Author Topic: Modifications to the feed drive system on a lathe.  (Read 4170 times)

Offline tonyfoale

  • Jr. Member
  • **
  • Posts: 12
  • Country: es
    • MotoChassis - Source for the chassis design book & software.
Modifications to the feed drive system on a lathe.
« on: January 23, 2020, 05:31:04 PM »
Sometime back I made some modifications to the drive system and head stock of my lathe described here

At the time of those modifications I was planning on fitting ball screws and motorising the feed shafts. So I did not bother ensuring that the original feed mechanism, specifically the change gears, remained functional.  As things have turned out I just did not have time to do that and I got fed up with always hand feeding the saddle and cross slide.  I decided to refit the feed gearbox and the change gears.  Photo 1. below shows the original drive system showing the change gears.

The small pinions used for direction change spin very fast. These are the two adjacent upper small pinions and the bronze bushes were worn as were the stubs that they ran on. This wear gave some roughness and a lot of noise to the drive, so I reamed out the bushes and made new stubs which eliminated any roughness and reduced noise.

However, the new drive belt was in the way of the big 120/127 gear pair. Photo 2. shows the clearance problem with the new drive system in place.

I could have fitted a longer drive belt and moved the idler pulley, but that was the drive side of the belt and I wanted to keep the belt as straight as possible.  The idler is only fitted to prevent belt flap at specific resonant and harmonic speeds and only causes a small bend in the belt. 

I saw another way of fitting that gear pair by moving them to the other side of its driving pinion and the pinion on the gearbox input as shown in fig 3.

That orientation gave rise to a lateral clearance problem shown in fig 4.

After a bit of head scratching I decided to change tack and look at driving it with a belt. I had a toothed belt that was close enough to investigate the feasibility.  As fig 5 shows it looked worth considering.  However, my leadscrew is imperial but I work in metric, so to be able to screwcut in metric I would have to retain the 127/120 multiplier of the gear pair being replaced but 127/120 cannot be divided down any more as integers. That would need two large toothed pulleys and there was no room for that. That meant if I used a belt I would have to use smaller pulleys and forget about screw cutting.  Well, that would be no problem.  I have a CNC mill which I can use as a CNC lathe by putting the work piece in the spindle, or I could use it in thread milling mode.  My son also got a new lathe recently so I could use his. That left me to only consider feed motions and the gearbox allowed a large range of feeds for a given input speed. fig 6. shows that there are 5 x 8 = 40 different feeds.

Forgetting about screw cutting made things much easier.  Unlike screw cutting, the feed does not need to be synchronised to the spindle, so in place of a toothed belt I could use a poly-V belt which give a smooth drive and would provide more than enough torque with a narrow belt.  Width was an important consideration as we will see.

My first thought was to retain the original 127/120 = 1.06 ratio of the gear pair being replaced.  That simply meant that the gearbox input pulley needed to have a pitch diameter 1.06 times the pitch diameter of the driving pulley. I made the top pulley from a piece of aluminium out of the scrap box, which seemed about the right diameter (very scientific). I measured the centre distance and calculated the length of belt needed and ordered one.  While waiting for that to arrive I roughed out the lower pulley. I did not want to fit an idler pulley and the centre distance was fixed, so it was important that the pulley sizes gave a suitable belt tension.  To achieve that I intended to machine the second pulley when I had the belt. 

When the belt arrived I offered it up and to my horror it was way to long. I checked the length and it was as marked and as I ordered.  I rechecked my calculation which is quite easy. If the pulleys are near the same diameter you average the pulleys' circumferences and add twice the centre distance.  My error was that I added the sum of the circumferences not the average. I had a choice between ordering and waiting a week for a new belt or making a larger lower pulley.  A larger pulley would cut all my feed speeds in half.  The highest feed was way higher than any that I would ever use so that seemed like it would work well because it would give a lower minimum feed which could be useful for fine finishes.  So I went for the larger pulley option but I had to make it from an old steel pulley because I did not have a suitable piece of aluminium. Fig 7. shows it all assembled.

