[peteeng]

Hi Lars - Ensure you have clear access to the carriage bolts so they can be re-tightened or checked without pulling the machine apart. They can come loose and you don't want it to be a mission to get at them. One design rule I have is that all bolts and screws have to be accessible without them being buried. Learnt that one the hard way. Plus if you have a blind hole make it deep so you minimise the issue of bottoming a screw. Peter

re: car spacing - The rule is "as far apart as possible" A good arrangement is that they are square so if the rails are 200mm apart they are 200mm apart in the other direction. Obviously the further apart they are the less travel you have if your rail length is fixed. But a compromise has to be met...

[joeavaerage]

Hi,
that is the advantage of having a saddle on each axis (X&Y). Should you need to get axis to the Y axis cars screws, the ballnut mounting screws then a half dozen
screws around the periphery of the X axis saddle and then the whole X axis lifts off and thereafter you have access to the Y axis saddle. Flip the X axis upside down and you have
access to the X axis saddle car and ballnut mount screws.

A single solid saddle between the Y and X axes means you have to assemble it like a 'layer cake'....not at all easy.....and worse once the layer cake is assembled you cannot retorque
any of the screws of the internal or bottom layers, exactly as peeteng is warning you against.

Craig

[peteeng]

Hi Lars - If you study images of commercial machines you will see how they resolve this issue. It is possible to design a machine that once built all bolts can be accessed. It does add another layer of thought to the Chinese puzzle but it will work out better in the long term. Peter

[joeavaerage]

Hi,
its as easy as falling off a log,......split the saddle into two, one on the Y and one on the X then bolt the two together.

I made my saddles square, so I could use just one drill setup to do all three saddles, and therefore they are all identical. It also meant that whe I rotated the the X axis saddle 90
degrees the Y axis saddle bolts line up perfectly.

This solution is so very VERY simple, and just the ticket if you are trying to do it yourself. If it were more complex then you'd have to get this work done on a jig borer in order that it be accurate,
and therefore pay bigtime for it to be done.

Craig

[Fatamorgana]

Something like this? I wont Get them square as Im locked to the long ballscrews etc.


Sent from my iPhone using Tapatalk

[Fatamorgana]

Always check how bolts will work

You will likely want flanged linear rail cars - they can be bolted from below

Thanks! Changed that now


Sent from my iPhone using Tapatalk

[joeavaerage]

Hi,
yes,that is similar....but I did not attach the ballscrew support to the saddle, but rather upside down and to the 'slab'. The ballnut mount is screwed down to the saddle.
The servo hangs out of the end of the 'slab', better called the X axis bed, but the servo moves with the axis.

You can now also see the premium of having a shallow gutter so that you can still mill the T slots without loss of rigidity.

By the way the through bolts you have shown holding the Y axis saddle and the X axis saddle together are directly under the rail where you can never get to them to install, torque or at need,
remove. Bring them to the outside so they can installed without interference. Even better should you want to disassemble for repair or adjustment you can access the screws easily and
then lift the entire X axis off to get to the Y axis underneath.

Craig

[peteeng]

Hi Lars - Can you put the bolt heads in so we can see the intended side of the car they will be? One side they will be buried the other side they will be open. If you use the flanged cars so you can use the "bottom" side they take a smaller screw, than if you use the top side. Always better to use the larger screw if possible as its stiffer. I don't understand the need for the spit saddle, one part is better than two? If you do go spilt I'd have more screws in it then 6 to hold it together. Plus I'd use dowels once settled.. Peter

[joeavaerage]

Hi,

I don't understand the need for the spit saddle, one part is better than two?

As I've said before if you make a single piece saddle then you have to allow screw access where the the top cars interfere with the screws of the bottom cars.
Additionally it puts an extreme premium on the need for accuracy otherwise the axes are not 90 degrees but 89 degrees 58 arc minutes....and that's just rubbish.
They need to be square to each other by a few arc seconds, and that is no not possible on ordinary equipment. If you have a split saddle it allows easy access
to the car/bullnut mount screws but it also allows you to 'wiggle' the axes to be absolutely orthogonal before clamping them up. I have a single dowel in the center
about which the two split saddles rotate.

As for the rubbish about one part being better than two???? How do you tram the head of a vertical mill? There are several bolts that you loosen and then you can rotate the head a little back and forth
until the head is trammed and then clamp them up again. That is how vertical mills have been made for decades if not centuries! That is by definition two parts with a restrained rotating joint between them.
Where is the difference between a vertical spindle to Z axis (rotating) joint and a X to Y axis (rotating) joint. Is one immune to forces to which the other is subject?....no.

This is an extremely practical solution, especially as Lars has a mill. If you are careful you can use the mill to do very accurate drilling provided you always work in one direction so no backlash
creeps in. It is for that reason that I made my saddles 'square' as it meant I needed only one drill pattern and that was possible to replicate near perfectly over all three saddles. When I set the X axis
atop the Y axis and measured with my granite square I found that the two axes were within about 10 arc seconds (of square) just by dropping the bolts in without any tweaking or wiggling.

