Sunday, 30 August 2015

Holbrook Minor Single Phase Conversion

Re-homing a feral lathe

I have been getting increasingly dissatisfied with my cheap Chinse lathe. It is a cheap copy of the original Emco 7x12 lathe (itself a cheap lathe) stretched to 9 x 20 purely by lengthening the bed and increasing the centre height. It then has a milling head bolted to the back, to also make it a rather poor milling machine. I spent a fair amount of time and effort converting the machine to CNC, but eventually decided that the mill wasn't really good enough, and converted a Harrison Universal Milling machine to CNC too.

Having decided that a better lathe was needed, I then had to decide which one. The Harrison M250 seemed like a good candidate as they are very compact for the capacity, and lots of them are being removed from schools where they have almost never been used. However I am not the only one for who that is a good set of features, and so the M250s (and the bigger M300 and M350) end up selling for quite a lot of money. I then accidentally found an advert for a Holbrook Minor, and after a bit of research on lathes.co.uk I decided that they had quite a lot of nice features that lend them to CNC conversion. They don't have any gears in the head, but instead use a variable speed drive and a 2-speed electrically-controlled gearbox. The feed is by a DC motor, rather than geared to the spindle too. But, mainly, I like the look of them.
As it happened, the lathe I found on Gumtree had been sold months ago, but then another one appeared.




Unfortunately the lathe had been stored outside for a while, and had managed to get a bit rusty. However lathes always look a lot dirtier and rustier in photos, so I decided to have a look. 

The Holbrook Minor is diminutive only in name. It weighs 750kg, so when I went to look at it I took a van and a 2-tonne engine crane.

It will come as no surprise that I bought the lathe, otherwise the blog entry would be a bit pointless. 

Loading it into the van was pretty easy, the place it was being stored had a telehandler. 


The picture shows the drive-train rather nicely. The flange-mount 3hp 3-phase motor bolts to a Kopp Variator.  there is then a short drive-shaft to a 2-speed gearbox with two electrically operated clutches and a brake. 

When I got the lathe home it was a simple matter to lift it with the engine crane, then drive the van out from under it. 

Actually, that's a lie. It was a horror of a manoeuvre, mainly because the 2-tonne engine crane (from KMS on eBay) appears to be made from cheese. I had the boom in the 1-tonne position, but with the van half-way out from under the lathe, this happened...


This left me with my newly purchased toy trying very hard to fall over and out of a van, and seemed quite keen to wreck the (hired) van in the process. Luckily a neighbour was at hand and he was able to quickly find a trolley jack, and another neighbour had left some bits of wood lying about. We were able to stabilise the situation with the jack, and then have a re-think about how to proceed. It turned out that the crane didn't bend any further in the 1.5 tonne position (and so it shouldn't, with a 750kg lathe) and that access for the crane was easier with the lathe half-way out of the van, so that this position could be used. 
The lathe was then lowered on to the legs of the engine crane and an attempt to roll the lathe into the garage mainly served to prove that the 2-tonne engine crane doesn't have 750kg castors either, as they collapsed. 
So, the lathe was finally rolled into the garage on several lengths of 1" pipe that I had plenty of. 



On a whim I decided to try the tailstock of the 9x20 Chinese lathe on the bed of the 10x20 Holbrook. It seems to demonstrate quite nicely the difference in scale between the nominally similar lathes. 


To access the Variator and gearbox for oil-changes and maintenance I needed some way to move the lathe away from the wall. The engine-crane had proved unsuitable for the job, so I made a frame the can bolt up around the lathe, hook into the hold-down bolt recesses, and then be used to jack the lathe up onto castors to roll it around the workshop. I am pleased to say that this actually works nicely. 



The eagle-eyed will spot that the lathe is actually hanging on a strap in the photos, I forgot to take any  after I had made the hooks, and these photos are of a test move to make sure that the (simple) screw jacking system works. 

Single-Phase Conversion

Once the lathe was home I almost immediately obtained a manual (including the circuit diagram) from lathes.co.uk.

The Minor was quite a high-tech lathe for the day (the contactors had a production date of 1968), and this is reflected in the unusual complexity inside the control cabinet.


