Cutting a slot in a screwhead

[Clive HartlandParticipant @clivehartland94829]

I am slotting some bronze screwheads for a friend. He modifies theodolites for non magnetic standards. All ferrous screws changed for bronze. He makes the screws and they need slotting.

I have a 2mm saw mounted on the mill vertical drive and screws held in the vice in a collet in a square johnson block.

What is best method of ensuring the slot is central,in the screw head.

I have a DRO fitted to the mill and readout system.

Hard to get a measuring tool into the setup to check. Any ideas please.

[Clive HartlandParticipant @clivehartland94829]

8mm screw, centering the circular saw to the head.

[JasonBModerator @jasonb]

I tend to turn the saw back and forth by hand while lowering it and touch it onto the top edge of the screw head, you can do it by feel or with a fag paper. Zero the DRO at this point.

Now feed down by half the saw thickness plus half the screw head dia and you should be on ctr line

[DC31kParticipant @dc31k]

Is there any advantage or ease to be gained by touching off on the top surface of your Johnson? Maybe a bright light could substitute if your fags are now paperless.

[Clive Brown 1Participant @clivebrown1]

I'd try putting into the collet a throwaway piece of material same diameter the screw thread.. Cut a slot and then re-adjust the position by trial and error until it looks right. Then cut the actual screws; but I haven't got a DRO!

[Ian PParticipant @ianp]

I find its a very trial and error process.

I too touch off in a similar manner to Jason but I use the upper and lower points (north and south) on the screw or workpiece held in a Stevenson block and then halve the DRO readings to find the centreline. Apart from the fact that it difficult to see or feel when the sides of the slitting saw teeth touch the workpiece, I find most of my slitting saws have some wobble so results are not always as I expect.

I sometimes true up arbor faces and spigot diameters in the lathe and without removing the arbor from the lathe, fit the blade and still get runout at the saw teeth. I could probably reduce the blade wobble by having a large diameter arbor that only leaves enough of the saw teeth exposed for the depth I intend to cut.

I find that with, say a 1mm wide x 75mm dia sawblade with a 38 diameter arbour if there is any wobble the slot position partly depends on which part of the blade circumference make first contact and makes a channel for the rest of the blade teeth to follow.

Ian P

[EmgeeParticipant @emgee]

On the last gudgeon pin remover tool I made, using an arbor mounted 2.75mm diam saw in a vertical spindle, I bought the slowly rotating 0.50mm thick blade down to touch off the top of the 4mm diameter part held in a horizontally fitted indexing/dividing head, set the Z axis to zero, moved off in the X direction until clear then dropped the blade to below the part, raised the blade to touch off the part and read the Z readout, moved off the part again in the X axis and raised the head 1/2 the recorded distance. Set the spindle revs to 300 and using the X axis moved the blade into the work by the 10mm required with a feedrate of 30mm/min.

Using this method evens out any runout on the blade and gives a spot on centre cut.

Emgee

Ian P beat me to it.

Edited By Emgee on 29/07/2022 20:54:12

Edited By Emgee on 29/07/2022 21:04:51

[John ATTLEEParticipant @johnattlee20632]

I would normally do exactly as Emgee has done with DRO. I find DRO on the Z axis very useful. However, I have done it by taking the blade out, touching off the appropriate surface of the arbor. I then raise the table by the appropriate amount (half the blade dia + half the head or a useful bit of round HSS bar 3/8 dia). Seems to work fine even with me!

John

[old martParticipant @oldmart]

I would use the Clive Brown method.

[Clive HartlandParticipant @clivehartland94829]

Thanks Gentlemen, the sacrificial pin of thread diameter seems the way to go. I have over a 100 to do!

As he is a long time good friend I cannot ask him to pay. Clive

[Clive FosterParticipant @clivefoster55965]

For this sort of job I prefer known error methods where a suitable test cut is made deliberately offset from the desired position and the actual offset derived by measurement and calculation allowing things to be adjusted to place the cut correctly on the actual workpiece.

Especially so if a DRO system is fitted as it makes compensating for the deliberate error in the test cut much easier than when using conventional methods.

For this job I'd make a dummy cheesehead "screw" and slice part of the head off leaving a D shape.

The position of the flat side of the D can be measured to appropriate accuracy hence establishing the position of the lower side of the screw slot produced at that vertical position of the slitting saw.

Given the diameter of the dummy "screw" head and measuring the depth of the D the offset needed to put the bottom of the slot to be cut in the proper screws exactly half the width of the screw slot below the collet centre line can be calculated.

I tend to break out the CAD system and accurately draw the proper screw head along with the D shape test dummy.

Clive

[John ReeseParticipant @johnreese12848]

JasonB stated the ideal method. It also works for setting a Woodruff cutter or gear cutter on center.

[JasonBModerator @jasonb]

It works for me as would the touching top & bottom then using the half function. DC31K's method of setting to the top of the collet block would also be easy and then just lower by half block height + half saw thickness.

[AdrianRParticipant @adrianr18614]

I use a DTI, zero it on the part then adjust the cutter height using the DTI. Add 1/2 the thickness of the cutter and part you are at the centre.

If you revolve the saw you can compensate for the wobble too.