This is a 'night and day' difference to my first mill where I had a single piece saddle and really struggled for accuracy, two piece....a breeze.

Craig

[joeavaerage]

Hi,

If you do go spilt I'd have more screws in it then 6 to hold it together. Plus I'd use dowels once settled.. Peter

I disagree.

I have two....read that, just two 8mm cap screws holding the two saddles together. Been like that for three years, perfect. About ten days ago I had a cock-up which saw the X and Y axes accelerate
away and crash into the end stops. The machine was homing and consequently the limit switches were ignored, surprised the hell out of me. No harm done....except the X and Y axes got a little
out of square as one of the screws holding the saddles together yielded. Took an hour to replace the screws and set the axes square again. It took a major crash to do that, the first in three years.

I invite anyone who has made a single piece saddle between two orthogonal axes to post some pics and run through the design procedure.

Craig

[peteeng]

Hi Craig - Here's how HAAS do it. They make 16 a day just for the VF-1 machine. Peter

https://youtu.be/1s7tPqFtBRE

[joeavaerage]

Hi peteeng,
I'm sure that's great for Hass, they have superbly accurate machines on which to make parts with jigs and fixtures to attain um accuracy.
That does not help us hobbyists whom are producing just one machine, and often on a much lesser machine to start with. In Lars case he has a nice rigid machine, but whats the bet its done
a million miles already. That is not to say that excellent work cannot be produced with it, but you need to approach it very carefully to avoid tricky and multiple setups. The split saddle is the
approach I used to make the best possible parts on a manual mill, remembering also that I'm not a trained machinist. I must perforce use techniques and ideas that lend themselves
to accurate parts WITHOUT relying on my skill or a superbly accurate machine.

The two saddles are both 275mm x 275mm x 20mm square. They are ground in a surface grinder and so are flat and parallel to within a few um. It takes only modest clamping force
between two such flat surfaces of that generous area to get very VERY good clamping indeed. To whit: it took a major crash to upset my setup...and that is with only two 8mm cap screws.

When, or if, you ever get around to making a similar arrangement....by all means use a single piece saddle and tell us all about it. I came across the split saddle idea from a friends Taiwanese made
knee mill where the X axis can rotate with respect to the Y axis.....and I thought 'that's f*****ing brilliant....its just what I need!!'. Using that as a basis I was able to devise a very simple, repeatable
procedure to get saddles with um level precision but on a manual mill and with me as machinist, a hobbyist machinist at best.

Craig

[peteeng]

Hi Craig - The industry has been making single piece saddles for at least 150 years, well before we had cnc machines. I'm sure Lars will figure a way forward. His Russian mill just needs a good checking over so its straight, square and true and I'm sure he could make very good parts on it. First preference with any machine build is to make accurate parts. Accurate parts implies they assemble accurately vs being wiggled into position... There's enough wiggle room in even snug clearance holes to align these sort of parts. Peter

[joeavaerage]

Hi peteng,

I've taken pictures of what I've done and reported how it works....I'm not talking about some abstract idea....I've made it. I'm not talking about what some manufacturer does or did,
I'm talking about what I have made.

I made a single piece saddle with my first mini-mill and I swore I would never repeat that same procedure. It requires multiple setups of the same part on different
sides (upper and lower) with perfect orthogonality....and that's just not realistic for a hobbyist on normal machinery. It requires a jig boring machine....and there are
plenty around, but none of them are under $150 /hr.

Then, after seeing my friends mill, I came up with a two piece saddle. Vastly easier to design and make and can be made to be as accurate as you can possibly measure without
huge skill or a jig boring machine.

enough wiggle room in even snug clearance holes to align these sort of parts. Peter

BS. Sure there is a hint of wiggle room in just about any bolt hole but to align the X axis 90 degrees to the Y axis requires that all 16 of the cars screws be loosened and the four ballnut mounts
be loosened and that they all have say 5um movement in one direction or the other. Not going to happen. Whereas two, four or even six through bolts loosened can be assured to provide the
few um of movement you might require.

As I posted earlier, when I first set my machine up I just dropped the bolts through and lightly torqued them before measuring if the two axes were square and to my surprise it was damn near perfect as is.
Did not need to wiggle it at all, just clamp them up tight. I made these saddles on a near new Taiwanese knee mill, a nice machine and in perfect condition, but neither is it an Okuma or DMG Mori, and yet
just by using it carefully I was able to drill the holes such that all I had to do was bolt it together, no wiggling required. A damned good result.

I'm recommending to Lars a design I have used and made with good success. Have you designed and built something comparable? Perhaps you'd post the pictures of it.

