'Didn't know about the 1st, 2nd, and 3rd cut taps. I've always used just one tap for through-holes, followed by a bottoming tap for blind holes. Doing it in steps makes sense and, as a bonus, would probably eliminate a lot of tap breakage.
For hand drills, I center punch, then use a drill a lot like the one Phil showed. If the hole is small, I'll chuck up most of the drill, leaving only a small portion exposed. This helps considerably with snapping and bit moving around before the prime cut is established. On metal, I'll almost always use a pilot hole, unless the hole precision requirement is lax. On plastics, I like to use a hot needle to pierce a pilot, prior to drilling. A nail, or other object works well too, depending on hole size.
Deffo do not tap dry, unless you are in a material that lends itself to self-tapping fasteners. Tap block recommended too, as the comments for being straight are dead on. Easy to snap. Incrementally tap, backing out to remove chips, applying cutting oil liberally. With taps, it's important to get the right size pilot hole to avoid shallow threads, or breakage of the tap.
For thin materials, such as sheet metal, I use a Unibit (step drill):
The reason is that, with a regular drill, the flutes will catch on the edge of the hole, distorting it or -- worse -- bending the workpiece. A Unibit will not catch and is much safer to use.
Contrary to public opinion WD-40 is NOT a lubricant.
The same company sells 3-in-One, so they got that covered. In any case, feel free to buy WD-40, Carpet Fresh, Lava soap, and 3-in-One by the case loads. Anything to boost the San Diego economy.
Contrary to public opinion WD-40 is NOT a lubricant.
Not only is it not a lubricant, it leaves a residue which can form a rubbery coating if applied repeatedly. The service techs at a company I worked for many years ago would spray WD40 all over the innards of any teletype that occasionally misprinted. This would cure the problem for a while, but eventually the gummy coating got so thick that it would stop working and no amount of WD40 would get it going again. Had to remove all the plastic parts and dip the entire unit in a solvent to strip off the coating.
Have you measured dimentions on the paper before and after gluing the paper to the metal?
All dimensions are definitely measured after printing (before gluing), as for measuring the templates after gluing, I am sure that I have on several occassions, but I don't remember the results. However, let me put your mind at ease, any problems that I have had with using templates, it was always my fault, due to inadequate alignment. That problem has been remedied. I notice that on many occassions, it was much easier to align the wiggler with cross marks as compared to others. I initially thought that the wiggler should be very close to the metal, and I always adjusted it by hand. However it is actually much better to have the wiggler a little ways away from the template surface to allow easy viewing of the "point". This is the reason that I am now using a feeler gauge to space the point tip to a clearance of 0.027". This clearance has worked very well for my current project. I have drilled many more holes since my last posting, and upto this point, all holes have been dead center to the naked eye, and alignment between mating surfaces is working out perfectly.
Do you have a particular brand of spray glue you like best?
That is a very good question. There really is not a very wide selection of spray on adhesives at my local hardware stores. When I initially started using templates, I used DAP Weldwood Multi Purpose Spray Adhesive, but the local hardware stores quit carrying it, so I had to switch brands. If I remember correctly, Weldwood was a lot easier to remove than the product I currently use, which is, Loctite Spray Adhesive General Performance. If I could find it locally, I would go back to Weldwood.
I use Krylon Easy-Tack when it's something I want to be able to remove easily without leaving a mess to clean up. Most hardware stores should have it.
Thanks for the tip. I never knew that Krylon manufactured spray glue. I will have to thoroughly check the Krylon paint products and buy a can if I can find it, just to give it a test run.
My brass holes did not turn out nearly as nice as my aluminum holes. However, in defense of my drilling methods, I still have not learned the fine intricacies of drilling brass.
idbruce, from this point on I will create a printable template. It makes so much sense that I don't know why I never did it before. Even though I use a digital vernier caliper I still have a problem with elongated holes
I think it will make tasks repeatable.
If you do print a template, start with unit references placed at the extents of the printable area. 1" squares are great. Print the stuff, then measure the image, ideally using an optical comparator, but at the least using a precision caliper and a lens.
Measure line thickness in both directions. This will vary by some small amount. Note it.
