What's new
What's new

cast iron straight edge and a surface plate

grannyknot

Aluminum
Joined
Oct 18, 2015
A couple of yrs back I bought a cast iron straight blank, I took it to a machinist who said He could mill it down to
.005" over the whole 24". All I want to use it for is checking cylinder heads an engine blocks, I placed the straight edge on a known good head and saw some pretty big gaps.
So I bought a used granite surface plate and tried it again, same thing.
Now I'm not sure if the surface plate is out or the straight edge is out and what I should do next to fix this.
I'm all ears.
Thanks,
Chris
IMG_0076.jpg IMG_0089.jpg IMG_0094.jpg IMG_0097.jpg
 
If you want to know, you need a reference. That comes from the plate. But a used plate must be recertified if you want to know.

I believe the idea is to scrape the straightedge straight once one has access to a known-good surface plate. alternatively, I imagine someone skilled could grind it a lot closer to straight than it seems to be.
 
One thing to try is checking it in a number of orientations and locations on your plate. If you see the same gaps everywhere, they are in the straightedge.
 
Some of the shitty vices actually suck a part down if that is how he milled it. It would create just that scenario
 
A couple of yrs back I bought a cast iron straight blank, I took it to a machinist who said He could mill it down to
.005" over the whole 24". All I want to use it for is checking cylinder heads an engine blocks, I placed the straight edge on a known good head and saw some pretty big gaps.
So I bought a used granite surface plate and tried it again, same thing.
Now I'm not sure if the surface plate is out or the straight edge is out and what I should do next to fix this.
I'm all ears.
Thanks,
Chris
View attachment 332248 View attachment 332249 View attachment 332250 View attachment 332251

 
Odds are, unless it's beat, the surface plate is pretty good and the straightedge isn't. Not surprising since it was milled. You need to determine that the surface plate is good, then scrape the straightedge until it's good too. For your needs it would be easier to have somebody grind it.
 
that straight edge looks milled on both sides.
find a 5" or 6" mike and measure the width of the straight edge every 2 inches.

then set it on the surface plate and use an indicator to measure the difference in height of the upper surface over its length relative to the plate.
then flip the straight edge over and record the numbers again.

then average the error produced by those 3 sets of numbers.

alternatively, set the straight edge on a different part of the surface plate and generate the numbers again.

also keep in mind the geometry can be complex. the first 6 minutes of this video will show you.
Check a straightedge for straightness: FLIP it the RIGHT way - YouTube
 
It looks like the stone has a calibration sticker on the corner. What is it claiming the stone to be, and how old is it? That would give you some baseline expectations. You can tell with a tenth indicator and a long arm if the stone is worn too much. Watch the vid :). Good luck

That sticker hasn't been of much help to me, maybe you guys can make sense out of it.
IMG_0074.jpg
 
I did the surface gauge trick on the granite plate, there is a slight dip in the center of the plate but couldn't find an error that was more than 0.001"
I measured the width of the straight edge with a mic and there were some big differences as you will see from the pics, then measured the profile from the plate to the top of the straight edge. The average width was 3.2596" the average height from the plate was 3.2719" for and average of 0.0123":ack2: But that average is being kind, there are sections where the difference jumps 0.006" with in the span of 4". I guess I'm never going back to that machinist.
So the next step is scraping?
Thanks
IMG_0121 (1024x768).jpg IMG_0125 (1024x768).jpg IMG_0132 (1024x768).jpgIMG_0132 (1024x768).jpg IMG_0133 (1024x768).jpg IMG_0135 (1024x768).jpg
 
