I’m getting to grips with my recently acquired Astra horizontal milling machine and I’m wondering if the cutter runout is acceptable. With a brand-new 4″ cutter (UK made) I measure just under 0.001″ runout at the cutter tips. The spindle runout is less than 0.0005″ in the centre, no detectable runout showing at the extreme ends without cutter or spacers mounted. The spacers are clean and are the original ones that came with the machine.
In use the cut doesn’t sound continuous. I’ve experimented with depths of cut, feed rates and speeds and although the cut progresses nicely and produces a clean finish I’m thinking that only a few teeth are contributing to the cut and may result in premature wear.
The idea of high feed/depth to overcome cutter runout doesn’t work with the Astra due to the rapid feed cross-shaft acting as a torsion bar when locked.
But I do worry about the preoccupation with precision on here, when there has been no observed fault or error measured in the parts made. In industry if the part was right that was what mattered. As an apprentice My lathe was a worn Harrison 10 on which I had to work to .00025″ for a job for RR, 400 of them. On the basis of some opinions on here the machine was only fit for scrap ? When the factory closed I would have been proud to have had it. Many a fine tune played on an old fiddle ! Noel.
Even as one of those fascinated by accuracy and precision for their own sakes … I must of course agree that in the ‘real world’ of manufacturing and Model Engineering [however one might choose to define that]: Near enough is good enough !
Things are a little different in microscopy [my own special area of interest], where microns are easily and necessarily measurable.
I agree that 0.001 runout is pretty good. One thing that might improve matters is to blue up the surface plate and check all the collar end faces for flatness. Blue will show up dings that visual inspection might miss.
Also keep the cutter toward the drive end of the mandrel if you can. Mid span looks balanced, but is the worst place for a cutter.
I also agree with Noel that the armchair engineer ratio on here is on the high side.
My main concern is undistributed wear on the cutter teeth, though at some point any cutter will need to be reground. I’ve always treated small end mills as semi-disposable but need to rethink my approach to horizontal cutters. At least most of the straight cutters are straightforward to sharpen using a simple index fixture operating on the teeth. I’ve heard of people grinding in-situ on the machine itself for the best runout.
I’ll take a look at the spacer ends.
Noel makes an interesting point about what matters is the finished work and the preoccupation with precision. Only recently have I started to look online and view forum posts regarding model engineering and wonder how anyone ever made anything in the past using a we’ll-worn lathe and rattling pillar drill. My first “big” lathe (other than a Unimat) was a round bed Drummond that had so much backlash and spindle play that it would now only be viewed as scrap. Yet I turned out hot air engines, steam engines and clocks for many years that all ran nicely and I never worried about the accuracy of the lathe.
Nowadays I’d be so full of dissatisfaction and self-doubt instilled by the Web that I’d never have got started in model engineering.
Carefully mark the lowest tip, loosen, wiggle it a bit, retighten and see if the lowest point is still the same. Check the key is not lifting it in that place. If it is a narrow cutter you may not need the key and can reposition the cutter but beware it stalling and spinning on the arbor.
I’ve tried grinding them on the machine but it didn’t eliminate runout, possibly because with the spindle not rotating the oil film got squeezed out (plain taper bearings). Since I didn’t have a tool and cutter grinder it had to do, and it did well enough.
I would look into why you can’t use a heavier feed and DOC; if the cutter runout is only 0.001 and you can hear an intermittent cut it suggests a very light feed.
Is the rapid feed one of those lever-operated arrangements?
I do a lot of horizontal milling on my Elliott model 00. When I got it, the next machine I acquired rapidly was a tool and cutter grinder as new horizontal cutters are expensive but old used ones are cheap and relatively easy to sharpen.
Cutters rarely run true enough to cut evenly on all teeth. There are many factors that make this difficult to achieve. A few things that spring to mind are: Is the arbor straight? How much clearance is there on the bore of the cutter? (shouldnt be a press fit, clearance is instant runout). How accurately was it ground? (The condition of the arbor used for grinding, its diameter and condition of the centres – even if a tapered arbor is used, how straight was the cutter assembled on it?) How accurately was the cutter indexed for sharpening (using an indexing finger which can spring out of the way to allow successive teeth to be ground means theres positional uncertainty) Then theres the grinding wheel. It will wear as you grind each successive tooth, however fine the final grind is (thinking of ally oxide wheels here).
It only takes a tenth or two on each of the above error sources and 1 thou is looking pretty good!
If you take deeper cuts with a higher feed rate you’re more likely to get more teeth cutting versus skimming off a thou or two.
Then theres the grinding wheel. It will wear as you grind each successive tooth, however fine the final grind is (thinking of ally oxide wheels here).
I wonder whether you could grind a cutter like you tighten the wheelnuts on a car – crossing diagonally from tooth-to-tooth rather than going round in a circle.
With a conventional grind, the first tooth is off a freshly-dressed wheel and the last tooth, at maximum wheel wear is exactly adjacent it, amplifying the difference.
The mill has a rapid feed lever and this is perhaps the weakest part of an otherwise sturdy design. There’s a bronze gear that engages with the feed screw and in rapid feed mode this acts like a rack and pinion arrangement. With the rapid feed locked (necessary to be able to use the hand wheel) the gear shaft is locked and the feed screw operates against the teeth of the now fixed gear. The problem with this is the gear is on the end of a comparatively long/thin shaft that springs with each cut if a heavy cut is taken.
