GME's New 80/20 CNC Build - My Design

I have started a new build of a 4 x 4 CNC router. First a little history.

Technically, this is my third CNC, but my second build. My first, a 4 x 4, was an 80/20 extrusion kit from Fineline Automation. Not a bad machine, but it was a nightmare to assemble, and did not have the features I later learned I wanted. The assembly part was mainly because of tolerances that were too loose in the extrusions (more on that later). I ran it for about two years, and decided to buy my second machine, a Saturn 2.

The Saturn 2 is a welded steel machine. It looked great and held a lot of promise in terms of accuracy, repeatability and rigidity. It was advertised as jig welded and stress relieved. After ordering it, I discovered other owners with some fairly dramatic negative issues with their machines. When mine arrived, shims between the gantry mounting and interface plates were a dead giveaway that something was very wrong. I won’t go into the ugly details, but suffice it to say that pretty much every thing that could be off was off by a lot and it would take a good deal of time, effort and expense to accommodate its defective construction. For anyone who is interested, you can find my thread about it. https://www.cnczone.com/forums/finel...cnc-posts.html I elected to return it. To Fineline’s credit, the owner took it back without resistance. So, the Saturn 2 was my 2^nd machine, but I didn’t get it set up for operation.

Now, I’m building my 3^rd machine, which merges what I consider the best of several different designs. Although I have MIG and TIG welders and welding skills, I decided not to go with steel tubing. Many will criticize my choice, but cost and logistics were a factor. Welded steel would require stress relieving and milling on a large format CNC mill. In my neck of the woods, the machine shop costs are considerable. As far as logistics, my shop is in my home in a daylight basement. Ingress and egress for large equipment, while doable (I have a cabinet and welding shop worth of heavy tools), is very difficult. I did it with the Saturn 2, but it wasn’t something I would want to repeat often. The problem is more about the physical dimensions than weight.

I will not be covering the electronics in this build. I sourced and built my own for my first machine, and it has worked well. For anyone who might the interested, here’s a rundown on the control side:

PDMX 126
PDMX 107 (Spindle control sister board)
ESS SmoothStepper
Gecko 203V stepper drives
Mach4

I’m using NEMA 34s for X Y Z and A (Y slave).


A note about 80/20 extrusions. While not widely known, 80/20 has cut and length tolerances for its extrusions. The cut tolerance is +/- .002” per inch and the length tolerance is +/- .015. Surprisingly, I’ve not seen any mention of the tolerances in the kit machine literature or any discussion about compensating. I’ve purchased my share of 80/20 over the years and have never found an end that was cut square. I’ve also found that there is variation from piece to piece in the same order, and pieces cut outside of the published tolerance. Some pieces that came with my first CNC kit were off as much as .006” in 1.5” (1530 profiles). Can’t find the tolerance on the 80/20 site? Look at the online catalog, page 850 in Machining Services. Not exactly easy to find.

Tolerances matter. A .002” per inch deviation translates to .136” (more that 1/8”) at the other side of a 68” run. 68” is the length of my cross pieces. Not good enough for my taste. You can’t force the extrusions to square across the width of the side frame, and the lengths varied some, which cause issues when aligning linear rails. I went with shims on my first machine and decided to not go that route again.

Oh, one other thing. Trimming the ends square on my chop saw with a non-ferrous blade was not an option. Unless you have full blade engagement, you get blade deflection. Been there and done that. I’ve even tried blade stiffeners. Helped a little, but still worse than .002”/inch. Makes sense, if you consider that at 5/8” diameter X 4” cut length end mill will deflect when cutting aluminum. If something that thick will deflect, it’s a no brainer that a 3/32” thick circular saw blade will also deflect.


The build:

I wanted a heavier and stiffer machine than the typical 80/20 kit. So, nearly the whole machine is constructed of 3030 (3”x3”) profiles. They are the thickest/heaviest version. The exceptions are the angle bracing for the legs, and the gantry, which is 40-8016 (80mm x 160mm) the metric equivalent of 3060 (3” x 6”). I used metric for the gantry to accommodate the use of a CNC Router Parts’ Pro Ballscrew Z Axis. Although I’m building some of the parts, I decide to buy the Z axis.