Measure the size of the squares in both directions. Work from the same side of the line, so your measurement is actually true to the square, not biased by the line width. eg: Horizontal measurement would be from the left side of the left line to the left side of the right line.
If using an optical comparitor, be sure and place it flat against the paper and view it with the centroid of your eye in line with it's center line axis. (straight on) If using calipers, use either the small opposing tips, or the inside of the longer blades. If using the tips, you can hold the caliper at a slight angle, placing the tips directly onto the paper. View through lens, normal to the paper (straight on again), using the outer edge of the tips against the inner edge of the lines, so you can know when the tip is occluding the line.
You will see a radius on the caliper tip, depending on wear and quality of manufacture. Use the edge beneath the radius, and align it with the edge of the printed line, again looking for occlusions to insure parallism and a precise measurement.
If using the blade of the caliper, work in the opposite fashion. Depending on your caliper blade tips, you might find the tips acceptable, but the longer edge surface is best. Place the outer edge of the line against the inside edge of the caliper, aligning as described above. Parallax on this is generally a bit more severe, due to the blades being thick. Move your eye across a small arc to insure the edge is touching the line, but not occluding it.
Analog calipers should be origined with a measurement reference prior to doing this. Perform a few repeat measurements, exercising about the scale you need for the paper against known sources. There is a human factor here you need to deal with, if you expect precision below .003 - .005"
Veriner types are the best for this, IMHO. They can be interpolated below .001, delivering a very accurate scale factor, particularly if the measurement is captured, then evaluated with the optical comparitor. (an old QA showed me that trick, and it's golden!! --Thanks Pete True. --Yes his name was true, and he was one hell of an inspector.)
Second best is 3.5 digit digitals, verified across a few precision reference so variances in their scale are known. Cheap ones will have variances. Know them, and it's not an issue.
Measurements performed this way and averaged can get you down to .002" or so, depending on all factors. That's as good as it gets without better gear.
Note these.
Measure the distances between the squares, note these.
Using one square as origin, do the math to establish what the dimensions are.
Print a couple more, perhaps moving the squares some to note differences and verify scaling is uniform on the print device.
Compare dimensions to the source digital data. There will be some variances. Divide to get scaling ratios.
Modify the source geometry with those scale ratios and print again. These measurements should fall within the noise level of your measuring tools, and that's as good as it gets.
Printers vary. Drum style printers will be consistent and repeatable to one dot in the direction of the imaging device. Generally, these scale 1:1, with the software not getting in the way, of course. In the direction of the drum, they will often stretch the image by a few percent. It's not much, but it will easily add up to .010" before you even know it!
Inkjet types have little variances all over the place, depending on the make and size.
Using this technique, it really doesn't matter how good your printer is. Even a low DPI device can be used to a fairly good precision. You draw what you want in your CAD / Illustration software. Then scale it to account for your device variances.
The result should be prints that are repeatable to a few thousandths of an inch, depending on dot pitch and printer paper advance repeatability characteristics.
If you can find a good pen plotter, use it! Those things can be scary good, sometimes repeatable to .001 or .002"
In my manufacturing time, I've done a ton of reverse engineering from plots, camera photos, scanned images (ugly), etc... This process rocks, and you will get printable things that very likely are more accurate and repeatable than you, or your machine is!
Most printers are kind of poor when printing to absolute paper space. The image will appear somewhere relative to one corner of the sheet. The image itself will very likely be significantly improved over that. This means you never count on the paper edge. Always cut, or pierce a hole through a known origin, ideally using registration marks for your board, or target object. I like to overcut a little, then use a light, or over-underlay technique to insure proper registration. This, of course, varies with the manufacturing conditions and target materials.
***A lot of you may remember the story about the old shop, pool games and cokes! Applying these general ideas to manufacturing processes can yield very surprising results!! The lesson I learned is the most important thing to know is what the process variances are. Once those are known, all sorts of compensation can be applied to make a fairly mediocre device or process yield high precision and repeatable results.
My measuring tools of choice are the optical comparitor, veriner calipers. From there, I'll downgrade to a nice set of digitals, and then there is the trusty analog caliper here on my desk. It's the most difficult, because it has the highest variances where one literally needs to learn to measure with it, and maintain that skill. Very important to exhibit consistency in applying measuring tools, or ones own failure to demonstrate repeatability will ripple through whatever it is we happen to be trying to do.
