X3 conversion w/ stock ACME screws????

[Matt McColley]

for a machine that will see light use, would it make sense to make a mount for the steppers that attaches to the existing bearing blocks (where the vernier and the hand wheels attach) and use the existing ACME screws?

Could Derlin nuts be used to take out back lash and reduce friction?

I'm thinking that I could get the rig up and running and then convert it to ball screws in the future?

I'm aware of the basics.... that ball screws are MUCH more efficient. But I don't feel the need to operate the mill at super sonic speeds. And if I could get, say, 500 hours out of it milling aluminum, before the ACME screws show wear in the centers, then I could justify the add'l cost of upgrading the screws.

I'm basically stuck, trying to do this project on a shoe string budget, because it's a hobby thing for me and I can't justify dropping a lot of bread on it.

thoughts?

opinions?

any body else out there done it?

[HimyKabibble]

It can certainly be done. Make the nuts adjustable (split the brass nuts), and get used to adjusting them periodically, and it'll work just fine. You should really design the whole conversion, and just assume the ballscrews will bolt in just like the leadscrews, and you'll be able to easily add them at any time.

Regards,
Ray L.

[cadmonkey]

Thoughts-
1) Yes it can be done
2) Delrin might help increase the efficiency of the screw - ACME's are much less efficient than ballscrews, but I don't know if a delrin block itself can handle the forces of metal machining with out deforming or otherwise becoming un-mounted.
3) You're stuck with trouble selecting appropriate stepper motors. You need more torque because of the lower efficency but this might be offset by the higher pitch of the ACME screw. But the higher screw pitch will require more speed and unfortunately the nature of steppers causes them to lose torque as the speed increases. Servos would alleviate that problem but then the price increases.
4) The stock bearing setup is terrible for backlash. I don't know if you can design up a mount that replaces the stock bearings with suitable replacements with the existing screws.

Nothing says it can't be done but there are obstacles - just making sure you're going in with your eyes open. There may be other items to consider, be patient and you should get some more replies.

[Kalvin]

Hi Matt,

I'm sure it could be done too. I read a thread somewhere around here where the person was actually getting pretty good results in regards to the backlash.
I would stick with the brass nuts also.


Now when those acme screws wear out....I'm sure I could find you at least one replacement set at a really cheap price. I converted mine to ballscrews like everyone else. (I bet there are acme screws just laying around.)

I will see if I can hunt that thread down for you.

Cheers,

[HackMax]

Hello Matt,

Like you I wanted to start out as low cost as possible so I built mine around the ACME setup. I used the same motors and other electronics that I would use with the ball screws so that in the future I could go that route. I based my entire system off Cadmonkey's design but when it came to attaching the motors to the ACME screws I ended up having to make the brackets and whatnot on the fly - all except the z axis.

I put the x axis stepper on the motor side of the mill, connected it to a solid coupler, and keeping the stock bearings and hand wheel setup. The y axis I did use bearings but the ACME screws have a 12mm diameter shaft to connect the motor to and I didn't have a lathe at the time to modify the shafts so I just got 12mm ID bearings and made the bearing and motor mount around those.

I don't know if the Z axis really needs the 2.6:1 pulley reduction in Cadmonkey's plans but I wasn't going to chance it and set that up as in the plans. The only thing I changed was the size of the shaft hole in the larger pulley to 12mm to fit the stock ACME screw.

What I found from doing this is I would get 60 IPM rapids on the x and y and 15 IPM on the z. Rapids aren't an issue using the ACME screws if you can handle these speeds. You may be able to get better rapids on the z if you go 1:1 since the stepper motor torque increases the slower it moves. I hope to go to 1:1 when I get the ball screw in the z.

I'm working on changing over the ACMEs now because the average backlash off all the axis was about .005". I cut a larger radius in aluminum and these seem to show the effects of the backlash even with the backlash compensation in the software. If you cut the metal (more pronounced in the conventional direction than the climb) you get a see-saw effect where the table moves in the available .005" leaving divots in the metal to sand out. I could get the backlash down to .003" but that wouldn't last too long and would soon climb back up to .005" If you can deal with this backlash I would recommend using the stock screws but I can't spend a lot of time sanding out divots to smooth the metal for chroming. Quality sandpaper is also upwards of $1 a sheet too.

