I've heard a few people say that when they push on the side of the column of the X3 with very minimal force, they measure a deflection of a few thou. Any idea where this is from?

The column is open "U" channel and this results in the some flex. A closed box would be a lot stronger.

Hi Rod,
While I agree that the closed box design would improve rigidity, there is definitely another source of the flex. If we assume that most of the flex is coming from the column and model the column as a cantilever U-beam, the deflection located at a load applied at 20 in (about 2/3 of the way up) from the fixed end would be:

y= Fx^2/6EI *(x-3a)

For the X3, the column is approx 5.5in wide and 3.75in deep. Wall thickness is about 0.5in. This is a conservative estimate, since there are ribs inside the column to keep it more rigid. This gives a moment of inertia of

I=1/12 * (0.5)(4.5)^3 + 2*[ 1/12 (3.75)(0.5)^3 + (0.5)(3.75)(2.5^2)]
= 27.32in^4

(it turns out, adding a closed box design would only give 14% increase in sideway rigidity, since most of the strength comes from the sides)

a=20in
x=a
E for cast iron = 13E6psi

Say we apply a load of 100lbs

Deflection would be given as:
y = 100*20^2 / 6*13E6 * 27.32 *2*20
= 0.00075"

Less than 1 thou for 100 lbs of sideward load. I've heard people say they get a few thou with just finger pressure, which is max around 20lbs.
So the deflection is not coming from the column, but probably the base. Maybe from the connection between the column and the base? I'm going to do some tests on my mill to see where the deflection is coming from, and build some sort of truss structure to improve the rigidity.

Hi Lagfish,
I got lost on your first formula so I am not even going to try to understand what you have said. :)
Just a thought though - have you allowed for the lever effect of the head as I think this would magnify the deflection more than the 1 thou you have calculated and could well bring it to the figures quoted.
I am all for finding and fixing the problem and will be very keen to see what you discover.

For your information I have seen a 200mm square RHS column fitted to the rear of an X3 column and it measured zero deflection with 10kg sideways pressure on the spindle. It did have a connection to the base so you could well be on the right track about the column to base connection.

Look forward to your findings.

I would like to see this also.

My first thoughts are whether the mill column was torqued down well on the base. The next is whether the base was bolted to a solid structure to keep it from twisting.

Most people posted that the deflection was one the front pushing the mill head up or column back (or at least I thought that). I would attribute that to the gibs and boxway. Mine is the same and very noticeable with the counterweight attached when the balance is changed.

I would like to see an alternative way of moving the z axis than the side bar system. Then the entire column could be one piece and since there is an empty base now with the CNC kits added. That would be a great place to fill in as well.

I would do this, but I lack the tools to do this with metal and concrete doesn't have enough mass to help there I don't think.

There is definitely more flex than just .00075. Where it comes from is the question. When the head is at the top of the column I can push the column back with just a moderate push and I can see it move. So if I can see it, I know its more than computed.

I tried to plunge a hole with a new 1/2" cobalt end mill in 6061 al the other nite. You can diffidently see all kinds of stuff moving around. I decided to drill a hole first, then end mill to size.

Rod:
I should have clarified my formula more. I used the moment of inertia formula for calculating the deflection of the column with a sideways applied load.
I = 1/12 *bh^3 + Ad^2

I'm not sure what you mean by lever effect of the head. If you mean that the distance between the endmill and the column creates a torque around the column, then this torque will only tend to twist the column into an S shape rather than causing a great amount of deflection. I also didn't put this into my calculation because people have noted deflection when pushing against the column rather than the endmill.

I previously thought you only saw considerable deflection when pushing the head sideways, but cjdavis noted that this is also the case when pushing it back.

One way I thought of determining the source of flex is to give the column a fixed load, and measuring the deflection at many points along its height. Plot this on a graph to get the shape of the deflection of the beam. If the shape is curved, then it means the column itself is flexing. If the shape is slanted but straight, it means the base is flexing. If the shape remains vertical but is translated, it means there is sliding between the base and the column. However, this would be hard to do because the total displacement is only a few thou, and my indicator can only measure 0.0005.

Anyway, attached are what I thought of to reinforce it. Basically a box frame with X braces on the back and sides, and it doubles as an enclosure. The green is the connection between the mill and the frame. If I used 3x3x0.125 steel tubing, it would run less than $150 for the steel. I calculated a deflection of less than 0.00001 with a 200 lbs load, which is most likely more than what the X3 can generate with cutting forces. I would weld the frame and drill holes into the column to attach to the frame via bolts.

If anyone with any experience with this could chime on this, it would be great. I have zero experience with this stuff, and just basing it off what I learned in school.

