BigBob
Plastic
- Joined
- Jan 24, 2010
- Location
- USA - Verona, VA
Short of having plates formally calibrated, anybody use these for preliminary surface mapping, and if so, how?
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Largest Manufacturing Technology Community on the Web
Stay Connected:
PRecision levels to evaluate surface plates
- Thread starter BigBob
- Start date
- Replies 68
- Views 26,948
Mebfab
Diamond
- Joined
- Jun 7, 2003
- Location
- Mebane North Carolina USA
Wont work. Spans too large a distance. Think of a bridge across a mountain gorge.
Try a long flat piece of iron on the plate with a tenths indicator stylus on the plate. SLide around. The long iron base will ride along on the indicator will pick up localized low/high spots
Try a long flat piece of iron on the plate with a tenths indicator stylus on the plate. SLide around. The long iron base will ride along on the indicator will pick up localized low/high spots
RC99
Diamond
- Joined
- Mar 26, 2005
Not sensitive enough...
The devices used, the likes of a Talyvel or autocollimator can measure angle differences the likes of 0.2 seconds or arc, compared to a level which is usually 8 to 10 seconds of arc between divisions.
The real Leigh
Diamond
- Joined
- Nov 23, 2005
- Location
- Maryland
Federal Spec GGG-P-463C (now obsolete) gives a measurement method that uses a sensitive dial indicator on a fixture.
See sections 3.3, 4.3, and others.
Download from http://www.atwaterkent.info/Images/FedSpec_GGG-P-463C_SurfacePlates.pdf
This requires a proper 3-point support for the plate, the design of which is given in the spec.
- Leigh
See sections 3.3, 4.3, and others.
Download from http://www.atwaterkent.info/Images/FedSpec_GGG-P-463C_SurfacePlates.pdf
This requires a proper 3-point support for the plate, the design of which is given in the spec.
- Leigh
Richard King
Diamond
- Joined
- Jul 12, 2005
- Location
- Cottage Grove, MN 55016
I use level all the time to test granite and cast iron plates. I have 2 levels, a Starrett 199 with a .0005/ 12" bubble and a King-Way 12" level with a .0003"/ 12" accuracy bubble.
I have been site testing plates with levels for 40 years and my Dad did it that way for 40 years before me.
I specialize in scraping Jig Bores and super precision machines. I recently scraped 2 Drake USA), 1 Studer (Swiss) and 1 Shigiya (Japanese) CNC Cylindrical grinders. In Final tests we ground test bars to straightness less then a .0001", one of the Drakes was .00002" before the off-sets were done. I used their Grade B plate to scrape my cast iron straight-edge. I had checked with my level and found an area that was flat to .00005" and used that area to blue up my straight-edge. My point being You can use a level to check a plate! There are better ways, but not necessary.
It's a good way to find "local errors" on a plate and best place to check my cast iron straight-edges when I am at a job-site scraping a machine. The way I do it is a lay out the plate using a straight-edge and pencil. I draw straight lines on the plate 1/2 the length of the level I am using. Cross hatching it.
I level the plate both directions before I test the plate in the center of the plate. I also check to be sure the plate is sitting on 3 points. They need to be placed in the exact position they were when it was lapped at the factory. Some plate manufactures use 25% and others use 30% on the 2 point end and in the center on the single point end. There is rubber pads or a X where they had mounted the pad. I check the levels flatness and test it's positioning error by flipping it end to end and pushing down on the ends to be sure the level is not high in the middle.
I get a pad of 1/4" graph paper and layout the plate on the paper. Then start on one end and watch the level, mark it on the paper and move the level 1/2 its length, press down on the ends, wait 10 seconds and record my measurement, move down 6" and do it again, etc. and check it again all the way down the pencil line, then do all the lines and the cross lines. I also do diagonal corner to corner check too.