Edit: One other thing I want to share. When performing measurements, get out of the habit of monitoring the display. Employ the same steps, looking for the same visuals, and or tactile responses for those devices that have them in amounts you worry about. Then look at the measurement, THEN consult the source.
We have a basic tendency to bias measurements to a result. Better to factor that out, repeating and averaging to get at the real data and variances.
Known references are typically precision machined blocks and tables, with surface variances in the fourth digit. .000x
You are scaring me When I used the phrase "Very Accurate" in the title, it was not my intention to strive for Ultra High Precision However, I am sure it can be obtained through the use of templates. If you drill your holes for loose fits, the process is much simpler.
I find it very useful to establish measurements basis that are solid. One can always downgrade, or account for stack-ups of various kinds. That's easy. What isn't always so easy is managing variances one isn't aware of, which is why I posted up the detail on the templates I did.
And there are people here who haven't done much. Little things, like where to measure from, what factors are involved, etc... might be quite informative. They were when I experienced them.
For the stuff we are discussing here, that process is likely a bit overkill, but maybe not. Depends. I used it a lot when reverse engineering things. And for mass checking of flat parts, a well sorted pen plotter, nice vellum and it's good to less than .010"
These days, we can scan to less than that, which is the process I would very likely employ today. Some of this is relative. When machining, I was regularly working under .005" With precision sheet metal, it varied from .060" to .010" -- .005" at times. (that's tough actually, if the part is bent)
To me, .060" is "tolerable" .030" accurate, and .010" and below, "very accurate" The aerospace and auto people I know would characterize it very differently, as would wood workers. YMMV
Edit: The investment here is basically a one time "lock in" on understanding the factors associated with the printer. Once done, constants ripple up through the other processes. Repeat those across the processes, compute stack-up potentials, and you have boundaries on repeatability. The numbers can be quite large quite easily! Human factors multiply over that too.
Expectations beneath those will require process refinement. Knowing where the "noise level" is means being able to match expectations to processes without emperical means consuming time and effort, potentially wasted if requirements / expectations land in the noise. Doing it all the way through also means eliminating "chasing your own tail" which can be a considerable time and materials savings, particularly when ramping up on new tasks. We are quite literally forced to understand the machines we use, which always pays off.
As you know I was just teasing you a little. Actually it was a very nice and informative addition to the thread. Instead of teasing you, I should have been thanking you for adding such quality information to the thread.
LOL On text, we sometimes cannot tell. No worries.
I should add I really enjoy the art of making things. So many techniques one can apply, and physical dexterity and skill factor in too. Your antics on building are something I enjoy reading here.
Edit: Here's something you may find entertaining Bruce. Why develop skills and processes like the ones I detailed here? Boeing. They have levels of certification for just about everything. Maybe everything. I don't know. What I do know is I ended up on a contract to build some wing support sections. That year I had convinced the shop owners to invest in a laser cutter to expand scope of work potential and to realize some very serious cost savings over the work horse turret punches.
Once we had gotten data transfer and G-code particulars sorted out, that thing got sales jazzed and they went out and literally got fist fulls of work! (That old school owner did the "here's some cash" thing, as he regularly did too.) So we end up with that contract, and there was no data! I phoned 'em up, the quote having been done on a rough size and material spec, asking for the detail so I could set things up for the guys to make parts.
No worries they said, we will ship you a part spec. I thought, "ship it? WTF?", but said thanks as they were in no hurry. A week later, I got this big tube dropped on my desk. "It's from those Boeing guys." The Manufacturing engineering group gathered around, all of us kind of grinning at the wonder of what was in that tube. Most people were sending us CAD files of some sort, the rest detail drawings. All ordinary. What did the Boeing guys send?
Size J 1:1 scale images of the part. They were developed in a CAD system, and photo-transferred to the rolled up drawings we got. They were huge! Something like 5' in length, 3' in the other direction. Not a number on them. Center marks, all the little drafting details were there, but no dimensions! I called them and asked about that, and they said we were not certified to use the data. Crazy! I told them it was going to cost a little extra to scale and develop the manufacturing data from the images. They said, "fine", or we could spend a week to gain certification. I didn't even want to ask at that point.