That would certainly stiffen things up a bit. Although 0.00001 with a 200 lbs. load for what is effectively a lightweight box section frame seems somewhat optimistic. My other comment would be to take a more holistic approach to the issue of cutter deflection under load, rather than concentrating all effort on one particular aspect.

Lots of good thinking though, keep going.

When you measure excessive deflection the best way to track it down is to put your DTI across every "joint" individually in order to find the culprit or culprits.

Just some thoughts and opinions.

Phil

Rod:
I should have clarified my formula more. I used the moment of inertia formula for calculating the deflection of the column with a sideways applied load.
I = 1/12 *bh^3 + Ad^2

I'm not sure what you mean by lever effect of the head. If you mean that the distance between the endmill and the column creates a torque around the column, then this torque will only tend to twist the column into an S shape rather than causing a great amount of deflection. I also didn't put this into my calculation because people have noted deflection when pushing against the column rather than the endmill.

I previously thought you only saw considerable deflection when pushing the head sideways, but cjdavis noted that this is also the case when pushing it back.

One way I thought of determining the source of flex is to give the column a fixed load, and measuring the deflection at many points along its height. Plot this on a graph to get the shape of the deflection of the beam. If the shape is curved, then it means the column itself is flexing. If the shape is slanted but straight, it means the base is flexing. If the shape remains vertical but is translated, it means there is sliding between the base and the column. However, this would be hard to do because the total displacement is only a few thou, and my indicator can only measure 0.0005.

Anyway, attached are what I thought of to reinforce it. Basically a box frame with X braces on the back and sides, and it doubles as an enclosure. The green is the connection between the mill and the frame. If I used 3x3x0.125 steel tubing, it would run less than $150 for the steel. I calculated a deflection of less than 0.00001 with a 200 lbs load, which is most likely more than what the X3 can generate with cutting forces. I would weld the frame and drill holes into the column to attach to the frame via bolts.

If anyone with any experience with this could chime on this, it would be great. I have zero experience with this stuff, and just basing it off what I learned in school.

I guess that if longer sideplates could be made, you could then add an additional plate on the back of the column, which would help keep the column from twisting on the Z-axis. The plate could be pinned in countersunk bolted to the back of the column. Even then, everything would have to be just so.

It appears that there is also little clearance between the acme/ball nut attachment on the rear strap.

What I was thinking of doing was making a brace that went across the front of the column and bolted to the sides with a tee shaped brace. Make it high enough to go over the table lip so it wont affect y travel. I was thinking aluminum, 1/2 inch or maybe 3/4 inch and that should be enough to prevent the column from tilting forward and backward, which seems to be the only flex that affects my cutting. Put some slots in the braces and it could be used to tram the column too.

Hi phil,
Thanks for the input. My theory on the frame is this: The flex can only come from 4 different places. 1. The head, which cannot be easily stiffened, since it moves. Although attacking ribs to it would probably make it stiffer. But my intuition tells me the head is not flexing very much. 2. The colum itself, which can be stiffened. 3. The connection between the column and the base, which can be stiffened 4. The casting that forms the base, which might be able to be stiffened by filling it with concrete or something.
The frame I have designed improves stiffness in 2 and 3, because it connects directly to the column and base.

Also, I added one too many zero to the deflection, I meant 0.0001", which is reasonable when you think about it, because you're adding essentially 2-3x the volume of the column in steel, which has a higher modulus than cast iron, and is oriented in such a way to provide high rigidity. I was also thinking of sandwiching the the beams with heavy gauge plate, like how on some furniture, as soon as you nail on what seems to be very flimsy cardboard backing onto an otherwise wobbly structure, the whole structure becomes stable.

Hi Pizzamaker,
I don't see why you would make it out of aluminum, since steel is cheaper, stronger, and easier to weld. I think your idea would work well too. But since i'm building an enclosure for my mill as well, I thought why not make the frame out of heavy tubing, which would only increase fabrication time slightly, and attach the mill to it to improve rigidity. Kill two birds with one stone.

MY GOD!
I can't Believe how anal retentive some people are on this site we are talking about a $1000-$1500 dollar hobby machine that's only designed to take light cut weener cuts.I think people should stop getting so fixated on numbers and get in the real world .Most parts people are ever going to make at home are going to be fine with a tolerence of +-.015mm.If your little weener machine start to panic a but with you 1mm deep cut maybe it 's too deep back it off and try again.