I just had a surface plate lapper come to one of my scraping seminars ( the hosts friend) and demonstrate how to test and lap a granite plate. He used set of electronic levels and a repeat-o-meter.
The level method is what the used before electronic levels were invented.
Watch the Swiss Wyler video on there site and they state this in the demo. Fred V. Fowler Company, Inc. - Precision Tools and Measuring Instruments
Repeat Reading Gage
As I said. Before Electronic levels were invented we use bubble levels. If your working in a laboratory doing to .00002" you my want to use a electronic level or one of the above products.
If you checking your shop grade A or B plate a level will do.
I have been site testing plates with levels for 40 years and my Dad did it that way for 40 years before me.
I specialize in scraping Jig Bores and super precision machines. I recently scraped 2 Drake USA), 1 Studer (Swiss) and 1 Shigiya (Japanese) CNC Cylindrical grinders. In Final tests we ground test bars to straightness less then a .0001", one of the Drakes was .00002" before the off-sets were done. I used their Grade B plate to scrape my cast iron straight-edge. I had checked with my level and found an area that was flat to .00005" and used that area to blue up my straight-edge. My point being You can use a level to check a plate! There are better ways, but not necessary.
It's a good way to find "local errors" on a plate and best place to check my cast iron straight-edges when I am at a job-site scraping a machine. The way I do it is a lay out the plate using a straight-edge and pencil. I draw straight lines on the plate 1/2 the length of the level I am using. Cross hatching it.
I level the plate both directions before I test the plate in the center of the plate. I also check to be sure the plate is sitting on 3 points. They need to be placed in the exact position they were when it was lapped at the factory. Some plate manufactures use 25% and others use 30% on the 2 point end and in the center on the single point end. There is rubber pads or a X where they had mounted the pad. I check the levels flatness and test it's positioning error by flipping it end to end and pushing down on the ends to be sure the level is not high in the middle.
I get a pad of 1/4" graph paper and layout the plate on the paper. Then start on one end and watch the level, mark it on the paper and move the level 1/2 its length, press down on the ends, wait 10 seconds and record my measurement, move down 6" and do it again, etc. and check it again all the way down the pencil line, then do all the lines and the cross lines. I also do diagonal corner to corner check too.
I just had a surface plate lapper come to one of my scraping seminars ( the hosts friend) and demonstrate how to test and lap a granite plate. He used set of electronic levels and a repeat-o-meter.
The level method is what the used before electronic levels were invented.
Watch the Swiss Wyler video on there site and they state this in the demo. Fred V. Fowler Company, Inc. - Precision Tools and Measuring Instruments
Repeat Reading Gage
As I said. Before Electronic levels were invented we use bubble levels. If your working in a laboratory doing to .00002" you my want to use a electronic level or one of the above products.
If you checking your shop grade A or B plate a level will do.
BigBob
Plastic
- Joined
- Jan 24, 2010
- Location
- USA - Verona, VA
Thanks, I will try the grid mapping and see how it looks.
So far I just moved the 0.0005/10in level 5 inches along a straight edge and back (got original repeat reading) in two places and see one graduation suggesting 0.001 in in ten inches.
I also used a B & S tenth indicator on a piece of steel around 6 inches long that has only a tenth variation either side slid under the same test indicator. When moving the stylus around my cast irong 10 x 15 plate (made by Jois and apparently an antique), I get under a tenth TIR. Flat??
Perhaps there is a large radius non-flatness, but the 8 inch level versus the 6 inch plate give me something to think about. I used the stylus extending away from the plate and right near it with the same results.
I did notice that just a slight rotation of the level gets me movement on the scale bubble. Preliminary leveling of the surface plate vs non flatness is also something I need to consider.
I will work on this some more. Great suggestions and I appreciate your inputs.
Bob
So far I just moved the 0.0005/10in level 5 inches along a straight edge and back (got original repeat reading) in two places and see one graduation suggesting 0.001 in in ten inches.
