[luke peter]

Hi Jayne!

Glad you are also reviving your efforts, and a little glad you took a long hiatus (it makes me feel a little less like I am dropping the ball ).

If I was buying motors today, I would get lichuan AC servos, the lc10-p ones work on pulse and direction like stepper motors do, and their price is reasonably reasonable... They also have ones that work on Ethercat, which is attractive to me in an elusive kind of way. They are motors, and drivers, and cables all together powered by 240v ac input (a search on Aliexpress will yield adequate options).
I am rebuilding Steves control system to use linuxcnc and a Mesa 7i96s card - still a work in progress, but will happily share the pain points once I am through with it.

[JayneV]

Hi Luke,

I will have to re-read your build thread to refresh my memory on “Steve” and catch up on what I’ve missed in my absence. I will be interested to follow your journey with LinuxCNC.

The AC servos sound interesting, I will read up on them to learn more about the pros/cons vs the steppers I was planning to use. Ethercat is an unfamiliar term that has popped up in some of my recent searches, another thing to read up on.

I only had a Quick Look on Ali express, it seems the servos are rated in W/kW and steppers rated in Nm. How would you select approximately equivalent rated servos and steppers to compare specs and pricing? For example, before shelving the project, I was planning to use Nema23 3Nm steppers, how does that translate to a servo rating?

[luke peter]

Hi Luke,

I will have to re-read your build thread to refresh my memory on “Steve” and catch up on what I’ve missed in my absence. I will be interested to follow your journey with LinuxCNC.

The AC servos sound interesting, I will read up on them to learn more about the pros/cons vs the steppers I was planning to use. Ethercat is an unfamiliar term that has popped up in some of my recent searches, another thing to read up on.

I only had a Quick Look on Ali express, it seems the servos are rated in W/kW and steppers rated in Nm. How would you select approximately equivalent rated servos and steppers to compare specs and pricing? For example, before shelving the project, I was planning to use Nema23 3Nm steppers, how does that translate to a servo rating?

The kW rating is correlated to a torque rating also, you will find the rated torque in the info in the description. Stepper torque will drop to near zero at around 500rpm, while servos maintain their rated torque to 3000rpm, so realistically anything over 1nm will be more than enough. I would get the 600w units for Steve, which has a 130kg gantry...

[JayneV]

Thank you, gives me something to think about. Still a long way off before needing to make a final decision on motors.

[JayneV]

Hi,

I have decided to build the modified Momus design instead of Carlos, mainly because this design is easy to fully enclose compared to Carlos. Containing dust and chips and attenuating some of the noise are a priority to help keep the garage reasonably clean and also not upset the neighbours.

Some of the details are still bugging me though. Firstly, I’m not at all happy with the z axis. The z axis in the attached image is a just a place holder and was quickly modelled in about 10 minutes for a visual reference. Will keep thinking about the details while the frame is being built.

The other detail concerning me is the mounting surfaces for the y axis rails. I don’t know how to ensure both sides are flat and coplanar. I’ve seen some people have used self leveling epoxy with a temporary bridge constructed to join the left and right walls so the epoxy creates a level surface across both sides. But then epoxy may not be the best surface to mount the rails to without compromising rigidity too much. I could use some help with ideas for mounting the rails as accurately as possible.

[peteeng]