I think a worthwhile modification would be to install a fixed feed nut which could be exchanged with the rapid feed unit as needed.
The original machines ran at a much higher speed than would be considered normal for a horizontal machine and to some extent this would probably disguise any runout. This particular machine has an improved range of speeds with back gearing.
I’d be interested to hear the experiences other Astra mill owners have of these machines.
I’m not sure what the manufacturing tolerances are on a cutter. All mine are Sheffield made by well-respected manufactures and I’ve initially been experimenting using new cutters.
When it comes to sharpening, how much would an aluminum oxide wheel wear between the first and last tooth, assuming just a light touch only on the tips of a lightly dulled cutter? Would a boron wheel (CBN) be any improvement?
There must surely be a clearance tolerance between the arbor and cutter or changing the cutter would be difficult.
Same goes for a vertical miller, there’s a runout on the cutter.
You clearly already know this, perhaps you need to consider than the web is also full of liars who claim they can do all sorts of things that they actually can’t just to wind you up.
You could use a diamond coated wheel (not the rubber ones ) wear is tiny.
The reason diamonds are not used to cut iron is that iron has an affinity for carbon. That is iron readily soakes up carbon which is why steel is often alloyed with carbon in varying ammounts. The diamond gets hot during the cutting process and dissolves chemically in the iron that it in contact with. Occasional use of a diamond wheel to finish sharpen a steel cutting tool will not do much damage, but if used on an industrial scale it would be very expensive. CBN type cutters are nearly as hard as diamond but are not affected and have a much longer life. Diamond is very good for use with tungsten carbide. For lapping with fine diamond grits, the use of a liquid is normal and there will be very little temperature rise so the diamond particals last well.
As mentioned, there can be too much worrying about small runouts by home machinists. When I was involved in Aircraft parts, it was normal to have a tolerance of +- 0.0005″ or +0.001″ -0, very few drawings were tighter.
….perhaps you need to consider than the web is also full of liars who claim they can do all sorts of things that they actually can’t just to wind you up.
Just wonder if the “rapid feed” is more for a “rapid return” particularly if it needs to be engaged to use the handwheels
A reduction in cutter diameter (reducing No of teeth) would result in a slower feed rate for the same given tooth load which may allow you to turn the handle at a managable speed.
Just looking at the L2/L4 on lathes.co.uk I would say that the lever is indeed a fast feed . However given the spindle speeds this is more of a second ops machine for use on non-ferrous metals such as cutting a slot in a previously machined component. Where you would set it up using the handwheels and then put the part into a fixture and whack the cutter through with the lever feed, rinse and repeat as many times as you can.
If I read it right the slowest spindle speed was 620rpm which is way too fast for steel. Take a 2″ dia HSS cutter and that wants to be run at about 200rpm in steel so will be considerably overspeeded though that will soon wear any high teeth down to size! Not sure what you are intending to cut with your new 4″ cutter but in steel that wants 100rpm and even 620 is a bit fast for non-ferrous
What are the actual speeds on your “improoved ” machine and are there any other automatic feed rates?
The only practical use I see for the rapid feed is in setting up long work off the DTI where it saves a lot of handwheel twiddling. Maybe it will be useful for other jobs but the cutter force is against the operators hand. It may be interesting to see how it works with some of the large slitting saws – it may work out well at low rpm as the feed could follow any eccentricity.
Low speed range gives from 35 to 100 rpm geared.
I’m no stranger to cutter runout, having started my working life as a general machinist in a works that had lots of well-worn milling machines with wonky arbours. The large roughing cutters on the geared machines had several tens of thou runout but the depth of cut and feeds were off the scale compared to what this machine will do. The motors in them were unstoppable – the key would shear or the work slip in the vice, but the motor wouldn’t flinch.
The only practical use I see for the rapid feed is in setting up long work off the DTI where it saves a lot of handwheel twiddling. …
Indeed, but the existence of rapid feed in conjunction with Jason spotting the machine has a high minimum spindle speed makes me wonder what type of work this horizontal mill is designed for? In my tiny mind, Vertical Mills are good for general purpose whilst Horizontal mills are better for repetition work. Verticals trade production speed for versatility, whilst horizontals are less versatile but quick at what they do well. Both can cut a groove along a beam, but if many beams have to be cut a horizontal mill will do the job faster. They can also do gang-milling, cutting more than one groove in one operation.
The combination of high spindle speed and fast feed lever suggest a mill optimised for churning several copies of the same non-ferrous parts out quickly. Maybe the discontinuous cutting reported by Mick is due to over-speeding rather than run-out; is the RPM too high for steel? If so, spinning too fast will wear the cutter out much faster than it having 0.001″ run-out!
Had a quick look for manufacturing tolerances on cutters and found this:
The numbers suggest milling cutters aren’t precisely made, and could be much worse than Mick’s 0.001″. I guess because the accuracy of the cut is determined by measuring on the job, not by assuming the cutter and machine are both in perfect condition. They never are!
Interestingly, the off-the-shelf tolerances met by most CNC shops are quite poor – about ±0.005″. CNC can do considerably better, but it costs more. Same is true of traditional methods – accuracy and precision are expensive. Therefore professional designers put considerable effort into avoiding the need for high accuracy, and maybe Model Engineers should do the same. Parts should be no better than they need to be!