There was very little lateral clearance for the belt which had to fit between the main drive belt and the pinion which drives the top pulley. It just fits but I had to thin down the upper concentric pinion. The last photo shows the tight lateral spacing.

So how does it all work?

Very well. I am pleased with it. Eliminating two gear engagements has further reduced noise, the belt that replaced those is very quiet and the whole system is very smooth. The feed reduction is a bonus that my belt length error demanded.  I measured the actual feed over a range of saddle movement of 5" and it came out almost exactly half the feed shown on the table of feeds.  It will slow a little due to belt slip when making heavy cuts, but that will be of no consequence. I do not even need to make that 0.5x calculation mentally because the feed steps are binary, i.e. each of the coarse steps upwards is double the previous, so I just have to read the next line down on the feed table.

For now I am happy to have my power feeds back, but the thought of ball screws and motors still lingers. Then a two axis CNC becomes almost trivial, but do I need a CNC lathe?  Probably not.

Offline Brass_Machine

  • Administrator
  • Hero Member
  • *****
  • Posts: 5502
  • Country: us
Re: Modifications to the feed drive system on a lathe.
« Reply #1 on: January 24, 2020, 02:17:37 PM »
Very good read... Thanks!

Science is fun.

We're all mad here. I'm mad. You're mad.

Offline celsoari

  • Full Member
  • ***
  • Posts: 131
  • Country: br
Re: Modifications to the feed drive system on a lathe.
« Reply #2 on: February 27, 2020, 06:19:24 AM »
nice work Tony

Celso Ari
Greetings from Brazil

Offline jackary

  • Jr. Member
  • **
  • Posts: 15
  • Country: gb
Re: Modifications to the feed drive system on a lathe.
« Reply #3 on: December 20, 2022, 06:15:41 AM »
Hi Tony,
Rather belated read of your experiments with your lathe headstock, it was very interesting. I agree with  your conclusions about rigidity and a simple direct polyvee belt drive and avoiding gear drives for smooth running. I came to the same conclusions with my Stepperhead lathe and that is smooth and viabration free. You may be interested in my rather long  description shown here

Offline BillTodd

  • Hero Member
  • *****
  • Posts: 1061
  • Country: 00
  • Colchester Essex (where the lathes were made)
Re: Modifications to the feed drive system on a lathe.
« Reply #4 on: December 20, 2022, 06:58:01 AM »
I believe 'gear head' lathes were a response to the general lower power available from early (low cost) motors and lower speeds used with tool steel/HSS tooling.  All the cnc lathes at the works are belt driven (even the biggest 1m/40" swing machines) .

It would be worth looking at replacing the feed gearbox with a motor too. Infinitely variable speed feeds are one of the best features of my HLV-H .

It's probably worth avoiding a cnc conversion, because unless you make multiple parts, it always takes longer to use.  (I have watched many a youtube video where the motivation for the conversion seems to be the belief that the cnc will make up for the operators lack of experience - they'd be better off just learning too use the lathe!)   

Offline hanermo

  • Jr. Member
  • **
  • Posts: 95
Re: Modifications to the feed drive system on a lathe.
« Reply #5 on: January 13, 2023, 12:19:01 PM »
Infinitely variable Very Very powerful feeds and most importantly huge spindle torque are probably the *best feature* of my CNC lathe conversion.

I commonly run my lathe as semi-manual with an MPG and adjustable feeds and speeds via 2 override industrial potentiometers for spindle speed and feed speeds.
Imho the cnc conversion is *vastly* better than my previous -excellent-, manual-only lathe with power feeds.

The infinitely variable spindle speed with huge torque was to my great surprise vastly better than the previous, already very good, mechanical multi-belt and backgear system.
In practice, I learned that often much higher spindle speeds, 2-3x, and more aggressive cutting, upto 3.5 mm depth of cut on 140 mm D tool steel, at 700 rpm, produced outstanding results, very quickly.

Lessons learned:
Huge spindle torque with variable rpm is a major, major, game changer.
About 3x the productivity, with 2x the quality, and 1/2 the work.

Analog controls via potentiometers or mpgs for feeds is amazingly useful and important.