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

[nlancaster]

Thanks for the feedback on the EG/HG GME!

[peteeng]

Hi All - Preload does not change the overall stiffness of the bearing. It does change the initial stiffness. Instead of thinking about preload think of it as internal clearance. For instance in ball or roller bearings they don't describe the preload of the bearing but the clearance of the bearing (although you can get negative clearance ball bearings for exactly the same reasons we use them on routers/mills). So if the bearing has a small clearance when the load comes on it wobbles or moves a small amount (or if its a conrod bearing for instance it would knock) before it stiffens up, especially if the load is changing sides. A zero clearance bearing starts life as zero but due to wear becomes a clearance bearing at some point. If the application requires positional accuracy this initial wobble is not good. If it's just transferring product down a line then clearance or zero clearance is fine. Routers and mills need accuracy so use medium or heavy preload. These do not wobble or knock as the load changes directions. Plus in the case of impacts a clearance bearing will knock and consequently brinell very quickly.

The consequence of preloaded bearings is that their friction coefficients are greater and potentially they can wear faster. So keep the lubrication up on them. From experience the light preload bearings used in twins or quads start life very stiff and you cannot feel any clearance in them when new. But after a couple of years you can feel the clearance in them if you seperate them out from the machine so you can wobble them individually, especially in the "roll" direction (across the rail). So med or heavy is best I think if you can get them. On my latest router I have used heavies so will be able to tell you how they go in a couple of years!!

Cheers Peter

Just to clarify bearing orientations. They are roll, pitch and yaw just like a plane...

[GME]

Hi All - Preload does not change the overall stiffness of the bearing. It does change the initial stiffness. Instead of thinking about preload think of it as internal clearance. For instance in ball or roller bearings they don't describe the preload of the bearing but the clearance of the bearing (although you can get negative clearance ball bearings for exactly the same reasons we use them on routers/mills). So if the bearing has a small clearance when the load comes on it wobbles or moves a small amount (or if its a conrod bearing for instance it would knock) before it stiffens up, especially if the load is changing sides. A zero clearance bearing starts life as zero but due to wear becomes a clearance bearing at some point. If the application requires positional accuracy this initial wobble is not good. If it's just transferring product down a line then clearance or zero clearance is fine. Routers and mills need accuracy so use medium or heavy preload. These do not wobble or knock as the load changes directions. Plus in the case of impacts a clearance bearing will knock and consequently brinell very quickly.

The consequence of preloaded bearings is that their friction coefficients are greater and potentially they can wear faster. So keep the lubrication up on them. From experience the light preload bearings used in twins or quads start life very stiff and you cannot feel any clearance in them when new. But after a couple of years you can feel the clearance in them if you seperate them out from the machine so you can wobble them individually, especially in the "roll" direction (across the rail). So med or heavy is best I think if you can get them. On my latest router I have used heavies so will be able to tell you how they go in a couple of years!!

Cheers Peter

Thank you, Peter. Appreciate the information.

Gary

[peteeng]

As an aside- The way the preload is achieved is by using bigger balls. The track and cars are all the same within the precision group. So technically you should be able to buy the correct balls and swap them out for the greater preload and or to upgrade a worn car. The balls should wear first. There has been a comment also about fitting the cars with heavy preload and I have not found that to be difficult. Also some have commented on stiff cars initially and I think the seals have something to do with that. They seem to ease up after a few runs. I use paraffin on the rails vs oil. Solvent based furniture wax is great. Mr Sheen used to be good but they have changed to a water based wax so it rusts the rails.... Being a solid it does not attract dust like oil does. Some of the synthetic motorcycle chain greases are excellent for the same reason. Peter

see attached for bearing load directions

[HuguesP]

the friction coefficient would still be the same, as the Materials don’t change. a preload as the name says it, act on the perpendicular force, so the force required to move is bigger F= fcoef * perp force.

[peteeng]

Hi Hugues - Yes you are right after thinking about it. The friction coefficient stays the about the same but the preload has to be added to the operational loads and this affects the life of the bearing. Take an MSA 20mm car with light preload. Its C=19kN. Its max preload is 0.02C so its 0.38kN or 39kgf. The heavy is 0.08C so its preload is 0.08/0.02*39= 4x the light eg 155kgf. These need to be considered in the cars life calcs. Ta Peter

[davida1234]

the preload has to be added to the operational loads and this affects the life of the bearing.

Hi Peter,
I think the operational force (the force to move a given bearing) stays within the bearing and only effects the driveline components.

[wmgeorge]

GME - Gary hows it going? Are you about ready to flip the switch. BTW thanks again for the help on Mach4, my machine has never ran so well!