I did the surface gauge trick on the granite plate, there is a slight dip in the center of the plate but couldn't find an error that was more than 0.001"
I measured the width of the straight edge with a mic and there were some big differences as you will see from the pics, then measured the profile from the plate to the top of the straight edge. The average width was 3.2596" the average height from the plate was 3.2719" for and average of 0.0123":ack2: But that average is being kind, there are sections where the difference jumps 0.006" with in the span of 4". I guess I'm never going back to that machinist.
So the next step is scraping?
Thanks
View attachment 332305 View attachment 332306 View attachment 332307View attachment 332307 View attachment 332308 View attachment 332309

scraping works :)...Some basic advice
#1 Get some Sandvik blades and a handle or make a handle.
#2 Get some Canode or Charbonnel Aquwash Amazon.com : charbonnel aquawash Yellow or orange and Blue alternatively Red and Black
#3 Figure out how much you need to remove If it is more than 3-5 thou, get someone to machine the SE
#4 Learn to rough scrape efficiently. and this can be tricky. One thing is that you need to use a blade with a large radius, use long strokes and try to create a hashed surface. Watch some of the videos. You will get the idea. The goal of rough scraping is to get blue spots everywhere and fairly evenly. Learn how to hinge the piece. When the piece is resting evenly on the surface plate it will rotate around two points about 30% in from the ends. drag the left and right end back and forth and watch where it wants to rotate.
#5 Develop the Russian approach to knowing when something is Gudenoff. In the roughing stage, you want to see at least 2 or 3 large spots every square inch of the entire surface of the piece.
#6 Learn where the blade removes the cast iron. Put a marker in the center of the curve of the blade and watch where the material is removed in relation to that.
#7 There is a kinda transitional phase between roughing and finishing. Once you are done with the overall roughing you need to lower the areas where you have more points touching. Switch to a blade with a 90mm radius. You will proceed with shorter strokes and evenly spaced strokes in evenly spaced lines. The stroke width and spacing should be about equal, and you should alternate passes in orthogonal directions. This will result in the basketweave pattern.
#8 When you get to the final finishing that is the phase when you only attack the highest points.
#9 Learn to read the ink. light blue nearly touching, darker is higher, the highest points will have a shiny center and dark atoll around it.
#10 You need a way to sharpen your blades to about 1200 grit to be efficient. Diamond lapping plates on a slow grinder work best.

I hope this will get you started in the right direction
 
scraping works :)...Some basic advice
#1 Get some Sandvik blades and a handle or make a handle.
#2 Get some Canode or Charbonnel Aquwash Amazon.com : charbonnel aquawash Yellow or orange and Blue alternatively Red and Black
#3 Figure out how much you need to remove If it is more than 3-5 thou, get someone to machine the SE
#4 Learn to rough scrape efficiently. and this can be tricky. One thing is that you need to use a blade with a large radius, use long strokes and try to create a hashed surface. Watch some of the videos. You will get the idea. The goal of rough scraping is to get blue spots everywhere and fairly evenly. Learn how to hinge the piece. When the piece is resting evenly on the surface plate it will rotate around two points about 30% in from the ends. drag the left and right end back and forth and watch where it wants to rotate.
#5 Develop the Russian approach to knowing when something is Gudenoff. In the roughing stage, you want to see at least 2 or 3 large spots every square inch of the entire surface of the piece.
#6 Learn where the blade removes the cast iron. Put a marker in the center of the curve of the blade and watch where the material is removed in relation to that.
#7 There is a kinda transitional phase between roughing and finishing. Once you are done with the overall roughing you need to lower the areas where you have more points touching. Switch to a blade with a 90mm radius. You will proceed with shorter strokes and evenly spaced strokes in evenly spaced lines. The stroke width and spacing should be about equal, and you should alternate passes in orthogonal directions. This will result in the basketweave pattern.
#8 When you get to the final finishing that is the phase when you only attack the highest points.
#9 Learn to read the ink. light blue nearly touching, darker is higher, the highest points will have a shiny center and dark atoll around it.
#10 You need a way to sharpen your blades to about 1200 grit to be efficient. Diamond lapping plates on a slow grinder work best.

I hope this will get you started in the right direction



Second thoughts on #3. Your mike measurements indicate 12 thou, but it is between the two sides, measure one side off the surface plate to get a feel of the deviations on each side. Start with just getting one side flat (the one that measures better), you can make decisions about parallel, and square later.
 
