The horror under the chuck

[ballen]

Hi Buck,

Because the geometry of the machine is very close to perfect....

I agree to that, the blue in shows enough hits IMHO to be filled with a coat of oil or grease and so not be a problem for a good number of years .. and then a next chuck top grind and so another under cleaning may be due.

I agree with this too, which is why I am planning to follow Richard's advice and NOT grind the table top. But I will stone it!

I normally figure 8 ore straight (fine side) hone the chuck sides for bugs and bruises and with an inward hone stroke hone a small bevel to all the edges, with taking most care to the pad facing edge corners with the stroke gong away from pad direction. hone direction important to remove any to not make a burr.

Yes, I'll pay particular attention to the edges and corners. But the vise sits a few mm inside the boundary of the pad.

.. the back bump rail is ground in place with a dish edge wheel to make square so I indicate square the chuck front edge because you can't grind that rail in place.

It looks to me as if neither the back or side edge of the chuck have ever been ground true. Both have gray paint on them, visible in the photos above. Is it right that both the edge of the chuck and the edge of the bump rail should be kissed/dusted in this?

Buck, I have an unrelated question for you or another J&S 540 owner. With the table off, I wanted to check that the lubrication of the ways was working well. So I turned on the hydraulic pump. But there was no lubrication coming from the way oilers. Is this to be expected? Will the table lubrication only flow if I turn on the hydraulic feed? Is it safe to turn on the hydraulic feed without the table attached on top? The hydraulic rams will move to the end of the travel, but will that damage them if they are not connected to the table/reversing valve?

Cheers,
Bruce

 

[michiganbuck]

Long time and brand new machine so never had the table off. Back bump rail kiss after install and chuck top grind done. back rail over hang the chuck top perhaps 2mm and kiss with dish edge wheel. Rail lip perhaps 1 or 2mm off chuck top. Painted sides of chuck are fine I would not touch them. A chuck such as a Brown & Sharpe 4x8" or 5x10" have a solid top edge face and bottom rail face that are ground true, they can be stood on side edge (on your magnetic chuck) to become a magnetic square (right angle)for grinding parts square off the part side.

 

[ballen]

Well I've got the chuck back on. Found a German equivalent of LP-3 (same contents, according to the MSDS) and used that to coat both the bottom of the chuck and the top of the mounting pad..

The process was slowed down when I mapped out the top of the chuck and found out that one corner was almost 300 microns = 0.30mm = 0.012" high. Thinking I had really screwed up, I first tried to slide a feeler gauge in between chuck and mounting pad to see where it wasn't seated. But it was seated. Then I took the chuck back off the table and measured it on the surface plate. Indeed, the chuck bottom and the chuck top were very far from parallel. Perhaps what happened was that over the years the rust growing underneath forced the chuck upwards, but the people using the grinder kept on dusting the chuck on top, oblivious to why it was moving upwards.

Here's the map of the top, in units of 0.01 mm = 0.0004" (so for example "10" means 0.1mm = 0.004"). The high corner is the front left one.

I've finished putting the coolant system back together, so tomorrow I will start the grinding. I'll go very slowly at first, working on the high corner, so that's a chance to get some practice on a small area first.

Cheers,
Bruce

 

[Richard King]

Just curious what the square cut marks are on the chuck surface is? A shadow or did you have it laying on scraped plate. Sort of goes bottom front left to back rear right...looks like scrape marks? Who knows about the reason it happens. a person could debate and try to figure out why and go insane thinking about it. how about did you measure the table top with no chuck? you could leave the indicator about 1/4" away from the top and slide a 1/4" gage block under indicator at say 10 places and see if it is messed up. The chuck maybe the problem. it would be interesting to know, but I still would not grind the table top, but if it is messed up you may have to think about it. Rich

 

[ballen]

Hi Richard,

Thanks for checking in.

Just curious what the square cut marks are on the chuck surface is?

Several weeks ago, I put the bottom of the chuck into an Evaporust bath to remove the rust. When I took it out of the bath, I put it (upside down) onto a ribbed rubber mat, the same kind which you find on the top of tool chests. I scrubbed the bottom with a brass brush, then put the chuck back into the Evaporust bath, and repeated. But I forgot to dry off the rubber mat, which was wet with Evaporust. So some Evaporust got onto the top of the chuck, and discolored it.

The bottom line: you are seeing the image of the ribbing on the rubber mat, transferred to the top of the chuck via the miracle of chemistry.