Also can sombody help me make a dicision i am builing a back deck for my house and i dont know what timber to use .i am thinking of treated pine but it seem to have to much expansion in areas with differing rainfall and think i will have trouble maintaining a consistand gap between the planks or should i use tallow wood this is a more expensive option and will offer less flex underfoot than treated pine ,i dont know !Help .
im gunna have a beer and think about this some more

This is a discussion forum, where people come to discuss all kinds of crazy notions. Just like sitting at the bar. It doesn't have to be correct, interesting or even of any use. It can be kinda like just shooting the breeze.

Most parts (you make) might be OK at +-.015mm but what do you do for those parts that need to be better. People are not necessarily lost because they are not going in your direction.

Phil

PS: A good teak decking would give even better durability and weather resistance than tallow wood, even more expensive though.

MY GOD!
I can't Believe how anal retentive some people are on this site we are talking about a $1000-$1500 dollar hobby machine that's only designed to take light cut weener cuts.I think people should stop getting so fixated on numbers and get in the real world .Most parts people are ever going to make at home are going to be fine with a tolerence of +-.015mm.If your little weener machine start to panic a but with you 1mm deep cut maybe it 's too deep back it off and try again.

Also can sombody help me make a dicision i am builing a back deck for my house and i dont know what timber to use .i am thinking of treated pine but it seem to have to much expansion in areas with differing rainfall and think i will have trouble maintaining a consistand gap between the planks or should i use tallow wood this is a more expensive option and will offer less flex underfoot than treated pine ,i dont know !Help .
im gunna have a beer and think about this some more

Hi stu,
Very funny comment about the deck:rolleyes:. When starting this thread, I knew someone would question why some go to great lengths to make sure their cheap hobby machine is as accurate as possible. I'll explain it to you:

1. Enjoyment. For me, there's a big satisifaction in taking something that is not made well or in poor condition, and realizing it to its full potential. This is why people buy old vintage cars and rebuild them.

2. Cost. Some of us can't afford to spend almost $10,000 on a Tormach. So we spend $1000 on a smaller mill and try to make it better. My budget for the whole project is $2500. Yes, if I considered the time spent on this project, I probably could've almost bought a tormach, but your time is free if you enjoy what you do.

3. Space. The X3 at 350 lbs is pretty much the limit of what I can fit in my apartment. I've seen a few ads selling huge used knee type mills for a few hundred more than what I bought my mill for, and believe me I'd love to get one.

Hi Lagfish,
I have only noticed the sideways deflection and that is with the jib locked down. I might have forward and back deflection as well but I didn't measure it in my test.

What I was trying to say is that the head will act as a lever back to the column. Sideways forces at the spindle are magnified if the column is twisting because of the distance between the spindle and the centre of the column.

This is probably of little value at the end of the day as we already have agreement that there is deflection.

My thoughts are to fix strengtheners to the top of the column as this is the point where maximum deflection will occur.

I do like your design because it strengthens the column to base connection.

The reason I'm choosing aluminum is that I have a bunch of scrap metal that would work. For the duty that it is doing, it's also plenty strong enough and easy to machine. I also don't intend to weld anything, because I don't have that option. However, if I did, your idea looks great. But because I'm not, my plan interfaces with the base and with the column, and should provide a nice bit of extra support for our $1500 hobby machine.

Also, one concern that I have heard mentioned with bolting the column to a wall or something that functions in a similar manner is this. When the wall/metal/whatever expands, if you column is bolted to it, the column will move in the same manner. While I am not sure how much the metal frame will expand, I would maybe just put a bit of thought into it and see if it might possibly affect column tram. If it was bolted to a wall, it would definitely be a consideration as a wall will move back and forth as it heats up during the day.

Also, living in New Mexico, which has similar conditions to a good bit of Australia, I would choose teak. I've made a few plant arbors out of it and it has held up great.

MY GOD!
I can't Believe how anal retentive some people are on this site we are talking about a $1000-$1500 dollar hobby machine that's only designed to take light cut weener cuts.I think people should stop getting so fixated on numbers and get in the real world .Most parts people are ever going to make at home are going to be fine with a tolerence of +-.015mm.If your little weener machine start to panic a but with you 1mm deep cut maybe it 's too deep back it off and try again.

Also can sombody help me make a dicision i am builing a back deck for my house and i dont know what timber to use .i am thinking of treated pine but it seem to have to much expansion in areas with differing rainfall and think i will have trouble maintaining a consistand gap between the planks or should i use tallow wood this is a more expensive option and will offer less flex underfoot than treated pine ,i dont know !Help .
im gunna have a beer and think about this some more

I enjoy improving something if I can. That is what this hobby is to me. If that means the machine, so be it. You don't have to read it, you certainly had nothing to add to this post, so I suggest you go back to what ever you were doing prior.