Morning Jayne - The design you have is similar to my Scoot. Your machine base will be plywood? Scoots is MDF and has been very good. In regard to the Z axis keep at it and don't get married to a design until you are very very happy with it. In regard to making the wall tops and rail foundations level and co-planer.
1) I don't suggest you use self levelling epoxy. It needs to be 5mm plus thick to self level properly and epoxy has a modulus of 3-5GPa. So its pretty much the same as laying a sheet of rubber on top of your walls
2) Buy a good quality long level and a good quality chinese engineering level. If you look at local eng levels they are very expensive. Mine cost around $60 and has been very useful. Its 300mm long.
3) Having been around boats and yachts and fairing for a long time its pretty straightforward to scrape and lap the top surfaces in. So pick a long flat piece of heavy aluminium channel the same width as a roll of sandpaper
4) Your build quality of the ply parts (in particular the squareness and parallel of the pieces) will determine the quality of the top of the walls. I recommend you get a local cabinet maker to cut the parts on a cnc saw or a good table saw. If you hand cut the pieces on horses with a hand saw you will be millimeters out at the top of the wall by the time you build it up. But thats no problem just more lapping to be done and patience if that's the chosen production route.
5) Using the levels and good straight edges survey the top of the wall and identify the high spots. Lap them down. In the images you can see I removed the melamine in some spots. The melamine is 0.9mm thick. Once you've lapped a bit decide which side will be the master side. Then use levels to shot across to the second side to see how thats going. Adjust the feet of the machine to correct "average level" as you go. Use a long straight edge lap to sand across the walls (thats the co-planar bit)
6) Use epoxy filler to fill lows unless you feel you can bring the surface down to the lows. But a skim of putty will make the process much quicker. Do not use auto body filler it shrinks. You can also use a black spray (just a flash) to "contrast" the surface occasionally to visualise what you have. You can buy powdered contrast at a panel beaters suppliers then you don't get overspray on your BMW. The epoxy putty is often a colour so you can wipe in with a flat metal spreader to fill the surface and contrast the surface at the same time. (Mines blue)
7) Once you get the surfaces level and flat to your satisfaction you can epoxy the metal top plate to that surface and it will be as good as it gets. Just remember that our early tool builders built iron and steel machines by hand (scraping and lapping surfaces) that could make things to the micron, so you can to.
8) I have found that 10mm aluminium is a good thickness for the top of the wall. Its stiff enough to stay nearly flat and not be too floppy that it will sag if you have a low spot.
9) In my current metal version of Scoot my top plate is adjustable so I just work along each side with the level and get it to a happy place. It has brackets top and bottom of the ribs.
10) On Frankie I used a lighter aluminium angle for the foundation and that was satisfactory as well.

Here's some photos. Peter

[peteeng]

Hi Jayne - Regarding reflexing the Y motors. I think this makes that area more complicated than needed. If you use 60mm motors (remember that you have two on that axis so twice the torque so equal to a 110mm motor) I think you will find they stick out same distance as your pulleys and covers. I have 112mm motors sticking out and they have never presented a problem. Here's Frankenrouter - similar to where your going. I used round rails as the stack height of the drive nut was the same as the carriage ht. But I think I'd stick to square in future the cost delta was minimal. I see you have 15mm spacers on your cars. Peter

Keep your walls narrow. 100mm is all that's needed to minimise the footprint. I like your gantry, that will be a really solid one. I think there will be enough room to not reflex the gantry drive if you use a 100mm or 70mm motor. If you use a 5mm pitch ballscrew you will have tonnes of force but will be a bit slower but on a small machine like this you won't notice. Get an aluminium mount like this one.

[peteeng]

Hi Jayne - I also see you have beveled the top of the wall "sides" to mitigate dust I suppose. But better to put them "under" the top plate. If you do this you save 60mm in footprint width (if its say 15mm ply). Also you have to think like a stonemason or a pyramid builder/bricky. Everything is built on top of each other so it can be self assembled and checked at each step of the way. Other wise you have to build things and correct things as you go... Cheers Peter

[JayneV]

Hi Pete,

Thanks again for your detailed reply. You confirmed my concerns about using self levelling epoxy. Bow that I think of it, I think we had a similar conversation about the epoxy in my first build thread.

I already have a very good long level. It will be good to coarsly level the entire machine and act as a straight edge across the walls. I also plan to purchase an engineers level. Will look into the Chinese options. Hare and Forbes sell a 300mm level for around $200. Expensive but not outrageously so.

