[TiagoSantos]

Well, I've posted part of my story on other threads when asking questions or something, but it was about time I made a build thread, if anythIng to keep my crazy ideas in one place..

It all started about this time last year when I started seeing DIY 3d printers and mini routers all over the place. I've always been fascinated with automation and CNC in particular, so I was hooked right away! I started buying parts and making plans for a crazy router/printer combo machine and soon ran into a sieg X2 and a 7x12 lathe selling locally pretty cheap. "Well those could be useful.."

Not only were they useful, I fell in love with making chips! In a couple months, I sold the mini lathe and bought an old 12x40 Clausing, and kept looking for a bigger mill.

At the time, I was working for a big film post production company. They had a film lab and a small machine shop that supported and maintained the lab. By chance, I happened to tell one of the engineers that I was looking for a larger mill and he told me the lab was shutting down and they'd be selling everything off. A few weeks later he called me one morning and offered me their RF40 for $400. I couldn't pass on that, even if I didn't like the round column.. It came with a little bit of tooling and a big metal workbench.

But the best was yet to come! In the following weeks, they gave me hundreds of dollars worth of bearings, metal stock, motors, servos, power supplies, a VFD, taps and other random tools. It was like Christmas every couple of days, as they dug through their stuff and called me to come pick up the goodies.

At this point, I had a gecko g540 based controller that I had bought for my initial weird router plans and some 425oz steppers. I quickly got the machine running with the stock leadscrews and used the quill as my Z axis, with a pulley on the stock fine feed screw. I wanted to make sure everything worked before I spent a lot more time and money on ballscrews, etc.

I used the machine like that for a few months, and started making x and y ballnut mounts for the Chai eBay ballscrews, which were on order. But I wasn't happy with the limitations of having the quill as a Z axis, 5" of travel wasn't enough and I found myself having to crank the head up and re-zeroing quite often. I know, I know, poor planning on my part, but if definitely wasn't great.

So I started making crazy plans again, was looking around for a big piece of steel column I could weld and machine into a column, and I already had some linear rails I could use to mount the head to the new column. But in a strike of genius (or pure luck) I decided to email one of the RF45 vendors to ask if they happened to have a spare column around.. No, he didn't, why the hell would he have a spare column lying around.. BUT he did have a brand new RF45 that was almost complete but missing a table, motor and leadscrews.. And he'd sell it to me for $500. I said no, I didn't need an entire mach... Wait, did you say $500? Oh.. Well, I guess I COULD use another head, and maybe the bigger RF45 base wouldn't be a bad thing.. Ah! I'll take it! I had heard good things about Pat at Machine Tools Warehouse, but this was incredible. His service was great and I know I'll be buying more stuff from him in the future.

And so we arrive at the current state of things. The plan now is to get the RF40 working with the ballscrews (more on that in a minute) and use it to do some major surgery on the RF45. I'll be converting the 45 head to belt drive and mounting the 3phase motor/VFD combo and finding a way to stick a ballscrew inside the new square column. Then I'll move the square column to the current RF40 base.

The second part of the plan will be to use the RF45 base, which has significantly more Y travel. My current 40 with the ballnut in Y is limited to 6" of travel, although I plan to cut away some of the base webbing and hopefully get closer to 8 inches. The 45 base has at least 9", possibly more! There is a problem though. Since I don't have a table for the 45, I'll have to adapt something somehow. The saddle on the 45 is bigger, so the table doesn't fit the bigger saddle. And the smaller saddle from the 40 has a different dovetail angle, so it also won't work on the new base. My current crazy idea is to, at some point, mount the RF40 table on the 45 saddle with linear rails, rather than using the dovetails. I have radial style rails that would be perfect for it.. And I'm tempted to do the same for the saddle, but the base is perhaps a bit too big to machine at home.

Lastly, to end this really long, boring post - with the stock leadscrews I was only getting around 40IPM before the steppers would stall randomly. The 425oz steppers were a poor choice, I now realize, as they lose torque very quickly as RPM goes up. But with the ballscrews, the increased efficiency and the fact they're around 5tpi (25mm pitch) rather than 10tpi (which means the stepper is spinning slower for the same IPM rate, where they have more torque), I was happily moving the table around at 120ipm last night. And I haven't even oiled the ways (I know, I know..), I couldn't resist trying it out. I will likely keep rapids at around 100ipm to have some margin, but I'm thrilled with the improvement.