As far as cutting speed it doesn't matter which setup you use. My normal cut with .5" 2 flute HSS end mill was 10 IPM at .05" DOC. I could probably go deeper or faster too but I hadn't gotten that far yet. The biggest hold-back is spindle speed.

Anyways good luck,

Rick

[cadmonkey]

Hackmax - Do you have any pics or drawings of what you ended up doing? If you could post them, or PM me for an email address to send to. If you don't mind I'd like to show some pics of how others have modified the plans to suit their needs - financial, phased implentation, etc. Thanks.

Also - with the ACME screws you can definitely go 1:1 on the Z, or even a step-up ratio. I know the 2.6:1 is overkill and I may end up going to a smaller ratio in the future to get better rapids. I rapid around 30 safely but that is a limit induced by my PC - it just can't send quick enough pulses for microstepping. With 16tpi ACME vs 5tpi Ballscrew, you've already got 3x the advantage. When you do install the ballscrew on Z if you go 1:1 then be conscious of backdrive when powered down - you may need to add a gas strut as I believe comes standard on the SuperX3 - this was part of my decision to go with the 2.6:1 - the natural cogging of the motor when coupled with a high ratio to the screw was sufficient to eliminate backdriving, even if I loosen the gib nearly to removed.

I toyed with the idea of doing Z on the ACME screw when I got the X and Y done since the Z didn't seem to all that have much backlash with the gearing of the front handle taken out of the equation, but since I had the Z ballscrew in hand since I had ordered a stock 6' length it wasn't worth the time to put AC bearings on the ACME. I do want to go to a preloaded double nut on the Z though - there's enough backlash with a single nut that I am not satisfied with and the best DIY double nut setup I can envision would loose some top end travel which I refuse to give up. The factory double nut is just longer than the standard nuts I used so it runs into the area that is at the bottom end of the travel that never gets used. That upgrade will wait for tax refund to come, doctor bills from the Cleveland Clinic to start showing up so I know how much they charge vs what my insurance covers and whether or not I am able to stay employed with my decreased performance under the diagnosis I have finally gotten from the high quality care in Cleveland. (eg - there are a lot of variables that need to be evaluated before I buy a $300 ballnut (IIRC) so that is VERY back burner). I am curious what size steppers you ended up using that are suited to ACME and ballscrew?

I know the stock 'adjustable' brass nuts were constantly in need of adjjustment while I was machining the parts and I ended up putting on digital scales to be able to machine accurate enough for the parts to come out with centerlines where they needed to be. And that was manual where speeds aren't as demanding as the speeds you'll see on CNC or loads with cutting forces involved.

[HackMax]

Originally Posted by cadmonkey Hackmax - Do you have any pics or drawings of what you ended up doing? If you could post them, or PM me for an email address to send to. If you don't mind I'd like to show some pics of how others have modified the plans to suit their needs - financial, phased implentation, etc. Thanks. Also - with the ACME screws you can definitely go 1:1 on the Z, or even a step-up ratio. I know the 2.6:1 is overkill and I may end up going to a smaller ratio in the future to get better rapids. I rapid around 30 safely but that is a limit induced by my PC - it just can't send quick enough pulses for microstepping. With 16tpi ACME vs 5tpi Ballscrew, you've already got 3x the advantage. When you do install the ballscrew on Z if you go 1:1 then be conscious of backdrive when powered down - you may need to add a gas strut as I believe comes standard on the SuperX3 - this was part of my decision to go with the 2.6:1 - the natural cogging of the motor when coupled with a high ratio to the screw was sufficient to eliminate backdriving, even if I loosen the gib nearly to removed. I toyed with the idea of doing Z on the ACME screw when I got the X and Y done since the Z didn't seem to all that have much backlash with the gearing of the front handle taken out of the equation, but since I had the Z ballscrew in hand since I had ordered a stock 6' length it wasn't worth the time to put AC bearings on the ACME. I do want to go to a preloaded double nut on the Z though - there's enough backlash with a single nut that I am not satisfied with and the best DIY double nut setup I can envision would loose some top end travel which I refuse to give up. The factory double nut is just longer than the standard nuts I used so it runs into the area that is at the bottom end of the travel that never gets used. That upgrade will wait for tax refund to come, doctor bills from the Cleveland Clinic to start showing up so I know how much they charge vs what my insurance covers and whether or not I am able to stay employed with my decreased performance under the diagnosis I have finally gotten from the high quality care in Cleveland. (eg - there are a lot of variables that need to be evaluated before I buy a $300 ballnut (IIRC) so that is VERY back burner). I am curious what size steppers you ended up using that are suited to ACME and ballscrew? I know the stock 'adjustable' brass nuts were constantly in need of adjjustment while I was machining the parts and I ended up putting on digital scales to be able to machine accurate enough for the parts to come out with centerlines where they needed to be. And that was manual where speeds aren't as demanding as the speeds you'll see on CNC or loads with cutting forces involved.