So, they got taped to a conference room table, and we got after it, using tacks to put little dents in the images, calipers, squares, and the usual geometry tools one uses for old school layout and planning. Took a few days and we had input the curves back into the CAD system. How to check before burning material? Enter that trusty pen plotter.
That process I wrote above ended up being the prototype for a whole bunch of goofy stuff we got, including a gasket one buyer slapped on a photo copier, then stuffed through the low res FAX, with an arrow pointing to it, "QTY 3 ASAP, Thanks!", LOL!! I still have that one, just because it's classic!
We ended up scaling every 2D device in the building, and tossed the metrics into a file so we could "build it no matter what or how we got an image of it", developing quite the niche. Faxes, photocopies, photos (those are rough, because of the perspective, and nobody takes one normal to the item), plots, bitmaps, hand "to scale" drawings, you name it.
The Boeing guys paid an extra $1000 or so on a part run that was about $250 otherwise, had they sent us 2D CAD data. Sheesh.
After calibrating the pen plotter, we plotted the CAD out with registration points as there was not a flat spot on the wing sections, aligned them, then followed the curve with a lens to catch variances. The tolerance? If our line was inside their line completely, the part was good, LOL!!! Our plotter took sheets, not roll feed, so we had to register the plots to one another, and the reference marks on the image we got. Thankfully, they did provide those.
The QC guys would have none of it, expecting that dimensional spec contract, so we verified the things, cut 'em on the laser and shipped them out, after physically laying them on the images we got, with the same instructions. If the metal touches the entire line, ship it.
So there you go. Funny how things end up sometimes. I am, on occasion, disturbed somewhat when I fly on an older airplane. Boeing has long since changed and updated their practices. CAD is the norm today, but it wasn't then.
Thanks for sharing, because I did find it interesting.
Before a good friend of mine passed away, he would always beat this phrase into my head:
Adapt and overcome!
It is truly amazing the number of solutions that can be acquired with a little thought process.
Our plotter took sheets, not roll feed, so we had to register the plots to one another, and the reference marks on the image we got.
I never had to do this from plotter sheets, but I once designed an apparatus that was approximately 36" X 48" and I needed a 1:1 drawing, so I printed multiple 8-1/2" X 11" sheets with registration marks and taped them all together. Surprisingly, it turned out much better than I anticipated.
That process I wrote above ended up being the prototype for a whole bunch of goofy stuff we got, including a gasket one buyer slapped on a photo copier, then stuffed through the low res FAX, with an arrow pointing to it, "QTY 3 ASAP, Thanks!"
YEP! That is a classic. EDIT: LOL. Sounds like something I might have done 30 years ago.
For those working with most soft aluminum alloys, they will undoubtedly be aware of the 'smear' that occurs when drilling, tapping, sawing and especially filing. My experience has been that this smear can be pretty much eliminated by simply using some methanol on your cutting tool, saw or drill. Squirting a bead of methanol on your file totally keeps it from filling with aluminum smear. Nice thing too, methanol is inexpensive, readily available, and does not leave much residue, whereas the cutting oils do.
I did a little more experimenting with the clearance between the wiggler and the material being drilled, and it appears as though 0.043" clearance is much better than 0.027".
Potatohead's story reminded me of my own experience with an aviation company and drawings....way back when I was 14 or so, I was really in love with the North American P-51 Mustang. I chose it as a topic for a school report....this was way before the Internets (1974). In my research, I found that North American had become Rockwell Aviation. I wrote a (what must have been) glowing letter to their marketing department telling them of my passion for the Mustang and asked if they might be able to send me some drawings and information for my report. After a couple of weeks had passed, two large (12x12x12), heavy boxes arrived on my parent's porch from the North American Division of Rockwell Aviation. Full sized BLUEPRINTS folded up in12x12 panels like roadmaps!! They had sent me full-sized copies of the remaining drawings they had on file from the early 1940's. I had full size drawings of one of the wings, the BOM and several large drawings of sub assemblies related to the wings. The second box contained some smaller 3-view drawings of the plane and a 2-3 inch stack of 8x10 pictures from the assembly line, the flight test line and the Army proving grounds. There were also a couple metal placards and tags from various spots on the aircraft. Needless to say, my final school project was very successful thanks to some kind soul at North American Rockwell. Sadly when my mom sold her house after I had left the nest, it never occurred to me to grab all those goodies.