Here is a photo of some, but not all, of the 80/20 for this build. Note the blue tape. I took measurements and recorded them on the tape. More on this later.

https://www.cnczone.com/forums/attac...d=418516&stc=1

Dimensions: The Y axis rails are 60” long. The cross pieces (what the spoilboard mounts to) are 68”. There are actually 2 lengths of Y axis rail, which are separated by 5” uprights. The second rail and uprights give the gantry height. This arrangement is different than the side plates commonly used on a variety of machines, including CNCRP Pro model.

My first machine had skate bearing carriages that rode on a steel plate on all three axes. The arrangement is similar to the CNCRP Standard kits. They were just okay for a hobby machine. They worked, but I wanted better. So, my Y axis will have Hiwin 25mm linear rails and the X axis will travel on Hiwin 20mm linear rails.

Okay, what to do about the 80/20 tolerances? My trusty PM 25MV benchtop mill and lots of time. (I didn’t have the mill for my first build). I installed glass scales and DRO on my mill. For anyone contemplating milling extrusion ends and looking for precision, I highly recommend a DRO setup or CNC mill. I only do hobby machinist work and have no formal training. I can say without reservation that I could not have pulled off this build without the DRO and tuning the mill (e.g., tramming side-to-side and front-to-back, adjusting gibs in all 3 axes, oiling and greasing). A mister also helped.

Here is a picture of the 80/20 support stand I built to facilitate milling the ends of the long extrusions. On top of the horizontal is a strip of UHMW, which is very slick. Since cutting the extrusion ends involves moving the extrusions, I needed something minimal resistance. I had some UHMW laying around; it worked great. Really slick stuff.

https://www.cnczone.com/forums/attac...d=418534&stc=1

Maybe you wanted to say a 80x160 for the gantry? 40x160 would be equal to a 1.5"x6.0".

Quote:

---

Originally Posted by davida1234

Maybe you wanted to say a 80x160 for the gantry? 40x160 would be equal to a 1.5"x6.0".

---

Yup. 80 x 160. Good catch. Thanks.:)

The starting point for milling was ensuring my mill was perfectly level, as measured on the bed of my trammed vise. I used a machinist’s precision level, and got it spot on. Then, I put a length of 80/20 in the vise and supported by the stand on the other. I used my machinist level to level the stand to the vise and perpendicular to the vise (stand leveled in all directions). It was a fussy and time-consuming process, but necessary for high quality results.

Now, it’s time to mill. I started by coating the extrusion ends with Steel Blue marking fluid. I used it to monitor the cutting process so I didn’t remove any more material than necessary. I used a 2 flute. 4” cut length. HSS end mill. HSS tends to be sharper than carbide and sometimes works better than carbide on aluminum. This may make machinists cringe, but after removing the marking fluid, I made 2 more passes at the same setting to ensure the end was cut square. Long end mills defect some. The extra two passes removed the slight additional material occasioned by end mill deflection. Yes, I know this causes rubbing and can be tough on tool life. However, it was that or accept that the ends would not be as square as possible. I went for square.

Here are some photos of a 1530 angle brace vise showing the marking fluid and its being milled off.

https://www.cnczone.com/forums/attac...d=418538&stc=1

https://www.cnczone.com/forums/attac...d=418536&stc=1

https://www.cnczone.com/forums/attac...d=418540&stc=1

You get a sense of how far off from 45 degrees the ends were. Note the makeshift work holding with angle plate and 1 2 3 blocks. I set my vice up at 45 degrees, but the body of the vice was in the way of cutting. I could have used the vise, but I would have needed to extend the extrusion out from the vise farther than I liked. Setting up the angle plate to 45 degrees worked fine.