With your thinking, we would have never been able to take an airplane from an idea and then turn it into the space shuttle.

Your wood comment was a waste of typing and webspace.:tired:

A column of uniform cross-section will twist uniformly over it's total length under these conditions. So if you want to be most effective in reducing this torsional deflection it would be best to stiffen the column over it's entire length.

Phil

[QUOTE=Rodm1954;426863]

My thoughts are to fix strengtheners to the top of the column as this is the point where maximum deflection will occur.

QUOTE]

if you twist the head hard with about 20kg force it stays in a little sideways position until you move it again. i used to do this for better flatmilling results :D
and also the material used in the colum is almost dustlike if you drill in it. its pretty weird stuff.

i pitty stueyde, actualy the forum moderators are at fault here, spreading negative should result in a (temp) ban.

I had a conversation with Richard from Syil America last night, and we actually talked about this a bit. The conversation ended with the consensus that with properly adjusted Z-gibs, you should have a sideways deflection of about .0005. I measured mine when I was on the phone with Richard, and I had a deflection of about .0025. After adjusting my Z gib, I was back to the .0005 deflection.

Wade

Could you post the details of this? Is there a special process that he recommended, or was it the make it to tight, then back off a bit adjustment?

I'll try and explain this in text, but I don't know how successful i'm going to be. I am planning on doing a small video on this but that may take me a few days.

Here goes:
If you are looking at the X3 head on, move to the dovetail on the right side of the machine. On the very top of the head you should see a big fat flat headed screw (about 3/4" diameter). There should be a matching one on the bottom of the head.

Unscrew the bottom screw about 3 1/2 turns. WARNING: AT THIS POINT, DO NOT MOVE THE Z AXIS OF YOUR X3 OR YOUR GIBB MAY FALL OUT OR BECOME VERY LOOSE. THIS WOULD NOT BE A GOOD THING!!!

Now, screw the top screw in as far as you can (Finger Tight, NOT Gorilla tight). On the Gibb, there is a little step it the top of the flat screw rests upon. Beneath the flat head of the screw, there is about 1/8" space then another flat part that is attached to this strange looking screw. It looks like a washer permanently mounted to the shaft of the screw. NOTE: YOU MAY NEED A FLASHLIGHT AT THIS STAGE TO LIGHT UNDER THE HEAD OF THE SCREW.

Now, unscrew the screw until the top of the washer comes to nearly the same level as the step in the gibb. At this point, loosen it 1/16th of a turn more, then tighten the lower screw. Your gibb should now be in adjustment.

I never new about adjusting the gibb until I talked to Richard. First non-sherline mill, what can I say? Because I hadn't adjusted that gibb in months (about 6 <wince>) my adjustment was a bit different. I just backed off 1/8" turn on the top screw once I tightened it all the way down. Then, when I tightened the bottom screw, I only went in about 1 turn (initially I loosened it about 3 1/2 turns).

So, after I did this (it took about 2 minutes) my left/right runout when from about .0025 to each side (total of .005) to about .0005 total which is what Richard said the tolerances should be.

Additionally, he said that if you see any black residue on your dovetail, that indicates premature wearing of your gibbs due to lack of lubrication. He recommends 2 drops of oil on the ways on each side of the machine as the first step to every new day of work on your X3. That includes X,Y and Z axis. The gibbs should be adjusted every 3-4 weeks on average use.

Hope this helps and actually makes sense.

Wade

Here goes:
If you are looking at the X3 head on, move to the dovetail on the right side of the machine. On the very top of the head you should see a big fat flat headed screw (about 3/4" diameter). There should be a matching one on the bottom of the head.


I just went and looked at my Grizzly X3 - there is no screw on the bottom of the head to match the one atop. All that is on the bottom is the tail of the gib protruding.

I've just played back and forth with the Z gib to get a good balance between rigidity and ability to move the head back up the column. I lock the gib during most activity but the screw is SOFT and I've had to coerce it back to straight to be able to use it effectively.

Back on topic - I plan to investigate packing an epoxy/steel rod mix in the column once I move to the Z portion of my conversion. I'd love to add some rigidity to the column. I just got the case hardening ground off my ballscrew today for the Y, tomorrow is the X, or maybe the normal lathework on the Y, or both...depends on if the snow lets up and I can move the car out of the garage to keep from spraying it with grinding fluid. Keep posting your ideas on these concepts.

Ohhh, dang. I guess the one I have is the Super X3. I guess there is a bigger difference than I though. If you have the standard X3, then apparently there is a difference in the Z gib. Does your manual say anything about adjusting the z gibb?