I plan to do the majority of cutting with a mitre saw. I have already adjusted the saw to o be quite accurate when I cut the aluminium parts for Carlos and was getting. Very good repeatability in the cuts, so I’m confident cutting wood will go well. The weak link will be me and when all the parts are assembled. I will need to take extra care to ensure everything is assembled accurately and keep everything square along the way as the machine comes together.

The top of the wall will be 12mm AL because I already have the material intended for Carlos.

Thanks again for your input, I tend to get bogged down over thinking things and the project never gains momentum.

[peteeng]

Hi Jayne - I think Einstein said something like if you have 60mins to solve a problem spend 55mins thinking about the solution then do it. So as long as the thinking is good quality and the design is progressing then you are making good progress. The building will go quickly if all the issues are sorted in CAD. So from my chair - the gantry & base concept is sound. Needs some tuning. Think about not reflexing the motors, sort the Z axis (or buy one from me! ) and then start building say July 1st.. Cheers Peter

[JayneV]

Don’t joke, I did seriously consider buying a z axis from you. Happy to discuss that option, perhaps in person(check your inbox)?

The reason I reflexes the motors is to save some space and make it easier to use close the machine without having a motor protruding. Should be easy enough to adapt the enclosure to accomodate the motors in a direct drive configuration. Will fiddle around with CAD some more.

[JayneV]

Hi Jayne - Regarding reflexing the Y motors. I think this makes that area more complicated than needed. If you use 60mm motors (remember that you have two on that axis so twice the torque so equal to a 110mm motor) I think you will find they stick out same distance as your pulleys and covers. I have 112mm motors sticking out and they have never presented a problem. Here's Frankenrouter - similar to where your going. I used round rails as the stack height of the drive nut was the same as the carriage ht. But I think I'd stick to square in future the cost delta was minimal. I see you have 15mm spacers on your cars. Peter

Keep your walls narrow. 100mm is all that's needed to minimise the footprint. I like your gantry, that will be a really solid one. I think there will be enough room to not reflex the gantry drive if you use a 100mm or 70mm motor. If you use a 5mm pitch ballscrew you will have tonnes of force but will be a bit slower but on a small machine like this you won't notice. Get an aluminium mount like this one.

Hi Jayne - I also see you have beveled the top of the wall "sides" to mitigate dust I suppose. But better to put them "under" the top plate. If you do this you save 60mm in footprint width (if its say 15mm ply). Also you have to think like a stonemason or a pyramid builder/bricky. Everything is built on top of each other so it can be self assembled and checked at each step of the way. Other wise you have to build things and correct things as you go... Cheers Peter

Hi Peter,

Sorry, I missed these two posts. In Adelaide for the weekend and only have my phone with me. I will definitely fiddle around with the y-axis motors and mounts to swap back to direct drive. On Carlos, the motors were direct drive and poking out the back, might explore that option since the back section of the enclosure (not shown) is fixed (fixed as in no doors but the whole enclosure could be removed if necessary) making it easier to accomodate the motors without leaving any gaps.

I already have the screws. X and Y are 10mm pitch and Z is 5mm. I have one of those motor mounts for the Z axis. Didn’t need it for X and Y with Carlos because all the parts were going to be CNC machined and the motor mounts were integrated into the frame structure. Those aluminium mounts would work well on the current design so I’ll order some more for the remaining motors.

The walls are 160mm wide. I have plenty of realestate on the bench in width. 160mm gives me plenty of space for the electronics to live inside the walls without being too crowded. It also leaves enough material on the face plate after cutting a 100mm hole for the cooling fan. The internal ribs have holes to allow airflow right through the wall to remove any heat generated by the electronics since it will all be inside an enclosed space.

The beveled top is straight out of the Momus plans. All the edges are beveled in the plans, presumably for aesthetics. It’s been a while since I’ve looked at the Momus plans but if I remember correctly, the wall side skins were put on near the end of the build which is probably why they are placed the way they are. What you say makes sense though. I will take another look at the wall skins. Unfortunately I left my iPad at home. I was hoping to remote in to my home computer to fiddle with Fusion while in the hotel at night, a bit hard to do on the phone. Perhaps I will go old school and use pen and paper, or pretend I am a real engineer and use a napkin while having some drinks at the bar.