Hey Greg I got started on it and it got big and is more appropriate to be placed in your thread. (My X3 CNC Conversion)

Edit: Oops almost forgot the rest of what you wanted to know. I used the same size steppers as you did because I knew I would be going to ball screws eventually.
Also I'm almost positive the z axis is a 2:1 from the hand wheel and is a 10 TPI screw.

Rick

[Matt McColley]

Gentlemen,

Thank you all for your insightful responses.... the chances of me pulling this conversion off are greatly increased by keeping the cost minimal.

CAD Monkey's plans are top shelf and I thank you very much for sharing the Solid Works files..... but I think I'm going the KISS route. Two reasons....
1.) lower cost and 2.) maximize my chances of getting it right.

Here's a link that shows how I'd like to do the X and Y Motor mounts....
http://www.jeffree.co.uk/pages/x3-to-cnc.html

One of the things I like a lot about this approach is that by sticking with the ACME screws and using a dual shaft motor, I can put the handles back on the shafts and be have easy dual use as both manual and CNC.

One of my concerns is that the weight of the steppers will put quite a bit of downward force on the neck of the factory bearing seat. Hmmmmm? At least two guys have done it this way, as Tony Jeffree's design is a modification of this one by Dick Stephen.
http://www.arceurotrade.co.uk/projec...cle_Part_2.pdf

But can these Brits be trusted? After all, they did abandon their offspring and jump on the SI train.

Other thrifty and neat feature I like are...

Jeffree's machined his ball screw adapters out of the ACME screws, which will require no mod to the bearing and motor mount if I upgrade.... Well, assuming I don't add an ATB, which I really should.

Stephen's machined out the existing ACME threaded brass nuts to mount his ball nuts. Another plus for thrift and easy upgrading.

As for dealing with back lash and adjustments.... here's my idea....

machine out the existing ACME threaded brass nuts to accept a ball nut, but then face mount one of dumpsterCNC's anti backlash/automatic wear compensation Delrin lead nuts. http://www.dumpstercnc.com/datasheet...ableadnuts.pdf

And since I'd be drilling/turning out the AMCE threads on the existing brass nut, if I had my ducks in a row and knew exactly what the thread details on the future ball nut were, I could possibly cut the threads into the brass nut up front and not even have to remove the brass blocks should I decide to upgrade, because the dumpsterCNC AB nuts would be mounted by drilling through the flange.

As for the Z-axis.... I really admire the CNC fusion design. I think it's well thought out and like the idea of not having to do any machining on the factory castings.

But there again.... I've got my own half baked idea... by going with a top mounting motor, driving an extension off of the existing ACME screw, couldn't I keep the gears on the bottom in place and again have a dual use (manual/CNC) set up?

But then again, I guess I could easily jog the head down by eye and use the quill for drilling/tapping. If the quill was only extended a small amount and then locked, would this be stable enough for milling?

