Union Tool & Cutter Grinder Restoration Shenanigans

[DiyAddict]

For years I've wanted some sort of bench top surface grinder for my tiny shop. Of course, I know there's no such thing (at least within my budget) - surface grinding needs large, bulky castings with the stiffness required to maintain any accuracy.

Nonetheless, the subject of small surface grinders comes up on the forums from time to time and the Union Tool & Cutter Grinder by Harrison is often mentioned. Indeed, Harrison even supplied a small 'Surface Grinding Fixture' for it, which consisted of an Eclipse Minor magnetic chuck and a mounting plate.

I just want to grind the odd parallel, make some flat india stones, clean up some milled surfaces etc. I know I'm never going to get a mirror finish. Also, I do have a number of tools and cutters that could do with sharpening!

So a few weeks ago, I finally bit the bullet and acquired this one:

It seems to be the later Harrison model, with the flat front and no 'Union' decal. It came with a load of accessories, and seems to have been looked after. The spindle seems to run true, and I was hoping to put it to work straight away.

However, it has a lot of wear. Running a DTI over the table gives a variation of 3 thou over about four inches. Also, the Y axis is very stiff and has a lot of backlash - around 25 thou. I think the lead screw is bent as well as worn.

So the plan is to scrape it flat, make a new lead screw and nut, and deal with any other issues with the tools I have at hand. This won't be a full restoration - I've no intention of painting it (mainly because I hate painting and am rubbish at it). I know @Pete. already has an excellent restoration of his Union T & C Grinder on this site.

At the time of writing, I've already done quite a bit of this, but am far from finished.

 

[DiyAddict]

The first test was to see if the main table had sagged - and the measurements weren't encouraging! The top table comes off with just a couple of bolts. The main table assembly slides off by removing the table stops and winding the rack all the way. It's held onto the saddle with what looks like a single piece of ground flat stock, around 1/4" thick:

More on that later. I put the main table upside down on my surface plate, and shimmed either side so that any deviations from flat would be roughly symmetrical. At this stage, I just ran a DTI along the flat way. The main table had indeed sagged (or worn) by around 8 thou either side. Not great, but not terrible either: the main table is 20 inches long.

To inspect closer, I unbolted the rack and removed the flat retaining plate. This is where a surprise happened. The retaining plate was bent - really bent:

The difference was around 4mm over the whole length. (Excuse my mixing of units - I tend to prefer metric for large measurements, but imperial for small ones). What could have caused a piece of 1/4" ground flat stock to warp in this way? When I removed it from the main table, the main table sag reduced from 8 thou to around 5 thou:

So I'm going to replace the retaining plate. There's no way I'm going to succeed in unbending it and ground flat stock is widely available. I decided to leave the main table for a week, so it could 'adjust' to the relief of pressure. Then the plan is to scrape it flat.

 

[DiyAddict]

While waiting for the main table to settle to its final sag, I removed the saddle and the casting with the Y axis dovetails. This latter is just held on with a single bolt, plus a tapered pin to prevent the table from pivoting:

Note the Lemon Pledge. From when I naively hoped I could just shine up the paintwork, add a bit of oil, and have the whole thing operating.

Anyway, It looks like Harrison intended to allow the whole cross-slide to pivot, but it's never mentioned anywhere in the manual, and it's difficult to think of a sharpening task that couldn't be achieved by rotating the grinding head or sub-table instead. In any case, this arrangement makes it super easy to inspect and measure the dovetails - you can just lift the casting off and place it straight on your surface table.

Here, I was fairly lucky. Any wear was even, and I was able to scrape the casting parallel in just a few passes. To check the dovetails for parallelism, I adapted a rather ugly device that I'd made some years back to make a similar check on an Adept Hand-shaper:

The dovetails were only about a thou out of parallel, so a few passes of step scraping sorted that out fairly quickly. So far so good. These are just initial roughing passes to get everything approximately square and parallel.

 

[DiyAddict]

More inspection. This time the main saddle. It's upside down in this pic, showing that it's secured to the dovetail casting with just a single fixed gib.

I'd never seen a fixed gib before, with no means of adjustment than shimming, but apparently they are not uncommon for grinders where the cutting load is comparatively low compared with milling machines or lathes. The helical gear for the X travel seems in excellent condition. No obvious wear to that or the cast-iron rack.

When I first took this machine apart, the screws holding the gib in place were loose. Not even finger-tight. As soon as you tightened them even slightly, the Y travel would lock up. Equally, the screws holding the nut in place had been removed. The nut remained in place by two dowels, but was able to ride up and down slightly. I'm not too worried by this. I can imagine these slight 'fixes' being applied to keep the machine working (I'm told it was part of a college engineering department). In fact, I'd rather have this situation than someone with unknown experience attempting a botched repair job. Although that last sentence does somewhat apply to me.