I apologize for the long post, but well, now you know as much as I do! Pictures to come at some point, I'm not a very enthusiastic photographer, but I'll snap something to make the thread a bit better!

Thanks for the tips, advice and inspiration so far, this place is incredible.

[TiagoSantos]

Finally added home switches to X and Y last night.. I procrastinated for so long as I knew drilling holes in the cast iron base for the Y switch would suck - so I didn't! I crazy glued a piece of aluminum to the base and mounted the switch on the nice soft aluminum.. I can finally home the machine and use it without havIng to creep up on the limits of travel and bumping the end stops every day!

Also hooked up the fogbuster to my new compressor, seems like it'll work great!

Plans for the weekend are to finish the enclosure/chip pan and move the machine into place ready to start making stuff. There are a lot of little things that are incomplete or not ideal, but I really just want to get going with converting the other machine.

I found out yesterday that DMM-tech is local to me, about 20 minutes from here. I exchanged a few emails and he expressed interest in seeing the machine in person to recommend the right servo setup.. I am tempted to upgrade to servos when I put the RF45 together. But it's $1000 plus pulleys and belts.. I guess I'll see how the rf40 does with the steppers for now, but I feel I am definitely pushing them to the limit.

[BAMCNC.COM]

Pictures?

[TiagoSantos]

I know, I know! I went downstairs to take pictures this morning before heading to work and ended up doing something else.

Ran into a bit of an issue - while I can move X OR Y at 120ipm easily (I cranked it up to 150 and it was still fine, 180 started stalling randomly) if I only move one axis at a time, I can't seem to move both simultaneously. One of the axis will stall while the other sort of keeps going, albeit slowly. This only happens with X and Y - I can move X+Z and Y+Z just fine, but the Z motor is only cranking the fine feed and doesn't have that much load on it.

Makes me thing it might be a power supply issue.. I'm running a 48v 7.3A supply, the motors are at 3A. Theoretically, it should be enough to crank both motors, though.

Any thoughts?

[TiagoSantos]

Just had a reply on the gecko forum about this and backlash compensation might be the problem. I'll try playing around with the settings, otherwise I may have to live with 2.5 thou backlash for a little bit.

I promise I'll take some pictures and a video tonight!

[wizard]

I know, I know! I went downstairs to take pictures this morning before heading to work and ended up doing something else.

Man did the above hit home. I can't count the number of times I've gone down into the cellar to work on something and hours later leave totally forgetting about it.

Ran into a bit of an issue - while I can move X OR Y at 120ipm easily (I cranked it up to 150 and it was still fine, 180 started stalling randomly) if I only move one axis at a time, I can't seem to move both simultaneously. One of the axis will stall while the other sort of keeps going, albeit slowly. This only happens with X and Y - I can move X+Z and Y+Z just fine, but the Z motor is only cranking the fine feed and doesn't have that much load on it.

Strange. Did the pile of goodies include any multimeters? If so I'd look at current draw off your power supply. This does sound a bit strange so there might be something else going on.

Makes me thing it might be a power supply issue.. I'm running a 48v 7.3A supply, the motors are at 3A. Theoretically, it should be enough to crank both motors, though.

Any thoughts?

Extremely stiff Gibbs or lead screw nuts might overload the the power supply if both run at the same time and there are no current limiting going on in the steppers drives. I'm not 100% convinced this is right though.

[TiagoSantos]

Hi Wizard,

Turns out the problem was with Mach3 and the backlash compensation settings. Someone over in the gecko forum pointed me in that direction and I found some old threads where Hoss discussed a similar problem and the solution. There's a setting in Mach3, in the general settings screen - something like shuttle acceleration or something - it was at 0.004 by default, setting it to 0.05 fixed the problem. No idea what else that setting influences, so I'll have to dig a bit deeper. I was quite happy with how the backlash compensation was working with the stock leadscrews and the vast amounts of play they had, so I expected it to work even better with the 0.0025" I have currently.

Anyway, as promised, here are some pictures! Keep in mind I am the most disorganized person you'll ever meet and I have the shortest attention span known to mankind. Also 2 years ago I had probably never used a tool on anything other than my bicycle, so excuse the less than ideal decisions I make along the way

To start with, some chips! This was when it still had the stock leadscrews. It worked well enough to make the parts I needed to mount the ballscrews and few other odds and ends I needed at the time, but it sure wasn't ideal:
https://www.youtube.com/watch?v=AtRMzUObZ9w

This the how well it moves now:
https://www.youtube.com/watch?v=6gaELclYUv8

Tormach Y axis way cover and Y home switch, both with my ghetto style attachment methods (remember, I hate drilling holes by hand!)


fine feed quill drive - that bracket was the first thing I did with the old CNC conversion, it started as a test piece of aluminum when after cutting a pocket and a circle I realized it would probably work to mount a motor! It sort of does..