I can not seem to find the information on using a tight wire for checking strangeness. you put your straight edge on a flat surface. place a piece of round stock, like a gauge pins with the same diameter, on top of each end of the straight edge. Then use a piece of .006" music wire (not sure on the dia.).one end of the wire is attached to the table then placed over the first pin down to the top of the other pin and wire is draped over the edge of the table with a 16 lbs. weight attached (not sure on the weight). this will preload the music wire and the sag of the wire will be near zero in 2 ft, the length of your straight edge. you can use feeler gauge stock, or what ever method you chose measure the distance from the wire to the straight edge. there are tables that show the sag of the wire in lengths of 30 ft if I remember right. this method will definitely check your straight edge. gauge pins have to be the same dimeter, wire has to have the same dia. the full length of the test length.

Corrections to the above diameters and weights........
See Forrest, I found it. Wire sag table.


.016 dia piano wire with 30lb weight

lenth of wire, sag and distance from end

4ft .001 at 2'

8ft .001 at 2',.002 at 4'

12ft .002 at 2',.004 at 4',.005 at 6'

16ft .003 at 2',.006 at 4',.007 at 6',.008 at 8'

20ft .004 at 2',.008 at 4',.011 at 6',.012 at 8',.013 at 10'

24ft .006 at 2',.010 at 4',.014 at 6',.017 at 8ft,.019 at 10ft,.020 at 12'

28ft .007 at 2',.013 at 4',.017 at 6ft,.022 at 8ft,.024 at 10ft,.025 at 12ft,.026 at 14'

32ft .008 at 2',.015 at 4',.021 at 6',.026 at 8',.030 at 10',.033 at 12',.034 at 14',.035 at 16'

36ft .009 at 2',.017 at 4',.024 at 6',.030 at 8',.035 at 10',.039 at 12',.042 at 14',.043 at 16',.044 at 18'

40ft .010 at 2',.020 at 4',.028 at 6',.035 at 8',.040 at 10',.046 at 12',.050 at 14',.052 at 16',.053 at 18',.054 at 20'

I hope this looks like something when I post it. Here goes.
 
I can not seem to find the information on using a tight wire for checking strangeness. you put your straight edge on a flat surface. place a piece of round stock, like a gauge pins with the same diameter, on top of each end of the straight edge. Then use a piece of .006" music wire (not sure on the dia.).one end of the wire is attached to the table then placed over the first pin down to the top of the other pin and wire is draped over the edge of the table with a 16 lbs. weight attached (not sure on the weight). this will preload the music wire and the sag of the wire will be near zero in 2 ft, the length of your straight edge. you can use feeler gauge stock, or what ever method you chose measure the distance from the wire to the straight edge. there are tables that show the sag of the wire in lengths of 30 ft if I remember right. this method will definitely check your straight edge. gauge pins have to be the same dimeter, wire has to have the same dia. the full length of the test length.

Corrections to the above diameters and weights........
See Forrest, I found it. Wire sag table.


.016 dia piano wire with 30lb weight

lenth of wire, sag and distance from end

4ft .001 at 2'

8ft .001 at 2',.002 at 4'

12ft .002 at 2',.004 at 4',.005 at 6'

16ft .003 at 2',.006 at 4',.007 at 6',.008 at 8'

20ft .004 at 2',.008 at 4',.011 at 6',.012 at 8',.013 at 10'

24ft .006 at 2',.010 at 4',.014 at 6',.017 at 8ft,.019 at 10ft,.020 at 12'

28ft .007 at 2',.013 at 4',.017 at 6ft,.022 at 8ft,.024 at 10ft,.025 at 12ft,.026 at 14'

32ft .008 at 2',.015 at 4',.021 at 6',.026 at 8',.030 at 10',.033 at 12',.034 at 14',.035 at 16'

36ft .009 at 2',.017 at 4',.024 at 6',.030 at 8',.035 at 10',.039 at 12',.042 at 14',.043 at 16',.044 at 18'

40ft .010 at 2',.020 at 4',.028 at 6',.035 at 8',.040 at 10',.046 at 12',.050 at 14',.052 at 16',.053 at 18',.054 at 20'

I hope this looks like something when I post it. Here goes.

