Grinding Dovetails

[Chips Everywhere]

I need to recondition one of my Hardinge HLVHs, I'm not looking for a perfect like new lathe. My goal is to follow the proper repair process to the best of my abilities and not half a$$ the repair. I was planning on sending it out for repair, but after getting quoted 13.5k for a partial rebuild, it had me thinking that I need to address some of these issues myself to reduce some of the cost a bit. If its out of my capabilities I will send out what I can't do in house.

The lathe in question was operated with coolant by enlightened operators with infinite wisdom, and so the hardened strips (fitted on later model machines) lifted from the cross slide. After a significant amount of research I found an epoxy to glue them back on, and I was able to put them back on within +/- 0.003" of the top of the dovetail.

Now the fun part, I need to grind the strips to remove that 0.003". I thought about using this fixture I made a while back. The fixture is made of aluminum with linear rails and I would use the lathe's compound with a THEMAC tool post grinder to grind the strips. The fixture would mount to the bed and has jack screws that allow for alignment. I am planning on making a dummy dovetail and mounting it to the front follow rest mounting holes, so that I can test grind it before I actually grind my cross slide.

The fixture will get inspected on a surface plate to make sure its not bowed or tweaked. I have limited grinding experience, so I don't even know what kind of wheel would be correct for my application. Any recommendations would be greatly appreciated. If this sounds like a bad idea, please let me know.

Before any grinding is done, I will be sure to prep the lathe against any grinding dust.

This is what I have so far:

 

[rimcanyon]

I’m not familiar with hlvh cross slides, so where are the hardened strips? On the face of the dovetails?

At first glance this looks like something you could do on a surface grinder.

 

[eKretz]

I think he's trying to grind the angled portion of the dovetail. Guessing there's no access to a surface grinder?

 

[Chips Everywhere]

I’m not familiar with hlvh cross slides, so where are the hardened strips? On the face of the dovetails?

At first glance this looks like something you could do on a surface grinder

The apron/saddle are one piece and it has Teflon under the saddle that you don’t want to damage either, so it’s a bit of an oddball to hold onto. Plus I only have a small surface grinder, 5x12

These are the harden strips:

 

[Chips Everywhere]

I think he's trying to grind the angled portion of the dovetail. Guessing there's no access to a surface grinder?

I’d like to grind the glued in steel inserts, but the dovetail would be a plus.

I don’t have a grinder big enough. I would need over 18” travel, preferably 24”

 

[eKretz]

You might be money ahead sending it to someone with a big enough grinder.

 

[MCritchley]

That’s a lot of messing around, you could have scraped that in a few evenings.

You’ll still need to scrap it in, it’s never going to have a good bearing.

 

[eKretz]

Hardened steel is a bit difficult to scrape.

 

[michiganbuck]

I think grinding a dovetail to everything straight and parallel to <.oo1 is pretty tough. The hold-downs better not distort the part, the opposite sides that are not functional may be needed to rest /fixture the part.
Likely one would map the part and mark it for + whatever along a to-be ground surface so as to take correctly from each surface...but even that is difficult to figure.
Dressing the wheel at an angle may be best, rather than trying to tilt at an angle.
A part hardened to around 45 to 48 would be very wear-free and still (possibly) scrapeable so the part can ground to about .001 -.002 and scraped the rest for the last fitting perfectly.

 

[Bakafish]

Looking at that glue line, I'm just curious, how do you account for the growth you are seeing? If anything the strips should have had some wear, if you were using the correct glue and technique to replicate the factory glue gap you should be low not high, right? I'm asking because it isn't clear to me how you are sure that it grew, and it deserves some scrutiny before undertaking irreversible material removal.

We only see one side of the cross slide assembly in your photos and you didn't explain how you got to that +-.003 number, maybe it is obvious to others more familiar with this arrangement, but I start asking questions when the solution being proposed seems like the hardest possible thing to attempt.

It's probably too late, but the time to do corrections was likely before you glued the strips back on as you could have determined the glue line thickness (set by precision filler material or the factory specifications) and if it needed to be bigger than the factory for whatever reason to scrape the soft beds of the strips prior to gluing to hit your numbers. If it were me and I suspected I hadn't got this right I'd be tempted to cook those strips off and do it again.

 

[stephen thomas]

If you are working on the sides of the DT, scrape. It will be time, efficiency, and quality ahead.

Your diagrammed set up does not seem to include the flats?
Could work, but a lot of room for mayhem making it worse, especially introducing twist from side to side.