Otherwise, the saddle had wear consistent with the table sag. I was able to get it approximately parallel and square fairly quickly, using a combination of my surface plate and a 12" camel back straight edge. And a carbide hand-scraper, of course.

 

[DiyAddict]

Here's the state of the Y axis lead screw and nut:

The lead screw doesn't look too bad, but it is bent, and you can see that the threads are much narrower where they are worn, than they are at the beginning on the thread where there is no wear.

The nut itself - a solid lump of bronze - is knackered. You can see how razor-sharp the crests of the threads are, when they should be flat.

Disappointing, but there are opportunities here. This grinder is imperial, where all my other machines are metric. Also, the lead screw is a right hand thread. On my lathe and mill, they are both left hand threads. So rotating the handle clockwise feeds the work towards the cutter. I don't know why Harrison did it the other way around on this Union grinder. Is this normal for T & C grinders? Seems a recipe for a dangerous crash to me.

It's really hard/expensive to source imperial lead screws and nuts in the UK. I'm tempted to go metric, where I could at least buy a trapezoidal nut easily, but where's the fun in that?

Instead, the plan is to make a new acme lead screw out of some EN8 I already have, and make the matching nut out of acetal. I've heard quite a few people praise the combination of steel and acetal, and some say it's actually longer lasting than bronze. I'm going to make them both left handed, so the Y feed operates the way I expect it to and hopefully avoids a few dangerous accidents.

 

[RobCox]

Looks like you're as naive I was when I got my lathe last year. A quick clean up, lick of paint and ready to go. 6 months later, complete disassembly, new bearings etc etc and it's now being used!

I've got one of these Union grinders. Very useful. I've used it for lathe toolbits, endmills (flutes and the ends), reamers, horizontal cutters and so on. I was going to try surface grinding with it. You need a spindle extension (which I made), then you find the drawbacks. The spindle is the motor spindle. No precision, no rigidity, chatter and a carp finish. I ended up with an Eagle surface grinder (which was in pieces, another mammoth project) which handles that requirement better.

For the tool sharpening tasks the Union is great though. Oh, and my Union is the only metric machine in my workshop. I'd take the easy option and go metric for the screws. I think the cross slide screw is RH for a reason, you tend to operate it with your left hand, so the behaviour turns out to be more intuitive for a right hander.

 

[DiyAddict]

Rob, nice to hear from you! I've really enjoyed watching your YT videos, and have got a lot of inspiration from them.

Yep, I have low expectations but a bona fide surface grinder isn't an option for me, even something as small as an Eagle or Herbert. I still reckon I'll end up with something that suits my needs - we'll just have to see where this heads.

I think the cross slide screw is RH for a reason, you tend to operate it with your left hand, so the behaviour turns out to be more intuitive for a right hander.

Ah- I hadn't thought of that. I've since looked at the operation of a few on YouTube. The Cincinnati no.2 also seems to have a RH cross slide screw, where the KO Lee seems to have a LH one. I'll think some more on this, but - spoiler alert - I've already cut the leadscrew.

 

[DiyAddict]

Time to tackle the main table. It had definitely sagged, rather than just worn, so would need around 5 thou taking off each side and each way to get it straight again.

Since it was bent, I had no reference surface to work to, so the first job was to choose one, and scrape that flat. With the main table upside down, I chose the very edges that run down either side. These are already machined, but are not bearing or mating surfaces.

To save some effort, I mounted the table on my mill with careful shimming, and machined off around 3 thou from each corner (The table was upside down, so the middle was 'low'):

Sorry about the blurry image but you get the idea. This left me with only 2 thou to hand scrape. Once that was flat, I could flatten the top surface of the table. These are the collar and lands that the top table bears on.

I laid the table down on the surface plate and took some measurements with a DTI:

These were as planned. I'd chosen the shimming for the first operation to minimise the metal I'd eventually have to remove from the ways.

I left the two edge lands alone, but machined a couple of thou from the central collar to minimise the scraping load. This one's less blurry:

Once that was scraped flat, I needed one more reference surface. This would be the one that the ways aligned to. I chose the side with the Tee slot where the table stops are fixed:

This was already pretty flat, and didn't need much scraping. Although not strictly necessary, I did my best to get it perpendicular with the other reference face, using a cylinder square and a torch:

 

[DiyAddict]

With the two faces and one edge all flat, I could finally start to tackle the ways. A few measurements on the surface table gave me an idea of what I was up against:

The figures marked are in tenths (0.0001"). Scraping these ways requires scraping into corners, so I decided to mill an undercut into the middle of the Vee way, and in the corner of the flat way:

For the Vee way, I used a 2mm HSS cutter. It felt a bit hairy but I've watched enough Stefan Gotteswinter to get my speed right and not overdo my depth of cut, and nothing broke. In the corner of the flat way, I was able to use a much larger cutter, as most of the flat way goes untouched by the saddle.