X axis motor, mount and bearing block. I reused the stock blocks and bored them out for the angular contact bearings. The locknuts suck and I need to double nut them, the set screws will not hold them in place..


Y axis stuff and X axis home/limit switch. One of the first items on the agenda once I can start cutting things is a set of switch enclosures, for the DMM proximity switches I'll eventually get.


Aaaand that's how it sits for now.



Hopefully I'll have time this weekend to rig up some sort of enclosure, at least enough that I can start making chips without making a huge mess. It'll be a struggle to make anything without finding another pair of M15x1 locknuts, I might have a go at threading a set on the lathe, I think a 26tpi thread *may* work..

[Dave [Swe]]

Hi all !
My first visit and post here, and I'm sure I will return often. ...
Starting to say I'm a technic freak from Sweden who had a dream for 15 years to have a CNC mill.
My work has only began and right now I'm drawing my Rong Fu RF 30 in Sketchup (thank God for that program )

Since Google kicked me in this way, I've spent some hours reading about several amazing projects !
Totally impressive !

My plan is to use stepper motors and ballscrews but I'm not pleased with the Z-movement. Including a simple pic from drawing where you can see my idea with a Z-Axis ballscrew.
Before I found some interesting projects with the Z-Axis here, I was convinced to use the drill-press original construction, but now I'm not so sure anymore
Hmmm. ...sorry to steal your thread this way, just got carried away

[SCzEngrgGroup]

If you hunt around a bit here on the 'Zone, you'll find a number of threads where people have done their quill drives in exactly the way you propose, though generally with the ballscrew and motor mounted off to one side. The thing you have to watch out for on the RF mills is the quill sleeve bearings tend to be rather sloppy, and get sloppier the further down the quill is extended. This makes for poor accuracy and poor cuts. But, the only alternative is to move the entire head, which means adding some means of positive alignment, like guide rails, parallel to the column, so you don't lose X/Y zero alignment.

Regards,
Ray L.

[109jb]

Hi all !
My first visit and post here, and I'm sure I will return often. ...
Starting to say I'm a technic freak from Sweden who had a dream for 15 years to have a CNC mill.
My work has only began and right now I'm drawing my Rong Fu RF 30 in Sketchup (thank God for that program )

Since Google kicked me in this way, I've spent some hours reading about several amazing projects !
Totally impressive !

My plan is to use stepper motors and ballscrews but I'm not pleased with the Z-movement. Including a simple pic from drawing where you can see my idea with a Z-Axis ballscrew.
Before I found some interesting projects with the Z-Axis here, I was convinced to use the drill-press original construction, but now I'm not so sure anymore
Hmmm. ...sorry to steal your thread this way, just got carried away

I see in your drawing that you are planning on running reduction drives and you state that you are planning steppers. I would recommend against that plan. Reason is because steppers lose torque at higher speeds. You will find that the torque loss is significant and you will a significant amount of speed. If you look, you will see that most stepper drive setups are direct drive and this is a result of this torque loss issue. I converted a Harbor Freight round column mill and experienced this on my z axis. The way I originally had it was with a reduction drive which resulted in 0.00025" of Z axis travel per full step of the motor. With this setup I was only able to get about 25 IPM reliably. I wound up changing it so that I was getting 0.001" per full step and was able to increase federate to 120 IPM.

On the Z axis, I too looked at different ways to do the ball screw on the Z axis and had a plan similar to what you showed. What I wound up doing was hooking the motor to the factory fine feed through a belt drive. I reversed the quill spring so that instead of pushing the quill up, it pushed the quill down. This setup worked very well. Most tooling will pull itself into the work piece and the spring insures that all of the backlash in the quill is taken out. The only time I ever could tell anything was when drilling and the drill would break through. The drill would pop through. This setup proved to be very satisfactory and very easy to implement.

