Looking for Sanford MG-612 Surface Grinder owners

Oops - error on my part! I should have said a "tilt" of 10 thou, not 3 thou. In any case, thanks for the encouragement. You did mention earlier about not tightening the chuck down too much, and I was careful to follow your advice.

Obviously I've still got a lot to learn, but at least I can now get some actual grinding experience (and work my up to mediocre!).
 
Try indicating the top of the coolant containing rail around the table and see what the runout shows there with table travel. I have no idea how a table could be ground .010" out from left to right.
 
Bob Korves said:
Try indicating the top of the coolant containing rail around the table and see what the runout shows there with table travel. I have no idea how a table could be ground .010" out from left to right.
Click to expand...


The top of the coolant rail on my Sanford MG is a painted surface. There are a few places where the paint has worn off and you can see a rough machined surface. The rail isn't finished nearly well enough to be trusted as an indication of table position.
 
Could also try to indicate the bottom of the table if there is sufficient access along one of the sides.
Edit: The .010" out really bothers me that I cannot say how it might have been caused, except perhaps at the factory. I would be all over it, if for no other reason than to see if whatever caused it may be causing other issues. I really have no idea how that could happen on purpose, post production, much less by an accident.
 
Last edited:
My Sanford MG is 60+ years old, and I don't have a real history on it, so no telling who did what to it over the years. For the time being, I am going to take the position that as long as I can get the top of the magnetic chuck flat with respect to the grinding wheel, that will be OK for now. I'm going to be grinding non-critical test/practice pieces for a while, so I expect that if there is a deeper problem it will show up soon enough.

On my initial grinding-in of the mag chuck, I got the measurements below on the surface. I took 3 sets of four measurements back to front, right to left; the top row is the back of the chuck surface, measurements are given in tenths:

0_________-15_________-15_________-15_________ -15

3_________-9_________-8_________ -11_________ -10

0_________-11_________-12_________ -11_________-11

I set the 0 for the above measurements at the back left of the table. Cleary there is a problem at the left edge of the table, but overall I would rate is as neither great nor horrible, but definitely in need of improvement. I decided to take one more attempt, taking the lightest possible cuts, and also taking much more time (I think I got a bit impatient with my initial grinding in, and went a bit too fast). After this second grinding-in, I got the much improved numbers below, and I am going to call this good for now.


0__________1__________2_________ 3_________1

-1_________2_________-2_________2.5_________0.5

0__________0__________1__________0_________0
 
First off are you using the spindle or the housing as the mounting point of the indicator stand? I tried using the housing the first time but found it didn't have a fine enough finish. The painted surface was uneven enough to give different readings. Just the pressure of the indictor point on the chuck was enough to cause the magnetic mount to shift slightly.

I am using an .0001" indicator that seems to be sensitive enough that a slight breeze could change the reading. When I removed the guard, wheel and hub there was enough room to mount the magnetic base directly on the spindle. Also I didn't let the point slide across the table. I set both the spindle and indicator at the 0 point at the left rear corner. Then I raised the spindle one full turn (.050 on the scale) and repositioned the table. The second test was at the right rear of the table. This corner measured .0002 lower than the left rear corner. I again raised the spindle and moved to the right front corner. This corner also measured .0002 lower than the left rear. The final corner to be measured was the left front. Using the same method this corner measured .0000.

In summary the table on my machine is currently .0002 lower on the right side than the left. To be sure the measurements were accurate I did the entire procedure 6 times. There was never any more difference than .0001 from the baseline. I did clean the table with a dust free rag and made sure it was properly secured to the table. The bolts are only slightly more than finger tight. I didn't use a torque wrench, but from experience I would say they are torqued to between 10 and 15 psi. As a test I did tighten them around 40 psi. It made an appreciable difference in the readings. The .0002 readings ballooned to as high as .0025. Before I ended the day I once again retorqued them to about 10 psi and the .0002 readings returned.

I'm sure given enough time and effort I could achieve .0000 across the board. However since the spindle dial increments only go down to .0005 I think .0002 is close enough for my purposes. If you refer to the factory specs 1953 sales brochure) for the table longitudinal and transverse error they were both a maximum of .0001 from the factory. I think .0002 after 64 years is acceptable.
 
Note that testing the mag chuck by mounting an indicator holder to the spindle area and then traversing the table back and forth and front to back underneath it, does not say anything about how flat the chuck top surface is. If the ways are worn more at the ends (very common), then the table may quite possibly be sagging down on the ends of its travel. The middle, however, will show a very good reading, because you are only checking it at one of the points that supports the table. This spherical geometry is common to find on surface grinders and milling machines. To get a more realistic reading, put 5 blanks, perhaps the size of a coin, on the chuck, one in each corner and one in the middle. Then grind them carefully and evenly, and then measure them for thickness. Compare that with what you got from the test with the indicator attached to the spindle. Another way to do it is to take the chuck off the grinder and test it mounted on a surface plate, with any rock in the corners shimmed up. Of course, then you will need to grind it again when you reinstall it on the grinder. A REALLY good straightedge can also be useful for testing. Sometimes ways will together be "twisted" from left to right, which can show odd results at the corners of the chuck. Never assume that the ways are flat, parallel, and co-planar over their length, instead test for it.
 
Last edited:
projectnut - I tried using an indicator stand on the spindle housing, but ran into the same problems as you. I ended up making a custom indicator holder which mounts directly to one of the screws for the housing cover. I did run the indicator ( a tenths indicator) over the chuck surface, but I found that I could get repeatable readings by tapping the indicator holder lightly with a pen to help reduce hysteresis error. Raising and lowering the indicator as you did is a good idea; I'll try that the next time I make measurements.

Oxtoolco did a video on the 5 block test:


I'll give it a try - should be good practice for me.
 
OK - five block test completed. Here are the results, with the blocks arranged as follows (where the top is the back of the table):

#1_________________________#2
______________#3_____________
#4_________________________#5

After grinding both sides, I measured each block with a Mitutoyo micrometer as follows (in inches):

#1 = 0.60315
#2 = 0.60275
#3 = 0.60280
#4 = 0.60280
#5 = 0.60290

Setting block #2 as zero, the difference in the blocks in tenths is:

#1 = 4.0
#2 = 0
#3 = 0.5
#4 = 0.5
#5 = 1.5

I'm not sure what's going on with block #1, but I'm pretty happy with these results overall.
 
There are all kinds of ways to have #1 that far off... even a small piece of dust or grit can cause it... You have excellent results for any machine.

Congrats on a great find!
 
You must log in or register to reply here.
Back
Top