Okay, it has been six years since I started this thread, and my method for drilling accurate holes is still basically the same, but a little has changed. However, since I have changed my method, my holes are now more accurate than ever. I am bringing this subject back up, because I am building a new machine, and every hole has been in perfect alignment with the intended target.
As mentioned, my method is almost the same as posted at the beginning of the thread, but here is what has changed....
1. Distance from template to bottom point of wiggler 0.015"
2. I no longer use a torch to help with template removal. Simply wet the template with a paper towel that is soaked with laquer thinner and the template will peel off easily.
3. After dimpling the template and metal with the tile cutting bit, I then use a modified DeWALT 1/4" pilot point drill bit to start all of my holes, which is basically the same as Phil's suggestion of using a countersink, but much cheaper.
The DeWALT (DWA1216) pilot point drill bit has patented core technology, which makes it more stout than a common 1/4" bit. I modify this bit, by cutting off the no slip portion of the shank, which makes it shorter and more sturdy, and makes it is just about perfect to work alongside the wiggler.
This modified bit easily centers itself within the dimple created by the tile cutting bit, creating a new dimple, in the metal, with drill bit geometry. Dependent upon the hole size to be drilled, I will either simply dimple with the pilot portion of the bit, or I will continue to drill until the 1/4" portion starts, or I will start with a 1/4" pilot hole, for much larger drilling.
EDIT: I should add that this method provides more accurate holes than starting a hole with a $20 1/4" Carbide Hole-Starting Drill Bit.
Comments
Jim
-Phil
Deffo do not tap dry, unless you are in a material that lends itself to self-tapping fasteners. Tap block recommended too, as the comments for being straight are dead on. Easy to snap. Incrementally tap, backing out to remove chips, applying cutting oil liberally. With taps, it's important to get the right size pilot hole to avoid shallow threads, or breakage of the tap.
The reason is that, with a regular drill, the flutes will catch on the edge of the hole, distorting it or -- worse -- bending the workpiece. A Unibit will not catch and is much safer to use.
-Phil
The same company sells 3-in-One, so they got that covered. In any case, feel free to buy WD-40, Carpet Fresh, Lava soap, and 3-in-One by the case loads. Anything to boost the San Diego economy.
Oh, and be sure to eat at Jack In the Box.
-- Gordon
Not only is it not a lubricant, it leaves a residue which can form a rubbery coating if applied repeatedly. The service techs at a company I worked for many years ago would spray WD40 all over the innards of any teletype that occasionally misprinted. This would cure the problem for a while, but eventually the gummy coating got so thick that it would stop working and no amount of WD40 would get it going again. Had to remove all the plastic parts and dip the entire unit in a solvent to strip off the coating.
Hmm, I wonder if that is why fish love it!
All dimensions are definitely measured after printing (before gluing), as for measuring the templates after gluing, I am sure that I have on several occassions, but I don't remember the results. However, let me put your mind at ease, any problems that I have had with using templates, it was always my fault, due to inadequate alignment. That problem has been remedied. I notice that on many occassions, it was much easier to align the wiggler with cross marks as compared to others. I initially thought that the wiggler should be very close to the metal, and I always adjusted it by hand. However it is actually much better to have the wiggler a little ways away from the template surface to allow easy viewing of the "point". This is the reason that I am now using a feeler gauge to space the point tip to a clearance of 0.027". This clearance has worked very well for my current project. I have drilled many more holes since my last posting, and upto this point, all holes have been dead center to the naked eye, and alignment between mating surfaces is working out perfectly.
That is a very good question. There really is not a very wide selection of spray on adhesives at my local hardware stores. When I initially started using templates, I used DAP Weldwood Multi Purpose Spray Adhesive, but the local hardware stores quit carrying it, so I had to switch brands. If I remember correctly, Weldwood was a lot easier to remove than the product I currently use, which is, Loctite Spray Adhesive General Performance. If I could find it locally, I would go back to Weldwood.