Wade

Just the X and Y - no mention of Z. That section must have been eliminated in favor of the page of pictures of the circuit board, the power switch from the inside, power light from the inside, etc. and the 2 pages of accessories. I doubt the Harbor Freight version has anything different.

I'm pleased with the operation currently, just guess the Super X3 is that much more - Super!

so I took some measurements today with an indicator and a bathroom scale, and with the help of an assistant.
With 35lbs of force on the side, the column deflected 0.005" sideways.
With 70 lbs of force on the back, the column deflected 0.006" forward.
This was measured 26" from the base, with the head taken off, and force applied directly on the column. So this has nothing to do with adjustment of the gibs.
I tried measuring the front and back deflection every 2" down from the top under the same load (70 lbs) to obtain a deflection curve. The data I got was very dispersed, but still suggested the deflection was linear, which means that most of the deflection is occuring at the base of the column.

So now that the large deflection is confirmed, the next step is to design something to reinforce the column.

Hi all,

I don’t have an X3 but I do have a similar sized bench top with a square column that would benefit from some reinforcement, so I’m following this thread with great interest.

I’m thinking something similar to this http://www.cnczone.com/forums/showpost.php?p=249849&postcount=5 inspired by Rodm1954 excellent thread http://www.cnczone.com/forums/showthread.php?t=29650

Beefing up the column on my mill will make the world of difference so I’m hoping this thread does not fade away with out conclusion. (group)

John

I redid the experiment today, with a lot better accuracy. I plotted the results, and attached the graph.
Basically, I put a bathroom scale on the top of the column at the back, and applied 70lbf, and measured the deflection every 2 inches down from 26 inch.
As you can see, the curve is nearly linear, and from this we can conclude that most of the deflection is coming from the support of the column. So if we can find a way to brace the support, we can increase rigidity drastically.

Thanks Lagfish

That is easy to understand information for me.

Stiffenening the column is a problem because the connection between the head and the leadscrew/ballscrew wraps around the column. The choices as I see it are to attach as support at the top and tie it back to the base or stengthen the column/base connection.

I favour the connection at the top because it adds more support to the column. This can be done on the side/s or rear of the column.

Is the support of the column the base? Could you post an image?

The connection at the top would strengthen both the column and the support. Although I think the column is rigid enough for a the stock motor. However, this would require a lot more material just to get it to the same stiffness as strengthening the support only. But strengthening the support isn't easy because of the design of the X3.

Looking at the mill, I'm not even sure where this deflection comes from. It's not from the column, because I measured the deflection curve and did calculations to prove this. It can't be from the bolts, because there is only about 300 lbf applied on the bolts from the moment created by the cantilever column, which is not enough to cause 0.006". I think it might be from the base casting itself.

I think it might be from the base casting itself.

That is a worry.
From what you say I am imagining that the base is flexing on the horizontal face. I asume you had the dial gauge on the X axis table so it really could be anywhere on that plane. My mill is bolted down so it is not easy for me to tip it up but do you think it is caused by thin casting and/or weak webbing?

What about an additional column support tied in to the vertical face at the rear of the base? The only other way I can see to do it is to get a large chunk of plate and bolt the mill to it then build a column support off the plate.

I saw a post about someone filling the whole base of their mill (forget what type) with a quartz-granite material and bolting the column through a steel plate on the bottom. Ambitious - there's a lot of cast iron in the way with how the X3 relays the Z shaft through that area. It was a consideration of mine for the future. Have to finish my CNC conversion first.

Perhaps a box steel frame could assembled going out from the base--to a riser, which parallels the column, and connects to the top of it, with shims, bolts, and pins. The riser itself could be a triangle of wielded box metal, sitting on extensions that go under the base.

Here are some FEM links, if anyone is interested.

The open ones are real bears to use. I just bought Rhino, and makes a nice mesh. I just have to figure out how to do the rest, which could only take me--months, but the links are there.

http://www.calculix.de/
http://web.mit.edu/calculix_v1.6/CalculiX/ccx_1.6/doc/ccx/ccx.html
http://www.freefem.org/ff3d/
http://www.geuz.org/gmsh/
http://www.geocities.com/Athens/2099/slffea.html

Yeah, I'm currently working on a truss box design, similar to the drawings I had on page 1. I have it drawn up in solidworks already, and when I have time, I will run FEA on Cosmosworks at the computer at school. I'm trying to figure out how much money I have to spend on steel before it is adequately rigid.
I did some rough hand calculations already based on pinned joints, and it's looking good.
I will also give some of those free FEA programs a try. Thanks for the input.

