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

I would like to check the column on my milling machine for left/right and forward/back alignment.

The usual tramming the head in via dial indicator is no good, or at least I have worked out I have a head parallel to the table but the column is not straight i.e. in the past I have corrected the column tilt to get the head parallel but all that does is mean I correct the not 90degree head to column by making a corresponding not 90deg column to base. So when I move up/down I have side/side or front/back movement as well - not good!!

So to correctly set-up I need to 1st have a plate or similar that is a reference 90degrees then move the head up and down and adjust the column angle to track that. Then I can adjust the angle of the head knowing both would then be 90degrees.

So.. how do I measure 90degrees - currently I tried a ground 90deg angle plate but that turns out to be far from flat so I have no faith in the angle accuracy either.

The only thing I've found to quote any accuracy is this starret gauge at £90 but perhaps is the best option...

https://www.starrett.co.uk/shop/precision/shop_tools/precision_steel_square/?item_id=BM1408

Any other options/thoughts?

Stick a 100 to150mm length of about 30-40mm bar in your lathe. Drill a hole through the middle and take a skim along the outer surface. Now face across the end and then turn a recess for all but about mm at the edges.

Aside from the bit held in the chuck which you can part or saw off, or ignore, this is now an accurate 'cylindrical square'.

Gently bolted down to the mill bed you can use the dial indicator against it to do the tests you need.

Keep it safe an you can use it for setting up parts, making jigs etc. Obviously there are various ways to make the upper end a bit neater or you could turn it between centres to make the whole length accurate.

Neil

The problem if you square up the column is that you may introduce an error in spindle movements. How do you know the machining for the column and the machining for the spindle are perfectly parallel to each other?

Martin C

If the spindle has a quill you can advance and retract, grip a bar in it with a dial gauge attached - via a linking piece if necessary - that contacts a machined surface of the column. If it's parallel, moving the quill in and out will show no deflection.

Neil, I like the idea and was going to say... but wouldn't the accuracy depend on the accuracy of the lathe?

But now i'm thinking about it.. if the cross slide is not 90deg you would still get a perfect 90deg since the part is spinning.. I can't see any downfalls... well of course I don't have a lathe yet(see my other post) and I think some large stock to provide a stable base so maybe 50mm.

Only other issue - not with the accuracy of the part but general use, my mill column will be out in both directions so tramming one direction it will move off the centre of the bar a bit and add error, but if something flat was used that would have to be trammed into the 'other' e.g Y direction else that would add error as well...

I will square the column first and then the head in 2 stages so both will be square. Square the column using something 90deg then the head my making it parallel to the table...

I think it would be a good idea to check that both ends of your cylindrical square are the same diameter, in case your lathe is turning a slight taper. Or am I misunderstanding the whole thing?

Rob

No you are right, I'm making the assumption the lathe is set up OK.

Neil

Try two cylindrical squares bolted to the table and then a suitable flat surface clamped to them (surface plate, plate glass, granite tile.....). That'll give you a surface plane to act as a datum to check if the column leans towards or away from you. Do the same after moving the cylindrical squares 90 degrees and check for leaning left or right. Once you're happy the column is square to the table, then start checking the head end.

It might help to know what milling machine it is.

So what mill did you end up getting?

If the quill is not parallel to the column it will be hard to tram both the quill and the column but what can be done will depend on the machine as a swivel head may have a bit more scope for adjustment than a fixed head.

You may also get different results as when milling you tend to have the head and quill locked, but when using the quill or raising/lowering the head to do a cut you won't have the slides locked so the natural tendency is for the head to nod slightly towards you. Could be chasing small amounts that are not worth bothering about and may be unobtainable.

Edited By JasonB on 01/12/2018 16:23:30

Men Ifr,

No machine is any more accurate than the machine which made it,there is no such thing as Apsolute accuracy ,even at top level. How far back do you want to go? back to the foundry ?

Unless you are making stuff for the Space Program [like I am,although they wont accept my offerings ],I dont think I,d worry too much about it. Neil gave good advice,but as I said,it depends on how accurate the thing was made,to begin with.

No machine is any more accurate than the machine which made it,there is no such thing as Apsolute accuracy ,even at top level.

That is why machine tool builders use skilled fitters - they recify any machining errors to produce machines equal to, or better than, the various machines that contributed the parts.

Nothing wrong with checking out the accuracy of a new (to you) machine & improving on it using time tested techniques.

Nigel B

Buy or borrow a decent precision square. The top and sides are perfectly square to each other and it will have a sensitive bubble with graduations. You can set the square on the table, against the column ways, against the mount for the head etc and compare measurements. The graduations will tell you how far out any surface is.

I have several 'precision' squares - only the usual stuff that is relatively cheap and does not come with a tolerance - I do not believe they are 90deg - one possibly 0.1mm out over 100mm which seems quite poor (for what I want) any-one else I know will not have a genuine precision square made to a tolerance.

Only the starret seems to have a good tolerance but is £££ but maybe worth it it it would get a lot of use (who knows...)

A borrowed diesel piston pin with one end on mill table and my trusted MC3 testbar in spindle of milling machine and a light-source convinced me that my mill was OK.

Only problem is that test bar has no pull in thread so I was always prepared for drop down.

It did not.

I currently have a SPG 9512 and have just bought the SP 2217-30B LV. I've had problems on the 9512 that I think are due to the head angle being out and me then setting the column to get the head parallel to the table. So both need setting up but I will focus more on the new machine once it arrives as it wil stay like that for many years to come so would like to get it right now.

I've been lapping and scraping the 9512 to practice for the bigger machine as well which has gone well but is time consuming.

The quill idea is worth bearing in mind for the 30B but the little 9512 does not have a quill.

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.

Edited By Pete Rimmer on 01/12/2018 19:22:25

The trouble with Neil’s square is that if the column is out in both directions, and you put a dti at say the back is any indicated dial movement a result of a tilt in the y direction or is it caused by the indicator rolling round the pillar(edit) in the x direction.

But if you go for the double square as indicated by Chris then you will eliminate this problem.

However, if the machine trams correctly then it will produce accurate work unless you are without a quill for adding cut, however the change to the geometry of any cut would probably be less than the deflection of the tool under load..

What type of machine is it? if it trams how does the problem affect your workpiece?

You need to give more information in a request such as this

Borrowing expensive precision kit as suggested earlier is not that easy for most mortals.

Edited By Zan on 01/12/2018 19:28:25

If you are intending to work to these high specs then you may as well go out and get yourself a halfdecent square. I have a 6" Moore and Wright 4006 which is made to BS 939 Grade B which is good for 0.01mm in 100mm. Cost you about £40. keep it in a draw and just use for checking tooling and setups, keep your cheapies for running a scriber against.

If you want better then get Grade A or AA but they will cost you and are really for lab use so Ok if you have a temperature controlled workshop.