Drilling holes

[ksouers]

OK, I'm stumped.

This morning I'm drilling holes in the cylinder blocks for the Elmer's Comber engine. I need the hole exactly centered in the back side of the block for the bearing shaft. The shaft is to be silver soldered into the block.

I have the end mill holder mounted and find the edges of the block and move to my mark and lock the axes. I also have a habit of double checking using my tapping center and a 6 inch scale. It looks fine.

I change out the end mill holder for the drill chuck, mount the drill bit and proceed to drill the hole a bit undersize for reaming. By visual inspection the hole looks off center. I measure from the edge of the block to the edge of the hole on both sides the axis with a caliper and find them different by almost .040

I remount the end mill holder and edge finder to recheck my edge measurements. It's spot on. So I double check the measurements of my parts. Yep, they're OK too.

I remount the drill chuck and drill another hole. Same thing, it's off by .040 WTF??
A light bulb goes off. I chuck the edge finder in the drill chuck and recheck my edges. Y axis is spot on, X is off by .010; should have been .100 but was .090. OK, so the problem seems to be the drill chuck.

I finished drilling the holes using the drill chuck to find the edges. When I went to ream the holes, the reamer wasn't centered over the holes either. Since the reamer is longer my intuition suggests I may be out of tram, or am I all wrong about that?

It's the first of the month, so I had just checked and trammed the mill and squared the vise last weekend. I'll recheck those in a few minutes just because I want to be anal. I'm using a Starrett edge finder, so I doubt that's part of the problem. The end mill holder is a Bison and I'm using the crappy drill chuck that came with the mill. The mill has an R8 taper. I'm using a 3-axis DRO, and I've been intentionally trying to find errors with it but so far have come up empty handed there. It doesn't "feel" right to me, but so far all I've been able to prove is that it's on the money.

Why would there be a difference of .010 between the end mill holder and drill chuck? And why only in one axis?
And why would a .010 error result in .040 difference? I would also expect any runout in the chuck to make a circle around the mark, not translate to one side.

Any ideas? Anyone?

 

[Mcgyver]

at the risk of being too simplistic, could it be backlash? That you picked up the edge, cracked the table over then reversed direction, hence put backlash into the equation. that might account for some of the .040 thou, but even the seems a lot of backlash

the following also won't account for the 40 thou, but you'd be better with the edge finder held in a collet, the clearance between the edge finding and end mill holder is a source of some error....by itself maybe minimal but its accumulated error that'll bite. all part of the debugging.

how are you starting the hole? best to lock the tables and start with a spot drill. even if the chuck has some run out, this rigid tool will form a cone pretty much perfectly in line with the spindle bearings. drill chuck run out, unless extreme, shouldn't matter much if the hole is accurately started with spot drill.

When i really matters, even if you use a reamer to bring about the final size, its better to true up the hole with a boring head before reaming. drills do not necessarily cut straight and as you noticed, the longer shank gives them enough flexibility to follow the hole. in other words they won't perfectly straighten a crooked hole - this by design btw, otherwise runnout of the drill chuck would make their accurate operation impossible

anyway, most of those are finer points presented for prosperity and does not explain the .040 thou, for that there is something more fundamental going on.

 

[ksouers]

Thanks for the reply and suggestions, Mcgyver.

Many decades ago I was taught to always "crank up the numbers". I don't know if they still teach that, but it's a habit that has persisted with me. My datum is always the back left corner. When coming up to a position, whenever possible, I always approach from the same direction. When over-shooting I back up a lot and try it again. This time was no different, when finding the edge and coming to the mark the table was always moved left and back, then locked. I don't think backlash was the problem, the table didn't move when I changed tooling.

One reason I like the Starrett edge finder is that it fits perfectly in the end mill holder, I don't have to tighten down on it. Friction is all it needs to stay put. I have a couple cheapies that aren't so great, one that won't even fit in the end mill holder.

I spotted the hole with a small drill (I think #40?), anyway, not a center drill. Just a touch. I had noticed on previous occasions that small drills turned truer than larger ones. By the way, final size on the hole was 3/16 and .225 deep. I don't think I could have bored that Position was more important than roundness anyway, I just needed a close fit on the shaft for soldering. I knew the reamer would bend and follow the hole, what I didn't expect was that it would be out of line with the hole that was just drilled.

I rechecked the machine, I guess it wasn't too bad but not what I expected. The vise was out by .0003 over 2 inches and tram was out by .0015 over 10 inches (I was expecting about .0005). So, I just got done straightening that all out again. I got lucky, tram is now dead nuts!! Won't stay that way for long. Same for the vise.