 

[DiyAddict]

So far so good. There would be a lot to hog off before I would need to do any blueing, but once that stage arrived, I had a problem: I'd need a straight edge capable of getting into the ways and, preferably, the same length as the table or greater. My only straight edge was 12", and the table is 20" long. And where the ways are recessed, it's very awkward to get into them for blueing.

At the time of writing, there's a guy selling the raw castings for an 18" Camelback straight edge on Ebay at a very decent price. I ordered one of these straight away as they're so rare to find in the UK. However, even after machining and scraping flat, it wouldn't be ideal for the Union's table.

I did have a two foot long offcut of marble from a fireplace makeover years ago. I wondered whether it would be possible to scrape it flat? If so, it would be the ideal size as a straight edge for the Union's ways.

Well it seemed to scrape well with my carbide scraper:

But the marble was very soft. It was hard to clean it down between scraping such that a load of marble grit didn't end up in the blue on my surface table. Also, I struggled to get the contact points any smaller than you see here. Maybe something to do with the grain of the marble. Also, given the marble was so soft, I'm not sure how many prints I could use it for before it wore down and was no longer flat.

Well it would be useful for giving me a print along the whole length of the Union's table, and should show up major hollow areas. It would be useful if all else failed, and in fact I did start to use it with good results. But then the Ebay raw casting arrived.

 

[DiyAddict]

Quick interlude on machining a large straight edge on a small mill (A Taiwanese clone of the Emco FB2). The casting was 500mm long, but the maximum X travel on my mill is only 400mm. Fortunately the casting came with sprues on the handle that I could machine flat to assist with mounting on the mill table.

I tried a large diameter flycutter to make up the extra travel, but the casting was sufficiently raw, and still with a few hard areas, I soon realised that this would take ages. Days probably:

Yes, I know the clamping arrangement is not a paragon of maximum rigidity. It worked surprisingly well though.

Actually, the FB2 clone has a round column, which means that you can make up the travel by swinging the head around a bit and doing it in two passes with just a standard carbide end mill. I used a four-flute for the top, then a two-flute for the edges:

Ok, the furrows created by the second pass were a couple of thou above those of the first, but I was able to get to the nice soft cast iron within just a few passes. I think I had to take around 4-5mm off the face in total. I later discovered that one of those 'corn cob' roughing end mills gave a much better surface finish on the edges than the expensive 2 fluter I initially used.

This was all the machining I did on this. Normally, you'd want a bevel cut into it to get into dovetails, but I needed a 90 degree edge to get in the flat way of the Union table. It didn't take long to scrape, and here it is, blued up ready for printing:

 

[RobCox]

Who's been a busy boy then! I left the sprues on my straight edge rather than saw them off.
I guess you don't have the room, but instead of that fragile 2mm endmill a 5/64 slitting saw on a horizontal mill would have done that job stress free!

 

[DiyAddict]

Onto the ways then. I've done scraping before, but I needed to learn a few new techniques for this one.

First the flat ways. The four lands for the rack & retaining plate are considerably higher than the flat way, which made scraping pretty awkward.

I had to learn pull-scraping to be able to reach everywhere, and found to my surprise that I quite liked it - easier to get accurate, short scrapes in some cases. I didn't even think of blueing until I was within a thou of flatness. In fact it was easier to 'blind scrape' then measure with a DTI on the surface plate, as that ensured I was staying parallel to the other faces.

Even once I was blueing, I still took it back to the surface plate for frequent measurements. Pretty soon I had it close:

I measured the flat way in four rows, to make sure I wasn't cutting a furrow, then wrote the results on any spare surface. Most of the time I'm within a couple of tenths. There are some larger differences on the bottom row, but that part of the way never comes into contact with anything, so it's not critical.

The scraping looks terrible in this photo! It's really nowhere near that bad, honest!

Once that was done, I scraped the four lands flat too.

 

[DiyAddict]

Then the Vee ways. This is where I needed the reference surface running along the edge, so I could butt up a DTI stand against it:

Here, I've managed to get within 2 thou by step scraping, so nothing going on in the middle yet. This is handy, as I'll be using what's left of the original surface to define the geometry, since it's impractical to take multiple rows of measurements like I could for the flat way.