And heres my really stupid question .... I realize that if there is power going to the steppers, they will be locked into position. But is there any way turning a de-energized motor can send pulsed back into the controller software to indicate position?

And my final cheap skate idea.... I put another thread up about using an Anaheim Automation drive pack (800W ps, 7amp hold, unipolar, bi-level drive that supports only full or half stepping). With the pitch on the ACME screws, could I get decent results (smooth, vibration free cutting with a resolution of .001) with half stepping?

Sorry to go on... that's a lot to process, I know.

Thanks again for your taking the time to read this and comment.

[cadmonkey]

Yes you could top mount the Z and keep the handwheels. Be forewarned - mentioning keeping the handwheels on converted machines can start a flame war around here... I'll just say this - keep in mind the wheel is at just the right height to knock you in the *^$s when you're running under automatic control... You can do manual work with the CNC software either by jogging or entering in G code on the fly. You can have the quill extended and locked a bit and still be rigid enough, just not too far, especially with larger cutters. If you want to send pulses back with steppers you need to add encoders onto the shaft and read those - just like servos do (though some servo drives don't route the encoders back to PC, they just throw a fault when they haven't seen the complementary amount of movement as has been commanded). I don't know what results 1/2 stepping would yield on the ACME screws as far as resonance and resolution go. Sorry.

[project5k]

a couple things that you might wanna think about, adding the hand wheels on the back end of the motors is a neat idea, i thought about this myself, till someone pointed out that the handels aren't ballanced, and will add vibration to the system when in cnc mode.

something else to think about, by keeping the z axis hand wheel, again ballance, and you will be adding more load to the z axis system, meaning that it will take more torque to turn the screw. I would also think that this would add a considerable noise component as the gears for the z axis arent exactly what i would call perfection.

Michael over at cnc fusion is a really great guy and really helped me out when i did my conversion. we emailed back and forth i dont know how many times, untill i fully understood what i was dealing with. The fact that he's basically my neighbor didnt hurt his running in the decision on what product to go with as well(hes just a couple hours down the road from me)

extending the quill down for milling operations is not something i would suggest, it adds more lever action to the whole machine and could introduce additional flex and possibly more chatter when machining. Mine is in the upright and locked position and thats where it stays.

feeding back to software from the motors would need to be handeled by an encoder, or some kind of DRO unit. I suppose if you really wanted to get off into it, you could work up the electronics to try and feed back the motor pulses, but the problem is going to be when you land between steps, or just hit a step but then rock back to the previous detent, this will introduce errors... not to mention that as speed changes, voltage out will change, and this could effect the sensing.

not to mention that depending on the drivers your using, feeding back when thier not on could cause the drivers to fail.

really the only way to do it reliably is with encoders or a dro. go and get yourself the mach 3 instruction book, it talks about doing just that.... http://www.machsupport.com/docs/Mach...all_Config.pdf start on page 52 of the pdf, section 4.10 linear encoders.

half stepping a 200s/r motor, with a 10tpi screw gives .00025 inches per step,
2*200*10=4000steps/inch
your main issue os going to be resonance in the motor. half stepping has issues at certain rpm's of the motor/screw/coupler system. resonance is a bad thing when talking about stepper motors, and should be reduced and or avoided if at all possible. it can cause lost steps, which translates to messed up parts, and or possably worse, much worse depending on how bad it is. (can you say crashing the endmill into the table?)

the other problem is gonna come in when you convert over to the 5tpi screws, it'll drop your resolution to .0005 per step. if this is an acceptable resolution, then ok, but if your doing some machining, say, a long low angle part, where the angle isnt parallel to an axis, say 5 or 10 deg off, then you may notice some resolution stepping. this probably wont be a major factor as we're only talking about a half a thou, but if you get into ultra precision stuff, or if theres some backlash in that axis' travel, it may become very evedent. on my system i'm running 200step/rev motors, 10 microstepping drive, 5tpi ball screws(cncfusion) and my theoretical resolution is .0001 and i'm more than happy, and i can still get rapids over 100ipm.