[GME]

I'm way overdue with an update. When I last posted, I was moving on to aligning the secondary Hiwin Y axis rail. First attempts did nothing but frustrate me. Here is a picture of what I tried first. It failed miserably:



I took a piece of 80/20 extrusion, fastened to the interface plate on the other side of the machine and rested on the interface plate shown in the photo. The dial test indicator was placed to align the linear rail parallel to the other one. After chasing may tail for an hour or so, it occurred to me that the extrusion I used might be flexing, thus throwing off my measurements in a significant way. Boy, was it ever flexing. Just moving from one mounting screw to the next caused the extrusion to move way out of position. I tried clamping it to both interface plates, which helped with the flexing, but created a different problem. If the linear rail wasn't perfectly parallel on the master rail when I clamped the extrusion, the secondary rail would go parallel. I fussed and fussed and fussed with it, and finally came to the conclusion that the what I was trying was never going to get me even close.

I did finally come up with a solution that works, but don't have any photos of it. You start out by making a triangle that spans the distance between the two linear rails. It needs to rest of the horizontal surfaces of the extrusions, but still be slightly shorter than the distance between the linear rails. Since I mounted my linear rails on the inside slots, it doesn't leave much room. The triangle angles don't matter. You don't need a 90 degree angle. What you do need is a straight edge to rest against the master rail. You also need perfectly rigid connections at each end of the hypotenuse and at the 90, or near 90 degree angle segments. Hold the triangle tight against the master rail, mount the dial indicator on the triangle on the secondary side, and screw the rail down from one end to the other. The triangle will allow you to get a bit over half way. Then you flip the triangle along the master rail edge, rezero the indicator and finish the job.

It took me days to get it all worked out and the rail set - I didn't have the luxury of working full time on it - my wife had an endless list of outdoor projects. Suffice it to say, it worked like a charm.

Note: I realize that pictures make all the difference in explaining what's going on. Written descriptions can be tough when trying convey the idea. So, if I've left more questions than answers, fire away. I'll do my best to make it clearer.sd

Gary

[GME]

Mounting the Gantry. Dan (NTL), this one's for you. I believe you were anxious to see what she looks like with the gantry in place. Here are a couple of photos:





Please pardon the mess! And no, I didn't use the dead blow mallet on the machine anywhere. I don't know why it's there.

I have to say, the gantry went on without a hitch. Everything lined up as it should I did discover an unplanned for, but easily fixable issue with the mounting bolt that go into the extrusion on the outside. While I had enough room to a hex key into either a cap screw or button head cap screw, it took to much effort to loosen/tighten them. I ended up using garden variety hex bolts, which I could easily turn with a wrench. Problem solved!

The photos highlight how rigid a connection I have between the interface plates and the gantry extrusion. Rock solid, yet adjustable. As you will recall from an earlier post, used slotted holes in the interface plates to give me plenty of adjustment for squaring the gantry. So far, I'm very happy with how it has gone together.

The astute observer will note that the profile linear rails are just a bit longer than the extrusions they are mounted to. I realized this when I order the rails. My supplier offered them in set lengths only. So I had to either cut them myself, or let them hang over a bit and accommodate it when I built my hard stops. I decided to go with the latter.

Gary

[GME]

Here's a quick shot of the CNCRP/Avid Pro Ballscrew Z axis I purchased. I bought the 12, which allows me to cut pieces nearly reaching the bottom the the gantry using long bits. Is the combination likely to occur? Probably not, but better get it now, rather than needing it and not having it later.



Gary

[GME]

I mentioned earlier about putting in the hard stops on the Y axis. Here are a couple of photos showing what they look like:







I used to pieces of aluminum angle for the job. One mounts to threaded holes in the ends of the extrusion. The other mounts to the first as shown. I milled the angle to fit and not look too thrown together. The way I have it set up, the bearing just reach the end of the linear rail when contacting the rubber bumper. This means I have pretty much every fraction of an inch for movement that the rails allow.

Had I to do it over again, I would have used longer Y axis extrusion. The ones I used are 60". If I were doing it over, I would have used something like 66" extrusions. Turns out to be a design flaw I had to accommodate. All-in-all, it came out okay. I ended up with almost 54" of travel, which, not coincidentally, is the length of my racks.



The above photo shows my hard stops on the X axis. Nothing special about it. I cut the width so that it the Z carriage passes over the mounted portion with plenty of clearance.

See that odd bit on top of the extrusion in the first photo? it's my proximity sensor flag. I mounted an 1/8" thick piece of steel flat stock to a milled chunk of aluminum, and mounted both to the extrusion. I went with a steel sensing flag after doing a bit if reading. First off, proximity sensors sense steel at their max distance. I liked the idea of having more clearance. What I found surprising was that thinner steel is better than thicker. I have no idea why thinner is better, but I'll accept it. If I recall correctly, the specs for sensors are based on 2mm steel thickness. We'll see, when I test it out.




The next two photos above show my proximity sensor arrangement for the Z axis. In the first photo, you can see the steel angle flag. I used 2 pieces of angle for it. One mounts to the rear of the fixed part of Z, The other is welded to the first. To ensure that the piece remained flat, I clamped them together and used my TIG to spot weld them together at the ends only. Just a couple of spots on the top and the same on the bottom. Too me way more time getting my torch, control and ground lead unwound than it took to weld. Ditto for putting them away.