Who knows about the reason it happens. A person could debate and try to figure out why and go insane thinking about it.

Sure, but if I couldn't see any possible explanation for this, I would do still more investigation before proceeding. The point is that the layer of rust I removed was at least a few tenths of a millimeter thick. It could have stretched the mounting bolts of the chuck, which indeed were VERY tight.

How about did you measure the table top with no chuck?

Surface plate. See the third and fourth photos of this post.

You could leave the indicator about 1/4" away from the top and slide a 1/4" gage block under indicator at say 10 places and see if it is messed up.

It's easier than that, because the chuck is so far out of parallel (between top and bottom)! I can just use a digital vernier caliper as a depth gauge to measure the distance from the chuck top to the top of the mounting pad. It varies by almost 0.30mm = 0.012" as I move around the chuck.

But I'll do it your way too, it will only take a few minutes, and is a good confidence-builder.

The chuck maybe the problem. it would be interesting to know, but I still would not grind the table top, but if it is messed up you may have to think about it.

I agree that right now it does not make sense to grind the table top, at least not yet. The print of the top of the pad from that earlier post is pretty reassuring. I'm happy that I checked that, and the geometry of the ways, because now I'm pretty confident where the problem lies: the chuck top/bottom are far from parallel.

Does it make sense to you?

Cheers,
Bruce

PS: once I am done grinding the chuck, I'll verify that it's a constant distance from the wheelhead over the entire surface. Then I'll repeat the geometry measurement I did before. I'll put my surface plate on top of the chuck, and map out it's surface with a micron-precision probe, and see if it's planar. If so, I am good to go.

 

[Richard King]

Hi Richard,

Thanks for checking in.



Several weeks ago, I put the bottom of the chuck into an Evaporust bath to remove the rust. When I took it out of the bath, I put it (upside down) onto a ribbed rubber mat, the same kind which you find on the top of tool chests. I scrubbed the bottom with a brass brush, then put the chuck back into the Evaporust bath, and repeated. But I forgot to dry off the rubber mat, which was wet with Evaporust. So some Evaporust got onto the top of the chuck, and discolored it.

The bottom line: you are seeing the image of the ribbing on the rubber mat, transferred to the top of the chuck via the miracle of chemistry.



Sure, but if I couldn't see any possible explanation for this, I would do still more investigation before proceeding. The point is that the layer of rust I removed was at least a few tenths of a millimeter thick. It could have stretched the mounting bolts of the chuck, which indeed were VERY tight.



Surface plate. See the third and fourth photos of this post.



It's easier than that, because the chuck is so far out of parallel (between top and bottom)! I can just use a digital vernier caliper as a depth gauge to measure the distance from the chuck top to the top of the mounting pad. It varies by almost 0.30mm = 0.012" as I move around the chuck.

But I'll do it your way too, it will only take a few minutes, and is a good confidence-builder.



I agree that right now it does not make sense to grind the table top, at least not yet. The print of the top of the pad from that earlier post is pretty reassuring. I'm happy that I checked that, and the geometry of the ways, because now I'm pretty confident where the problem lies: the chuck top/bottom are far from parallel.

Does it make sense to you?

Cheers,
Bruce

PS: once I am done grinding the chuck, I'll verify that it's a constant distance from the wheelhead over the entire surface. Then I'll repeat the geometry measurement I did before. I'll put my surface plate on top of the chuck, and map out it's surface with a micron-precision probe, and see if it's planar. If so, I am good to go.

I was talking about removing the chuck and mounting the mag base or clamping an indicator rod on wheel guard and measuring the top of the table top directly. from the print we know the table top is high around the T-slot. You may want to use an old fashion dial indicator as sometimes those electronic one are to sensitive. I prefer using a .0005" B&S Best test then a .0001" Interapid. I have had 2 electronic Indicators and only used them scraping Jig Bores. Thanks for this thread, it's very informative for us readers. Rich

 

[ballen]

Hi Richard,

I was talking about removing the chuck and mounting the mag base or clamping an indicator rod on wheel guard and measuring the top of the table top directly. from the print we know the table top is high around the T-slot.

Are we looking at the same print, this one? I don't see evidence that it's too high around the T-slot, only that the parts of table which did not corrode away are still planar.

You might be referring to the print of the bottom of the chuck, which was taken off a surface plate, and looks high around the T-slot. But I think that's because the area around the T-slot was not corroded as much as the other areas beneath the chuck.

