Cost effective way of measuring 90degrees very accurately on Milling machine

[Martin WhittleParticipant @martinwhittle67411]

Posted by larry phelan 1 on 01/12/2018 16:30:48:

No machine is any more accurate than the machine which made it

So would you like to explain how we got from chipping wheels out of stone to modern precision machinery?

Martin

[Men IfrParticipant @menifr84251]

0.3 in 100mm it’s unacceptable for me

[Men IfrParticipant @menifr84251]

Duncan, yes good idea.

[Men IfrParticipant @menifr84251]

Duncan, yes good idea.

[Nick HulmeParticipant @nickhulme30114]

Posted by Men Ifr on 01/12/2018 14:21:43:

Any other options/thoughts?

Yes,
State the make and model of the mill you wish to transform into a silk purse, with some it's an uphill battle, with others it's a lost cause.

[Neil WyattModerator @neilwyatt]

Posted by Nick Hulme on 01/12/2018 21:36:52:

Posted by Men Ifr on 01/12/2018 14:21:43:

Any other options/thoughts?

Yes,
State the make and model of the mill you wish to transform into a silk purse, with some it's an uphill battle, with others it's a lost cause.

 

It's an SPG 9512, a tilting column mill (like the original X2) so the 0.3mm in 100mm error is one of how it is set, not one of how its built.

It appears the head has also got knocked out of true to the column.There is a cure for this that doesn't need a square. Remove the head completely, the front and back are held together by four bolts. Loosen these and using paper to shim up the front part of the head with both parts on a level, flat surface, it should be possible to align it accurately with the rear part and nip up the bolts. You can then reassemble and tram as normal.

Neil

Edited By Neil Wyatt on 01/12/2018 22:31:50

[john carruthersParticipant @johncarruthers46255]

For an accurate square I use old optical prisms for 90d and 45d.
Even the worst prisms these days are quite good +/- 2 arc min surface 1/4 wave, while the better ones can be +/- 2 arc sec with a surface better than 1/20 wave.

**LINK**

[JasonBModerator @jasonb]

Just by way of comparison 2 arcmins is about 0.05mm/100mm

To get similar to a Grade B square you want 0.36arcmin or 21arcsec

[Michael GilliganParticipant @michaelgilligan61133]

Posted by JasonB on 02/12/2018 08:46:10:

Just by way of comparison 2 arcmins is about 0.05mm/100mm

To get similar to a Grade B square you want 0.36arcmin or 21arcsec

.

So the 2 arc second prism [John's 'better ones'] would be plenty good enough

… But the price may be off-putting.

MichaelG.

.

P.S. … There was a DIY AutoCollimator described in MEW

[Neil WyattModerator @neilwyatt]

Posted by Michael Gilligan on 02/12/2018 10:37:56:

So the 2 arc second prism [John's 'better ones'] would be plenty good enough

… But the price may be off-putting.

 

http://www.astroboot.co.uk is selling four porro prisms for less than a fiver, extracted from broken binoculars.

You could just grab some tatty porro prism bins from a car boot.

Neil

<edit> with a flat sheet of glass and some patience you may be able to find the best matching pair by putting them face to face and looking for newton's rings.

Edited By Neil Wyatt on 02/12/2018 11:39:00

[SillyOldDufferModerator @sillyoldduffer]

Posted by Martin Whittle on 01/12/2018 21:20:21:

Posted by larry phelan 1 on 01/12/2018 16:30:48:

No machine is any more accurate than the machine which made it

So would you like to explain how we got from chipping wheels out of stone to modern precision machinery?

Martin

Martin's comment suggests an alternative approach that does not require a high-precision square. Precision tools are useful in a workshop because they save time, and are easy to use. But it's possible to go back to first principles. One that might work in this case is dividing. Main disadvantage is it requires attention to detail and perhaps a number of attempts.

Mount a DTI in the quill on an L shaped boom as shown:

Position the DTI at North and set it to zero. Use a set-square to determine 'South' with reference to the end of the table. Flip the set square over to halve the error, even though exact alignment with south isn't that critical. Swing the DTI to that position and note the difference. Now tilt the head to halve the DTI reading.

By repeatedly halving the error it's possible to improve the alignment. It is not necessary to have a precision set-square,

On my Chinese mill there's a knack to tightening up the bolts when tramming in the normal right-left direction. Although in theory it shouldn't happen, tightening the bolts can disturb the front-back tram slightly. Adjusting the tram can be very frustrating and I avoid disturbing it if I can.

The accuracy of tram by this method depends on the length of the boom, the accuracy of the DTI as a comparator, and the flatness of the table. The skill with which the head can positioned (with shims if necessary) is likely to the limiting factor, at least it is in my workshop! Even if the table happens to be tilted, then the head will still be set at right angles to it. If the table is bent, then a regrind is necessary.