I have a Syil CNC'd Super X3 and I'm getting .005" deflection left and right (0.010" total )on the head when I apply 5-10lbs of pressure to the side of the head. I tried wwendorf's instructions for adjusting the Z gibb without any luck. I've checked the column bolts, the two large nuts that lock the head rotation, and I've tried locking the Z gibb and the head still displaces enough that you can see it by eye.

When I got the mill I removed the Z gibb and lapped exactly as Syil recommended and it seems to fit just fine. This gibb adjusts differently than the x and y. Is there some other mechanism in the gibb slot that I might have messed up?

I noticed something:

Looking at Arc Eurotrade's tear down manual (I don't kave an X3), it appears that given certain z/zed motor arrangements, it looks like the strap doesn't go all the way down the back of the column. If that's so, perhaps there's a little room for a pinned/screwed plate to triangulate the back of the column. It won't help a lot with twisting, but it might help with side loading.

The area I am thinking of has two grounding screws in it. The strap seems not to be able to go down there because the spindle would hit that table.

I would think that whatever you can do to make the column stiffer would be best done at the column base. That area has a lot of hole in it, that could be replaced by metal. The base could be stronger there too.

The support plate might need to be a box which fits around the z-motor.

Guys

I think maybe I was near the start of this thread last year- when I posted some (frustrating) deflection values for my SX3. There were a lot of comments following, but I just wanted to add/revisit a couple of things.

Firstly, at least in my case, the deflection at the spindle in the X axis is higher than the deflection measured at the equivalent Z position on the column. That implies that there is a twist component involved here. Maybe it is actual column twist, but in my experience with it, a much larger % is related to the Z dovetails and gib. It is sort of catch 22 on that, since if it is very tight, you get stiction and lose some Z axis precision in fine motion (i.e. 3d contouring). If it is looser, then you get more deflection of course. And, because the spindle is below the Z axis bearing center of rotation, you add this second twist moment. So now you have 2 twist axis issues added to any core deflection in the column or base.

I dont think either of these components are going to be trivial to 'fix.' I suspect that there is more core twist in the column than theory would suggest. Im not going to jump on the 'bash China' bandwagon, but the materials in a $1200 retail mill are not going to be best of breed, and that column is all of 100lbs max. And though I havent done any FEA on the design, it is not clear to me how the shallow horizontal ribs will contribute in a significant way to anti-twist. But Im a composites engineer, and the rules there are different.

In my intial measurements, I found that deflection was worse with the head near the top of travel- which implies some bending or twist in the column. But, it was not gone with the head near the bottom of travel- not even 10% reduced as I recall. So my perspective is, with a greatly reduced arm/moment when the head is near the base, deflection there is likely bearing rock/twist.

Regards,
Rob

Now that I finally got some free time to work on the mill again:)
I did measurements just now with the indicator attached to the column at the same height as the tool.
I applied 50 lbf at the tool with gibs fully cranked, and measured around 0.001at the tool.
I loosened gibs all the way and got more than 0.010, with 50 lbf
I tightened to a level of tightness where I can still move the head with my hands, with it mounted on ballscrews, and I measured around 0.003 with 50 lbf.

My conclusions are these:
The head itself is rigid enough, as 50 lbf only deflects it around 0.001. This is good, because making the head more rigid would be hard to do.
Most of the deflection is coming from the gibs/dovetails as extrapilot said, but can be fixed by tightening the gibs. This also increases static friction (stiction), but could probably be overcome by using stronger steppers. How much torque could be estimated by measuring how much weight I need to push the head down, with the gibs tightened to amount to give an acceptable deflection.

I didn't measure actual column twist, but this could be reinforced by attaching a stiffening frame to the column like I mentioned before.

I made some FEA analysis on a frame I designed, and I'll post up some results in a bit.

so I've been working on a frame to stiffen the X3 column. I drew it in solidworks and did the FEA analysis using Cosmosworks. This is my first time doing FEA, so I don't really know what I'm doing. I also don't have any ways of verifying the FEA results with calculations. Please note that I'm not a professional engineer, so everything here is knowledge applied from what I've learned in school and also self taught. So if any real engineers want to critique my design / process, or give me advice, please do.

First image is the mockup of the mill with the frame.

I simulated the forces felt by the frame by applying 100lbf to a plate extrusion in the place of the mill base. I fix-restrained the frame by where the tool would be located, to simulate the torque that would be generated. I also put a sliding restraint on the bottom of the frame, as the frame would be sitting on the floor, but not fixed to it. The second image shows the restraints and forces.

The third image shows the stresses in the frame when applying a transverse (milling in the y direction) force of 100 lbf to the plate.

