Town Woodhouse milling machine re-commission and cnc conversion

[daleyd]

Forgive me for being stupid but why does everyone use ballscrews for cnc?
Why does accuracy matter? Can't a cnc device read off DRO type scales and move everything related to the readout? That way play and backlash etc wouldn't matter or would it?

My understanding is that cnc machines can mill in all directions (or at least ideally they should be able to)- if you use a conventional lead screw you couldn’t do this, as some directions you will be climb milling and run the risk of the tool dragging the workpiece in and wrecking it due to the backlash.

 

[Fastdruid]

Forgive me for being stupid but why does everyone use ballscrews for cnc?
Why does accuracy matter? Can't a cnc device read off DRO type scales and move everything related to the readout? That way play and backlash etc wouldn't matter or would it?

It's not for accuracy, acme can be ground to equal precision and you can use DRO on both.

Two main reasons. First is efficiency, an acme is about 40% efficient while a ballscrew is about 90%. If you take my previous example, it takes ~15Nm to lift the table on my Harrison but the actual required torque is only ~2Nm! This affects everything, it needs bigger motors, it's harder to overcome the stiction as well as the constant friction and you end up with either far lower speeds or *much* bigger motors.

Second is backlash, you _can_ eliminate backlash from acme but it comes at even more of an efficiency cost. Ballscrews can easily have negligible backlash _without_ significant efficiency cost.

Backlash is particularly important with CNC because you're going to struggle to make anything round or stick to any kind of tolerance with any backlash. It's not quite so important with manual because you're almost always just doing straight cuts and if you need to do a curve put it on a rotary table. There is a certain amount of backlash elimination you can do in software but it can really affect things.

 

[minimutly]

Well yes...and no as long as they are correctly specified. See page 34 here https://www.hiwin.com/pdf/ballscrews.pdf

> The static load is the load which will cause the balltrack to have a plastic deformation exceeding 0.0001x ball diameter. To calculate the maximum static load of a ballscrew, the static safety factor Sf of the application condition should be considered.

If again we take 400Kg to be a reasonable load with a single FSC ballnut https://www.hiwin.com/pdf/ballscrews.pdf

A 25mm 5mm lead ballscrew has a Static Load Co of 4612 (Kgf), so we can see that actually 400Kgf x 2.5 =1000Kgf which is < 1/4 of the Static Load Co.

This is a good link, thanks, just what I needed.

 

[minimutly]

This is a good link, thanks, just what I needed.

I take that back - too much information - my head is fried!

 

[minimutly]

Today's (well the last few nights tbh) progress.

 

[minimutly]

Today's (well the last few nights tbh) progress.
View attachment 280810

So this bit goes into the knee on one side, there will be a row of vertical bolts at the front, another further back, and I've made an infill further back so that there is no bending moment on the bolts. The sticking out bit is where the in/out carriage clamp runs. Once the part is fitted I will finish machine this area.
Talking of machining, I made this out of a big lump of cast iron rescued from a job some years ago - in fact I wasn't really sure it was cast iron before I started. By tonight I've had a bellyfull of the black messy stuff...
More needed on the other side of the knee.

 

[fizzy]

Coming on very nicely. I am so glad I resisted scrapping it despite all the philistines telling me I should!

 

[minimutly]

Coming on very nicely. I am so glad I resisted scrapping it despite all the philistines telling me I should!

Bah, what do they know. I will admit to considering buying a knackered turret mill and fitting the TW turret to it, but I'm glad I didn't.
Repairing it is one thing (still a way to go), but the cnc thing is way out of my knowledge base, and adding a load of options I'm not quallified to answer, yet.

 

[minimutly]

Here's a question for the scrapers on here - @Pete. and @pressbrake1 - do you scrape the dovetails on the knee, or the ones on the column, or both? Right now I don't have the equipment to grind them, so can I machine the ways, scrape and fit?