I also used a B & S tenth indicator on a piece of steel around 6 inches long that has only a tenth variation either side slid under the same test indicator. When moving the stylus around my cast irong 10 x 15 plate (made by Jois and apparently an antique), I get under a tenth TIR. Flat??
Perhaps there is a large radius non-flatness, but the 8 inch level versus the 6 inch plate give me something to think about. I used the stylus extending away from the plate and right near it with the same results.
I did notice that just a slight rotation of the level gets me movement on the scale bubble. Preliminary leveling of the surface plate vs non flatness is also something I need to consider.
I will work on this some more. Great suggestions and I appreciate your inputs.
Bob
RC99
Diamond
- Joined
- Mar 26, 2005
wow I would never have thought in a million years that a machinist level would be sensitive enough to check a surface plate... I thought it was all in the old days auto-collimator stuff and these days very sensitive electronic levels and laser interferometers
The real Leigh
Diamond
- Joined
- Nov 23, 2005
- Location
- Maryland
sfnative
Plastic
- Joined
- Dec 4, 2013
- Location
- San Francisco
A spirit level is as accurate a tool as you need
Electronic levels are just more convenient.
Spirit levels are self-proving by swapping end for end.
A spirit level can be calibrated to a known angle by using appropriate shims or gage blocks.
A spirit level can measure angles over different spans by attaching movable feet to the underside.
For the adventurous, even the sensitivity of the vial can be adjusted by flexing it.
That's a hint for making your own supra-precision level quite simply and with arbitrary resolution.
No reason to pay $1000+ for a pair of devices you can build in a few hours for $25.
Money aside, the real advantages of making your own are in the arbitrary resolution/accuracy and in the certainty of your calibration.
Electronic levels are just more convenient.
Spirit levels are self-proving by swapping end for end.
A spirit level can be calibrated to a known angle by using appropriate shims or gage blocks.
A spirit level can measure angles over different spans by attaching movable feet to the underside.
For the adventurous, even the sensitivity of the vial can be adjusted by flexing it.
That's a hint for making your own supra-precision level quite simply and with arbitrary resolution.
No reason to pay $1000+ for a pair of devices you can build in a few hours for $25.
Money aside, the real advantages of making your own are in the arbitrary resolution/accuracy and in the certainty of your calibration.
The real Leigh
Diamond
- Joined
- Nov 23, 2005
- Location
- Maryland
This is generally a discussion board for professionals engaged in business and following best business practices.
While we encourage participation by serious amateurs, that is not meant to encourage hobby-grade methods.
- Leigh
sable
Titanium
- Joined
- Feb 7, 2013
- Location
- midlands,UK
Anyone come across a "Fell" level ,I have one and it is basically a heavy duty Bullseye level which I am pretty sure is (was ) made for checking surface plates amongst other things.
sfnative
Plastic
- Joined
- Dec 4, 2013
- Location
- San Francisco
Grading methods
???
Ever shim a lathe bed?
Ever scrape out a dent in your mill table?
I suppose with the advent of CNC and regularly scheduled checks written to the calibration experts there is no reason to ever understand the foundations of accuracy and repeatability.
???
Ever shim a lathe bed?
Ever scrape out a dent in your mill table?
I suppose with the advent of CNC and regularly scheduled checks written to the calibration experts there is no reason to ever understand the foundations of accuracy and repeatability.
RayG
Hot Rolled
- Joined
- Feb 21, 2012
- Location
- Victoria, Australia
Here's how to calibrate a surface plate with sensitive differential levels. They typically cite sensitivities of 0.1 arc-second, which is a bit beyond the normal "precision levels", but there is no reason why you can't use machinist levels for flatness checking, it just depends on how accurate you want the result to be.
