CHAT ABOUT CHATTER by the late Rev Brian Favell

In 1941 I was 18, and a member of the Rugby Model Railway Club. It had the care of a lathe belonging to someone in the forces who never got back to claim it for years. This lathe was a Drummond 4" round bed machine on a cast iron stand with a socking great flywheel about two feet diameter, worked by a treadle. Collectors of ancient Meccano Magazines may remember regular ads in it for Drummond lathes. These always had a picture of either a 3.1/2" flat bed machine complete with back gear and screwcutting gears, or a 4" round bed lathe with no back gear. That was the one. If my memory serves me it cost £10: sounds cheap but in the thirties that was three weeks wages for an ordinary working man; five weeks wages for a miner.

For the price, and bearing in mind the limitations of small lathes in those days, it was a good machine (the 3.1/2" one cost nearer £30). By the way, I did the machining for a 2.1/2" Austere Ada on it. It had two weaknesses apart from the lack of backgear (which didn't matter on the treadle drive because one could vary the speed easily).
The first was the standard weakness of small lathes in those days: the mandrel nose was horribly small and therefore not all that strong.
The other one was to do with the saddle and cross-slide. The saddle could be rotated around the bed which allowed one to adjust the tool height exactly. At the front side of the saddle and vertical to the cross-slide was a hole little if any more that 1" diameter, into which fitted a cast iron peg in one piece with the cross-slide base. This made it possible to set the cross-slide at any angle to the lathe centre-line and, as I said, to easily adjust the tool height. The bad news was that the slide was sort of overhung; not rigidly supported.
The trouble was that it wasn't at all rigid. The cross-slide could flex downward under load, and bearing in mind the height of the tool post, that was fatal. The trouble was that when the tool was under load it would flex forward under load, causing it to dig in. Then the tool breaks free and repeats the process ad infinitum.
This is the mechanics of 'chatter'.

A year or two later, in the course of my training (I was a Drawing Office Apprentice) I had best part of a year on centre lathes: first an old Dean, Smith and Grace with a three foot diameter faceplate, then with a newly overhauled 6" Smart and Brown with a big DC motor with a small version of a tram controller. I learnt a lot - by trial and error - about sharp tools and correct cutting angles, getting good finishes and avoiding chatter. So now here are some of the results.

Although modern small lathes are immeasurably tougher and more rigid, tools still manage to dig in and cause chatter. Nowadays cross slides are more rigid, being fully supported right across the saddle. Nevertheless tool posts of whatever type still tend to flex forward into the work when under load with the same result - chatter!

Industry has produced the right answer with capstan lathes and automatic machines. These universally have on the cross-slide two tool posts: the front one as normal, and the back one arranged to hold the tool upside down. If you consider this, with normal lathe rotation (top towards you i.e. anti-clockwise looking on the faceplate, you will see that on the back toolpost the loading is upwards, causing the post to flex away from the work. Result - no chatter, even with a parting tool which is the worst possible case.

As a matter of interest, the Swiss go the whole hog with their Bechler automatic lathes, which run backwards by our normal standards. They don't have a back tool post: the front tool post holds the tool upside down which again flexes away from the work. They also have an array of tool posts ranged around the chuck, all of which when advanced into the work flex away from it. This matter of running backwards has another advantage which they make full use of.

These machine have a sort of tailstock which have a live (i.e. driven) shaft mounting an ordinary circular die or tap. The shaft is driven faster the the main mandrel, hence when it is advanced into the work a right-hand threaded die screws itself onto the work. When the thread is the right length the forward pull trips a little catch which stops the tailstock shaft, rapidly unscrewing the die off the work (the principle works just as easily with a tap). All this has the further advantage that the mandrel does not change speed: therefore screwing can be done at the same time that another tool is turning another diameter or face. Hence not only do they not need a Coventry type self-opening diehead, the floor-to-floor time is much faster. I have seen brass bits for plugs and sockets turned on Bechlers with a floor-to-floor time of under five seconds, where a conventional English auto would take twenty or thirty seconds . . .

Anyway, back to our model engineering problems. If you are turning something small, try a rear toolpost. If you are having chatter problems with something large diameter, try running your lathe backwards with a tool upside down in your normal tool box.

Just a couple more things. If you have modern tools with interchangeable tips all your cutting angles should be right anyway and you ought to be able to get a perfect finish. If you don't, look round for some other reason. For instance, I have just commissioned a new (Chinese made) lathe, and found problems getting a good finish turning some small mild steel parts. So I miked up my work and found that not only was there a poor finish, but it was two or three thou oval! Therefore I opened up the screwcutting gear cover and carefully tightened up the two locknuts at the back end of the mandrel. Not much: perhaps about ten degrees rotation, while checking that the freedom of rotation of the mandrel was not impaired in any way. BINGO! With the same tool at the same setting, suddenly a beautiful, highly polished mirror finish.

Don't get the idea I'm complaining about this new lathe. Yes, it was cheap, but the workmanship is excellent - all ways ground, etc. And wonder of wonders, though barely more than £500 it has an all-geared head! This is such a luxury: having had a geared head I wouldn't be without it. With any turning job involving screwing or largely differing diameters you change speeds a dozen or more times, and to do it in two seconds with just a couple of small levers is such a joy. By contrast, it doesn't have a Norton box on the leadscrew, but one doesn't need to change the feed speed at all often. Of course it would be nice to have a tumbler reverse, but one can get along without it. And if the lack becomes at all troublesome I might make one some time . . .

Another afterthought. Modern lathes have high speeds: use them! My previous lathe was a Warco 500; 6" centre height plus a mill/drill. The lowest mandrel speed is 160RPM. Not too bad, but I was turning the wheels for Giles' No. 2 on it. The blanks were 12" diameter circles (flamecut) and 1.1/2" thick. They just fitted on the machine . . . . but 12" diameter? . . . 38" circumference at 160RPM? Yes, 505 feet per second. Perfectly OK with modern tipped tools. I couldn't take deep cuts, but with 1/2mm cuts we got there in reasonable time and shifted the metal. You see, the limitation on this sort of work is the power of the motor - 3/4HP in my case. If I could have got the speed down to 40RPM, would I have been likely to take cuts more than 2mm deep on a 6" radius? Surely not! Shallow cut at the faster speed still meant the same amount of metal taken off in the time. And believe me, I made swarf! Two dustbin bags from each wheel!

Moral: if you are turning steel or brass at (say) 1" diameter, that gives a cutting speed of about 260 feet per minute at 1000RPM, or 440 feet per minute at 1700RPM, while the new tipped tools are good for 500FPM. Why bother to go slower? You can push it to the point where steel swarf comes off blue or brass sprays all over the landscape. And when you have taper roller bearings in your headstock as many lathes do today, they are no limitation on speed.

In the old days of that Drummond 4", treadle driven at a maximum of, say, 500RPM you couldn't go faster. But if you've got a 3/4"HP motor and a substantial lathe you don't have to go slower . . . .

Brian Favell

(1923 - 2006)