 

[Agroshield]

...do you scrape the dovetails on the knee, or the ones on the column, or both?

On the repair you are doing, it might be good to put a small radiused relief cut at any sharp internal corners (and chamfer/break any external corner that interlocks with the internal one). You do not want a stress riser to promote a crack. For any repair part that will be permanently in place, consider a thin layer of suitable epoxy before final squeeze of the bolts.

On the alignment, you have to start with the column: the ways of that have to be a flat plane with parallel, straight edges. And that flat plane needs to be square in two axes with the turret seat (or the turret seat made to be square with the rescraped column ways).

When you wind the knee from top to bottom of its travel, the column ways should confine movement strictly to linear movement along the Z-axis. Any aspect of their geometry that causes X- or Y- change or any rotational movement around Z- is undesirable.

It is a little easier to visualise if you imagine the machine laid on its back: you want the column ways as close to a surface plate as possible, with the added challenge that the front and back edges must also be straight and parallel.

 

[Fastdruid]

Talking of machining, I made this out of a big lump of cast iron rescued from a job some years ago - in fact I wasn't really sure it was cast iron before I started. By tonight I've had a bellyfull of the black messy stuff...

I do like machining cast iron....but can't stand the mess it makes!

 

[minimutly]

On the repair you are doing, it might be good to put a small radiused relief cut at any sharp internal corners (and chamfer/break any external corner that interlocks with the internal one). You do not want a stress riser to promote a crack. For any repair part that will be permanently in place, consider a thin layer of suitable epoxy before final squeeze of the bolts.

On the alignment, you have to start with the column: the ways of that have to be a flat plane with parallel, straight edges. And that flat plane needs to be square in two axes with the turret seat (or the turret seat made to be square with the rescraped column ways).

When you wind the knee from top to bottom of its travel, the column ways should confine movement strictly to linear movement along the Z-axis. Any aspect of their geometry that causes X- or Y- change or any rotational movement around Z- is undesirable.

It is a little easier to visualise if you imagine the machine laid on its back: you want the column ways as close to a surface plate as possible, with the added challenge that the front and back edges must also be straight and parallel.

What was that character in Vicar of Dibley yes, yes, yes, no...
The original casting broke because the whole lot fell over, not because of a crack forming in a corner, so no need to go crazy on stress relieving imho. However it's not difficult to do so I will give it some thought.

On the column ways being straight and true, and the fronts being flat I'm quite lucky in that there is very little wear (that can be seen anyway), so given I have no way of checking it I think just cleaning up the rust on the surface is all I can do. I'm open to suggestions on what kit I need to check this, and where to find it.
The ways on the knee consist of a fixed, cut dovetail on one side (much of which has been violently removed), and a square shoulder (a short piece of which I'm replacing), against which a shimmed dovetail is bolted in place, which I have. So this side will be easy. The dovetailed side much less so, hence the question about scraping the dovetails. I am thinking of getting some silver steel rod and making a parallel so I can measure any out of parallel before pinning and bolting the replacement dovetail in place...
Or do I need to make a piece that fits the space in between the knee sided perfectly, and then setting the new dovetail. But surely this would need the dovetail scraping in?
Maybe I should have kept the column on it's back and used that to set the new pieces? I guess I can still do that while upright?

 

[Agroshield]

The column is easier on its back simply because gravity is working in your favour and you can use a machinist's level to tell you a lot of things.

Maybe minimal kit is a level, a straightedge, a (big) micrometer, two ground rods and some gauge blocks. The very top and very bottom of the column ways should be unworn as the knee never touches them. The level across these the short way at the top will get the whole thing level. It should still be level the short way at the bottom, or there is a twist. The level on the straightedge on gauge blocks top and bottom the long way will level it the other way and check for twist in that axis. The straightedge alone will show you any convexity or concavity in the ways, held both parallel with Y- axis movement and X-axis movement. The micrometer and ground rods will allow you to measure the dovetail. Absolute number does not matter but should be the same through full length of travel.