The fourth image shows the resultant displacement in the frame applying the same load. As the plot shows, the displacement at the center of the plate is around 0.001".

The fifth image shows stresses when applying a forward load (milling in x direction) of 100lbf.

The sixth image shows displacement when applying the same load, showing a displacement of around 0.0008"

The material is 1020 hot rolled steel
Total weight of the frame is 282 pounds
At CAD $0.72/pound for HR extrusions at my local steel supplier, and accounting for 10% extras/mistakes, estimated total material cost is $223.

So according to the FEA analysis, the frame drastically improve stiffness of the mill, at an acceptable cost. If nobody sees anything wrong with the design, I'm going to begin fabricating this either tommorow or monday, and it should be completed in a week or two.

i pitty stueyde, actualy the forum moderators are at fault here, spreading negative should result in a (temp) ban.

Banning stu for being too negative would be even more negative, in my opinion. Then again, do negatives add or multiply? If they multiply, the result is positive...:)

Het verbieden is niet het antwoord

sorry but im daily confronted by people that do nothing whole day and go annoy other people by by boredom. so im quiet sensitive when somebody posts negativity. especially in threads where users scientifically approach problems.

Bertmiller,
They only annoy you if you take it on board. Just some harmless opinions from somebody that looks like they were talking to the world through the bottom of of large tumbler.

sorry but im daily confronted by people that do nothing whole day and go annoy other people by by boredom. so im quiet sensitive when somebody posts negativity. especially in threads where users scientifically approach problems.

Actually, Bert after re-reading stu's post http://www.cnczone.com/forums/showpost.php?p=426779&postcount=13 I would describing him as being a rude elitist, rather than negative. He is basically saying if you want the performance of an expensive machine, buy one. From his perspective, people are expecting too much from light machines and then whining about their disappointment. To stu that is being negative and he finds it annoying.

Not everyone can afford a heavy machine, and thinking about simple measures to improve a light machine are certainly worthwhile for those that can't afford better machines.

Some posters come across as really negative, but still make a useful contribution, whether they intend to or not. A poster by the name NC CAMS springs to mind. He grinds camshafts for a living and keeps telling hobbyists that they shouldn't even try to do it. I want to ask why he has to protect them from failure - help them or shut up?
Even so, he actually does give a lot of good advice. For instance: http://www.cnczone.com/forums/showpost.php?p=329887&postcount=6

Then you have posters that don't contribute to the discussion, they just harass others. These are the ones that may be banned. I don't think stu fits in with the latter category. Stu may be a pompous ass, but what he said actually makes sense if you have money. It certainly makes sense if you are running a business - don't buy machines that are too light for the job.

He is rude, but he is also negative. He says that you should get over this problem and buy another machine. If thats not being negative, then I don't know what is.

I can't see the point in posting such a message. He should just have left this topic alone, and we should just ignored his post.

This topic is about someone who spotted a problem, and wanted to know where it comes from and probably how to fix it. I see no problem with that. I would certainly not start throwing crap at people for it.

Anyway. This is not part of the topic, and this discussion should probably end soon.

I don't have an X3, but I have the X2 and I can also see a lot of deflection when taking heavy plunge cuts, but I have some thoughts on strengthening the collumn with some sheets of steel. I reduced the deflection a lot by lapping the ways on my machine, and tightening down the gibs. Most of the slop can actually be found there, and I believe that you can make these machines strong enough with some time and good ideas.

There have been quite a few people who have worked on the x2 and come up with great ideas for it. Hoss springs to mind, and another gentleman who used the column with a bigger travel table. He went all out if I remember correctly. There was also somebody who did something similar to what I suggested, and he designed it so you could use the brace to tram the column. It was really slick. Someone did the same thing with a taig and it looked and worked great.

Now back to the x3. Right now I am converting my new-to-me emco to dc. After I am going ahead with my initial thought of making a support that bolts to the column below the lowest travel of the head and to the base of the machine. It will not support the head up high or prevent twisting, but since the biggest issue I have is the column leaning forward, that is what I would like to rectify. I will also incorporate adjustment screws to help tram the column and head, since it a PITA to do it with the 4 bolt down screws. I have also started the epoxy fill of the column just like Bob Warfield and a few others have done to their IH mills. Just as a first impression, WOW. Its not completely filled yet and already weighs considerably more and sounds, like Bob mentioned, dead. It doesn't vibrate like it used to. It was also pretty simple with just some sand, small small roundish pebbles, and rubber shavings. The pebbles are along the lines of the zeospheres that are being used in the epoxy build thread, just not quite as small, and the rubber shaving came from a container of dried up Herculiner. I also intend to add some beef to the column bolting area but haven't quite figured out what. I also figured I would see how it works with the support before doing anymore. Oh yeah, the really cool hockey puck feet are also being made, and they help with vibration too. And have a much bigger footprint for contact with the bench.