If you don't have a DTI the set up can be done with a rule, in which case a magnifying glass helps. It's not necessary for the rule to be graduated because you're halving the error, not measuring it. Dividing is surprisingly accurate and it's one of the ways in which we escaped the Stone Age.

Dave

[Men IfrParticipant @menifr84251]

All, Just for a bit of background reading that I though I'd pass on I found this:

http://totallyscrewedmachineshop.com/documents/Testing%20Machine%20Tools%20%28Dr.Schlesinger%29.pdf

I'm trying to do the column adjustments, the problem with this document and all others I have seen they assume the 90deg square tolerance << target tolerance so <<0.02mm in 300mm! you need a good square for that!

Also one of the most useful measurements I cannot do normally see fig 1 in the sheet -, run out of the spindle because you would need an arbor with tolerance << target tolerance 0.01…

1 caveat may be if you had a fairly good arbor say 0.01 or better then indexing angular position in the spindle and repeating the measurements say 4 times at 0 90 180 270 deg may allow you to calculate the machine run-out by subtracting arbor run-out. On my current mill I just measure run out on the upper and lower taper part of the R8 to give me an idea…

[Men IfrParticipant @menifr84251]

Oh Page 50 in pdf (48 from original book) for vertical mill

Dave – I'm struggling to see what you suggest, is north and south just the front and back of the table? I'm not sure what the set square is for it looks like you are setting the head parallel to the table with a DTI?

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 02/12/2018 11:36:34:

Posted by Michael Gilligan on 02/12/2018 10:37:56:

So the 2 arc second prism [John's 'better ones'] would be plenty good enough

… But the price may be off-putting.

http://www.astroboot.co.uk is selling four porro prisms for less than a fiver, extracted from broken binoculars.

.

If those are accurate to two arc seconds … BUY THEM ALL

MichaelG.

[Men IfrParticipant @menifr84251]

By precision square I mean one of these type of things. I don't know your location but if you were near to me in Kent I could loan you mine.

Arceuro do one that is claimed to be 4 arc seconds, which is plenty good enough for what you need.

Thanks for the offer Pete but I'm in Warwick

I think the frame levellers are just that – I need to measure 90deg and their tolerance (I believe) refers to detecting and angle from 90deg horizontal.

[JasonBModerator @jasonb]

Am I missing something but those prisms look so small to me you would have a job measuring with them over a meaningful length.

I don't think there is any cheap way to do it, a bit like needing a better machine to make a good one you need measuring equipment that is better than the tolerance you want to achieve. Do you have an idea of the sort of numbers that will be acceptable to you?

I made this video a few months back an dnever got round to publishing it. Shows the SX2.7 being trammed with a dti over 150mm diameter, followed by checking quill then head movment over 50mm in Y and then in X.

Tram 0.01mm over 150mm

Y quill 0.04mm over 50mm Y head 0.01mm over 50mm

X Quill 0.01 over 50mm Y head 0.02mm over 50mm.

Bear in mind that the action of the rack pushes the quill and that you are using it unlocked, same with the head you can't measure movement without locking it so in both cases figures may be a bit better. Grade B square 0.01mm in 100mm rating. I should also say that I removed the column from the base to lift it onto the bench so did spend about 10mins resetting that, nothing done to quill, head or its gibs.

Machine has cut perfectly well for the things I have done on it but I'm not going to the moon anytime soon.

[Pete RimmerParticipant @peterimmer30576]

Posted by Men Ifr on 02/12/2018 13:09:50:

By precision square I mean one of these type of things. I don't know your location but if you were near to me in Kent I could loan you mine.

Arceuro do one that is claimed to be 4 arc seconds, which is plenty good enough for what you need.

Thanks for the offer Pete but I'm in Warwick

I think the frame levellers are just that – I need to measure 90deg and their tolerance (I believe) refers to detecting and angle from 90deg horizontal.

Frame squares have all four sides finished perpendicular to each other. They are very versatile for doing alignment work.

[JasonBModerator @jasonb]

But what do you call perpendicular? There is no spec for the accuracy of the sides so at that sort of price I would not expect them to be any better than a £40 Grade B square, maybe less.

The OP has said he does not want to buy squares that don't state the spec so why buy that?

Edited By JasonB on 02/12/2018 13:55:48

[Ron LadenParticipant @ronladen17547]

Posted by JasonB on 02/12/2018 13:46:07:

Am I missing something but those prisms look so small to me you would have a job measuring with them over a meaningful length.

I don't think there is any cheap way to do it, a bit like needing a better machine to make a good one you need measuring equipment that is better than the tolerance you want to achieve. Do you have an idea of the sort of numbers that will be acceptable to you?

