Thinking about where I need to improve - measurements

I’ve been thinking about the engine I’m making. It’s very much of the nominal dimensions make the things fit inside the other things.
That’s ok for a first attempt. But I’d like to be better than that. I’ve recently got a wiggler for the mill and used it to drill a number of holes in the wrong place.
I think the reality here is that I’m not allowing well for the end float in the table. I could strip it and remove some I suspect. (I’ve done this on the mini lathe with some brass washers)

or as I read endfloat / backlash is not a problem I could learn to work round it. I’m not thinking about adding in DRO’s to measure actual moves as opposed to ‘screw’ moves.

I strongly suspect that the old hands on here could use my machines to a much finer tolerance.
Can anyone pass me some guidance?

I have a vernier and a micrometer and I understand how to use them. (I think!)

Any wisdom welcomely recieved!

My take on the backlash issue is always approach the positions from the same direction.

If you use the wiggler to find the LH edge of the part, you will be moving the table from right to left. When you move into the part from that position (you did remember to allow for half the diameter of the wiggler) you will still be moving the table right to left and dimensions should be ok.

However if you overshoot, and need/want to move back towards the LH edge (i.e. move the table left to right), then simply go past the point you want and come back to it so that you approach the position with the table moving right to left.

Same principle applies front to back.

I used to do my milling on a machine without a DRO. I learned to cope with backlash, but, with a job of any complexity, trying to convert all the desired measurements to hand wheel rotations and vice versa and holding all of these permutations in my head or scribbled on bits of paper that were never where you needed them tended to be an error-inducing, time- and material-wasting drag.

If you spend money on anything to upgrade your mill, a DRO of some sort at least beats everything else hands down.

Some mill tables have a facility to add bed stops at chosen intervals so once set you will always be moving the desired amount by just moving gently between stops.

The Victoria Universal mill I had used a tee slot on the front edge of the table with adjustable stops fitted as required.

Emgee

What Peter says always work in the same direction which is usually from the l/h edge and far side so rear left corner is your 0,0 datum.

Awesome, I suspected some one would have some immediately actionable changes I could make! I’ll work on setting out from one corner to minimise backlash

Hi Nick,

If you don't have them already, fit locks to your slides and always lock the axis you are not moving - at the very least it will help prevent you accidently knocking a handle and moving something you didn't mean to.

If you have adjustable dials at both ends of your X-axis (and both my mills have) then Joe Pie has the solution you need for backlash - just watch this video (it makes more sense when he actually demos it btw)

Eliminate Backlash in the X-axis

Note that Joe also talks about 'climb' milling. If you have a a lightweight mill (and Joe doesn't) this is probably not something you should try...but the two-dial tip works well.

Regards,

IanT

Some very useful info here, thank you.

I use Peter's method even when the backlash is only 0.002". Always back off a greater distance than your machines backlash and it helps to write down the start and finish figures before each set of moves. Setting stops if they are available also helps. I might machine a test piece rather than risk the actual job if complications look like arising.

When I did that, as JB posted, I always used front left as my datum Remember that when moving from one point to another there is always 2 errors - the start and finish - and these could add to any previous errors if you measure from different starting points.

Moving only from that one datum avoids extra errors creeping in, and often means just progressing from left to right etc.

Before I mastered the art of avoiding backlash, I often drilled a hole and popped a drill in the drilling before taking out the backlash and setting my dials to zero. That worked for the accuracy I deemed good enough. DROs and edge finders have made things so much easier/simpler.

When using a wiggler to set a datum face such as an angle-plate for a much lower work-piece, often I clamp a small plate or parallel to the face - just use a little G-clamp - and "wiggle" to the projecting mating face of that, as a proxy for the work-piece face.

This takes only a few seconds extra, but puts the table movement to the machined features, in the same direction as that needed to find the edge. (I do back off and approach again to verify the edge.)

Cheap DROs would speed up your work and be as accurate if not more accurate, they are easily good enough to about 0.05 of a mm or even less

About 25-30 quid an axis

+1 DRO transforms using a mill because dials allow human error to creep in. Forgetting to compensate for backlash, losing count when a dial has to be rotated several times to traverse a long distance, miscalculating the number of turns needed, and the target reading, especially when subtracting, plus confusion if interrupted!

• DROs don't change until any backlash has been taken up, eliminating backlash as a source of error.
• DRO measures over long distances, knows where it is at all times, and doesn't require any mental arithmetic. Operators don't have to remember dial turns so far, and are distraction proof.

Not being a natural machinist, I often plan sequences of operations to minimise tool changes and reduce the chance of errors accumulating and reduce the chance of mistakes by following a written plan.

I often mark the job or use existing features to provide check and new references, for example I might bore a hole referenced from the front-right corner, and then use the hole as reference for features related to the hole. On a complex job, it's worth resetting the mill from established references periodically to stop error accumulating. The extent of pre-planning needed varies with talent and experience; I do much less now I'm familiar with my machine, because I've learned to apply the various techniques described above by others semi-automatically.

Don't attempt to fix backlash by attempting to eliminate it. On most machines it's impossible. Worse, tight adjustment causes excessive wear, and backlash soon reappears - the fix is untrustworthy. Worth reducing backlash occasionally to save time, but better to get into the habit of compensating for it or fit a DRO!

Dave

Don't know about the rest of you lot but I'm definitely not the man I was and anything that reduces the brainpower required means I can work longer and better before the senility sets in

Edited By Ady1 on 21/06/2021 10:16:58

Ady, that rubber glove on your head must be the cause.

For ‘old school’ methods … This is a gem : **LINK**

http://www.opensourcemachinetools.org/archive-manuals/precision_locating_and_dividing_MRS_No135_1914.pdf

MichaelG.

It is. To be a good machinist in those days you clearly had to try a bit harder, and have very good close vision.

When using a wiggler, and then needing to reverse direction, to a dimension, backlash must be eliminated.

Example.

Moving table forwards to bring fixed jaw of vice into contact with wiggler.

Lockm table.

Rotate handwheel in opposite direction until backlash is out, and handwheel stops rotating. (Beware of over tightening! )

Set dial to half diameter of wiggler probe, unlock,l and move the table...(When the dial reads Zero, the centre of the spindle coincides with the face of the fixed jaw )

If using DRO, it may be necessary to set Zero, and then, having released the lock, move the table by half wiggler diameter, lock and reset Zero..

You can then move the table by the distance required. A drill point will be the distance that you offset from the Zero. Don't forget to take account (halving ) of the diameter of the cutter if milling to produce a shoulder.a set distance from the fixed jaw.

Howard

That's why I use a proxy face, Howard, if I can't pick up the work-piece's own edge:

Apart from 2 or 3 iterations to verify the reading, it means moving the table in the same direction to the "wiggle point " then to the wiggler centre, then to the first machining point..

No reversals and trying to determine the rather wooly end of the backlash.

I do that whether using the dials or DRO.

If one is looking for micron accuracy for point to point moves using backlash, do remember that the feed screw must be in perfect order to achieve perfection. The backlash can vary quite considerably - from either end to the middle of a worn feed screw.