I guess I'll try it again and see if I come up with the same error.

 

[Mcgyver]

don't use a small drill, won't work, can wander all over before it bites. centre drills are for drilling centre holes for turning in the lathe - it'll work better than a small drill, but not as well as a spot drill. Use a spot drill for starting holes in the mill, the lathe, the drill press. You may not have picked up on it, but i said in the previous post that when using a spot drill, run out in the chuck won't matter, it'll make a neat cone, the angle being the same as a drill bit's (unlike a centre drill) so the bit will start without chatter perfectly in the centre of the cone

 

[ksouers]

Thanks for all the advice, Mcgyver. It is greatly appreciated.

I went out to solder up the good blocks and decided to try to salvage the bad ones. No joy there.

I also thought I would see if I could repeat the previous error. I have a tapered center I use for tapping, it's just an old end mill that's been ground to a point. I put that in the tool holder, ran the spindle and touched a point on a block of brass, just enough to make a mark.

Then I swapped out for the drill chuck and used the center again. It was over the previous mark, or at least close enough I couldn't tell. So I swapped out the center for the same drill bit I used before and touched off. It was in the same spot.

I also went edge hunting. The drill chuck found it .011 before the tool holder did. Now I don't know which one to believe, though I'm suspecting runout in the drill chuck caused the early detection.

I guess I've got some more experimentin' to do.

 

[Airhead]

This sounds like the problem I had with my mini-mill (Grizzly #8689, which I think what is being called the X2 now a days). What I found was that the head was assembled with the spindle and dovetail out of line with each other. The result is that when you tram the spindle to the table, the dovetail column is no longer perpendicular to it. This causes the different X positions for different heights of the head.

Gook luck,
Rick

 

[ksouers]

Rick,
Very interesting.
Thanks. Yes, that does sound like it might be what I've experienced, different geometries for the two parts of the machine. It might also explain why I didn't notice the problem until using larger (longer) drills.

How did you determine that was the problem? How much was it out of line? What was done about it? Can it be adjusted out with the gibs?

I got mine from Harbor Freight about 4 years ago, so I doubt they'll replace it.

 

[ksouers]

Well, it looks like Rick may have identified the problem.
It looks like the spindle is not parallel with the column, but rather at a slight angle along the X axis.

Here's what I did:

I turned an abrupt taper to a point on a piece of drill rod (silver steel) that is about 9 inches long. I clamped a piece of flat brass to the table, mounted the tapping center in the mill holder. With the mill head all the way down to the table, running the spindle I popped a small mark on the brass.

Without changing anything else, I ran the head all the way to the top and swapped out the short center for the long drill rod. Here's what I found.

I haven't measured that error yet. I'm still trying to figure out what to do about it.

One thing I think I'm going to try: The spindle box and dovetail are separate pieces. I'm thinking (hoping?) there may be enough slop that I can wiggle the two pieces into alignment. Just how I'm going to do that, I don't know. The bolts are inside the dovetail, so there isn't going to be any "adjust it till it's right then tighten up".

Of course, any suggestions or comments are welcome. Please.

 

[Mcgyver]

way to nail it Rick. So far as suggestions go, i guess I'd get rocks thrown at me if i said buy a bridgeport ?? ;D ;D

I'm not familiar with that machine, unless it has a tilting head i don't know that its readily adjustable, seems like it a long way out . Whenever you encounter a machine tool alignment issue, the first thing to do is a lot of carefully checking. Don't remove a sliver of metal until with accurate tools and indicators you're carefully quantified the problem and no exactly where it resides. This will often involves disassembly and some ingenious set ups with indicator, square and reference flats to show where the problem is. The benefit of so thoroughly checking the problem out is that it occasionally occurs the something else was the problem or that there is an easy fix is identified. Given that it looks so far out, lets hope that is the case!

if it turns out the spindle bearing bore is in fact that much out of alignment with the z dovetail, what I'd do, because I'm demented this way, is to scrape it properly into alignment. Its not a quick job, and you'll have to learn some new skills but its the right way to create machine bearing surfaces that are straight, square and mate properly. I'm surprise more guys aren't scraping these mills; there's no free lunch and there are reasons they're sell for a fraction of a Bridgeport....which is both good and bad. The good is that the cost is reasonable, bad in that the bearing surfaces are less than perfect or at least there are a lot anecdotes out there to that effect. The good news on the bad is that the situation can be rectified if you have the desire and patience.