Yeah, that kinda works for checking, but because of the flexibility of the setup the precision and repeatability are a bit iffy, maybe +-0.0005" or so. Spotting and scraping can get you to within 10 millionths of the precision of the surface plate, in other words, the difference between the surface plate and the scraped surface is the thickness of the ink layer. Most shop grade surface plate will be flat within a couple tenths over 2 feet, inspection and lab grade will half that tolerance with each grade. For measuring the SE a tenth DTI on a height or surface gauge will prove the SE within a couple of tenths. A scraped SE can be flat and straight to 0.0001"/ft or better.
 
Yeah, that kinda works for checking, but because of the flexibility of the setup the precision and repeatability are a bit iffy, maybe +-0.0005" or so. Spotting and scraping can get you to within 10 millionths of the precision of the surface plate, in other words, the difference between the surface plate and the scraped surface is the thickness of the ink layer. Most shop grade surface plate will be flat within a couple tenths over 2 feet, inspection and lab grade will half that tolerance with each grade. For measuring the SE a tenth DTI on a height or surface gauge will prove the SE within a couple of tenths. A scraped SE can be flat and straight to 0.0001"/ft or better.

I guess that I'm missing something.... The original question was I have a 24" straight edge that I would like to check the straightness. Having a surface plate it is questionable as to which one is "true". The method presented should easily check the straightness at a very low cost within 0.0005". If the straight edge is outside of the specifications required by the owner then the methods for correction can be pursued.
 
I guess that I'm missing something.... The original question was I have a 24" straight edge that I would like to check the straightness. Having a surface plate it is questionable as to which one is "true". The method presented should easily check the straightness at a very low cost within 0.0005". If the straight edge is outside of the specifications required by the owner then the methods for correction can be pursued.

you can check the surface plate with a DTI and a surface Gauge to within a tenth or two. I posted the video from Suburban earlier. BTW I have nothing against the wire method, just wanted to point out its shortcomings. The wire method is invaluable if optical methods are not available. The wire method is still valid to get a surface or a bed way within a thou tolerance over a distance.
 
Last edited:
From my experiences, surface plates with no visible damage will wear where the most traffic is and that's typically only in high high usage shops. Small shops using the plate a couple of times a day don't wear it out. The shops running millions of parts and very large parts 24/7 tend to do the most damage, and TYPICALLY, those same shops need to have traceable inspection and will keep their plates calibrated. When they do start to wear, they'll get them resurfaced and be back in business. Small plates are cheaper to replace, but typically don't see near as much volume of parts. These of course are all assumptions, but I'd be more concerned about the casting then the granite.

I don't know that I would blame the machinist who cut the casting, unless you were very clear with what you were trying to accomplish, dimensionally. If you just told him, "Here, mill this casting flat for me." then he did that. A cast iron straight edge is a very simple tool, but making one involves more than just pouring a casting and milling it flat (There's a reason why new finished straight edges cost thousands and plain castings cost hundreds).

Raw castings need to be stress-relived to be stable. The most common method is to do it in a heat treating oven, but other methods are employed as well including vibration and aging. The idea being that the material straight out of the sand might be cool, but it's still not used to it's new shape and will move in small amounts until it has settled. Stress-reliving accelerates that process.

After that, you machine it to get it into it's final shape, and the level of precision required in the finished part has a big bearing on how it is machined. Many companies often see iron castings only as a mass production part made with special fixtures, and don't put much thought into the tolerances when machining a single piece alone. When you clamp a straight edge to the table or in vices, you need to be sure you are not twisting or bending it and take light passes. Even then, it's not abnormal to need to give it a second heat-treating session as you have now removed significant material from the original shape that was contributing to it's stability. The action of cutting the metal away can also induce stress in the part. After all that, milling is still going to leave a few variances here and there which need to be cleaned up by grinding or scraping.
 








 
Back
Top