Finished up the design for the gantry risers a couple days ago. I'll take the design to a machine shop iv'e dealt with before to see how much they want to waterjet some precise holes and cut a couple shallow pockets. Shouldnt be too expensive. Especially since the file is already in .dxf format, most of the work is already done. The owner showed me their waterjet machine last time I was there. The operator sets the workpiece on the bed and the machine autolocates it with a laser and cuts the uploaded file. I was impressed. I'll ask them about surface grinding the two faces that will mate together. This will get me a perfectly flat and stress free attachment from my linear rails up to the gantry. If its within the budget..., if not I'll get some practice with epoxy joinery.
Re: 2nd build - For a friend - Clearpath Servo selection
Over the last week I've spent a lot of time reviewing power curves for a few different Clearpath servos. Back when I first started this, I created an account on Teknic and was promptly contacted via email by one of their engineers offering assistance in sizing a motor. Based on what I told him, which at the time was a close guess on some things, he recommended a https://www.teknic.com/model-info/CPM-SDSK-3421P-RLN/
enhanced version. The enhanced version has a resolution of .06 degrees as compared to the standard version resolution of .45 degrees. So i figured out that the standard version's encoder is 800 counts per revolution; 360 degrees/ .45 = 800. The enhanced version is 6000 counts per/rev: 360/ .06 = 6000. The gear reduction of the NEMA 34 mounts from CNCrouter parts is 3.2:1. It conveniently works out to about 1" (.9817") of gantry travel for every 1 motor shaft revolution based on a their pinion gear and rack. So using the standard version Clearpath servo encoder count of 800 counts per rotation and a 3.2:1 gear reduction I can achieve a positioning resolution of .0012" (.9817" inches traveled per rotation/800 counts per rotation). The enhanced version is about 10x that (.982"/6000 = .00016"). The standard version seems plenty accurate enough. I think... I ran across a post about another Clearpath servo user on Reddit who built a CNCRP 4896 kit with them. He said the servos were fast and powerful but he had increased their resolution which slowed them down to about 300 ipm. (????)
I'll have to contact Teknic and ask them about it.
Just got off the phone with Teknic about their ClearPath servo's. Excellent customer support BTW. I wanted to double check their recommended motor selection now that my design is a little more complete and, ask them about higher resolution settings translating to slower speeds. Their engineer ran another simulation for me again and verified what I thought was a better choice of servo motor. Initially he had recommended SDSK 3421P enhanced motor but it looked to me like the 3421S was a better match based on its higher torque capability. It has a lower RPM but I only need 1000 RPM at the motor to hit 1000 ipm travel on the router with the 3.2:1 gear reduction. And in reality, most moves arent rapids and will occur in the greatest part of the torque curve. Actually, RMS torque with the SDSK 3421S doesnt begin to decrease until just before 1000 RPM.
I also cleared up a couple things about motor resolution on the Standard vs the Enhanced models. The standard model has an 800 count per revolution encoder which means you can position to 800 different points in a complete revolution. With my gear reduction I can achieve .0012" positioning resolution with the standard model. The enhanced model has 6400 counts per revolution and obviously can position at 6400 different points in one complete revolution. But within the motor parameter menu you can change the input resolution. For the standard motor you can select 200, 400, 800, 1600... on up to 25600 pulses per revolution. You can still only position to 800 different points but according to the engineer the motor still sees all the pulses and this translates into smoother cuts. The trade off is, if the motion controller cant send pulses fast enough it will slow down the travel speed of the machine. Setting the input resolution of a Standard model motor with an 800 count encoder to 1600 essentially doubles the pulse rate required to maintain the same travel speed. The motion controller I plan to use operates at 400kHz and is fast enough not to incur a speed penalty with a higher input resolution. Much more confident now in my motor selection.
Decided to make my own data and power cables so i spent a couple hours today selecting the right Molex connectors. It was made a lot simpler due to the excellent documentation in the Clear Path user manual. They had the Molex part numbers that turned out not to be 100% correct but close enough to get into the right series and find the right ones. The connectors are only pennies! I spent less than $15 on the connectors and ordered extra. However i had to buy the Molex crimp tool, $54, and a removal tool, $13, which added a lot. With shipping it was just over $100 for the connectors. I still needed wire. That was expensive! It didnt help that the shortest length available was 100ft. At the max I only needed 55ft of 16/2 and 22/8. I remembered i might have some on hand. Checked in my garage and there it was; 16/2, about 100 ft. Still needed the 22/8 but i should be able to source some from a friend locally.