After getting the ends milled square, it was time to measure to get all like pieces the same length. I made comparative measurements to determine which piece was shortest. Then, I used the shortest piece as the part to measure against to determine how much needed to be removed from the rest. On the long pieces, I put an outrigger on my support stand and added a hard stop. That way, I could zero my DRO to the desired ultimate length and cut each piece until I reached zero. Sounds good in theory, but it didn’t work in practice. I found that the stop flexed, and no matter how careful I was placing the stock in the vise and moving to the stop, the results were inconsistent. Failing the use of a hard stop, I resorted to cutting the measured amount each piece was too long. This is where the blue tape came in. I wrote on the tape how many thousands need to be removed for each piece.

Assembly next. I did not take photos until I got the legs and braces on and cross pieces going in.

Here is a photo of the frame with the first cross piece in place. Linear rails, the gantry and stock for gantry mounts/interface plates are on the floor under the frame.

https://www.cnczone.com/forums/attac...d=418542&stc=1


Here is a closeup photo showing some aluminum angle temporarily bolted to the extrusions. I used the angle because there is quite a bit of wiggle room when assembling 80/20. The angle ensures that sides and tops align properly. 1. Attach the angle, 2. tighten the connectors, 3. remove the angle, and 4. final tightening of the connectors.

https://www.cnczone.com/forums/attac...d=418544&stc=1


Here is a photo with a Woodpeckers precision 45-45-90 triangle clamped to the extrusions. I clamped it for photo taking purposes. I was gratified to find that everything went together perfectly square. Square ends make square frames.

https://www.cnczone.com/forums/attac...d=418546&stc=1

You will note the angle supports hanging out in the air. I had to put them in place on the legs before adding connecting pieces. I quickly found that if not perfectly positioned, they will cause the upper frame members to flex up or down, if only by a few thousands. I decided to get the frame assembled and then
finalize with connecting the braces.

Here are additional pictures as the frame went together.

https://www.cnczone.com/forums/attac...d=418548&stc=1https://www.cnczone.com/forums/attac...d=418550&stc=1

I forgot to mention the hard stop I made to get the short piece to precisely the same length. Here's a photo:

https://www.cnczone.com/forums/attac...d=418552&stc=1

You may have noticed that the cross rail in the front is set back some. I did that to allow me to cut dovetails off the front of the machine. I took measurements from the Z axis assembly and spindle. I determined that the spindle will stop past the the front most extreme of the frame, but not enough to comfortably cut dovetails. The setback was to provide 2" of cushion. Here is a photo of installing the rail with setback.

https://www.cnczone.com/forums/attac...d=418554&stc=1

After installing the cross rails, I cross measured with a tape measure to ensure that the frame was square. It was as perfect as I could get reading off a tape measure. No adjustments required.

I then proceeded with installing the risers and top Y axis rails. Here is a photo of the frame:

https://www.cnczone.com/forums/attac...d=418556&stc=1

I cross measured again. Still as perfect as I could have hoped for.

Next up was doing the final leveling of the machine. In the photo above, you will note that the angle braces are installed. I fussed and fussed with the machinist level, but could get it level. I thought I would have to tear the whole thing down and start over when it occurred to me the angle braces might be the source of they problem. Turns out they were. I loosened them all up, and went back to leveling. I was within .0005" all around with the braces loosened. Then, taking the utmost care with positioning, I reinstalled the braces. It worked. My leveling held.

The exercise told me that I got my lengths right and the ends square on the mill. and I avoided any twist during assembly. Besides taking great pains to get things right, using something as simple as the angle to get the extrusions aligned paid off. That the extrusion ends were square was not in question. I measured each end for square before it left the mill. I also confirmed lengths before I started assembly.

Next up: Gantry interface plates.

Gary

It’s fun to see a perfectionist at work. That knows what he wants and how to get there. I especially like the dropped bed. That way your not mounting the X axis extrusion up on a high perch swaying in the wind. That was my approach too. Except I’m a hobbyist wood worker so my base was wood, and the table was a torsion box design 5” tall and on a 6” grid made out of 3/4” MDF, it is HEAVY.