Here is my build thread for my round column mill

http://www.cnczone.com/forums/bencht...ill_drill.html

[109jb]

If you hunt around a bit here on the 'Zone, you'll find a number of threads where people have done their quill drives in exactly the way you propose, though generally with the ballscrew and motor mounted off to one side. The thing you have to watch out for on the RF mills is the quill sleeve bearings tend to be rather sloppy, and get sloppier the further down the quill is extended. This makes for poor accuracy and poor cuts. But, the only alternative is to move the entire head, which means adding some means of positive alignment, like guide rails, parallel to the column, so you don't lose X/Y zero alignment.

Regards,
Ray L.

Ray is exactly right. The round column mills do have some slop in the quill that needs to be addressed. The way I did it was to replace the quill lock handle with a smaller bolt and a lock nut. The bolt was small enough to go through the "clamp blocks" without engaging the internal threads. I then snugged up the bolt/nut enough to take out the slack without actually locking the quill or creating too much drag. This worked well to combat the slop that exists. Not perfect but very good and if you plan for a light cleanup cut it worked well.

[cmnewcomer]

Dave,

I too tried using the existing quill drive and found it unreliable for what I was doing so I put it inside the head and drove the quill directly similar to your idea. I originally used a stepper with a 0.25" shaft and it eventually broke. Not sure if I had over tighten the belt, had a faulty stepper, or it was too much to drive the quill but I went with a 0.375 shaft and have't had a problem since. You can see the setup here http://www.cnczone.com/forums/vertic...etrofit-4.html.

Best Regards.

Carl

[TiagoSantos]

Yesterday was quite the adventure. I had built the machine on top of the sheet of plywood that would eventually become the base of the enclosure/chip tray. I guess initially I thought I'd just brush up the paint around the machine..? I don't know what I was thinking.

So I had to lift the machine off the base, take the plywood out so I could paint it, etc. I've lifted the machine plenty of times with my engine hoist, but for some reason this time it just wasn't working out. It wanted to tilt back, and by the time I realized how badly balanced it was, I had no way to drop it back where it was. What followed was 35 minutes of incredible acrobatics and contortionism, to lift the machine up, push down on the Y axis motor mount to help balance it, pulling the plywood from under the machine, sliding the heavy steel stand from under it, grabbing the aluminum extrusion stand I had built a while ago and sliding that under the machine, dropping it down, etc. All in all, incredibly dangerous and stupid, but by the time the machine was up, I couldn't leave to get help.

Anyway, after that was finally done, I took the plywood enclosure pieces into the garage to start painting. I went to the marine hardware supply place to look for some type of marine paint that would work, the guy recommended polyurethane based topside paint. Told me where to look and said they'd all be around the same price, $34.99 for a quart, just had to pick a color. Well, as luck would have it, "Teal Blue" must not be a popular color, cause it was marked down to $16.50 Didn't even know what color that would turn out to be, but I didn't care, half price is half price.

I diluted a bit of the paint with some xylene to prime the wood yesterday, and this morning gave it a first coat of paint. Looks like it'll work and be pretty tough, but sure takes a long time to dry. Doesn't help that it was below freezing overnight and just a couple of degrees above this morning.. I wanted to put the enclosure back together tonight but it won't happen. I might see if it's dry enough to give it another quick coat of paint and then work on putting it together during the week at some point.

[TiagoSantos]

Like 109jb, I played around with adding a ballscrew to the quill and even scrounged a ballscrew from an film scanner at work, but ended up driving the fine feed with a belt. Seems to work fine, I also reversed the spring which I'm sure helps. That's a nice tip about snugging up the quill lock, I'll have to look into doing that!

[Dave [Swe]]

Thx guys !
Amazing knowledge here, and really like to see different solutions to similar problems as the ones I find on my own.....

About the issue with lack of torque in the stepper motors at high speed, I´m aware of that problem, but at full-step 2ms drive at 500 pulses per second, my machine will move 12,5mm/sec, and that is much faster than I want to manually use a simple end-mill......
How fast are you running your stepper motors ?
Of course the transport will take some more time to do, but I gain precision and torque with the 5:1 ratio.
With direct connection with my stepper motors I would look at 0.025mm / step and that is way too much.

About the idea with the "sloppy" quill bearing, and to use the quill lock mechanism to tighten it up, I had a crazy idea about changing these two parts out and put 2 regular ball bearings in, one on each side, angled towards the center of the quill.
Just to be able to tighten it without addind too much drag.....
To make this work properly, there should of course be bearings on the opposite side , and with proper adjustment and idea, the dream would be to tighten it very hard with almost no drag at all....