I have "scraped" precision hard steel surfaces with an angle grinder or die grinder and a small cup wheel.
You "stab" the spots. Do not swing the tool in a "grinding " motion. Adjust the angle you are holding it, dress the wheel as necessary, and get in the mindset of steadily tickling off the high spots. Be extra cleanly when re-spotting between passes. The stray grit is especially hard on CI spotting tools. As is the hardened steel substrate if you don't deburr it with a flat hard stone.

You need to establish the flats first & check and control for twist. Say over dowels with a level. IOW, map them, make a plan, then work both sides toward that flat goal. Technically, they should be co-planar and that will be the easiest way to keep track of everything. I prefer to scrape the topslide first, since it is an easy check & spotting to a surface plate, to use for the gage for bearing across the bottom ways. It is still a good idea to check the bottom ways with other means, periodically. Especially since the ways are longer than the topslide bearing area.

smt

 

[triumph406]

:

View attachment 400244

The problem I see is that the carriages are too close together in both X and Y If it was me I'd have them spaced further apart in both X and Y directions. And then use carriages with the highest preload you can get.

And of course you will have to be absolutely square with the X axis, otherwise parts you face will not be absolutely flat across the face.

Good luck, I think if your carefull it could well work out as you desire.

Sending it to Scaffer would likely be faster and cheaper. But then what's the fun in that!?

 

[Chips Everywhere]

I will try to answer some questions.

My plan was to only grind the hardened strips, if I could grind the dovetail sides that would be a big plus, but I can scrape them in so the chances of me screwing up are significantly less.

I am using the dovetail top as a reference, its a ground surface and that is what I used to get my 0.003" measurement. I checked on 2 other Hardinge lathes and this surface is parallel to the dovetail hardened strips. On the machine I glued the strips on, both strips start at 0 on the same end and on the other end, one strip is +0.001" and the other is -0.002", hence the 0.003" that I have to remove to address the issue.

@michiganbuck

I agree that trying to grind the dovetail straight may be out of the question, but if I could grind the hardened strips then that would put me in a comfortable spot to address the dovetail sides through scraping. I have to till the grinder to clear the spindle nut. I can tilt the grinder and keep the compound perpendicular to the bed for easy of measuring the Z travel, though its not pictured on this one:



@Bakafish

After talking to several rebuilders, including one who did some grinding on my other HLVH, the common practice is to glue the strips back on and then grind them. Trying to glue the strips back on parallel is quite difficult, I'm shocked that I even got as close to 0.003". The strips definitely have wear, and I am going to end up sitting lower than factory when this is all done. The cross feed nut will need to be refitted to compensate for the material removal of the cross slide. I am using the top dovetail surface as a reference to measure my low and high spots along the travel.

@triumph406

I have the slide fixture made from a previous project. I can modify it to accept the square block that the compound mounts on to, it does have higher preload trucks and the rails are H from what I recall.

@stephen thomas

I am aware of this hard scraping method, but I wouldn't want to try it for the first time and experiment on this.

 

[Bakafish]

Thanks for clarifying several points, the way you phrased it led me to believe you were trying to lower the overall height, but it sounds like you are just trying to get the surfaces parallel (to a non-bearing surface that we hope is an accurate reference.)

I am aware of this hard scraping method, but I wouldn't want to try it for the first time and experiment on this.

As straightforward as your proposed method sounds, I think there is a lot more chances of doing unintentional damage than what he was proposing. You are basically building your own miniature surface grinder using an aluminum bed, linear rails, and an awkwardly mounted tool post grinder. As I'm a novice I appreciate that the constrained motion offers some reassurance over free hand work, but being too clever and over-engineering a solution usually results in a learning experience rather than success. Grinding seems to have a lot of unintuitive complexity that argues against this working how you would hope, and your sled isn't as coupled to the work as would seem prudent.

Try the simplest and most reversible solutions first. A few minutes practicing the spot grinding technique on a test strip will tell you a lot about how feasible it may be, it seems plausible though, and potentially slow enough that any damage can be mitigated. With the right abrasive selection the amount of material removal can be held in check so that, worst case, you aren't really much worse off than currently. The spindle grinder on the other hand seems like it could ruin things in one bad pass.

The other thing that concerns me is the fear of going backwards often leads to false choice limitations. You did a lot of work/research to get those strips back on, but you painted yourself into a $15k corner. If it were me, I'd think about cooking off the strips and replacing them with Turcite or some other hand scrapable surface. Nothing stopping the next guy from paying the money to reinstall the strips if they are so inclined. From what I've read in other posts, the hardened strips were a late edition that seem unnecessary and problematic anyway.

 

[Chips Everywhere]

Thanks for clarifying several points, the way you phrased it led me to believe you were trying to lower the overall height, but it sounds like you are just trying to get the surfaces parallel (to a non-bearing surface that we hope is an accurate reference.)