[peteeng]

Hi Jayne & Others - Here's more stuff for your napkins. I have made machines like the Momus in two ways. 1) with the C shaped ribs and 2) with a base and walls separate. I suggest that the separate approach is best from experience. When I made the machine with the C shaped ribs (Frankie) I cut them on a CNC so they where quite accurate. If you cut these manually I doubt you will get the machine base and the top of the walls to be level and parallel. If you make the base you can survey it and level it and correct it then build the walls and place them on the base without the top plate and then lap the top of the ribs so they are flat and level, both sides with them close together. Then install them at the edges and re-survey them. Then proceed to lap the tops and get the foundations set up. The other aspect is weight. If you build a medium size machine with the C ribs the base becomes quite heavy to deal with by yourself. If you have to break the machine down to move it, three parts is better then one.

The next issue is what hold down style are you going to use? thread insert grid or T slots? I prefer a threaded insert grid. I have read about the T slots tearing out, you have to use a lot of small screws or you have to glue them in. Now if you go the insert path you don't want a bottom sheet on the machine base. Even if you use T slots the base is stiff enough without the bottom sheet. The inserts let a lot of sawdust through and if the base is "sealed" you will never get that out and it will fill up and increase in weight and be a nuisance. So leave the bottom open and let the sawdust fall onto the bench or if you do sheet the bottom leave a side/s open so you can stick a vacuum cleaner down there occasionally. I don't like T slots mainly because the screws then stick up really high whereas with the inserts the screws always just have the head exposed so minimises crash potential. Even then I nick them occasionally (or worse) so If I had high screws I'd do much worse... Peter

[JayneV]

Hi Peter,

Thanks for the tips. The walls are separate to the base it’s only the front and rear skins that are a single piece. Front is U shaped and the back is rectangular across the full height and width of the machine with a window in the middle. The window will have a removable panel incase I ever want to work on something longer than the machine.

I have a whole bunch of t-slot that was salvaged from our old solar panel installation which I had considered using with Carlos, but the wall thickness of the slots is too thin and I found it is easy to distort if the fasteners are done up to tight. Threaded inserts seem like an effective and easy option that will perform better than the t-slot I have on hand. I have modelled holes along the bottom of the sidewalls for the purpose of sticking a vacuum hose in there to clean out any dust build up. Another option is for the side skins to only cover the wall without extending down to the bottom of the base, but I kind of like the fully enclosed appearance. Could also make a removable or hinged panel to allow easy access to the inside of the base. I’ll sketch up some alternative options.

With the threaded inserts, I thought about 3d printing blanking plugs for every hole and only removing a plug when an insert is used for securing a job. Constantly removing and replacing plugs might get old really quick though.

[peteeng]

Hi Jayne - Masking tape works for the holes. Peter

[JayneV]

I started playing around with the simulation feature of Fusion to compare different z axis configurations. I don’t really know what I am doing but I’m having a bit of fun with it. I figured if the simulation is run with the same settings between different designs, then I could at least get a rough idea that say design A is stiffer than design B, or C is stiffer than both A and B even if the actual numbers in the results don’t have any real world significance.

The attached image is the first result on a design with the z axis at full extension down from the gantry. If am understanding the results correctly, they are saying the lower edge will displace by 0.331mm with the load applied where the arrow is. That is assuming perfect bonding between the various mating parts in the assembly (I think?). It’s a start. So now if I change the z plate material from 6061 AL to say plywood and run the test again, I should get a result that would at least indicate which material choice is stiffer, presumably aluminium if all else remains the same and only the material is changed.

I am finding simulation interesting, just wish I had some actual engineering background knowledge to understand how to properly use that tool to aid the design process.