The power and data cables from Teknic were over 400. So making them myself saved about $300!
With the money saved by making my own cables, i decided to upgrade the servos to the Enhanced versions. Its only $50 more per servo and seemed like an easy decision for such a critical part. With the Enhanced servos and making my own cables i am still about $300 under what i had budgeted for standard servos and cables.
Still waiting to hear from the machine shop about my RFQ for the gantry risers.
Yesterday i called Mike at Precision Epoxy. He was away from his home in Atlanta and wouldnt be back there until Friday. He only takes checks. By the time he tells me how much and then add the time for the USPO to deliver a check to
Atlanta, it looks like it will be about 2 weeks before i get my epoxy. Wish it was a little easier to get his product.
I had a chance to talk to my cousin who is a Phd chemist and has a lot of experience with the West system brand of epoxy. He felt like a lower molecular weight epoxy would self level better than the West system brand. With the delay with Precision Epoxy i was considering using West but now I will wait. Im still ahead of schedule.
Aluminum extrusion from 80/20; requested a status update today
Mike at precision epoxy; sent a follow up email.
The gantry riser machining quote from the machine shop; called today and was told to expect a quote today.
The Molex connectors arrived today from Allied Electronics. And I'm expecting some nice 22/6con shielded control cable that I found on ebay to arrive via UPS today.
Made some progress today. UPS Freight called me early in the morning and coordinated a delivery time. They said the driver would need help unloading it. It weighs 230lbs. And asked if a full size semi could get down my road?
I'm always a little hesitant to order aluminum extrusion because its not cheap. I've used it in a couple machine builds now and you cant deny its straight, strong and goes together easy. 8020 makes about every conceivable fastener to bolt it together. The rack and the linear bearings bolt right into the T-slot's. The prices are reasonable but all the hardware and machining and the shipping adds to the overall cost and the final amount is always more than anticipated. But for a CNC router its a bargain at twice the price!
After I opened it and made sure it was all there, I bolted together just the outside parts of the frame and assembled them on the machine base. I centered the frame on the base and marked the inside and the outside edge on the top of the base. This was mainly to make sure my epoxy pour was sufficiently wide enough. But also to locate the frame edge so some anchors could be welded to the base and attach to the frame.
Mike from Precision Epoxy got me the quote for the epoxy. For 1.2gallons of his steel casting epoxy, a mixer and a couple gallon mixing pails and shipping it was $234. Thats about twice what I had thought it would be. There might be cheaper epoxies but it would be a gamble on my part to go try something else and this CNC isnt for me so its got to be right.
I was kind of on the fence about using the self leveling epoxy for the base. I thought we had welded it pretty flat and if there was some gaps I could shim them up level. But as soon as I put the extrusion on top of the base and got down and eyeballed the gap I was glad I decided on the self leveling epoxy. There were 3 points where the frame contacted the base. The rest of the perimeter had about a 1/32'nd to a 1/16th gap. I plan on pouring the epoxy to 1/4" thick so this should level out nicely.
Last edited by 1Jumper10; 07-13-2016 at 07:20 AM.
Reason: clarity
With the frame having arrived I got started working on the parts that would anchor the frame to the base. I used some 2" x3/16 wall angle iron and cut up a dozen 2" pieces. I measured where the hole should be after the frame was sitting on a 1/4" epoxy bed and drilled out a 21/64 hole. This was just a little bigger that the normal mounting hole in the 40 series extrusion and should me some extra clearance to account for minor level variation after the epoxy goes down. It didnt take long with the sawzall and drill press.
Next I started pre-fitting the cross pieces in the frame. I just pre-fit them at first since I was planning on disassembling it once the anchors were in place. But the double anchor fasteners were such a pain to assemble that I decided once it was together, it wasnt coming back apart. Normal 8020 double anchor fasteners go together easily but mine were special I had a bunch of 8020 parts on hand already from a project that I did about 15 yrs ago. But they were for 15 series(Imperial) and I was using 40 series (metric). I checked with the engineering firm that distributes 8020 here in my part of Michigan to see if this would work. They said it would as long as the 40 series extrusion was counter bored for the 15 series anchors. So all my 40 series extrusion came counter bored for the 15 series anchor. It worked. Barely. The double anchor nut doesn't provide enough clearance center to center and causes the threads of the cap screw to impinge on the center parts of the extrusion. This causes the screws to bind. I had to tighten each anchor about a 1/4 turn at a time. All 48 of them. This slowed me down but it was good because I remembered to slide in T nuts for my anchors before I had the cross pieces in
After the cross pieces were in and tight I double and triple checked the frame for square. It was dead on. Thats the beauty of the double anchor fastener and 8020 extrusion: as long as the factory makes square, accurate cuts, when you tighten it down it will be square.