Craig

[peteeng]

Hi Craig - I have no argument with what you have done. It works for you. I'm just stating that industry has done it for some time a different way. I have built several types of saddles, sorry no pictures to hand, any precision machining I require I contract unless it's something I want to have a play with. Peter

[peteeng]

Hi Lars et al - I have designed hundreds of saddles but only a few have matured. Saddles are important but at the same time not. You can even get crossed bearings so the saddle is not needed. I have tried to use crossed bearings a couple of times but driving them is a problem, you end up with a saddle again. Maybe you can solve that design conundrum if you want to explore this type of bearing. Peter

https://www.aliexpress.us/item/32568...2usa4itemAdapt

Last time I played with these the PMI brand did not have screw holes on the sides. So I could not attach anything to the bearings. These ons have threaded holes on the sides and I can see I could put a bridge across those to hold a drive nut. Interesting and doable... I'll revisit this next coffee napkin sketch

[Fatamorgana]

Hello and thanks for all the feedback!

Im currently working with the saddle x/y, I have made it alot smaller and moved the through bolts out to the sides, 6 in total, also have put inn m8x25 bolts for the flanged cars.

Im really struggeling to see all the benefits of turning the rails upside down as this make me loose alot of travel x direction, I have to put in about 440mmx80 steel just to give the rails something to attach to, the mill cant reach out there, Im looking for a travel of 450mm, and for that to happen I have to make the x-axis 880mm wide, just to use 450 of it for milling.

Any points of direction here? Im have worked my way through all of the cnc builds both professionals and hobbyist, and most are build the traditional way with rails sitting the the classic way, so as I said Im struggling here.

[joeavaerage]

Hi,

Im really struggeling to see all the benefits of turning the rails upside down as this make me loose alot of travel x direction, I have to put in about 440mmx80 steel just to give the rails something to attach to,

Wrong. The travel is the length of the rails, less the length of one car and less the center to center distance of the cars. For example:
rail length =800mm
Center to center between cars=250mm
length of one car=80mm

Max travel =800 - 80 - 250
=470mm

If you want more travel you have to get longer rails OR reduce the distance between the cars. Longer rails will mean a longer 'slab' of steel, the rails will after all have to be bolted to it whether they face upwards of downwards.
You try to retain as much separation between the cars as it aids stbility.....but reduces the travel foe a given length rail. The saddle need only be long enough to contain the cars, ie 250mm + 80mm =330mm

Travel is unaffected by whether the rails are upright or upside down. With rails underneath the X axis bed you do not have to provide way covers, whereas if you have the facing upright you will
require way covers for both X and Y axes. I promise you to design and build reliable way covers is a PITA.

Im looking for a travel of 450mm, and for that to happen I have to make the x-axis 880mm wide, just to use 450 of it for milling.

Yes that is correct, most machines are like that. Look at your Y axis, the slab is at least as long as the rails is it not? Then why would you try to make the X axis shorter?

If you have 800mm long rails then you need the slab of steel 800mm long the mount them....no getting around that is there? Whether you have the 800mm slab with the rails mounted on top, and a saddle bolted
to the cars and the vice on top of that.....OR.... you have the 800mm slab (conveniently with T slots milled into it) and the rails bolted upward from underneath you still need that 800mm slab.

Any points of direction here? Im have worked my way through all of the cnc builds both professionals and hobbyist, and most are build the traditional way with rails sitting the the classic way, so as I said Im struggling here.

There is no one standard way. My first mini-mill had the X axis bed facing upwards. There are two reasons I did not like it, first was that all the swarf falls straight down and onto the ballscrew...but worse
was now I had to putt the whole axis bed in the mill upside down and drill to accept the Y axis cars and ballnut....and that drilling had the be perfectly orthogonal to the X axis rails which are on the otherside....a setup nightmare.

My new mill I decided to put the X axis upside down, that is to say with the rails bolted upwards from the underside. This meant that they are largely protected from swarf etc. The saddle is now quite small
275mm x 275mm and very much easier to put ion the mill and drill for the cars and ballnut mount. The axis bed is 700mm long, 240mm wide and weighs 115kg. Note also that you have to put it in the mill
aligned with the Y axis....ie you need a bloody big machine to do it.......more often than not you'd have to pay someone to do it for you whereas I could do the 275mm x 275mm saddle myself, rather than contract it out.

Craig

[joeavaerage]

Hi,
you second and third pics are wrong....you cannot have the rails hanging out like that...they have to be bolted to something and something stiff. The rails are nowhere stiff enough to be
used unsupported.

also have put inn m8x25 bolts for the flanged cars.

The flanges of the cars are predrilled and tapped. You have to go with how they are made. They are bloody hard, there will be no drilling and re-tapping them. If I'm not mistaken
cars for 20mm rails are 5mm dimeter, cars for 25mm rails and 30mm rails are 6mm diameter.

Craig

[Fatamorgana]

Hi Craig.

Ok, that makes sense to me, trying to se all aspects of this, not easy, thanks for enlighten me.

Regarding the bolt for cars it seems they are m8 according to the chart that I have attached here.

I have made some changes to the design, it was mostly to illustrate my issues that I made those changes. My concerns for making the millingtable/x-axis so massive was if the cars could withstand the "hangout" and the weight of the x-axis when the table is on the outermost side. I have attached some pictures of it. One design I have removed alot of material to compensate for this, but that also have its toll on stifness I would guess, so Im kinda of unsure here of wich design to go for.