleration. Some controllers have jerk control and acceleration control. Jerk control results in much smoother motion then just controlling acceleration...
I hadn't heard of snap crackle and pop. At uni it was called jounce (snap) camshaft designers are obsessed with jerk. I suppose camshaft designers are not needed much now...
Heaps of potential acceleration. Check the math... errors and omissions are not the responsibility of the author, their descendants or assignments. Especially second scandinavian cousins....etc etc...
Hi All - I'm back and gearing up for the build so this now turns into the build log.... I've ordered the castors for the bench and tuned up the bench design. Today I order the cut ply from the cabinet maker. He can cut this very true vs me with a fence saw (I sold my bigger router a few weeks ago). I nested the bench out and it fits 3 sheets very well. I'll adjust a couple of sizes to make better use of the sheets. YaG has gotten bigger as time goes on and now it gets bigger again. I've decided to make it a full sheet width deep ie 1200mm deep. So start on the bench asap then YaGs machine base then the machine. I'm missing having a big router at hand already!! Peter
Hi All out there - The castors and machine feet have arrived. The plywood for the bench is cut up and ready to pick up but my truck didn't want to start 2 weeks ago. Had to have it towed to my mechanic who got it started but it would not run so had to tow it to a diesel mechanic who identified that the speed sensors for the fuel system had died so 2x$490AUD for the sensors and it still didn't start!! Then he didn't have the tools to retime the timing chain (next code from the black box) plus the sensors may have been A-OK but can't send them back to Renault.... What a drama!! Now the van has been moved to another mechanic that has the tools to time the engine. And I wait to find out costs and when it will be done.... I'm going back to petrol and looking for new truck, a Toyota Cummuter will be good. Had one before and clocked 500,000k no worries...... Renaults only done 110,000k should be just run in.. Peter
Levelling castors look a bit small but will be fine. Next time will get bigger ones... a la Craig
Hi All - I have a little project with a hiking trailer I'm making. The belt (towbar) needed stiffening so I made a little stiffener for it from some scrap carbon fibre on Brevis-1. Now I paint it red because I have red epoxy paint in stock....
The trailer is in its 3rd iteration and quite mature. The belt is in its 4th design and now I'm happy with the lot. Need some diversion therapy sometimes....
Hi Boys & Girls in CNC Land - For some time I have been intrigued why mounting the rails on the diagonal of the gantry results in less deflection then when mounted on the front face. In some cases it's been 50% stiffer. I haven't had time to investigate this properly until today.
The drawback is making a more complex saddle but that also has the advantage of getting the drive screw on the top of the gantry if you wish. So this morning I made two models using a simplified car and rail and a 100x100x6mm (1200mm long) steel gantry. On one the rails are on the face on the other on the diagonals. The saddle is a 50x50 solid steel member. The lever arm is 200mm below the gantry and is 1000N (100kgf) So the face mounted deflected 0.267mm and the diagonal is 0.156mm which is 42% stiffer. That's definitely something to chase. So then I looked at the reactions and their trajectories. The car contact forces on the 20mm rail are higher in the face mount. This is because they are closer together I believe. The supports reaction is horizontal and through the section centre for both. So by the time the load gets out to the columns the twisting must be pretty consistent. I ran non linear and linear solutions and there are no non linear effects in this model ie NL and Linear deflections where identical. My other conjecture and I can see this in the video but hard to pick in the images is that because the two rails are mounted on the same face and they have the same moment direction the front face is twisted much more then when they are the diagonals. So I believe it's a localised effect of 1) More local contact loads due to the smaller distance and 2) being on the same face the local loads warps the face more.
This can be corrected by making the front face thicker if you want to mount the rails on the front. But then this would be more weight (maybe) then an L shaped saddle (maybe) depends on how you do it.
Takeaway - Mount the rails as far apart as possible and prefer on different surfaces to uncouple their combined effect.
My next question is - Is a triangular gantry better then a square or rectangular one?? Cheers Peter
Hi Everyone - Got onto the triangle gantry but how to compare apples to apples? Do I build equal inertia sections? Equal weight sections? all too complicated so I used the same envelope. So the triangle is basic 100x100x6mm. The gantry length is same as last time 1000mm sorry last entry I said 1200mm but Its 1000mm. I applied 1000N 200mm down from the gantry. So the bearing arrangement geometry is very similar SQ vs TRi.
So compliant to rigid:
1) 0.267mm SQ front face 100%
2) 0.158mm TRI diagonal 59%
3) 0.156mm SQ diagonal 58%
4) 0.147mm TRI on face. 55%
So although the TRI section has less geometric inertia its stiffer. Most interesting. Next question how do I build this? Peter
So I thought about a semi circle and added it to the list. The same envelope and thickness. It deflects 0.147mm as well so its an equal stiffest design. There are some commercial machines like this. I think this will be my composite design gantry...Peter
Re: Designing a new router called Brevis-HD
Reply with Quote
Peter