Once I was happy with the frame geometry I started laying out the frame-to-base anchors.
I clamped the frame securely to the base and checked the diagonals again for square. Still good. I laid out the 12 anchors evenly along the frame. When they were where I wanted them, I C-clamped them lightly to the extrusion and they were then welded carefully to the base. We worked in a criss-cross pattern to minimize any chance that we might pull the frame in some unwanted direction with the welds. We used a stick welder for this since its easier to get better penetration with less heat. We also shielded the the aluminum from the weld splatter as best we could. It got a little smoke on it but it should clean off easily. I didnt take any pics of the weld process but here are a couple after the anchors were on and the frame removed. Got some extra helpers and we lifted the frame off the base and set it aside. I thought it might be hard to get off the base but it wasnt. It was a tight fit and it had to be lifted off evenly or it would bind but it came right off. There was no play at all when we were welding them on and I thought I might have to pry it off or loosen the anchors but we didnt. Very happy with how it fit. Snug with no stress.
Tomorrow I'll start building the dam on the base for the epoxy. The anchors will have to be worked around for this but I think that having the anchors fastened/welded directly to the base is better than fastening them through a layer of epoxy.
Also I finally heard back from the machine shop on the gantry risers: $460 for the set and 3 weeks to deliver Going to have to see if I the budget is there. I'm over on some things but under on others.
Got out some Bondo today and filled in around each anchor. They are not welded completely around and I was afraid that the epoxy would leak out under them. So i sealed them up. Sanded. And added more bonded and sanded some more. Which was completely unnecessary but somehow seems mandatory when you start sanding body filler ???
Im still trying to come up with a dam to contain the epoxy. I've thought about modeling clay, weather strip, spray foam, Gorilla glue, silicone-corn starch putty... but I cant decide which would work best? At this point I'm kinda leaning toward spray foam and fiber board.
I found blue painters masking tape to be excellent. Just get wide stuff (2" or more) and tape it along with a free top edge to contain the epoxy. Stiff enough to hold shape, epoxy doesn't stick to it.
I tried other tapes, such as gaffer tape, but they didn't hold their shape. I also used wood, but you want something that epoxy doesn't stick to well, so usually need some coating.
I strongly recommend using a fast setting epoxy to seal your dam. Just go around all the joints and put a bit on. If you don't, all your expensive slow curing epoxy will end up on the floor fast.
I wish I could just tape around the perimeter but the steel tubing has a waterfall edge that I want to stay back from. This wont be possible on the outside of the base but I can manage it on the inside. And I was thinking about priming the form with regular epoxy to seal up any holes in the dam. Its probably a good idea and I've seen it mentioned by at least one epoxy manufacturer so I probably will. If I just keep it around the joints and the anchors and at the bottom of the waterfall it wont affect the flatness on top or the surface area for bonding of the main epoxy. I sandblasted it again yesterday so I'm confident I've got a good rough surface to bond to. Thank you for your input Pippin88. This has been a pretty one sided conversation so far
Re: 2nd build - For a friend
Reply with Quote
I had a bunch of 8020 parts on hand already from a project that I did about 15 yrs ago. But they were for 15 series(Imperial) and I was using 40 series (metric). I checked with the engineering firm that distributes 8020 here in my part of Michigan to see if this would work. They said it would as long as the 40 series extrusion was counter bored for the 15 series anchors. So all my 40 series extrusion came counter bored for the 15 series anchor. It worked. Barely. The double anchor nut doesn't provide enough clearance center to center and causes the threads of the cap screw to impinge on the center parts of the extrusion. This causes the screws to bind. I had to tighten each anchor about a 1/4 turn at a time. All 48 of them. This slowed me down but it was good because I remembered to slide in T nuts for my anchors before I had the cross pieces in 
Going to have to see if I the budget is there. I'm over on some things but under on others.
??? 