I made this video a few months back an dnever got round to publishing it. Shows the SX2.7 being trammed with a dti over 150mm diameter, followed by checking quill then head movment over 50mm in Y and then in X.

Tram 0.01mm over 150mm

Y quill 0.04mm over 50mm Y head 0.01mm over 50mm

X Quill 0.01 over 50mm Y head 0.02mm over 50mm.

Bear in mind that the action of the rack pushes the quill and that you are using it unlocked, same with the head you can't measure movement without locking it so in both cases figures may be a bit better. Grade B square 0.01mm in 100mm rating. I should also say that I removed the column from the base to lift it onto the bench so did spend about 10mins resetting that, nothing done to quill, head or its gibs.

Machine has cut perfectly well for the things I have done on it but I'm not going to the moon anytime soon.

Thats a damn good set of readings for a new machine that has not had any adjustments to the head, got to be happy with that.

[Martin ConnellyParticipant @martinconnelly55370]

That's a small machine with a heavy head. The weight of the head will cause the column to bend by a different amount when the head is high compared to when it is low. Is this going to cause a headache when trying for accuracy? My solution for working with a round column machine will solve the problem for this machine, use a 16mm parallel shank ER collet holder in a 16mm R8 collet. You can slide it down for short tools and either use it all the way up for longer tools or take it out and use R8 collets alone. No head raising or lowering required.

Martin C

[Pete RimmerParticipant @peterimmer30576]

Posted by JasonB on 02/12/2018 13:55:05:

But what do you call perpendicular? There is no spec for the accuracy of the sides so at that sort of price I would not expect them to be any better than a £40 Grade B square, maybe less.

The OP has said he does not want to buy squares that don't state the spec so why buy that?

Edited By JasonB on 02/12/2018 13:55:48

I wasn't referring to the particular one in the ARC link just that frame squares in general are made perpendicular on all sides close enough that they come fitted with an insulator for handling them lest you heat the top edge and push them out of square with heat from your hand.

Now that the class of machine has become apparent I guess it's all immaterial since it's not a very rigid design anyway.

[Neil WyattModerator @neilwyatt]

Posted by Michael Gilligan on 02/12/2018 12:56:53:

If those are accurate to two arc seconds … BUY THEM ALL

MichaelG.

I can't promise that! But in a batch of four I would expect at least one to be good enough to tram a mill.

I am concerned that although the OP's mill clearly needs aligning properly, expectations may be a little high.

An alternative to use with even a wonky square is to look at the readings with the square on the left, then on the right and adjust until the change is the same on both sides when the head is moved up and down. The column should now be vertical.

Now possible to set the head perpendicular by tramming, but I think the problem faced is that the head.can only be adjusted off the mill.

This is why I suggested the 'shim with paper' methodology. This must be more or less how they are set in the factory, however, they have the advantage of doing this before they are painted. When I discovered the head of my mill was out, it was evidenced by a crack in the paint where the two parts join.

Obviously using paper, a flat surface and painted castings isn't ideal, but I can honestly say it worked OK for me.

I described the process for the old-style X2 in this article from MEW 201:

http://www.arceurotrade.co.uk/reviews/MEW201_p52_58_X2_Mill.pdf

Page 5 describes what I did and there's a photo on page 7.

I strongly suggest trying this, followed by conventional tramming (as in Dave's photo) before investing in any additional equipment.

Neil

[SillyOldDufferModerator @sillyoldduffer]

Posted by Men Ifr on 02/12/2018 12:49:05:

…

Dave – I'm struggling to see what you suggest, is north and south just the front and back of the table? I'm not sure what the set square is for it looks like you are setting the head parallel to the table with a DTI?

Yes, front-back is what I meant by north-south. Sorry about the confusion – I wasn't being consistent in my terminology.

The set-square is just to make sure the DTI is positioned close to 180 degrees around from its start position, ie exactly opposite.

See also Neil's post at 14:44 – I'd assumed any front-back shimming could done without taking the head off.

Dave

[Neil WyattModerator @neilwyatt]

Posted by SillyOldDuffer on 02/12/2018 14:59:28:

See also Neil's post at 14:44 – I'd assumed any front-back shimming could done without taking the head off.

Dave

In Part 1 of my article in MEW 199 I described how I did the front-back tram by scraping. Despite my crappy scraping, the much increased contact area helped rigidity, but not as much as a bracing plate.

Problem is the fixing screws are nearly in a straight line:

Adding a bracing plate locks the column solid, front to back:

Allowing you to do stupid things which the original mill wouldn't have had a chance of doing:

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Pete Rimmer on 01/12/2018 19:20:03:

By precision square I mean one of these type of things. …

Arceuro do one that is claimed to be 4 arc seconds, which is plenty good enough for what you need.

.

That's a very impressive specification for the price, Pete

… well spotted

MichaelG.

[Author]

Posts