You may want to use an old fashion dial indicator as sometimes those electronic one are to sensitive.

I already did this, it's described in an earlier post in this thread. The non-corroded parts of the table were all a constant distance to the wheelhead. The corroded parts were a few hundreds of a mm low. Here is the quote (post #15 in this thread): "First step in assessing the condition of the parts: I've put a 0.01mm = 0.0004" indicator on the wheelhead and run the table back and forth and in/out. The high points are right on the money over the entire surface. The low points are running into the pits and valleys. These look worse than they are, about 0.01 - 0.02mm low. So I think the table and ways are in factory condition, apart from the rust pitting."

Thanks for this thread, it's very informative for us readers.

I appreciate having someone with your expertise and experience, and others with similar backgrounds, commenting on this for a newbie like me. This group is the only opportunity that I have to get such good advice. Without it I'd be lost.

Cheers,
Bruce

 

[ballen]

To anyone who is still following this saga:

I didn't do any grinding today, but did spend a couple of hours getting the coolant system cleaned out and operating.

The total capacity is about 66 liters. I first filled the coolant tank with 63 liters of clean water then ran the pump and flushed out and replaced the water, until I got rid of the grit and dust that had accumulated during 30 years of storage. Then I added in the 2.5 liters of coolant (synthetic, recommended concentration 4%) and ran the system some more with the feed pipe spraying directly into the drain/intake, to mix the water and coolant thoroughly.

Then I experimented with the flow rate and the drainage. If I open the valve more than about halfway, the catch basin fills up almost completely, and the drain can no longer keep up. A bit less, and here's what it looks like:

I'd be grateful for some suggestions about the volume of the coolant, and the placement of the stream relative to the grinding wheel.

The bottom of the copper pipe is a bit worn away. Is it worth trying to reform that into a clean horizontal slot, or is that just wasted effort?

Thanks in advance for the advice. I know this is all beginner stuff, but the pointers and tips I get here help me a lot.

Cheers,
Bruce

 

[ballen]

Well the time has come to end this thread, at least for now. I ground the chuck as advised in this thread.

Downfeed increments were 0.004mm = 0.00016" Total amount removed was about 0.30mm = 0.012", so approximately 75 full passes over the chuck. I was VERY happy that the J&S 540APR has auto-crossfeed and auto-downfeed capability. It took about 6 minutes per pass, so about 7 1/2 hours to grind this. I watched this closely at the beginning, and by the end was just checking up on it every 45 minutes. I suspect that I could have taken passes that were at least ten times as deep without doing damage, but was conservative and followed the advice I was given here.

Here are a handful of photos taken in the course of the grinding. At the start, fussing with feeds and flows

At around the halfway point:

I took a pause further on for a family outing, so shut everything down:

Finished. The surface is so flawless that just wiping it off with a paper towel and spraying a bit of oil on it reduces the quality. I like grinding!

I'm now confident that if needed I could now grind the bottom of the chuck without doing any damage. The chuck never got above room temperature during this process. In fact there were only a few visible sparks.

The wheel showed no tendency to clog during this. In fact I removed the first 0.25mm of material before I even dressed the wheel, and it didn't make any difference that I could see to how the wheel cut. I am sure that without the lubrication/coolant that would not have been the case. The wheel was made by Dorfner Schleifmittelwerk in 1986, and is a 200 x 20mm, 2A46 K10 VAX white aluminium oxide, vitrified bond, 46 grit, at the harder end of the "soft" range, next letter is "middle hard", middle open structure (not "very open").

Cheers,
Bruce

 

[CharlyDE]

Hi Bruce,

you made good work for preparing your flat grinder for startup.

I do like, that your measured "level curves" were confirmed while grinding the magnetic table.

Now, you have a new capability for making parts in your shop. I'm curious to see such parts, you will grind in the future.

Cheers,
Karl

 

[ballen]

Hi Richard,

I still think it would be worth testing with a bubble. Also if you think your floor is bad. Lay a flat bar on the floor and set the level on it shimmed to read level and move the grinder table and see if it moves. Your assuming to much here, I have discovered in machine tool rebuilding you can't assume anything, you have to prove it. Many cases if you have a week floor you set the machine leveling jacks or feet on a 1/2" x 12" x 12" plate to spread out the weight.

I decided to set the machine on a more solid base, along the lines of what you suggested, but my plate is 70 x 90 cm and 12mm thick, and weights about 120 pounds.