Bruce
-Phil
Just don't get caught by the fish and game. It's illegal here in Utah
Thanks for the tip. I never knew that Krylon manufactured spray glue. I will have to thoroughly check the Krylon paint products and buy a can if I can find it, just to give it a test run.
Bruce
My brass holes did not turn out nearly as nice as my aluminum holes. However, in defense of my drilling methods, I still have not learned the fine intricacies of drilling brass.
Bruce
I think it will make tasks repeatable.
Measure line thickness in both directions. This will vary by some small amount. Note it.
Measure the size of the squares in both directions. Work from the same side of the line, so your measurement is actually true to the square, not biased by the line width. eg: Horizontal measurement would be from the left side of the left line to the left side of the right line.
If using an optical comparitor, be sure and place it flat against the paper and view it with the centroid of your eye in line with it's center line axis. (straight on) If using calipers, use either the small opposing tips, or the inside of the longer blades. If using the tips, you can hold the caliper at a slight angle, placing the tips directly onto the paper. View through lens, normal to the paper (straight on again), using the outer edge of the tips against the inner edge of the lines, so you can know when the tip is occluding the line.
You will see a radius on the caliper tip, depending on wear and quality of manufacture. Use the edge beneath the radius, and align it with the edge of the printed line, again looking for occlusions to insure parallism and a precise measurement.
If using the blade of the caliper, work in the opposite fashion. Depending on your caliper blade tips, you might find the tips acceptable, but the longer edge surface is best. Place the outer edge of the line against the inside edge of the caliper, aligning as described above. Parallax on this is generally a bit more severe, due to the blades being thick. Move your eye across a small arc to insure the edge is touching the line, but not occluding it.
Analog calipers should be origined with a measurement reference prior to doing this. Perform a few repeat measurements, exercising about the scale you need for the paper against known sources. There is a human factor here you need to deal with, if you expect precision below .003 - .005"
Veriner types are the best for this, IMHO. They can be interpolated below .001, delivering a very accurate scale factor, particularly if the measurement is captured, then evaluated with the optical comparitor. (an old QA showed me that trick, and it's golden!! --Thanks Pete True. --Yes his name was true, and he was one hell of an inspector.)
Second best is 3.5 digit digitals, verified across a few precision reference so variances in their scale are known. Cheap ones will have variances. Know them, and it's not an issue.
Measurements performed this way and averaged can get you down to .002" or so, depending on all factors. That's as good as it gets without better gear.
Note these.
Measure the distances between the squares, note these.
Using one square as origin, do the math to establish what the dimensions are.
Print a couple more, perhaps moving the squares some to note differences and verify scaling is uniform on the print device.
Compare dimensions to the source digital data. There will be some variances. Divide to get scaling ratios.
Modify the source geometry with those scale ratios and print again. These measurements should fall within the noise level of your measuring tools, and that's as good as it gets.
Printers vary. Drum style printers will be consistent and repeatable to one dot in the direction of the imaging device. Generally, these scale 1:1, with the software not getting in the way, of course. In the direction of the drum, they will often stretch the image by a few percent. It's not much, but it will easily add up to .010" before you even know it!
Inkjet types have little variances all over the place, depending on the make and size.
Using this technique, it really doesn't matter how good your printer is. Even a low DPI device can be used to a fairly good precision. You draw what you want in your CAD / Illustration software. Then scale it to account for your device variances.
The result should be prints that are repeatable to a few thousandths of an inch, depending on dot pitch and printer paper advance repeatability characteristics.
If you can find a good pen plotter, use it! Those things can be scary good, sometimes repeatable to .001 or .002"
In my manufacturing time, I've done a ton of reverse engineering from plots, camera photos, scanned images (ugly), etc... This process rocks, and you will get printable things that very likely are more accurate and repeatable than you, or your machine is!
Most printers are kind of poor when printing to absolute paper space. The image will appear somewhere relative to one corner of the sheet. The image itself will very likely be significantly improved over that. This means you never count on the paper edge. Always cut, or pierce a hole through a known origin, ideally using registration marks for your board, or target object. I like to overcut a little, then use a light, or over-underlay technique to insure proper registration. This, of course, varies with the manufacturing conditions and target materials.