Looking forward to more phots and descriptions

Hager

Gary on my supporting table at the bottom I have 4 pieces? Is that your plan also, so the Y bottom and X bottom are all supporting and bracing that area?

Quote:

---

Originally Posted by wmgeorge

Gary on my supporting table at the bottom I have 4 pieces? Is that your plan also, so the Y bottom and X bottom are all supporting and bracing that area?

---

Bill,

I'm not sure I understand your question. Are you referring to the leg set? If so, the answer is no. I am relying on the angle braces to provide rigidity. As far as I can tell at this point, they will do they will do the job. If you look at the CNCRP Pro 4 x 4 machine, the leg design is similar to theirs. They have a fore to aft extrusion one one side, to accommodate hanging the control and VFD boxes. I have my boxes on casters and don't plan to mount them, so I decided to forego buying an extra length of extrusion. The 3030 extrusions are pretty expensive, so I saved some money. I hope I don't regret the decision.

Gary

https://www.cnczone.com/forums/attac...d=418600&stc=1



The gantry interface plates and gantry support are next. Here's the design. The thinner plate is 3/8" thick It mounts to the underside of the gantry with 6 bolts that are independent of the lower plate. The angle plate assembly also mounts to the 3/8" plate from the underside. I changed the original design, which called for a 3/4" upper plate. 3/4" seemed like overkill. The through holes on the horizontal angle plate piece will probably not have through holes. In the original design, the support plate bolted to what is now a thinner plate. However, 3/8" is too thin to bolt into. So, I'll be bolting the support plate from the underside of the 3/8" plate into the 1" x 2" support piece using blind threaded holes. I hope my description makes sense.

Note that I've used slotted hole in the 3/8" plate. This is to allow me to readily adjust the gantry for square. The 3/4" plate has slotted holes for in it where it attaches through the 3/8" plate and into the underside of the gantry.

Gary

Yup you want good bracing, this is my first extrusion CNC, but I went way overboard with 1/2” 6061 bracing. But it’s still solid, LOL

Quote:

---

Originally Posted by GME

Bill,

I'm not sure I understand your question. Are you referring to the leg set? If so, the answer is no. I am relying on the angle braces to provide rigidity. As far as I can tell at this point, they will do they will do the job. If you look at the CNCRP Pro 4 x 4 machine, the leg design is similar to theirs. They have a fore to aft extrusion one one side, to accommodate hanging the control and VFD boxes. I have my boxes on casters and don't plan to mount them, so I decided to forego buying an extra length of extrusion. The 3030 extrusions are pretty expensive, so I saved some money. I hope I don't regret the decision.

Gary

---

I mean like what the Saturn 2 and CAMaster have the lower leg set of braces,,,, I think you will find the movement of the gantry will flex that aluminum used for legs, even below the diagonal bracing. This machine is made from heavy steel and welded.


Attachment 418608.

Quote:

---

Originally Posted by wmgeorge

I mean like what the Saturn 2 and CAMaster have the lower leg set of braces,,,, I think you will find the movement of the gantry will flex that aluminum used for legs, even below the diagonal bracing. This machine is made from heavy steel and welded.


https://www.cnczone.com/forums/attac...d=418608&stc=1.

---

I understand. You may be right (but I hope not). I don't have the additional extrusions at this point, and if I did, so I'll give it a go with what I have. If it works okay, great. If not, I'll have to either take the whole thing apart, or jack the frame up, remove the leveler mounts, and fit the new pieces in. I'm hoping the angle braces work, like they do on the CNCRP legs. I suppose another option would be to fabricate a second set of angle braces that are longer. I would only use them on the sides without the cross pieces. I could use roll in t-nuts, which I have on hand, If I were to go that way, I would not have to take the machine apart or jack it up. I'll be guided by how the machine behaves in operation.

Thank you for the suggestion.