Here it is hanging in the air with the plate about to get pushed underneath:

Now down on the plate in the front:

and in the back:

I was not expecting this, but the three leveling screws are quite tubular, so I turned some thin aluminum "pads" to go under them.

It's now about as level as I can read with my best south bend lathe level. Walking around the machine has no effect on the bubble, in spite of my wood-on-hard-foam floor.

Cheers,
Bruce

 

[ballen]

I decided that my grinding skills have gotten good enough, and the time has come to clean up the underside area. I talked with a couple of people in the UK who rebuilt J&S 540s for a living, and they told me that they routinely grind the table pads to clean them up. So I ground about 0.005" = 125 microns off the top of the pad, using a freshly dressed 46J wheel, lots of coolant, and light passes. (I had to redress it several times because the corrosion and goo clogged it.)

Here is the reground top (the coolant smears make it look worse than it is):

Note that the duct-tape splash shield was needed because the smart engineers who designed the normal shield made it about one inch (25mm) too high for a 200mm wheel to reach the table pad . I was very tempted to use some sheet metal shears to address that, but restrained myself


The head-to-table distance is constant, as it should be. Here is a short video showing this with a 1-micron Mahr Millimess indicator:
c

Then I wanted to check that the grinding had not introduced or removed stresses that twisted the table. To do this I put a 300 x 450mm surface plate on the table. It's supported at three points with 1-2-3 blocks on sheets of paper so that it will not rock.

and I can read the surface height to 1 micron resolution with a Heidenhain MT30 height gauge:

I have drawn a coordinate grid with 5cm = 2 inch spacing on the top of the table with a sharpie marker:

I measured the height to the surface plate at the 40 grid crossing points, put them into an Excel spreadsheet, and used the LINEST function to fit this to a 2-dimensional plane. The result is that the deviations from an exact plane have rms value of 1.1 microns (43 millions of an inch) and the largest deviations from a plane are -1.5 and +3 microns. Since the calibration papers that came with my surface plate show deviations of about 2 microns, I am confident that the table surface is flat to within a couple of microns, and grinding the table pad did not twist or bow the table or screw up the geometry of the grinder.

Next step is grinding the bottom of the chuck.

Cheers, Bruce

 

[cash]

Is the surface plate you have on the table have the faces parallel to each other?

I have never seen this kind of test done so I am just trying to figure out what it will tell you?

In a surface grinder you need to make sure the entire surface is FLAT. So lets say you have a high spot in the middle, or the ends fall off. The surface plate will NOT show this.

I am just asking as to see if there is something I am missing.

 

[Richard King]

I have seen this type of test when I taught a scraping class at Gallmeyer & Livingston years ago But they did it on 3 leveling jacks (to get the 4 corners the same before testing) and not 1 2 3 blocks on paper towels. They also would hang a precision ground angle plate designed to be bolted on the wheel guard. Then they indicated the 4 corners zero and used a height gage to measure for error or the table top. I am assuming Ballen is using his math skills to plot the surface plate much like we use on a CMM when measuring something that is already flat. Then averaging out the plane like + 1, - 1, etc.

He is measuring the top plane that he knows is flat because of his chart the plate MFG. included with it when new. He made way to much work not using jack screws in my opinion. I would have had a bad headache doing it his way. Everything looks good and it is obvious he likes to experiment or walk his own path after analyzing all the advice. That happens when one doesn't have a Journeyman with you when your learning. I as assuming he reground the chuck after grinding the table top ? If not you will have to. As doing the chuck first you Ground the error in the saddle and table wear or scraping (can't recall if you scraped it now and haven't reread the thread) from off the chuck. Grinding the error off the table top first and then grinding the chuck makes it parallel and now I bet the chuck is not parallel if you did not kiss the chuck bottom or top. Rich

 

[ballen]

Cash, Richard,

The surface plate is not parallel between top and bottom. It is about 20 microns thicker at the top left corner than the bottom right corner.

Those are not "paper towels" shielding the 1-2-3 blocks, they are pieces of ordinary writing paper. I did not want to scratch the 123 blocks or table, and all I need for this measurement is stability for an hour. A few square inches of paper between flat surfaces with 37kg sitting on it works well for this.