Gauging Quality, Episode 1 - Surface Plates - YouTube
And the Taylor Hobson Flatness Measurement of Surface Tables with Talyvel
0.1 arc-second equates to 0.48 microns per meter, so a level with a sensitivity of the order of 0.0005 mm/meter would suffice. ( 6 uinch/ft )
Most machinist levels are in the 0.02 - 0.05 mm/meter sensitivity range, which isn't sensitive enough to do a proper surface plate calibration.
Even if you had a sufficiently sensitive level you would still have to manually convert all the angles into elevations and map out the union-jack pattern to do a proper calibration...
But, when all is said and done, a machinist level can check for flatness, maybe not accurately enough for calibration perhaps, but probably good enough for a quick check?
One of these days I'd like to finish my Repeat-O-Meter project.. If anyone has built one, I'd be interested in seeing some details.
Ray
Gauging Quality, Episode 1 - Surface Plates - YouTube
And the Taylor Hobson Flatness Measurement of Surface Tables with Talyvel
0.1 arc-second equates to 0.48 microns per meter, so a level with a sensitivity of the order of 0.0005 mm/meter would suffice. ( 6 uinch/ft )
Most machinist levels are in the 0.02 - 0.05 mm/meter sensitivity range, which isn't sensitive enough to do a proper surface plate calibration.
Even if you had a sufficiently sensitive level you would still have to manually convert all the angles into elevations and map out the union-jack pattern to do a proper calibration...
But, when all is said and done, a machinist level can check for flatness, maybe not accurately enough for calibration perhaps, but probably good enough for a quick check?
One of these days I'd like to finish my Repeat-O-Meter project.. If anyone has built one, I'd be interested in seeing some details.
Ray
DMF_TomB
Diamond
- Joined
- Dec 13, 2008
- Location
- Rochester, NY, USA
a 0.0002" per 10" (0.002mm/1meter) is commonly sold (roughly 4 arc seconds for 1 full level division)
.
1) at 1/4 of a level division thats 0.00005" per 10" or 0.0005mm / meter is commonly measured
.
2) if you put a smaller block (1-2-3) block down and set level on it you can read an area 3" long
.
3) the surface plate must be level. if off level in one direction it will influence readings. this is why level vials generally have a up and down adjustment AND a side ways adjustment. a level with a Vee base on a cylinder will generally show different level readings as it sits off level at 90 degrees. thats why the little 2nd vial at 90 degrees to main vial is extremely important. generally the better levels (Swiss, German, etc) have the 2nd vial.
.
4) anything that shows finer than 0.0001" per 10" is pretty good. a surface not clean will show massive out of level difference. a piece of 0.001" shim per 10" is 5 full level divisions off (off scale).
.
4) optical lens resolution thats is the ability to distinguish individual lines rather than a grey blur I have measured many difference optical instruments at 1 to 4 arc seconds. 1 arc second is 0.0006" at 10 feet. i have seen many instruments than claim calibration to 1/10 arc second and yet if you measure a target at 10 feet repeatability of readings generally are no where near 1/10 a arc second.
1 arc second at 100 feet is 0.006" at 100 feet. if you look at a 1/100" ruler or scale at 100 feet i have never been able to see individual lines on the ruler let alone read it. repeatability of readings at 100 feet maybe 0.020" but 0.050" is more likely. most optical instrument makers exaggerate their calibration. if the crosshairs on the instrument is 4 arc seconds wide how can you measure less than 1/4 of a crosshair width ? (1 arc second). generally the crosshair is about 2x wider than optical resolution so the customer can see the crosshair easily. finer crosshairs are hard to see and read.
........ we use to call it the "bullshit" from the reasonable specs that can seen and test for
.
1) at 1/4 of a level division thats 0.00005" per 10" or 0.0005mm / meter is commonly measured
.
2) if you put a smaller block (1-2-3) block down and set level on it you can read an area 3" long
.