It is difficult to give a short summary of what is needed and there is not one single reference with which to gen up on this. If I had to cite a few, it would be Connelly's Machine Tool Reconditioning (http://totallyscrewedmachineshop.com/documents/MachineTool-Reconditioning.pdf), Moore's Foundations of Mechanical Accuracy (https://archive.org/download/FoundationsOfMechanicalAccuracy/Foundations_of_Mechanical_Accuracy.pdf) and Schlesinger's Testing Machine Tools (http://www.etech.net.au/Testing Machine Tools.pdf).

 

[minimutly]

I'm thinking of making up a 60 degree setting tool to the width of the knee "wings" , less an allowance for a feeler gauge or something suitable. I have an engineering spirit level, and maybe I'll get a steel bar and get it ground.
I have a copy of Conellys machine tool reconditioning.

 

[minimutly]

Not much progress happened on this to be honest - I've done some tidying up on the side plate, need to re mount the knee now so I can finish machine that, then fit the plate and pin.
I have tried to straighten the knee "screw" - without much success tbh - the thing was bent right next to the thick section that the timken sits against, so I ended up with a set in it...
I have been reading up on cnc conversions - joined the mycncuk forum. I have learnt that the cost of these conversions can run into thousands, and whilst I really would like to be able to make dozens of widgets, or modify mini heads without at least some of the heartache and time spent, out laying two or more grand on a project seems crazy. Also, driving the knee has pros and cons, and I can't get the issue of drag/spring Vs looseness out of my head. Added to this weight of the whole thing, the cost, and having it all over my workshop in bits is putting me off the idea.
So, my current thinking is to go 3D, with normal xy, but making an attachment for the rear of the turret, with a motor and 30 int socket or collet mount, in an up/down carriage, driven by ballscrew directly. This means I can crack on with rebuilding the mill, have the use of it and the shaper while I devise the y head and carriage longer term. Sure, swapping the shaper for the cnc head will be a pain, but there we are. I can buy a dro with pickups for both my millers, so swap between them as required as well. I think...

 

[minimutly]

Not much progress happened on this to be honest - I've done some tidying up on the side plate, need to re mount the knee now so I can finish machine that, then fit the plate and pin.
I have tried to straighten the knee "screw" - without much success tbh - the thing was bent right next to the thick section that the timken sits against, so I ended up with a set in it...
I have been reading up on cnc conversions - joined the mycncuk forum. I have learnt that the cost of these conversions can run into thousands, and whilst I really would like to be able to make dozens of widgets, or modify mini heads without at least some of the heartache and time spent, out laying two or more grand on a project seems crazy. Also, driving the knee has pros and cons, and I can't get the issue of drag/spring Vs looseness out of my head. Added to this weight of the whole thing, the cost, and having it all over my workshop in bits is putting me off the idea.
So, my current thinking is to go 3D, with normal xy, but making an attachment for the rear of the turret, with a motor and 30 int socket or collet mount, in an up/down carriage, driven by ballscrew directly. This means I can crack on with rebuilding the mill, have the use of it and the shaper while I devise the y head and carriage longer term. Sure, swapping the shaper for the cnc head will be a pain, but there we are. I can buy a dro with pickups for both my millers, so swap between them as required as well. I think...

A little progress has been made - the "straight" side has been repaired with a cast iron insert, bolted into the knee with m10 socket head bolts. I've turned it around on the miller table and machined the wedge shaped side.

 

[minimutly]

The insert will be more of the scrap cast iron spar..

The plan is to make the insert fit the machined surface, and make a steel "t" piece to bolt over and through it, so bracing it back to the side of the casting.

 

[minimutly]

This is the knee with one side repaired..

 

[fizzy]

Coming on nicely lovely to see.

 

[minimutly]

Progress so far:- I'm going on the scraping course next week and will be taking this with me....
Fingers crossed I get the machining done today to fit it in the correct location in the knee