as for drivers, i'm running gecko 203v's, and i couldnt be happier. I love the "vampire" drives, cause you, and i'm quoting gecko here, you cant kill 'em. all sorts of safeguards, short protection, and stuff like that. if your not super on your game, you can blow up some of the other drivers out there, and all you did was sneeze hard. I gotta say, gecko makes a good product, and the 203v is my personal favorite. Keep an eye on the gecko site, they sometimes put thier stuff on sale(i happend to catch them on sale when i got mine)

something else, if your not already familiar with the terms resolution and repeatability, learn what they mean and the differences between them. i know that this was a hurdle when i first started into all this, and knowing the difference has helped me to become a better cnc operator.

[Matt McColley]

I'll have to do some more homework and mull over "recycling" my AA drive pack and half stepping. I don't really understand what the difference is between unipolar and bipolar, nor between bilevel and chopper drives.

So I'm wondering if I could source stepper motors that would work with this AA drive, and then, if the resolution or resonance became an issue, could be set up to run off of a micro stepping drive like the Geckos. Assuming I didn't miswire and fry the motors, I could at least test out the AA drive pack that way.

I still like idea of setting up for dual use (CNC/manual) on both X and Y. Perhaps I could rig up the hand wheels in such a way that they could be quickly removed/installed.

Any body out there able to shed a little light on the topic of encoders? Specifically, are they expensive? Do most BOBs have hook ups for them? Do they mount on the back shaft? and if so, would I still have room left ot mount the hand wheels.

I do have quite a bit of experience jogging around on a CNC table router, as well as writing G-codes on the fly and entering them in at the command line. I find it to be a royal pain in the neck and that's why I'm so interested in setting up for "dual use".

I am going to punt the idea of keeping the Z linkage to the hand screw in tact.

Is it an expensive proposition to wire up a touch off? and do the populare apps like Mach 3 provide for that input? If I could easilly get my exact tool height entered after a tool change, then jogging or tapping in G codes to position Z would be a breeze.

Also, are there any big advantages or disadvantages to driving z from up top? There's obvioulsy going to be some drilling of good size holes involved, and you lose the ability to get any mechanical advantage via. the motor connection. Any others that I'm missing.

Any comments on mounting the X & Y steppers directly to the factory bearing seat?

Any one out there ever use the dumpsterCNC AB nuts?

thanks again,

I sure do appreciate your responses.

[project5k]

The easest way i can explain unipolar vs biploar is like this. Imagine you have a coil, like one in a motor, and lets call the connections to it a and b. you can apply power to it with a + and b -, or with a - and b +. in unipolar you pick one or the other and use it exclusivly. in bipolar, you use both all the time. thats the super simple version...

chances are that you can find motors that you will be able to use for both, and if you do mis wire something, i would think that it would be more likely to kill the drive than the motor.

if you can figure a quick release system for the handles, then i really dont see a problem with it, there are some that might "give you the business" about it, but its your machine, and if thats what works for you, then i say go for it. In my mind i picture somethng like a race car steering wheel release mech.

my experience with encoders is fairly limited, but what i do know is that they do need to mount somewhere on the shaft, i have seen far end encoders, and motor mounted ones. if you mount it on the motor, then i would think that it would go onthe back end shaft. If it were me, and i were considering getting and using encoders, i would just get the motors with them already installed, just to simplify my life. The other option would be to get the encoders, and mount them on the far end of the screw, but this has its advantages and disadvantages. Encoder pricing will depend largely on how many divisions it has, or how accurate its supposed to be.

one thing that you might consider is using a jogging pendant. I just recently saw one online for about $90 that was usb interface, and would allow you to jog each axis down to the single step.. I'm seriously considering getting one of these for my machine, as i sometimes would like the ability to operate in a "manual mode"

i wouldnt think that it would be very expensive to set up a touch off tool height system, i havent tried it yet, but i think its a great idea. Perhaps someone else can give us some input on this.

I dont see a major problem driving the z screw from the top, and if you wanted to get creative, i bet you could still set it up with some mechanical advantage...