3) the surface plate must be level. if off level in one direction it will influence readings. this is why level vials generally have a up and down adjustment AND a side ways adjustment. a level with a Vee base on a cylinder will generally show different level readings as it sits off level at 90 degrees. thats why the little 2nd vial at 90 degrees to main vial is extremely important. generally the better levels (Swiss, German, etc) have the 2nd vial.
.
4) anything that shows finer than 0.0001" per 10" is pretty good. a surface not clean will show massive out of level difference. a piece of 0.001" shim per 10" is 5 full level divisions off (off scale).
.
4) optical lens resolution thats is the ability to distinguish individual lines rather than a grey blur I have measured many difference optical instruments at 1 to 4 arc seconds. 1 arc second is 0.0006" at 10 feet. i have seen many instruments than claim calibration to 1/10 arc second and yet if you measure a target at 10 feet repeatability of readings generally are no where near 1/10 a arc second.
1 arc second at 100 feet is 0.006" at 100 feet. if you look at a 1/100" ruler or scale at 100 feet i have never been able to see individual lines on the ruler let alone read it. repeatability of readings at 100 feet maybe 0.020" but 0.050" is more likely. most optical instrument makers exaggerate their calibration. if the crosshairs on the instrument is 4 arc seconds wide how can you measure less than 1/4 of a crosshair width ? (1 arc second). generally the crosshair is about 2x wider than optical resolution so the customer can see the crosshair easily. finer crosshairs are hard to see and read.
........ we use to call it the "bullshit" from the reasonable specs that can seen and test for
The real Leigh
Diamond
- Joined
- Nov 23, 2005
- Location
- Maryland
As to "regularly scheduled checks to the calibration experts"...
That's the way professionals do it.
You cannot certify your own facilities or equipment.
That must be done by a calibration lab using equipment and procedures traceable to NIST (nee NBS).
I assure you I'm well-versed in the "foundations of accuracy and repeatability".
- Leigh
Tyrone Shoelaces
Diamond
- Joined
- Apr 19, 2006
- Location
- Manchester, England
On a side issue anybody with a 0.0001" per 10 inch level had better have plenty of time and patience. I used to have one, it got used every blue moon. I sold it after a while. You had time to go for a coffee before the bubble settled. Regards Tyrone
DMF_TomB
Diamond
- Joined
- Dec 13, 2008
- Location
- Rochester, NY, USA
level sensitivity
i have seen plenty of 0.0002" per 10" (0.02mm per m) in China. you are right they are sensitive. for rough leveling i always start with a 0.004" per 10" (80 arc second) 0.005"/foot level. At this rough level can measure 1/4 level division or 0.001" per 10" it is the primary level to rough in equipment. it's range of 0.001" per 10" to 0.016" per 10" is a practical rough leveling range.
.
a 0.0002" per 10" level can easily be off scale at 0.001" per 10" or you cannot measure more than 5 lines off.
.
i have a square frame level where i added a 0.004" per 10" level vial in addition to the 0.0002" per 10" level vial. That way the same level could read 0.016" per 10" down to 0.00005" per 10".
.
the 0.0002" per 10" vial is too sensitive most times. the heat from a flashlight can easily move the bubble. the slightest spec of dust can throw the reading off scale. yes the level bubble can be slow responding. i have read in the moore book that having a room with air temperature of 1/10 of a degree F and letting objects to be measured sit for a day so the become even temperature are basic requirements when trying to work in the millionths or even 10's of millions of an inch
.
i have seen plenty of 0.0002" per 10" (0.02mm per m) in China. you are right they are sensitive. for rough leveling i always start with a 0.004" per 10" (80 arc second) 0.005"/foot level. At this rough level can measure 1/4 level division or 0.001" per 10" it is the primary level to rough in equipment. it's range of 0.001" per 10" to 0.016" per 10" is a practical rough leveling range.
.
a 0.0002" per 10" level can easily be off scale at 0.001" per 10" or you cannot measure more than 5 lines off.