The second photo immediately above should my sensor mount. Nothing special. It's a piece of 3/8" aluminum plate that is threaded for the sensor. I may redo it. I can recut/weld a new flag that moves the sensing surface much closer to the front. Then, I can have sensor threads on both sides of the holes and use locking nuts on both sides.

Just a quick comment about the Z axis sensor. The CNCRP/Avid Z axis design calls for mounting the sensor in the top of the axis and picking up the top of the moving part of the axis as the sensing surface. I prefer a bypass sensing system, so I made one. Not knocking the AVID design. It is certainly cleaner looking than mine. Then again, I am more interested in function than looks. Also, FWIW, I can adjust how high my Z axis goes when homing. No sensing moving the full 12", if I don't need to. If I need more height, I can just move the flag/senor arrangement to accommodate it. Much more flexible.

Gary

[GME]

GME - Gary hows it going? Are you about ready to flip the switch. BTW thanks again for the help on Mach4, my machine has never ran so well!

Bill, see above updates. I am replacing all my motor wiring with better cable. My stepper cable had mylar shielding only. Works, but not the best. I'm replacing it with braided shield. Finding cable with a small enough diameter to fit an XLR plug as been a real trip. I wanted to use IGUS CF6, but its diameter is too large. I am also changing over to CF6 for my VFD to spindle cabling. Although overkill, I've gone with 14 awg instead of the usual 16 awg for 2.2kw spindle. I previously changed out my connector and it will handle 14.

Just curious, but is the CNCRP/Avid stepper motor cable marked with mfg and model number? They also use an XLR plug. I'm wonder what they found that fits the connector. They advertise braided shielding and 18 awg wire, which is what I was shopping for.

just as soon as my stepper cable arrives and I wire the plugs, I'll be flipping the switch. (I already have the CF6 for the spindle).

Glad to hear you like using Mach4. Same experience as I've had with. Never looking back.

Gary

[ger21]

While not specifically VFD cable, here's some 18ga that's .39" dia.
https://www.wireandcableyourway.com/...ex-190-cy.html

[GME]

While not specifically VFD cable, here's some 18ga that's .39" dia.
https://www.wireandcableyourway.com/...ex-190-cy.html

Thanks, Gerry. The cable you linked to seems very good quality. However, at .39" in diameter, it won't come even close to fitting into an XLR connector.

I probably wasn't clear. I have the VFD cable. Igus CF6, 14 gauge. My problem was cable for the steppers. Most all of it with a braided shield is in the .31"+ diameter range - too much diameter for an XLR plug. You can cut the end off the rubber boot on the connector to get it in, but it's too big for the internal plastic clamp. The connector won't screw closed. Guess how I know. The mylar shield stuff is smaller and fits very well, but not great for continuous motion, or so I've read. It looks like CNCRP/Avid found a source that fits, but I'm not finding it.

[ger21]

Igus CF881?

I don't think you really need to use shielded cables for steppers.
I'd use Igus CF880.

[wmgeorge]

Thanks, Gerry. The cable you linked to seems very good quality. However, at .39" in diameter, it won't come even close to fitting into an XLR connector.

I probably wasn't clear. I have the VFD cable. Igus CF6, 14 gauge. My problem was cable for the steppers. Most all of it with a braided shield is in the .31"+ diameter range - too much diameter for an XLR plug. You can cut the end off the rubber boot on the connector to get it in, but it's too big for the internal plastic clamp. The connector won't screw closed. Guess how I know. The mylar shield stuff is smaller and fits very well, but not great for continuous motion, or so I've read. It looks like CNCRP/Avid found a source that fits, but I'm not finding it.

The CNCRP Plug and Play wiring had connectors prewired at each end and was shielded. Those steppers only draw 3-4 amps or so and its not continuous and #18 wire is good for at least 5 amps.

My Hitachi VFD is rated at 11 amp or was set at 11 amps, #16 wire would carry that without issues as my spindle was 8 amps. Generally speaking 1 Hp on 240 3 phase would draw 3 amps per phase so 9 amps should be about 3 Hp. All general rule of thumb ratings.

[ger21]

Years ago, I recall the founder of Gecko saying that 20-22 awg is all you really need for steppers.
My 5 amp Nema 23's have 22awg leads on them.

[wmgeorge]

Ok from what I can tell its 18 gauge, 300 volt, 80 DegC cable and in one of the shots you can read the product number. You can download and enlarge as needed.

[GME]

Years ago, I recall the founder of Gecko saying that 20-22 awg is all you really need for steppers.
My 5 amp Nema 23's have 22awg leads on them.

Thank you, Gerry. My steppers are 7 amp NEMA 34s. I found some charts online. Looks like they agree with Gecko's founder; 20 awg would do it. That gets closer to a workable fit in XLR connectors.