 

[ksouers]

So far as suggestions go, i guess I'd get rocks thrown at me if i said buy a bridgeport ?? ;D ;D

OK. Problem solved!

Well, that was easy :big:

No rocks thrown here! I wouldn't mind having a Bridgeport at all. But I really don't have room for it at the moment. And I don't think I can get it past the wife, somehow I think she would notice. Although when this problem reared it's ugly head, the thought did cross my mind. At least maybe upgrading to the SX3. I have half-heartedly been keeping an eye out for a Bridgie for a couple months. There was a J-head with single phase 3/4 HP for sale nearby for around $2100USD, but I think it's gone now.

There's no tilting head. In fact, there are no adjustments on the head itself. Tramming is done by tilting the entire column left or right. If you look at the first picture, the wide shot, at the bottom of the column is a circle. In the center of the circle is a large bolt, and a large nut on the back. So, you have to loosen the nut then tilt the column. However, when it's loose it allows the head/column to also tilt forward slightly, throwing Y-axis out of tram. Then when you tighten up the nut it pulls the head towards the left about .0003-.0007 depending on the mood it's in. So you have to anticipate that by over-tramming just a bit. Sometimes I get lucky, but usually I miss it by about .0005.

The female dovetail is just a cast iron spacer that the head bolts to. There are four bolts holding it on. Here's a shot of a replacement part from Little Machine Shop.

Spindle Mount

I'm thinking there might be enough slop in these bolts that I can twist the head around till the spindle is parallel with the doves. I think I'll probably end up with a set up similar to what Bogs did for his spindexer. Once I get the spindle parallel to the X-axis I can wedge a piece of drill rod in the dove to check it's alignment.

I thought about scraping the bearing surfaces, very briefly. 1) I don't have the skill. 2) The machine is only worth a couple hundred dollars, I can spend a month or six scraping and measuring and fitting and tuning and tweaking and I'll end up with a machine still only worth a couple hundred dollars. I figure I can spend a lot of time making the thing into a Bridgeport. Or can just buy a Bridgeport.

I want to get the thing in as best shape I can, but don't really want to over-do it either. It's a great little mill, considering what it is. I've been quite happy with it's performance.

 

[ksouers]

Oh, I almost forgot. I measured the two marks with a caliper. Sorry, best I could do. I came up with about .052, plus or minus a couple.

 

[tel]

Might 'ave been easier to just measure it with the table dial?

 

[BobWarfield]

I agree, the four bolts are your salvation here. Here is a Grizzly manual as well for your mill with an exploded view:

http://www.grizzly.com/images/manuals/g8689_m.pdf

A little trigonometry and you can figure out how much things are off and how much play is needed to put it right. If there is enough clearance now, you are in fat city. If not, you'll have to enlarge the holes a bit.

Have you given thought to how you'll properly align things before torquing down the bolts again?

Typically I would think to use a cylindrical square on the table being swept by a DTI in the spindle while you move the head up and down. Be sure to check the Y error while you're at it. May as well shim if necessary, though it doesn't look like there is much of that going on.

Looks fixable to me!

Cheers,

BW

 

[ksouers]

Might 'ave been easier to just measure it with the table dial?

Might 'ave

Didn't think about ???

Though I had in my mind I didn't want to change anything till I had a better idea of what to do about it.

 

[ksouers]

I agree, the four bolts are your salvation here. Here is a Grizzly manual as well for your mill with an exploded view:

http://www.grizzly.com/images/manuals/g8689_m.pdf

A little trigonometry and you can figure out how much things are off and how much play is needed to put it right. If there is enough clearance now, you are in fat city. If not, you'll have to enlarge the holes a bit.

Have you given thought to how you'll properly align things before torquing down the bolts again?

Typically I would think to use a cylindrical square on the table being swept by a DTI in the spindle while you move the head up and down. Be sure to check the Y error while you're at it. May as well shim if necessary, though it doesn't look like there is much of that going on.

Looks fixable to me!

Cheers,

BW

Thanks Bob.

I know that one of the things I have to do is make sure it's square with the dove before I put it back on the machine. There is no way to adjust it while it's on.

I'll have to make sure the column is square with the table first, then to check the head I guess just do again what I did this morning.

I'm not sure what you mean by a "cylindrical square". If you mean a cylinder where the end is absolutely square, then that is a possibility.

 

[ksouers]

Ok, here we go again.