.
i have a square frame level where i added a 0.004" per 10" level vial in addition to the 0.0002" per 10" level vial. That way the same level could read 0.016" per 10" down to 0.00005" per 10".
.
the 0.0002" per 10" vial is too sensitive most times. the heat from a flashlight can easily move the bubble. the slightest spec of dust can throw the reading off scale. yes the level bubble can be slow responding. i have read in the moore book that having a room with air temperature of 1/10 of a degree F and letting objects to be measured sit for a day so the become even temperature are basic requirements when trying to work in the millionths or even 10's of millions of an inch
DMF_TomB
Diamond
- Joined
- Dec 13, 2008
- Location
- Rochester, NY, USA
level picture
picture of level i added a less sensitive level vial too.
.
the 0.004" per 10" level vial has adjustment knob pushing against oring. calibration by reversing level takes about 10 seconds time.
........ this roughing level vial get part close (0.001" per 10"), then and only then is main level vial at 0.0002" per 10" read
picture of level i added a less sensitive level vial too.
.
the 0.004" per 10" level vial has adjustment knob pushing against oring. calibration by reversing level takes about 10 seconds time.
........ this roughing level vial get part close (0.001" per 10"), then and only then is main level vial at 0.0002" per 10" read
Attachments
Tyrone Shoelaces
Diamond
- Joined
- Apr 19, 2006
- Location
- Manchester, England
On the other hand I once got to borrow an " Hilger & Watts " electronic level. You could use it in two executions, 1) 0.001" in 10 inches or 2) 0.0001" in 10 inches. This was done with a flick of a switch. As I recall the readings were almost instant in both executions.
Regards Tyrone.
DMF_TomB
Diamond
- Joined
- Dec 13, 2008
- Location
- Rochester, NY, USA
precision level
the precision frame level i bought in China for $20, normally they are $30-$40 I believe the store was not selling a lot of high precision stuff and wanted to sell it at cost.
.
the USA Starrett level i added to it cost i believe $80.
.
The one i saw in China were ok in quality and could work in colder shops. many machine shops are not heated and i have seen plenty that were close to freezing 0 Celsius. i am not selling they could do the best precision work that way but many were basically a garage type shop with a big roll up front door.
....... many electronic tools are not sold their as much as other countries, as they do not work well in cold weather. i saw no electronic levels in 2008 other than carpenter levels with a laser.
........ the lower cost in master precision levels means they could buy one often. the slightest rust or damage and they could get a new one. it is my belief that precision spirit levels do not cost that much to make but are marked up to get a high profit for items that are sold in such low quantities.
......... i was able to buy a 4", 8", and 12" long block levels and the frame level all in total for less than the a electronic level and all were 0.05mm/m to 0.02mm/m sensitivity. to be honest i valued the 0.05mm/m levels as more useful than a 0.02mm/m or a 0.01mm/m level.
.
the precision frame level i bought in China for $20, normally they are $30-$40 I believe the store was not selling a lot of high precision stuff and wanted to sell it at cost.
.
the USA Starrett level i added to it cost i believe $80.
.
The one i saw in China were ok in quality and could work in colder shops. many machine shops are not heated and i have seen plenty that were close to freezing 0 Celsius. i am not selling they could do the best precision work that way but many were basically a garage type shop with a big roll up front door.
....... many electronic tools are not sold their as much as other countries, as they do not work well in cold weather. i saw no electronic levels in 2008 other than carpenter levels with a laser.
........ the lower cost in master precision levels means they could buy one often. the slightest rust or damage and they could get a new one. it is my belief that precision spirit levels do not cost that much to make but are marked up to get a high profit for items that are sold in such low quantities.
......... i was able to buy a 4", 8", and 12" long block levels and the frame level all in total for less than the a electronic level and all were 0.05mm/m to 0.02mm/m sensitivity. to be honest i valued the 0.05mm/m levels as more useful than a 0.02mm/m or a 0.01mm/m level.