[Martin KyteParticipant @martinkyte99762]

Or leave a saw width witness on the head of the screw and adjust the saw by eye using a loupe. Personally I cut screw slots with a knife file positioning the file with my thumbnail. as the cut progresses a small sideways force is sufficient to correct any positional error. This method also produces a slot with tapered sides so the driver fits correctly.

regards Martin

[Clive FosterParticipant @clivefoster55965]

Touch off methods fundamentally assume that there are no infelicities in workpiece mounting, tool grind or cutting behaviour that may affect the relationship between touch position and the edges of the actual cut. It is of course a given that the tool and work be sufficiently stiff not to deflect when the (gentle) touch is made. Like many folk I find the "just tears cigarette paper "or "drag a feeler gauge out of the gap" methods give a better feel for precise touch off than simply bringing the tool down onto workpiece or reference for those jobs that have to be so right that even Inspector Meticulous may smile, briefly.

Nice thing about the known error method is that all infelicities are eliminated because measurements and settings are referenced to an actual cut.

Its also generally quicker.

Eye up cutter for a suitable sweep across a sacrificial component or set up dummy to make the D.

If you have one set DRO to zero, use a different memory if need be so you don't loose important datums.

Measure depth of D and height of flat above table. (Verify that the results are consistent. I'm a great fan of paired check measurements as the habit has saved me significant time when a set-up wasn't what I thought it ought to have been.)

Move cutter so as to bring it to the calculated correct position.

Make a second check cut on another sacrificial or dummy component. If its wrong figure out what your error was and start over, otherwise go into production.

There may be some parallel universe where slitting saws are universally renowned for their stiffness, flatness, accuracy, good behaviour in cut and general tolerance of all but the most grievous mishandling. Where ever it is it's far enough away that the light hasn't got to my workshop yet!

Touching off an affordable slitting saw in an affordable holder is a prime example of the sort of job where I want an easier way to be accurate. My slitting saws and holder came from ArcEurotrade a decade or more ago so they are decent and were good value for money but lack the ultimate accuracy of high end industrial products.

Clive

[HopperParticipant @hopper]

When touching off the blade against the top surface of the job, it can help to paint the top surface with a felt pen to make the initial touching off mark more visible.

[larry phelan 1Participant @larryphelan1]

I did a similar job some time ago, using a chuck mounted on my rotary table.

Set the saw height by eye, made a cut across the blank, then turned it 180 deg and made a second cut.

Any error was soon revealed and corrected.

Could use a collet, might be kinder to the screws.

No DRO to hand, dont have one, dont want one.

[SillyOldDufferModerator @sillyoldduffer]

Slotting screw heads one at time would soon get tedious if a large number had to be done. Has anyone tried 3D printing a jig to do this sort of work?

Something like a pair of these sandwiching multiple screws so that all their heads are slotted in one pass of the slitting saw:

My CAD example shows slots 6 on the assumption that the two halves could be held together somehow in an ordinary vice, but for serious work the jig halves would be just shorter than the mill's travel, and held in a purpose built holder bolted to the table.

Making the two halves of this jig in metal would be time consuming rather than difficult, but 3D printing makes producing them easy, even if the holes were tapered and ridged to grip the screws better, and the two halves were jointed together.

Dave

[Clive FosterParticipant @clivefoster55965]

Given the incredible sensitivity of the Mk1 human eyeball to tiny errors of centration and concentricity I suspet Daves device would prove disappointing in use. The flexibility of 3D printed plastic may inherently limit the accuracy of holding to a level below what would be desired and setting up so things are truely parallel to the amchine table travel may prove fraught.

Machining in situ from alloy bar would seem a more reliable method. Production CNC jockeys do such as a routine task for multiple part set ups.

Clive

[noel shelleyParticipant @noelshelley55608]

Using the vertical head drill a series of tapping size holes (10) and tap in say a piece of 1/4"thick material with top face set true. Change drill for milling cutter of the slot size. Screw in blank screws and mill the slots. Lock all ways not needing to move. This asumes that all screws are identicle ! Noel.

[JasonBModerator @jasonb]

The biggest problem with Dave's jig is that it will hold the heads horizontal if in a vice so you have to use very small dia milling cutters with a slow feed so unlikely to be any faster than a saw. Most of the screws I make have slots of 0.8mm width or less. Unless you have a horizontal mill with rack feed then it would be faster

I usually do them in the spin indexer which is quicker to change screws than having to use spanners on a collet block

[Glyn DaviesParticipant @glyndavies49417]

I mounted a cheap 3 jaw chuck to an angle plate and leave it permanently set up on my D-W miller. No DRO, so I turn a piece of scrap 1/4" rod to form a pip whose diameter equals the thickness of the slitting saw. Then align by eye. If using small countersunk Allen screws, I slot them to allow them to be removed if the hex socket becomes rounded.

[Clive FosterParticipant @clivefoster55965]

+1 for Jasons suggestion of the Spindexer as being faster to change workpieces than a collet block.

I'm surprised how little I use my collet blocks, primarily for that very reason. A box / tube spanner long enough to easily loosen the collet nut when held in my usual vices went on to the to-do list about 5 minutes into my first use of a collet block. Its been on the list about 20 years so far!

As has a base conversion thingy so my Spindexer bolts straight to the Bridgeport table rather than needing clamps. Only 19 years tho'. However I did square off the two long sides in alignment with the spindle so I can set it true (enough for all normal purposes) by setting a parallel in one of the table slots and pushing the Spindexer base up against it.

Clive

[Author]

Posts