At the top-left of the main box is the transformer that produces 24V control power for the gearbox, 110V of the contractors, 300V for the DC feed motor and 50V for the low-voltage light from a 440V 3-phase input.  Below that are the reversing contactors and the main on-off contactor Then an overload trip unit, and a rack of fuses. The large finned double-bank Selenium Rectifier converts the feed-motor armature current to DC, the one in the door is the field-coil bridge rectifier and the one at the bottom-right is the 24V bridge for the gearbox clutches. 

The variable speed DC motor achieves variability by passing the 300V AC through a Variac on the control console. 
At least I think it is a Variac, the wiring diagram shows a loopy version of a potentiometer, but the abbreviation "VR1" seems odd for a Variable Transformer. 


To run the machine in my garage I needed to make all that work on 240V single phase. The first thing to do was to re-wire the spindle motor in delta configuration. Unusually I needed to make jumpers for this, as the motor has an unusual circular terminal layout, and the star-mode jumpers could not be adapted. 


Star (Or Wye if you are American) 480V


Delta 240V

With the motor now converted to 240V operation I could use a Variable Speed Drive (VFD) to generate 3-phase. I found a suitable 2.2kW drive from a vendor on eBay for £83. 
I tried to retain as much of the original control wiring scheme as possible. This may seem odd when the plan is to convert to CNC, but I wanted to test the lathe out first. 
The original Forward/Reverse joystick contacts were wired to the VFD FWD/REV terminals and I added a 240V coil contactor as the no-volt release main contactor. The lathe came to me with a key-operated switch but no key, so I replaced that with a 22mm dual push-button start-stop to control the contactor. 

To control the 2-speed gearbox I installed a DIN-rail mounted 24V PSU (also from eBay) and again this was wired through the original control wiring. 

The DC motor originally ran from a 300V output of the transformer for the field coils and the 240V between neutral and Phase 1 for the armature. I wired both these to 240V single-phase power through a couple of modern 35A silicon rectifier bridges. I didn't need 35A, but that size from Maplin have handy Faston terminals and a central mounting hole. 

It proved to be rather easier to re-wire the suds pump for 240V as that was just a matter of moving wires:


Rather than use a small VFD for the suds pump I chose the much cheaper route of using a capacitor to synthesise a third phase. For simplicity this was mounted in the control cabinet. So the coolant switch now just switches live + neutral  and then back in the cabinet again the third wire to the pump is connected to one side of a Motor Run Capacitor and the 240V Live to the other side of the capacitor and the first terminal. 

And here is how the control cabinet now looks:


There is quite a lot more room in there now. Time will tell if there is enough for a PC and servo drives to run the lathe in CNC mode. 
At the top, right to left: The Lovato 2.2KW VFD, then (on DIN rail) a 24V PSU, 240V contactor, a pair of rectifier bridges and the suds-pump capacitor. 

After one false-start where it turned out that the current through the power-on lamp was enough to keep the contactor engaged (I had misunderstood the wiring scheme slightly) the lathe now has all the features that it had originally, but runs from 240V single-phase. 

Before anyone comments: Yes, I need more individual earth cables. 







Monday, 3 August 2015

Rivett lathe slideway refurb

Rivett lathe slideway refurb


As documented in a previous instalment of this blog about 18 months ago I made an oak stand for the Rivett 608 that I bought in an eBay accident. What I don't mention in that entry is that when I got the lathe working I then realised that the slides were badly worn. So I took it apart to investigate.

The Rivett 608 has a rather strange design. The saddle is on the front of the bed, sliding in a dovetail. The reason for this is unclear, but Rivett were full of curious ideas like that. 