Here's a close up showing the wear on the Vee way:

You can see the scoring along the worn part, and some flaking marks that a previous owner must have added to retain oil. The two arrows show the original machined surface, where the mating Vee on the saddle doesn't touch. It shows I could have gotten away with a 3mm end mill when creating the undercut. This would have made life a lot easier, because that thin unworn edge was quite a job to scrape away without overly scraping the main face.

Well on the way here, just the hollow in the second Vee way to finish. If this was a milling machine or surface grinder, I could be thinking about the pub now, or at least knocking off for the day with the warm glow of a job well done. But since it's a tool & cutter grinder, I have a WHOLE 'NOTHER TABLE to scrape!

 

[DiyAddict]

I'm sure you're all as sick of scraping now as I was, so I'll try just to stick to any entertaining aspects of the top table. But first, here's a soul-destroying picture of how bowed it was:

At least there are no valleys and ways nestled within it - just flatten either side. Nonetheless, I was so sick of scraping, I took it to a local shop with a J & S 540 to see if he could surface grind it flat. No dice - the 540 has a travel of 19" I think, but his magnetic chuck was only 16" and he wasn't prepared to take it off. Quite reasonably.

Things then descended into one of those cycles of trying alternative methods that take much longer than if I'd just gotten on with hand scraping. Here I am attempting to flycut the lion's share off with my mini mill:

This could have worked, but it turned out that the top table on my Union has a really gnarly hard area in the middle that just made the flycutter skip over it, or dig in, depending on I know not what. Or care.

So I just had to man up, defer my beer, stick the radio on and get the hand scraper out. Some time later, it was done:

With that side flat, I was able to save a little time milling the other side to shave off a couple of thou:

then scraped the rest. So here we have the final, fully scraped and - more importantly - straight, flat and perpendicular table assembly:

A thing of beauty, surely.
Paul

 

[Martin Hocking]

Bravo paul!Great stuff, keep going!

 

[DiyAddict]

Thanks Martin. I'll take all the encouragement I can get right now! Hoping to have more to report later.

 

[indy4x]

@DiyAddict definitely keep going, finding it really interesting

 

[DiyAddict]

Right then, with the initial scraping complete, I'm into the fitting stage, which will inevitably mean more scraping, but I'll try and keep it short.

The single fixed gib has been bothering me for a while. This is the one that holds the saddle onto the knee casting - here it is upside-down:

I wonder how this came on new machines? If it's not a precise fit, the whole table will either rock, or seize up when the allen bolts are fully tightened. Maybe a small amount of rock can be tolerated for tool grinding? I know the gib strip was originally just machined, not ground or scraped. So either it was machined to exacting tolerances, or Harrison were happy with a little play, or they fitted it with shims.

There were certainly no shims on mine, and as I mentioned it before, you couldn't fully tighten the bolts without locking up the Y movement completely. And if I'm eventually to persuade this machine to perform surface grinding, I really think smooth movement with no play is going to be essential.

The fixed part of the saddle that the gib bolts to is just machined at a right angle. There's no undercut and it wasn't straightforward to machine one, so I cut a generous chamfer on the gib itself to assist with fitting:

Without the chamfer, it would have been very difficult to assess whether the gib was properly seated as I scraped away the various mating surfaces. I then made sure the gib itself was flat by scraping:

You can see gib was initially bowed, so fixing it to the saddle with the bolts would have bent it, altering the geometry of the dovetail part. But it didn't take much to get it fully flat:

I could then attempt to fit it by scraping in the dovetail face fully. It was a long way out initially:

To make it match precisely with the mating surface, I blued up the knee casting (which I'd already scraped flat a while ago), then bolted the gib as tight as I could without locking up, then moved the knee against the saddle to create the print each time.

As I got closer to full contact, I added a DTI to measure play between the knee and the saddle. Eventually, I got a good surface, with play of just a couple of tenths with the gib fully tightened down. Here's a pic where I'm well on the way:

I'm pleased I got the whole assembly to mate well with minimal play, but I can't help feeling I've gone about it all the wrong way. It was so labour intensive - there must be a better method. I suspect shimming was anticipated when these were designed.

 

[DiyAddict]

Fitting the saddle to the table now. For this I could use the table as a straightedge, knowing that the Vee way would align everything for me. Once again, it was difficult to get into the ways to spread the blue, but a wallpaper seam roller, with a bit of foam rubber glued on worked very well:

The first time round, I blued the Vee ways, and was delighted to find almost full coverage on the matching saddle vee ways.

However, the saddle ways weren't quite matched to the table, as I'd experienced a tiny amount of rocking when I first placed them together. The rocking was minimal - just a couple of tenths. The first print showed this up quite clearly:

A few more rounds, and they were nicely matched - no more rocking: