HOME
What's New

 

Workshop Stuff

Workshop

Moving the mill

Spindle Noises

ShumaTech Digital Readout

ShumaTech DRO Continued

ShumaTech
DRO-350 Repairs


South Bend 9" Lathe

South Bend 405 Lathe Bench

Grizzly Mill Revisited

Surface Grinder Rebuild

Surface Grinder Continued

Grinder April 6, 2008

Grinder April 20, 2008

Grinder August, 2008

Grinder September, 2008

Grinder November, 2008

Grizzly G3103 Mill
Rebuild

Grizzly G3103 Mill
Rebuild - Part 2

Moving the Shop

Moving the Shop 2

Bringing Home a Sheldon 12" Shaper

Sheldon 12" Shaper 2

Sheldon 12" Shaper 3

Sheldon 12" Shaper 4

Sheldon 12" Shaper 5

Sheldon 12" Shaper 6

Sheldon 12" Shaper 7

Sheldon 12" Shaper 8

Sheldon 12" Shaper 9

Way Alignment Tool

 

 

Email Jim


 workshop.gif

Sheldon 12" Shaper - pg. 5
September 28 to October 14, 2015

After finishing scraping the tool head and swivel block and reworking the lock, I decided that I had to paint the tool head. I really hadn't planned on stripping and painting anything on this machine at present, aside from shooting primer on the missing front panel and the repair I made on a side cover. However, after all of the work I had done on the tool head, looking at the two different colors of flaking gray paint and all of the chips on the tool head made me decide to spruce it up a bit. Since I have started down this slippery slope, I will also paint the other parts I have repaired and put primer on. At this point, I don't plan to paint the whole machine. I would rather get some time in using it. I've had it since the end of May and only have about two hours total time of using it.

I used some aviation stripper, then wire brushed, then sanded the tool head, taped it off, and shot a couple coats of primer on it. I really can't tell if the darker paint below the light gray with a bluish tint was another coat of paint or maybe a darker primer. I like the darker bluish gray better than the lighter color, but many of the Sheldon's I've seen pictures of match the lighter color. While I had considered having the original paint matched and going with some Benjamin Moore Urethane Alkyd Industrial Enamel as I had done with my South Bend lathe, I decided against it and purchased some Steel Gray paint from Tractor Supply. I also purchased some black and some blue. I would attempt to come close to the darker color by mixing the paint myself. I've used this paint before on one of my tractors and it holds up pretty well. It doesn't seem to chip as easily as some of the other enamels I have used.

After I sanded the tool head with 220 grit paper and put the second coat of high fill primer on the tool head, I had some time to kill before calling it a night. My next scraping job would be to scrape the boss that the table support rides on. I also needed to scrape the bottom of the support itself. I hadn't decided which surface to start with at this point. To get started with this job, I first needed to map the surfaces I would be scraping. I set up a magnetic indicator base on the bottom of the table and ran the arm down toward the boss. I attached the DTI and measured the boss for being parallel with the travel of the table and cross rail ways. I measured in three positions: The rear side of the boss, the front side and the center. All three measurements were within a couple tenths of agreeing with each other across the length of the boss, but the boss seemed to be out of square in two directions. The boss was high on the left side by 0.0025". This was over 12 inches of the 16.5" boss. Not terrible, but not to the Sheldon spec of 0.001" in 12" and I wanted to get it closer to 0.0001" if I could.

tool_head_paint1
sheldon_table_support1
With the old paint stripped off of the tool head, I gave it a couple coats of high fill primer. Once it dried for 24 hours, I again sanded with 220 paper.
I measured the top of the table support boss from left to right. The boss was low on the right by about 0.0025" over 12". The boss is 16.5" wide.

The next step was to measure the front to rear surface of the 1 9/16" wide (Y axis) boss in relation to the top surface of the table. I had previously measured the table in relation to the stroke of the ram and after adjusting the ram gib, had proven that the ram and table were in the same plane within 0.0001" over the length of the 12" table. Now I wanted to get the boss to match the ram and table as close as I could. I set my camelback straight edge on the table and attached the mag base for the DTI on the under side or the straight edge. I could now slide the straight edge across the table and read the relative height of the boss. This boss was angled a bit more than I figured it would be. About 0.002" in 1 9/16". That's a lot.

I decided to check the table support to see if the bottom surface was 90° to the mounting surfaces. I was surprised that I couldn't detect any light between the vertical leg of the square and the two surfaces that bolt to the front of the table. The support base is square with the mounting surfaces, but the boss it rides on is not square with the table. I wouldn't have been surprised at less than 0.001" over a little more than an inch and a half, but 0.002" was a lot. I needed to give this some more thought.

I tried a pivot test on the table support bottom surface against the surface plate and it pivoted at the center. This showed that the surface was slightly convex and high in the center. This is no surprise as sliding surfaces usually pick up more swarf on the outer ends than they do in the center, and consequently the wear is greater at the ends of the shorter sliding surface. Because I was still unsure what to make of the measurements I taken on the boss, I decided that I would start scraping the table support first. The support is pretty scarred on the bottom. There are also some wear lines that are not parallel with the bottom rectangular surface. It appears to me that the support had been mounted cocked on the front of the table at some point in the shaper's past. I also noticed that the way wiper felts on the support base were worn out and not doing their job of keeping swarf from getting between the support and the boss. Knowing that wiper felt replacement is often overlooked, I had purchased enough to last me quite a while. I'd replace these once I scraped the bottom of the support.

sheldon_table_support2
sheldon_table_support3
Using the straight edge to support the DTI, the boss appears to be out of parallel with the table by 0.002" across the 1 9/16" section of the boss.
Amazingly, even though the bearing surface of the table support is scarred and worn, the attaching surfaces are square with the bottom.

The first print I took of the table support base showed color in the center, but more color on one side than the other. This didn't necessarily mean that one side is more worn as I could have just put a little more pressure on one side while I was marking it. When a piece is high in the center, it will rock as you try to print it. Before I started scraping, I mounted the inked table support to the table. I would try to wear off the print I had just taken against the surface of the boss to see if my print matched the front to back slope of the boss. This should confirm or refute that the table support base is in the same plane as the boss it rides on.  However, like anyone who has tried to bring a machine into being square with itself has found out, there is always another issue to contend with and sometimes they don't make a lot of sense.

The table support base is L shaped. The lower leg of the L that sticks out has rounded corners and the side that is flush with the upright leg has square corners. If I mounted the support with the leg with rounded corners facing the operator, the base over-hung the operator side of the boss by about a 5/16". If I turned the support the opposite direction, it over-hung the inboard side by 1/4". With the rounded corners facing the ram, I could possibly place a shim between the table and support in order to center the base on the boss, but I am at a loss as to why this should be necessary. Judging by aesthetics alone, I assumed that the side with the rounded corners was meant to face out toward the operator so the rounded corner wouldn't catch your pant's leg. If my reasoning is correct, then there's a rather large misalignment of the table support to the boss. As I said, it doesn't make a lot of sense.

It has been said that scraping is easy, but knowing what to scrape takes a lot of detective work. The same seems to hold true for figuring out what the machine designer had in mind.

I am trying to play detective and am coming up with no good explanation for what I am observing.  On the somewhat plus side, the fact that the support base overhangs the boss and leaves one side of the boss with no contact could account for more wear on the rear of the boss than the front, if the L shaped portion was usually mounted facing the ram. I admit that it's only a guess, but it makes a little more sense than anything else I can come up with.

I needed to scrape the support base no matter what the cause of the misalignment, so I thought I'd start with this surface and think about the alignment of the support base to the boss a bit more.

Being only a novice scraper, I have been doing some more reading and watching some videos about scraping with the Biax in order to improve my technique. After reading about how your should keep the distance between each scrape mark about the same as the width of the scrape mark, I realized I hadn't been doing this as much as I could. I decided to try to slow down the speed and increase the length of the stroke to see if I could rough scrape a little more accurately than I had in the past. I used #2 speed on my Biax 7ELM with about a 1/2" stroke. With the scraper angled at 45° to the long side of the support, I moved the Biax over the length of the surface. I ended up with scrape marks separated by spaces that were about the same width of the scrapes. This was an improvement over my my previous technique. It took a couple passes to scrape the color from the first print. With each new print, I again reversed direction and ended up with a bunch of X marks on the surface. Slowly, I gained more bearing points. However, I found that with the stroke speed slowed down, I needed to hold the Biax handle tighter against my body to counteract the push of each slower stroke. If I didn't support the Biax tightly enough, it jumped around quite a bit.

sheldon_table_support4
sheldon_table_support5
To have an additional way to check the table support boss being square with the table, I decided to scrape this surface first.
A few cycles with the Biax and some cuts with a hand scraper to break up bigger points, I began to get some color across the bottom surface.

After a few more cycles of scraping with the Biax, I switched to a hand scraper to start working on the finish passes. I know that it is possible to finish scrape with the Biax, but I do a better job with a hand scraper. Once I was pretty close to the bearing I wanted, about 25 points per inch, it was time to relieve the center 40% by a small amount. Even though there are way wipers on the table support, in my short time of using the shaper I found that a fair amount of swarf lands on the boss that the support slides against. The swarf will get under the wipers and wear the outer ends of the support first. Hopefully by relieving the center portion of the support base, the table support will stay flat for a longer period of time. Relieving the center section involves some guess work since you cannot reliably print the area you are relieving without a specially made template. To relieve the center 40%, I cut in a cross-hatch pattern, 45° one direction, then switched directions for the next set of passes, with each scrape going the full width of the surface. I scraped four cycles in each direction without printing, while offsetting my scrapes by about a quarter inch each cycle. I tried to use the same pressure for each scrape. If I have done this correctly, when the ends of the table support finally wear down so that the center 40% starts bearing on the boss, I would like the center surface to have a nice array of bearing points. I don't know if I will achieve this, but that is my goal.

In the picture below right, I have the first few passes of relieving done. As you can see by the stray bearing points in the center section, I still have to go a bit deeper. I am now beginning to finish scrape the center section by using shorter strokes and just cutting down the high points as marked by my camelback placed perpendicular to the base. Printing this way is not that accurate, but it's better than scraping blind. I am using a good deal of pressure on each stroke to increase the depth of the relieved area. Once I get the relieved area finished, I also need to start splitting some of the points on the outer ends to get a little more even bearing.

sheldon_table_support6
sheldon_table_support7
A few cycles with a hand scraper and I am starting to get some pretty good bearing. The next step will be to relieve the center 40% of the surface.
The base is 7.0" wide, so I relieved the center 2.8" (40%). I still need to scrape a bit deeper as evidenced by the stray spots in the center.

Toward the end of the evening, I decided to recheck the measurements I had taken on the boss in the Y (front to rear) axis with a level. I had seen a little too much discrepancy in the DTI measurements I had taken when I had tried to move the straight edge fore and aft on the table. I rechecked the table for being level and found that it was off by a half thousandth over the 10" wide table. While the shaper not being level, wouldn't affect my DTI measurements of the boss, getting the shaper level to gravity would allow me to use my box level to confirm the measurements I had taken with the DTI. Because the shaper didn't stay leveled over a couple week's time probably meant that the thick rubber padded I had added to the leveling adjusters were too flexible. While in the process of readjusting the table level, I noticed that I could move the level's bubble by stepping on the flange that the adjusters bolt to. I guess this confirms that the rubber pads were allowing the machine to move around a lot more than I wanted. I needed to fix this before I went any further.

sheldon_sheldon_level12
sheldon_level13
The rubber pads have been removed from the shop made height adjusters. I'm now thinking that I should have wider adjuster feet and secure them to the cement floor.
It is easier for me to level a machine by leveling diagonally. Having the level aligned with the adjusters makes it easier still. The tape marks allow me to repeat my level placement.

I raised the shaper and removed the adjusters. I removed the rubber pads and reinstalled the adjusters. Although the 2" X 3" 3/8" thick steel pads on the adjusters are about the same size as the commercial adjusters I purchased for a much heavier machine, I am wondering if a larger pad would be better on the shaper. The force of the moving ram is unlike any other machine tool I own. A wider steel pad attached to the cement floor would probably be an improvement. However, the new adjusters will have to wait. I needed to take some measurements, so I now I needed to level the shaper again. As I have said before, I like to level diagonally. The problem I encountered with the shaper is that the leveling pads are inboard from the corners of the rectaganal body of the shaper's base and can't be seen while I am trying to position the level on the table. I have found that it is easier to have the level line up with a imaginary diagonal line between the adjusters. To help me position the level, I measured the layout of the adjusters. 27" wide and 22" deep. Since the table is 10" wide and 12" long, I did some quick math. 22/27 = X/10, X = 8.148" or about 8 5/32". So I needed my diagonal tape strips to intersect the corners of the 10" wide table and then intersect at 8 5/32" along the Y axis. With the tape strips now corresponding to the adjuster locations, it was somewhat easier to level the shaper.

sheldon_table_support8
sheldon_table_support9
Table support sitting on the boss in the direction that I would think to be correct except that the front of the support base overhangs the boss
If I reverse the support, it overhangs the rear of the boss, but not by as much. While I can correct for this, it doesn't seem right.

With the shaper leveled again, I could now check the table support boss with the level as well as with an indicator. This would make it easier to bring the boss into level with the table. I mounted the table support again to measure how far out of alignment it was. Facing the direction that I assumed it was meant to mount, it overhung the front of the boss by about 5/16". If I reversed the support, it overhung the rear of the boss by about 1/4". Since I wanted to use the table support as a final check while scraping the boss, I decided to machine up a shim to center the support on the boss. I had no quarter inch flat stock, so I machined the shim from 3/16" and added a couple 1/16" washers until I could purchase the correct stock. As shown in the picture below left, the shim and washers centered the support on the boss. I am still at a loss to explain why the table support doesn't align properly. I did some looking online and found a few images of Sheldon shapers with the table support overhanging the front boss. Apparently mine is not unusual in this respect. I still don't care for the setup. I think that the more surface the support base has in contact with the boss, the more evenly it will wear. I also think my plan of turning the support with the leg to the inside and using a shim will work fine. It should help keep the newly scraped surfaces from wearing unevenly.


sheldon_table_support10
sheldon_showing_table_support
I made a shim from some stock I had on hand and added a 1/16" thick washer behind each leg to get the support centered on the boss. .
I found a picture online showing another Sheldon shaper with the table support overhanging the boss. Apparently this issue wasn't limited to mine.

To scrape the boss, I would need to correct for it being angled in two directions. The high point, viewed from the front of the machine, is the front left corner. I used my camelback straight edge to mark the boss. It showed color only in the center of the boss. I forgot to take a picture of this, so the first shot is after a couple passes with the Biax. I am using a lot more downward pressure as I come to the left front of the boss. I have a couple thousandths to scrape from the left front end. I scraped until I had a few spots of color over the whole boss, then started step scraping to lower the left side of the boss. I also used a little more pressure on the front side of the boss to start getting rid of the angle in the Y axis. In a couple hours, I had brought the surface to within a thousandth high on the left (X axis) and within a couple tenths front to back (Y axis). Since I have been scraping with more pressure on the left side of the boss, the difference between the peaks and the valleys is as much as 0.0004". This makes it a little tough to measure the slope of the boss as the indicator bounces as I crank the table and attached indicator across its length. I decided to spend a little time refining the surface so I could get some better accuracy in my measurements. This might be considered wasted work as I still have a slight angle to the boss, but I felt that I need to get a more accurate idea of how much to scrape to get the boss parallel with the table's travel.

sheldon_support_boss1
sheldon_support_boss2
Since I forgot to take a picture before I started scraping, here's a shot after a few cycles with the Biax and a couple cycles with a hand scraper.
The boss is still about 0.001" high on the left. I will now scrape enough cycles to get the surface smooth enough to map out the slope accurately.

After another few cycles of hand scraping, I had lessened the height between the points and the valleys and had increased the point count a bit more. I again measured the slope of the boss in both directions. I was now measuring a slope of 0.001" over the full 16.5" of the boss with it high on the left side. Over the 1 9/16" width, I was pretty close to level. It was time to do one more step scrape of the boss. I decided to set up the camera and take pictures of the process of applying color to act as a guide. I rolled marking fluid on the left side covering about 1/5 of the surface, then scraped it off with the Biax. Then marked 2/5 and scraped that. I repeated this until I had marked and scraped the whole surface. When scraping for bearing points, I would try to leave spaces between the scrape marks, but the object of step scraping is to evenly lower one end of the surface, so I kept the scrape marks close to each other. I also normally reverse directions on each cycle, but I am more consistent scraping from right to left and with moderate pressure can scrape 0.0002" per pass. Since I needed to lower the left side by one thousandth, five passes should get the boss close to level. I put together an animated gif showing the process. I did forget to take a picture of the fourth scraping pass, but otherwise, it shows the process pretty well. I also added two shots of measuring the surface. The indicator shows about a ten-thousandth of slope in 16", but it really isn't quite that close. I went from a very good points count to a very sparse points count and there are sections of the boss where there are two or more tenths height difference within a quarter inch. I will do a bit of hand scraping to get an idea of where my high and low areas are, then rough and finish scrape from there. At least I now have a boss that is in the same plane as the movement of the table.


 click for
                      step_scraping animated gif
sheldon_support_boss3
Step scraping the table support boss to remove one thousandth inch of slope over 16.5". Click image for animated gif (2.5 megabyte file size).
After the step scraping passes, I went from a decent points count to next to no points at all, but the boss is now pretty much level with the table.

As usual, making a print of the newly step scraped surface is a bit of a let down. It now didn't have much in the way of bearing points. I made a couple passes just picking the tops off of the few points I did have until I could get an idea of the high and low areas. I then made a couple passes at 45° with each scrape covering the entire width of the boss. I refined the surface a bit more and decided that it was time to put the level on the boss. I knew I was close to level, but was pleased to see that the level's bubble was  aligned with the center indications. Less than a ten-thousandth out over ten inches. Not too bad at all.

sheldon_support_boss4
sheldon_support_boss5
After one pass with the hand scraper, I am beginning to see where my high and low areas are located. I have some more scraping to do.
Double checking the plane of the support boss against the plane of the table on a leveled shaper. I couldn't have asked for a better reading.


Checking the boss over the short side was a bit trickier. With less than two inches to rest the level on, the measurement wasn't going to be as accurate as measuring on the longer axis, but I was just looking for confirmation that the measurement I had taken with the DTI was close to what the level showed me. Again, the bubble lined up with the center marks on the vial. Now all I needed to do was to print the boss again and start refining the surface.

sheldon_support_boss6
sheldon_support_boss7
With the level set across the 1 9/16" boss, some inaccuracy is expected, but the level shows that this plane is pretty close to level with the table.
Another couple of cycles with the hand scraper and I am beginning to see some color across the whole surface. Finish scraping is the next step.