Since I do not have a good way to "jack" the surface plate by 10 or 15 microns, to level it with respect to the table pad, I have "removed the tilt" in software (using the Excel function LINEST which I am using for two-dimensional least-squares linear regression). This uses the set of 40 measurements to determine the slope of the surface in the two horizontal directions (respectively, these worked out to 0.02 microns per millimeter in the transverse direction, so 9 microns of tilt over the long axis, which is 450 mm long, and 0.0656 microns per millimeter in the cross direction, so 9.8 microns of tilt along the 150mm of the cross axis. After subtracting those tilts from my data points, the residuals that are obtained (which would be exactly zero if the surface plate was perfectly flat and the grinder table was moving in a perfect plane) show the deviations from planarity. In this case it's a root-mean-square (rms) deviation of a bit more than 1 micron with a maximum error of +- 2.5 microns. I think at least half of that min/max error is the surface plate itself.

If there is interest here, I can post the spreadsheet for others to use. It's a nice way to check surface grinder geometry without having to jack the surface plate to be parallel to the bed in both dimensions.

Cheers, Bruce


(PS to Cash: in a surface grinder, you need to ensure that (a) the table is moving in an exact plane and (b) the distance from the wheel to the table is constant. My measurement was using a surface plate to check (a). I checked (b) by grinding the table pad on the machine then checking that after grinding the table-to-wheel distance is a constant.)

(PS to Rich: when I first did this in August or September last year, I tried to level the surface plate with respect to the table motion. But I don't have jacks or shims that are fine or precise enough for this. My finest shim stock is 8 micron metal tape, and it is very fragile! About the best I could do was move the plate in increments of around 0.001" or 25 microns, and I needed to be more than a factor of ten more precise than this. So it was easier to put the plate into place as I did (top parallel to the table motion to about 20 microns) and then simply take the measurements. Typing those values into a spreadsheet and removing the tilts along the two axes took 15 minutes, whereas I fussed trying to level it for more than an hour and just got frustrated.

There is one bit of geometry that I have not yet checked, and am not properly equipped to check, which is that the vertical motion of the grinder head is exactly perpendicular to the plane defined by the table motion. That vertical motion must have some tilt in both the transverse and the cross directions. I plan to measure this, but currently I don't have right-angle standards that are precise enough to do it properly. That's next on the project list. I assume that this is why Gallmeyer & Livingston needed a precision angle plate for checking the geometry of their grinders.

I am currently grinding the bottom of the chuck to get it flat and remove the corrosion. Then I'll mount the chuck on the table and grind the top. All of this is getting much easier with practice and experience.)

 

[Richard King]

That same software is used in CMM's and what I meant. Paper is paper....I was making an observation and wasn't trying to argue. You should add the chart spreadsheet somehow so others can see your method. Did you re-grind the chuck? Nice job. Rich

 

[ballen]

That same software is used in CMM's and what I meant. Paper is paper....I was making an observation and wasn't trying to argue. You should add the chart spreadsheet somehow so others can see your method. Did you re-grind the chuck? Nice job. Rich

Rich, the "nice job" from you means a lot, thank you! Yes, I am sure that a CMM can do exactly this type of fitting. I'm happy to post my spreadsheet, but not sure where to put it. Suggestions?

I reground the chuck once in October (see pictures earlier in this thread here: http://www.practicalmachinist.com/v...or-under-chuck-324747/index3.html#post2843259) but I found that it had moved somewhat since then, so I decided to clean up the mess underneath, thinking that corrosion growning there was responsible. In fact there was a different explanation, now sorted out, but I decided that my skills were good enough to do this properly now. I've working on the bottom of the chuck now, after that's done and back in place I'll need to clean up the top, just a few microns should do the trick.

Cheers,
Bruce

 

[dian]

where do you work out? that plate is 100 mm thick, right?

 

[ballen]

where do you work out? that plate is 100 mm thick, right?

Yeah, 37kg = 81 lbs. I stage it carefully. I first put the plate onto a table just next to the grinder. Then I get everything all set up. I wear work gloves with leather pads so that if I catch a finger there is no damage.

If it was OK to drill some 12mm = 1/2" holes into the sides of the plate, I would epoxy in some threaded inserts and then just lift it with the chain hoist. What's the general opinion, is that OK or will it screw up the plate?

 

[Mark Rand]

I'd say that it will have no effect on the accuracy of the plate. Granite and gabbro will have been stress relieved at high temperature for many years after their initial formation. After that, they have sat for millions of years at lower temperatures. They tend to be very stress free, which is one reason to use them for surface plates. #

PS, last week I was lifting a 450mm x 600mm x 75mm surface plate on and off a 600mm high casting on a bench, while scraping it. I could only do three cycles each time before going into the house for a cup of tea and a rest...