Gary

Gary it might not be an issue with your light weight aluminum gantry, but I know when mine is running fast I am glad for all the bracing I can get. Maybe if its needed, IF its needed you could attach the lower brace rail with aluminum plates to the side. Looks very nice so far.

80160 is not exactly lightweight.

Quote:

---

Originally Posted by ger21

80160 is not exactly lightweight.

---

+1. Not nearly as heavy as the Saturn 2's steel gantry. I needed my engine hoist to lift it. At least I can manhandle the 80160 without a hoist, or better still, ask my wife to help.

Quote:

---

Originally Posted by wmgeorge

Gary it might not be an issue with your light weight aluminum gantry, but I know when mine is running fast I am glad for all the bracing I can get. Maybe if its needed, IF its needed you could attach the lower brace rail with aluminum plates to the side. Looks very nice so far.

---

Thanks, Bill. You're right. I could attach a brace with side plates. However, to do so would require putting the extrusions either higher or lower than the existing ones. I could fit a side plate on the outside, but the existing cross brace would prevent adding a plate in the inside. I suppose I could add some aluminum angle on the inside, but I'd have to mill it to get the faces square. Every time I buy aluminum angle, it's always way out of square.

Gary

Looking good Gary!

Also if you're worried about the legs you could always bolt a piece of MDF to the base to stiffen it up and you'll have a nice shelf to use for storage under the table.

Looking forward to the rest of the build,
Dan

If the lower leg braces being referred to are the horizontal pieces (one with cable drag chain on it,) then those do almost nothing for stability/strength. They just keep the legs parallel. If the legs are anchored down (even if it's just friction with a rubber pad,) they really aren't necessary. You really need triangular structures for strength. The gussets in GME's build look like they add more strength than those cross pieces on the yellow one.

Quote:

---

Originally Posted by skrubol

If the lower leg braces being referred to are the horizontal pieces (one with cable drag chain on it,) then those do almost nothing for stability/strength. They just keep the legs parallel. If the legs are anchored down (even if it's just friction with a rubber pad,) they really aren't necessary. You really need triangular structures for strength. The gussets in GME's build look like they add more strength than those cross pieces on the yellow one.

---

I like your answer better. Not sure what you are referring to when you said, "one with the cable drag chain on it." I don't have a drag chain mounted yet, but when I mount it, it will be on one of the upper front-to-back extrusions.

Gary

Quote:

---

Originally Posted by GME

I like your answer better. Not sure what you are referring to when you said, "one with the cable drag chain on it." I don't have a drag chain mounted yet, but when I mount it, it will be on one of the upper front-to-back extrusions.

Gary

---

I was referring to the photo of the yellow router.

Got it. Thank you. I appreciate your weighing in.

Gary

Quote:

---

Originally Posted by skrubol

I was referring to the photo of the yellow router.

---

The "Yellow Router" or CAMaster is a very well made and selling commercial duty unit. Not a DIY or hobby model. If those legs keep the other parallel I would think that would be a big deal! Most commercial routers have those and are welded steel. Look whats on the market.

Quote:

---

Originally Posted by wmgeorge

The "Yellow Router" or CAMaster is a very well made and selling commercial duty unit. Not a DIY or hobby model. If those legs keep the other parallel I would think that would be a big deal! Most commercial routers have those and are welded steel. Look whats on the market.

---

It's most likely to prevent the router from getting damaged during transport and to help keep things square while (and after) welding. Ask any mech engineer, rectangles are weak, triangles are strong.

Quote:

---

Originally Posted by wmgeorge

The "Yellow Router" or CAMaster is a very well made and selling commercial duty unit. Not a DIY or hobby model. If those legs keep the other parallel I would think that would be a big deal! Most commercial routers have those and are welded steel. Look whats on the market.

---

You could argue that the only reason it works is because it's welded steel. Parallelograms are nowhere near as rigid as triangles.

I think CAMaster has it pretty well figured out by now, after how many years?