***A lot of you may remember the story about the old shop, pool games and cokes! Applying these general ideas to manufacturing processes can yield very surprising results!! The lesson I learned is the most important thing to know is what the process variances are. Once those are known, all sorts of compensation can be applied to make a fairly mediocre device or process yield high precision and repeatable results.
My measuring tools of choice are the optical comparitor, veriner calipers. From there, I'll downgrade to a nice set of digitals, and then there is the trusty analog caliper here on my desk. It's the most difficult, because it has the highest variances where one literally needs to learn to measure with it, and maintain that skill. Very important to exhibit consistency in applying measuring tools, or ones own failure to demonstrate repeatability will ripple through whatever it is we happen to be trying to do.
Edit: One other thing I want to share. When performing measurements, get out of the habit of monitoring the display. Employ the same steps, looking for the same visuals, and or tactile responses for those devices that have them in amounts you worry about. Then look at the measurement, THEN consult the source.
We have a basic tendency to bias measurements to a result. Better to factor that out, repeating and averaging to get at the real data and variances.
Known references are typically precision machined blocks and tables, with surface variances in the fourth digit. .000x
You are scaring me
Bruce
And there are people here who haven't done much. Little things, like where to measure from, what factors are involved, etc... might be quite informative. They were when I experienced them.
For the stuff we are discussing here, that process is likely a bit overkill, but maybe not. Depends. I used it a lot when reverse engineering things. And for mass checking of flat parts, a well sorted pen plotter, nice vellum and it's good to less than .010"
These days, we can scan to less than that, which is the process I would very likely employ today. Some of this is relative. When machining, I was regularly working under .005" With precision sheet metal, it varied from .060" to .010" -- .005" at times. (that's tough actually, if the part is bent)
To me, .060" is "tolerable" .030" accurate, and .010" and below, "very accurate" The aerospace and auto people I know would characterize it very differently, as would wood workers. YMMV
Edit: The investment here is basically a one time "lock in" on understanding the factors associated with the printer. Once done, constants ripple up through the other processes. Repeat those across the processes, compute stack-up potentials, and you have boundaries on repeatability. The numbers can be quite large quite easily! Human factors multiply over that too.
Expectations beneath those will require process refinement. Knowing where the "noise level" is means being able to match expectations to processes without emperical means consuming time and effort, potentially wasted if requirements / expectations land in the noise. Doing it all the way through also means eliminating "chasing your own tail" which can be a considerable time and materials savings, particularly when ramping up on new tasks. We are quite literally forced to understand the machines we use, which always pays off.
As you know I was just teasing you a little. Actually it was a very nice and informative addition to the thread. Instead of teasing you, I should have been thanking you for adding such quality information to the thread.
So thank you!
Bruce
I should add I really enjoy the art of making things. So many techniques one can apply, and physical dexterity and skill factor in too. Your antics on building are something I enjoy reading here.
Edit: Here's something you may find entertaining Bruce. Why develop skills and processes like the ones I detailed here? Boeing. They have levels of certification for just about everything. Maybe everything. I don't know. What I do know is I ended up on a contract to build some wing support sections. That year I had convinced the shop owners to invest in a laser cutter to expand scope of work potential and to realize some very serious cost savings over the work horse turret punches.
Once we had gotten data transfer and G-code particulars sorted out, that thing got sales jazzed and they went out and literally got fist fulls of work! (That old school owner did the "here's some cash" thing, as he regularly did too.) So we end up with that contract, and there was no data! I phoned 'em up, the quote having been done on a rough size and material spec, asking for the detail so I could set things up for the guys to make parts.
No worries they said, we will ship you a part spec. I thought, "ship it? WTF?", but said thanks as they were in no hurry. A week later, I got this big tube dropped on my desk. "It's from those Boeing guys." The Manufacturing engineering group gathered around, all of us kind of grinning at the wonder of what was in that tube. Most people were sending us CAD files of some sort, the rest detail drawings. All ordinary. What did the Boeing guys send?
Size J 1:1 scale images of the part. They were developed in a CAD system, and photo-transferred to the rolled up drawings we got. They were huge! Something like 5' in length, 3' in the other direction. Not a number on them. Center marks, all the little drafting details were there, but no dimensions! I called them and asked about that, and they said we were not certified to use the data. Crazy! I told them it was going to cost a little extra to scale and develop the manufacturing data from the images. They said, "fine", or we could spend a week to gain certification. I didn't even want to ask at that point.