Tel, I did the table translation. It came up with .052 as well. 'Course I was using the same uncalibrated eyeball to line up the points as I did with the caliper.

So, I started disassembling the head. I said to myself, "Self, it's already out of whack. All you can do is make it more out of whacker. Give it a roll."

I scribed a couple witness marks on both parts, top and bottom, so I could see what I was doing and loosened up the bolts. The parts were stuck, so a smack with a dead blow loosened them up. I started fiddling with them. Twisted around a bit. Tightened up and put the head back on the column.

This is what I got on the third try.

That looks pretty darn close to me! The original two pop marks are to the right.

I've still got to put the rest of the machine together. I'll tram it again and check out things much closer and see what I get.

But, after that attempt? I'M GOING TO THE CASINO!!!!

 

[Mcgyver]

good for you Kevin! that looks very close

best if you can mount an indicator on the spindle mount and by moving the head up and down, compare readings against a quality square (the good ones will be 2 tenths over the length of the blade) very lightly clamped to the table. drill rod is not necessarily straight (or round believe it or not) and however it is held in the spindle introduces error which is amplified over the distance. you are obviously putting a lot of effort into doing an accurate tram so you want the spindle/column alignment to be able to keep up.

you mention some challenges on everything moving a bit when you clamp up the column to base. to a degree this is to be expected - the layperson thinks of metal as a solid immovable, but we know every force that is applied causes it to change its shape. Having said that, a lot of movement like this is indicative of the two surfaces not being flat - when they are not flat and they are squeezed together it distorts one or both. This is a situation where scraping is a fantastic and easy technique to get each of the surfaces very flat so there is minimal movement as you tighten down. you mentioned you don't have the skill for scraping, baloney is say, you just don't know how to....yet You have plenty enough mechanical aptitude to figure this problem out and scraping is no harder than filing or hand tapping, so not this evening, but at some point consider it.

As you are in diagnostic mode, order some prussian blue. comes in a little tube, like oil paint which is basically what it is except is doesn't dry. spread some thin enough on a surface place so you cans still see the features of the plate (you do need a SP) using a vynil glove. touch one of the mating parts to the plate, gently move at back a forth a few times, a 1/2 an inch or so. lift it up and have a look. if there's lots of blue all over, the part is flat - thats how its check to a very high level of accuracy. check both. bet they're not even all over. i also bet once you do this if it will occur, hmmmm, if shaved a little metal where its blue, and kept repeating, it could get it perfectly into the same plane and the darn thing would clamp up accurately. now you know how to scrape the only complication is to make sure it ends up dead nuts on square to the y plane (you clamp a square to it and keep checking with an indicator as you scrape)

 

[Airhead]

Sorry I didn't get back in time to be more help! It looks like you got it figured just fine though.

Yeah, you just have to keep trying until you get it right. I went so far as to clamp a couple of pieces of scrap to one part so that they hung over the other one. Then I tapped holes for 'push bolts', kinda like the bolts to adjust the tailstock on a lathe. It still took several tries to get it right. Mine just had tapped holes in one casting and clearance holes in the other. I don't know how they allign them when they put them together. Well, maybe they don't!

Mcgyver's right about drill rod. In fact this was when I found out just how crooked a piece of 1/2" drill rod can be. Amazing!

Glad you got it fixed


Rick

 

[ksouers]

Rick,
Thanks for the suggestion. You were right on the money.

I don't know about being "fixed", but from the rough estimation it looks a lot closer to what it should be. I'll find out how bad it really is once I get a square and DTI on it.

 

[ksouers]

Mcgyver,

Thanks. Sometimes it pays to take a gamble. I'll find out how close I got when I finally get to measuring things. Right now my only concern was being able to complete my parts for the Team Build. I think I'm close enough for that, but I'll still take some measurements to be sure.

I hope I didn't give the impression I was against scraping. I'm not. I just didn't think the effort would be "cost effective" on this machine, given the effort involved and the time the machine would be out of service. Also, I'm quite afraid I would just totally screw it up!!

I'm familiar with the basics and would like to try it some time. I don't have a surface plate, so I may give the 3-plate method a try. That should give me something that I can work with.

That actual bearing surfaces of the mill don't appear to be that bad, but really the only way to tell is testing them with a registered surface. What is pretty bad are the actual dovetails themselves. They are wavy and not quite parallel. They would benefit the most, but would also be the hardest to do.

Well, plenty of time for that later.

Thanks again for the suggestions, and especially the morale support.