Refining the surface from the cross-hatch rough scraping I had done went pretty quickly. I started by scraping off as many high points as I could get with half inch strokes using the hand scraper. I kept scraping at a 45°angle and reversed directions of the angle on each cycle. I continued this for a few cycles, then switched to scraping off individual spots. If I had a group of spots angled to the left, I scraped through them to the right and vise versa. It took about five more cycles to get to the picture on the lower left. The boss now had areas of 35 or more points per square inch in some areas and as few as 25 per inch in others. While I would like to see the points count to be a little more consistent, my number of points per square inch is in excess to what's necessary for this surface. Even though I have been scraping for almost a decade, I still have much to learn about getting a consistent pattern and points count on what I scrape. However, I have gotten a lot quicker at refining surfaces since I started work on this shaper. To be able to take a surface from less than one point per square inch to 25+ in an hour or so is pretty quick for me.

sheldon_support_boss8
sheldon_shaper2
I am pretty close to done scraping the boss. I have a few sparse areas and will scrape another pass or two, but am pleased with my progress.
Taking a test cut. With one side of the CI block surface ground, I planed the other side. The planed surface came out pretty close to parallel.

With all of the scraping done, I finished up painting the tool head, table support, and the vise parts. My paint mixture is a little darker than the paint on the rest of the shaper and looks a little more blue when not in direct light. In the picture above right. the tool head is the same color as the table support, but they look very different in the picture. Enough about the paint. Time to take a test cut.

Since the vise still needed to have the rails cut, I needed to find an alternate means of mounting the 4" X 6" block of cast iron. I would have preferred to mount it directly to the table, but didn't want to use hold downs that would need to be removed as I planed the whole upper surface. I ended up using a two piece milling vise. It was too tall, so I needed to put parallels below the cast iron block I was cutting. Not really the best way to check the accuracy of the cut, but it would let me test whether the tool head ways worked better and whether the lock was now working as designed. I set the ram travel for 7" and set the ram position to have 1/2" over-run on either side of the 6" block.

I started my cut with the table moving left to right and got used to the rhythm of the clapper box lifting at the end of each cut and snapping closed just before each cut began. About half way through the 4" feed in the X plane, the sound of the clapper got a bit louder and the timing of the noise changed. At first I didn't understand the change in sound and rhythm, but soon discovered that the ram stroke was lengthening and was now up to 8". The force of the clapper closing on the longer stroke made it sound a bit louder. By the end of the cut, the ram stroke had increased to 10". What the heck? Something wasn't right. However on the plus side, I had zeroed the tool head dial and locked the tool head before I began the cut and when the cut was done, the tool head had not moved. I wound the table back to the starting position and set the tool head for another cut. 0.002" this time instead of 0.005" as for the first cut. I reset the ram travel to 7" and began the cut. This time, the ram stroke started increasing at about 1/4 the way through the 4" cut. Rather than stop the machine and reset the ram, I just let it cut so I could check the test block for the top surface being parallel with the bottom surface. The stroke had increased to 11" by the end of the cut.

With the top surface of the block planed, I removed it, put it on the surface plate, and measured it. Even with the less than accurate two piece vise, the plane of the top surface was within a half thousandth of the ground bottom surface across the worst diagonal measurement. The other diagonal was within a couple ten-thousandths. It appears that scraping the tool head, swivel block and gib has helped. The tool head lock also seems to be working as designed now. I only have to lightly snug the lock screw to keep the tool head from creeping down and making the tool bit cut deeper into the work. So I seem to have succeeded in fixing the problem that I set out to repair, but have discovered another issue that needs some attention. Such is the way it seems to go with old iron. On one hand, I'm happy that I have been able to take care of something that needed attention, but I am now playing detective again to see if I can deduce why the ram stroke is increasing.

Next installment: trying to figure out the increasing ram stroke.

Shaper 1
Shaper 2
Shaper 3
Shaper 4
Shaper 5
Shaper 6
Shaper 7
Shaper 8
Shaper 9


© Fager October 14, 2015