So, they got taped to a conference room table, and we got after it, using tacks to put little dents in the images, calipers, squares, and the usual geometry tools one uses for old school layout and planning. Took a few days and we had input the curves back into the CAD system. How to check before burning material? Enter that trusty pen plotter.
That process I wrote above ended up being the prototype for a whole bunch of goofy stuff we got, including a gasket one buyer slapped on a photo copier, then stuffed through the low res FAX, with an arrow pointing to it, "QTY 3 ASAP, Thanks!", LOL!! I still have that one, just because it's classic!
We ended up scaling every 2D device in the building, and tossed the metrics into a file so we could "build it no matter what or how we got an image of it", developing quite the niche. Faxes, photocopies, photos (those are rough, because of the perspective, and nobody takes one normal to the item), plots, bitmaps, hand "to scale" drawings, you name it.
The Boeing guys paid an extra $1000 or so on a part run that was about $250 otherwise, had they sent us 2D CAD data. Sheesh.
After calibrating the pen plotter, we plotted the CAD out with registration points as there was not a flat spot on the wing sections, aligned them, then followed the curve with a lens to catch variances. The tolerance? If our line was inside their line completely, the part was good, LOL!!! Our plotter took sheets, not roll feed, so we had to register the plots to one another, and the reference marks on the image we got. Thankfully, they did provide those.
The QC guys would have none of it, expecting that dimensional spec contract, so we verified the things, cut 'em on the laser and shipped them out, after physically laying them on the images we got, with the same instructions. If the metal touches the entire line, ship it.
So there you go. Funny how things end up sometimes. I am, on occasion, disturbed somewhat when I fly on an older airplane. Boeing has long since changed and updated their practices. CAD is the norm today, but it wasn't then.
Thanks for sharing, because I did find it interesting.
Before a good friend of mine passed away, he would always beat this phrase into my head:
I never had to do this from plotter sheets, but I once designed an apparatus that was approximately 36" X 48" and I needed a 1:1 drawing, so I printed multiple 8-1/2" X 11" sheets with registration marks and taped them all together. Surprisingly, it turned out much better than I anticipated.
YEP! That is a classic.
Bruce
For those working with most soft aluminum alloys, they will undoubtedly be aware of the 'smear' that occurs when drilling, tapping, sawing and especially filing. My experience has been that this smear can be pretty much eliminated by simply using some methanol on your cutting tool, saw or drill. Squirting a bead of methanol on your file totally keeps it from filling with aluminum smear. Nice thing too, methanol is inexpensive, readily available, and does not leave much residue, whereas the cutting oils do.
Just sayin'
Cheers,
Peter (pjv)
I did a little more experimenting with the clearance between the wiggler and the material being drilled, and it appears as though 0.043" clearance is much better than 0.027".
Bruce
Nice story. Too bad you forgot about those documents. I would imagine that they would have some significant value today.
Bruce
As mentioned, my method is almost the same as posted at the beginning of the thread, but here is what has changed....
1. Distance from template to bottom point of wiggler 0.015"
2. I no longer use a torch to help with template removal. Simply wet the template with a paper towel that is soaked with laquer thinner and the template will peel off easily.
3. After dimpling the template and metal with the tile cutting bit, I then use a modified DeWALT 1/4" pilot point drill bit to start all of my holes, which is basically the same as Phil's suggestion of using a countersink, but much cheaper.
The DeWALT (DWA1216) pilot point drill bit has patented core technology, which makes it more stout than a common 1/4" bit. I modify this bit, by cutting off the no slip portion of the shank, which makes it shorter and more sturdy, and makes it is just about perfect to work alongside the wiggler.
This modified bit easily centers itself within the dimple created by the tile cutting bit, creating a new dimple, in the metal, with drill bit geometry. Dependent upon the hole size to be drilled, I will either simply dimple with the pilot portion of the bit, or I will continue to drill until the 1/4" portion starts, or I will start with a 1/4" pilot hole, for much larger drilling.
EDIT: I should add that this method provides more accurate holes than starting a hole with a $20 1/4" Carbide Hole-Starting Drill Bit.