Hmmm.....dont know if anyone understand what I mean.....I will try to draw this in Sketchup.....

[TiagoSantos]

0.025mm per step is just under .001", and that's also what I'm running. Most drives will micro step though, but if you need much better than 0.001" accuracy, you might need a different machine altogether! Personally, I think if I get 0.001" out of my machine, I'll be pretty damn ecstatic, but it might take a bit of work to get there. 0.001" is an incredibly small dimension, and I'm not doing aerospace work. Actually, I think a lot of aerospace work has bigger tolerances than that, not everything has to be that precise I'd be surprised the a pair of wings on a fighter jet are within 1/8" of each other :P

Like I said above, I'm using the 425oz/in steppers that are notorious for having lots of stall torque, but lose torque very quickly with RPM. With my old leadscrews, which were 10tpi (2.54mm pitch) I couldn't go any faster than 40 or 50IPM. The stepper was spinning so fast that torque was dropping off right above that. With the ballscrews, which are 5mm pitch, I can move the table around at 120IPM easily. The steppers are spinning right around the same RPM as they were before, where they still have some torque left, but the table is moving twice as fast. Technically, I doubled the gearing so the table needs twice as much torque to be moved, but the ballscrew efficiency probably makes up for that.

Anyway, the 5:1 ratio seems like a lot to me, but it depends on what ballscrews you'll be using - if you're using some crazy 10mm pitch screws or something like that, most of the calculations we've been making with the readily available 5mm screws will not necessarily apply

[Dave [Swe]]

My ballscrews has 5mm pitch and my goal is to get an accuracy of 0.02-3mm.
If this machine allows this or not will be cleared in a few months
Main issue as I see it is the Z axis.

Sorry I cant translate the metric units I use, but my stepper motors are 3,1Nm.

I guess the motors are strongest in a full step ?
They do full step revs at 200p (150rpms), and 1/2 with 400p(75rpms) and so on.....?

[TiagoSantos]

1Nm = 141oz/in or something near that, so 3.1Nm is around 437oz/in, so probably similar to the steppers I have.. They might be a bit stronger at full step, but I believe the torque speed curve affects things a bit more.

Anyway, I just did the conversion for the speed you are hoping for - just under 30IPM. If you think you'll be happy with that, then go for it - with the 5:1 ratio, I think that is a reasonable expectation. It really depends what you'll be doing, I know when my machine was set to 40IPM (about 17mm/sec) it was painfully slow to watch it do anything, but I am very impatient in nature! The X axis travel for instance is about 20 inches - so it took about 30 seconds to go from one end of the table to the other. Doesn't sound like much, but.. When you're watching the machine cut a part and it spends 50% of the time going through the air instead of actual cutting, it's incredibly frustrating!

The good thing about doing a belt drive is it won't be a big deal to change the pulley ratios, just a pulley and a belt for each axis, so experimenting will be easy

[Dave [Swe]]

Yes, the ratio is simple to change later....true !

Spent a little too much time thinking tonight.....according to my wife..... ^^


I wonder if alignment bearings would make this quill stable.....
Just tested it in Sketchup and with 3 bearings below and 3 bearings on top, of course adjustable to take all "sloppyness" away....

I will loose about 40mm in Z movement, but I still have at least 70mm left and maybe this would be acceptable.....

Dont know how to upload photos to this forum, so I just use my cellphone i a few moments instead.....

Edit:

[TiagoSantos]

What are you planning to use the machine for?

70mm is less than 3 inches of travel.. It will take very, very careful planning and setup to not require moving the head up or down while machining a lot of parts.. If you use R8 collets, you probably won't have room to go from an end mill in an R8 collet to a drill bit in a chuck.. You'll definitely need a set of screw machine length drill bits (shorter) and possibly use an ER20 collet holder for your end mills, so they're about the same length as a drill bit on a chuck.

The alignment bearings are a neat idea, but make sure you won't miss the travel, 110mm is pretty short as it is!

The reason I ended up buying a dovetail column machine (the RF45) was exactly that.. I was going around in circles with so many crazy ideas to fix the shortcomings of the round column mill/drill, when I finally realized I'd be better off selling the round column machine and buying a machine that fit my needs a bit better. I lucked out on getting the dovetail machine (or most of it) cheap enough that I could keep the RF40 to make the parts I need for the full conversion, but I was ready to sell it.