This whole little project cost me less than 60 bucks, just a bit less than an big burly mean machine capable of gobbling up iconel and Ti. And it was satisfying to boot.

With my factory CNCed SX3 I am doing some very accurate milling with Z keeping a tolerance of better than 0.01mm along a flat surface skimming laminate off 0.5mm stainless sheet over a length of 250mm and width of 90mm using a sold carbide 3/8" carbide slot mill. 0.2mm DOC in fiberglass.
For what the machine is, once the gibs are adjusted ACCURATELY and everything is tweaked, I reckon this is pretty good.
For the first fifteen minutes when I start to run this 4 hour job I lean down on the head with 2 fingers to tweak by what I estimate to be 0.007mm until the temperature stabilizes, then walk away and leave it.
Each part runs in 2 minutes 45 seconds. 20 parts an hour for 4 hours.
Makes me $80/hour. That's value I reckon.
By the way, none of this would work until I completely redid the gas strut balancing correctly. See these 2 links.
http://www.cnczone.com/forums/showpost.php?p=438154&postcount=3
http://www.cnczone.com/forums/showpost.php?p=397011&postcount=4
My base is fixed on the standard cabinet top, which is pretty tightly boxed, so I think this adds to the base stiffness considerably.

Yes. Sure it's not super stiff, but take one fine finishing pass after rugged fast roughing and it's OK with me. I used a 10 ton machine that had more head tilt, problems than this.

As for head tilt I move the head tilt, (mostly cos of my tight gibs) when I move down I over shoot 0.02mm then backoff 0.02mm and this unloads the head so it sits nice. If the gas strut over compensates the weight then give up. !! I lube the slides with teflon stuff I've had 25 years. SLICK50.

So you filled your X3 with the epoxy compound? Based on what you have done so far, to you think the full gallon kit is what is needed for the completion of an X3 fill?

I have been reading Bob's results and they look very promising indeed. Between the support design above using HSS tube and the addition of the epoxy in the tube frame and Mill, that might make for the most rigid X3 design.

I've been planning on remachining my column in order to accept precision linear rails and bearing blocks to bolt the head to, there by eliminating (to a certain extent) the slop from the gibbs. Anyone see any issues with this idea??

That still does not take care of the deflection problems from the column but it can eleviate slop in the gibbs.

I've been planning on remachining my column in order to accept precision linear rails and bearing blocks to bolt the head to, there by eliminating (to a certain extent) the slop from the gibbs. Anyone see any issues with this idea??

That still does not take care of the deflection problems from the column but it can eleviate slop in the gibbs.

That would be a good idea. You're probably aware of 5-bears research's site. He modified a table to accept linear rails. http://www.5bears.com/cnc09.htm
The only issue I see is adding the rail would offset your tool a bit, so that your tool is no longer at the center of your y travel. But if you convert it to CNC, you can make the Y travel longer anyway by using a longer leadscrew and/or drilling a hole at the bottom of your column where the leadscrew would hit. Infact, I have some heavy duty precision rails lying around from a big robot, but I might save those for a router project...

I already converted my X3 to CNC however I could always offset in Y by adding a shim bracket to the motor mount bracket. I have a set of THK HRW27 rails lying around that I was going to use on this project.

Nice thing also about implementing linear rails in the Z axis is lowering the stiction of the axis and increasing efficiency of the drive system.

Hmm, My reply seems to have gone missing. Apologies if this turns up twice.

I'm the guy RodM mentioned in an earlier post.
http://members.westnet.com.au/pjsconsulting/images/cnc/big_spine_x3.jpg

For more info: http://www.cncathome.com/X3.html

Cheers
PK

OK, sorry it took so long for a reply. I actually was too lazy to go an look for a real epoxy compound so INSTEAD I used a can of rhino truck bed lining instead. I mixed it up with sand and 25lbs of glass beads from Harbor Freight. I let it sit in the sun for about a week, 100 degree days, and when I came back on my next days off everything had hardened nicely and the base was probably 55lbs heavier and because the rhino lining is rubbery it made the base sound dead. It would not resonate at all. The only thing is the top looks like crap because I was too impatient to slowly add layers and instead poured the base in one go. So there are a lot of air bubbles at the top. Because I added so much sand and glass beads, the rest of it is nearly rock hard, it feels slightly softer than a hockey puck. I havent had a chance to run anything, but the damn base was a heck of a lot heavier.