It appeared that a previous owner had re-ground the external faces of the bed, but had not done anything to the internal dovetail. 
Unlike every other dovetail in the world the Rivett saddle dovetail bears on all 5 faces. This was achieved by very careful scraping by someone who was so proud of his skills that he signed the back of the serial number plate. 
So, the effect of the external grinding was that the outer vertical faces no longer contacted the saddle, transferring the forces to the internal bits between the feed shaft and leadscrew and the sloping dovetail angles. 
I managed to re-scrape the saddle so that all three vertical faces were simultaneously in contact (if the Blue was applied thickly enough) but I got rather discouraged by my attempts to scrape the internal angles. It was looking like it would be very hard indeed to check those faces for angle and straightness and parallelism to the top surface of the lathe. 
Another problem is that the Rivett dovetails (in another departure from accepted practice) are perfectly sharp-cornered, with no relief groove as is normally seen. Which makes scraping even more difficult. 
So, the lathe stayed in bits for a few weeks, and then a Ner-a-Car project came along, and weeks became months. Then months became a bit over a year.
In the interim I bought a new lathe, a Holbrook Minor for conversion to CNC. I made the decision that I wouldn't allow myself to start on the Holbrook until the Rivett was back together and capable of making parts. 

I decided that what I needed was a planer, like this nice one on YouTube. It is almost certain that that is how the dovetails would have been originally cut. 
However, I couldn't find one locally, and then had the idea that perhaps I could simply plane it by hand. It works for wood, after all. But the problem with that would be holding the cutter in line with the lathe bed. Some sort of linear guide would be needed. I considered a number of possibilities then it finally struck me that I was over-complicating. If I want to be parallel to the top surface of the lathe bed then I should use the top surface of the lathe bed...

So a plan was born. I scavenged around the workshop for a bit of flat plate that I could put a hole in the middle of, but then my eyes fell (figuratively) on an old Myford motor plate. After a skim flat on the mill and the addition of some adjustable brass guides and a spare lathe compound slide, I had the device ready. 

The first thing to do was to get the angle right. I did this by mounting a dial indicator in the toolholder, and winding the slide up and down in various places until the reading did not change. 


Once the angle was set I replaced the DTI with a lathe tool, and set to work. 


I even made an unexciting video of the process. A bit of light-scraping was needed afterwards to get the surface texture right, but the fit turned out to be rather astonishingly good. 




The saddle adjusting gib had already been shimmed twice, and there was no way that it was going to be suitable for the new saddle position, so I set out to make a new one. The gib is 12" long, about 1" wide and 1/4" thick. Unfortunately the only cast iron bar I could readily find was 35mm diameter, so there was a lot of machining to do. 

First I flattened off one side to about half way. 



I had to progressively fit and remove clamps as the cutter moved across the work. 
Then to hold the increasingly thin strip down I drilled and tapped a lot of M4 holes, and bolted it to a faced-off lump of scrap plate. 




I was then able to mill the bar to the correct thickness and taper. (CNC makes this sort of thing really easy. G1 F50 X-350 Z8.05 sort of easy. At this point I was suddenly glad of the excess width that I had inadvertently bought, there would have been no room for screws with a more conservatively-sized bit of material. 

With the thickness approximately right I was then able to cut down to width with a dovetail cutter to get the required angle. (my mill has a rotating but not "nodding" head. This process would be easy on a Bridgeport or similarly adjustable machine)


You might have noticed that the gib appears to be rather longer than 12". In fact it is 350mm, and would have been longer still but for the fact that 350mm is the limit of travel of my milling machine. I always get nervous around tapers, and like to make them long, try for size then cut off the excess. 
In this case the gib slid through about 1" further than I thought it should from the measurements, so I was glad of the extra length. 

The adjusting screw slots were next on the list, and those turned out to be a bit of a challenge because they are in the end of a long thin (and fragile) bit of iron. 
I could have switched the mill to the horizontal head, but that is a fair bit of effort, and I hit on a way of doing it with the lathe which actually worked rather well. 


And here is the finished gib installed on the saddle. It is on the limit of adjustment, as it is only ever going to get looser. (No, you can't really see it. )



The next problem is that the saddle sits low in the bed. It always would have done, regardless of the method used to sort out the dovetail wear. I think that I can get enough adjustment in the saddle angle screws for that to fit, but the offset is enough to make the leadscrew bind in its very tight slot, and to make the power-feed bevel housing foul the bottom of the lower slot. So a little more work to be done there, but I think I have workable solutions.