As straightforward as your proposed method sounds, I think there is a lot more chances of doing unintentional damage than what he was proposing. You are basically building your own miniature surface grinder using an aluminum bed, linear rails, and an awkwardly mounted tool post grinder. As I'm a novice I appreciate that the constrained motion offers some reassurance over free hand work, but being too clever and over-engineering a solution usually results in a learning experience rather than success. Grinding seems to have a lot of unintuitive complexity that argues against this working how you would hope, and your sled isn't as coupled to the work as would seem prudent.

Try the simplest and most reversible solutions first. A few minutes practicing the spot grinding technique on a test strip will tell you a lot about how feasible it may be, it seems plausible though, and potentially slow enough that any damage can be mitigated. With the right abrasive selection the amount of material removal can be held in check so that, worst case, you aren't really much worse off than currently. The spindle grinder on the other hand seems like it could ruin things in one bad pass.

The other thing that concerns me is the fear of going backwards often leads to false choice limitations. You did a lot of work/research to get those strips back on, but you painted yourself into a $15k corner. If it were me, I'd think about cooking off the strips and replacing them with Turcite or some other hand scrapable surface. Nothing stopping the next guy from paying the money to reinstall the strips if they are so inclined. From what I've read in other posts, the hardened strips were a late edition that seem unnecessary and problematic anyway.

I appreciate your input on the grinding fixture, I agree that it may not be as simple and straight forward as it may seem or wish this that it was. These strips prove to be more problematic than useful when it comes to rebuilding.

I gave it a fair amount of thought before gluing the strips back on, explored other options, even thought about replacing them with turcite, but I don't think that would be the solution for this machine. The well known rebuilders that service these machines keep the strips and regrind them, in my research I've yet to see someone use a substitute for the strips. They obviously have all the tooling to get the job done correctly and efficiently, unlike me . I'll end up sending out the apron.

I am quite curious what kind of work can be produced with a bolt on grinder like this, it maybe completely unsatisfactory, but down the road I'll probably mock up a grinder and find out. Thanks everyone for your input.

 

[Bakafish]

I really hope it works, and you document it along the way. The idea itself doesn't sound bad, my worry is that some of the people I most respect in this field seem to have a learned trepidation with the arcane nature of grinding. When people with a lot of experience regularly have trouble with conventional work on a high end machine, with a mag-chuck and coolant, best case in other words, it dissuades me from considering ad hoc applications. But I'd love to be wrong about that, if you pull this off it opens a lot of possibilities.

On the grinding method, I think it would be an advantage to find a way to mount the rail assembly to the 'reference surface' you chose, the top of the dovetail. I know it's a precision machine and the ways should have a tight relationship to the surfaces you are trying to level, but the closer everything is to the cut, the less chance for impercision. Ideally (in my mind anyway) you mount a known parallel plate and rails to that surface, and then don't move it for both cuts, which means you only rotate the grinding head 180º, the carriage follows the same profile for both sides.

 

[stephen thomas]

Bakafish makes the point that bothers me about imagining that grinding with that set up will improve geometry over existing to attain, say any better than .001 everywhere at some risk. To me the issues would be 1.) twist from moving the grinder set up between sides 2.) the flat ways do not actually have to be co-planar so long as the long axis of each is parallel - but is convenient if they are. IOW same height & flat to each other/no tilt. 3.) without coolant will you have enough patience to grind without creating flatness or surface quality issues from differential warming of the materials? 4.) will you have enough patience so that differential warming does not pop either of the strips loose?

Differential warming is the example Grand Rapids used to make in ads for the reason why they finish scraped, rather than ground, the ways on their grinders. As did other machine tool builders.

It does sound like the divergence you relate is measured over small, isolated areas. So if you can contrive the geometry of the set up, if could be successful. I think you asked about wheel spec - i'd prefer a very thin rim cup wheel. But if you use your set up, don't use harder than an I wheel, porous, and dress it fast (coarse). 46 grit.

smt

 

[michiganbuck]

Measuring can be tricky because one has to know/check that the baseline on both sides of the male dovetail is parallel and the same height.
This is easy if the opposite side is true but tricky if the opposite is just a casting and the base is ground (or machined) in two setups, perhaps due to the side of the grinder(machine).
*The the Ops diagram looks like the part will be true all around (all sides). That is a big plus for easy neasuring.

Measuring and Designing Dovetail Slides

Dovetail slides that must be machined accurately to a given width are commonly gaged by using a gage pin of known diameter and measured as indicated by the dimensions x and y

www.engineersedge.com

 

[4GSR]

You need your linear rails to straddle the dovetail cross slide and have a cross rail above with your grinding spindle mounted to it. That way it stays fixed and aligned to the dovetail cross slide when you set up to grind each hardened way strip.