PDA

View Full Version : Precision layout



Mr Ron
12-23-2011, 12:39 PM
I am trying to do a layout where the dimensions are 45.788" for example and other dimensions larger than the largest measuring instruments I have. The largest steel rule is only 18" long; the largest vernier caliper is 12". Is it possible to use the steel rule and mark off the dimension in 18" increments, using a magnifing lens and scriber to get close to the 3 decimal number? If it is off a couple thous, I can live with that. The layout is for a CNC router I'm building. The material is plywood, actually MDO, a resin impregnated material, very stable dimensionally and strong. The dimension in question can vary by a 1/16", but I have to be able to read the dimensions, as other parts must mate up, such as bolt holes.

portlandRon
12-23-2011, 01:14 PM
Would be hard but not impossible to do with a 18" ruler and come out to within three decimal points.

If I understand correctly you said they have to be within a 1/16" if that is true just get a good steel tape measure. The printed lines on these can be rather wide so a trick to keep thing as accurate as possible is to also mark on the same edge of a line - example always mark on the edge of the line closes to the end of the tape.

Scottike
12-23-2011, 01:26 PM
The best solution would be to buy a 48" steel rule, but at 125.00 for a cheap one, I can understand why you wouldn't.
Shop around for a tape measure blade with 1/32" graduations on the first foot of length (probably a 1/2" wide x 12' blade)
Take the steel tape measure blade and secure it (contact cement) to a piece of steel (prefered) or aluminum angle stock that's the length you need. (48" ?)
Don't trust the hook on the end of the tape, remove it and position the end of the tape with the 1" mark precisely aligned with a scribe line 1" from the end of your stock.
That should let you layout within 1/64" with a little care.
I'm sure the more experenced around here can offer you with a more precise solution. So I'll be curious to read their suggestions.

mklotz
12-23-2011, 01:28 PM
How about facing three rods slightly less than 12" in length. Measure them with your vernier and calculate the length needed to make your 45.788" measurement. Face a fourth rod to that length. Lay all four rods in a 45" length of angle iron with the rod ends projecting from both ends of the angle iron.

Stepside
12-23-2011, 01:30 PM
A quality tape measure should do the job. Start at the 1 inch mark (burn an inch) and add an inch to the desired length. Also stay on the same side of the marks.

When I started at my last teaching job I waas sure that it was the dumbest bunch of students I had ever met. Not one of them could measure within a consistant 1/16 of an inch. Then I found out I couldn't measure either. I got out all the combination squares and rules in the shop and set my vernier height gauge at 2 inches. Lo and behold not one ruler/scale measured 2 inches. I found out later that a angry student the year before had carefully cut off somewhere between 1/32 and 3/32 inch from every ruler in the shop.
By starting at the 1 inch mark they could all measure.

The end of tape measures are where the most wear is going to be. Rulers from combination squares are going to be damaged from dropping on concrete .

Use Trig and the tape measure to obtain any angles needed.

Paul Alciatore
12-23-2011, 01:56 PM
+/- 1/16"? I would definitely use a tape measure. Actually, I would probably use three or four of them and go for the average of the lot.

Do use the one inch mark as the start and add one inch to the distance.

Do use a magnifying glass to make reading it easier.

0.788" is an odd decimal equal to 50.432 1/64s or 12.608 1/16s. That's about half a 1/16 past 3/4" and should be easy to eyeball with a magnifier.

But do use more than one tape.

wierdscience
12-23-2011, 02:02 PM
Length of all thread rod,something like 3/8 or 1/2 and some nuts.Jam two nuts against each other,run two more down and jam them together 12" from the last set and so on.Use your Vernier to accurately set the distance between each set.The closer you fit that distance the less cumulative error will be present in your final result.

Long lengths of all thread rod can be had at the local big box store in the electrical hardware section usually near the Uni-strut rack.

If you wanted to refine it you could pick up a cheap pair of dial calipers and screw them to the end of your new gauge so long as you accurately set the zero point.

tdmidget
12-23-2011, 02:02 PM
You need to determine what your tolerance is. You say you can be off "a couple " of thou. Is it .002? .005? .010? There might be someone who can layout to .010 but I've never seen it done. Maybe one of these wizards posting previous to me will come do it for you. Basically there is no such thing as layout to those tolerances. If really needs to be to .002 then you need a machine that can measure it.

philbur
12-23-2011, 02:25 PM
Surely if you have a positional requirement as loose as 1/16" but need corresponding parts to within a couple of thou then you should make an appropriate jig, or just machine them as a pair.

Phil:)

garagemark
12-23-2011, 02:34 PM
How about two known blocks, say 18" long (since he can get that measurement) and "hopscotching" end for end until the rule would be used for the final dimension?

I too would use a tape measure. I can cut hardwood lumber quite accurately with a good tape (not the dollar store brand). It is fairly easy to get within a 64th, and a 32nd would (should) be child's play.

Mark

Paul_Chretien
12-23-2011, 03:11 PM
Very interesting post. I assume that there are other parts that you plan to make in this size range (greater than 36").

If so, it maybe expedient to have comparable measuring devices beginning with a precision 48" steel rule as well as trammel points and spring-style calipers. BTW you could make your own trammel points and calipers. The trammel and calipers would be used to transfer measurements to and from the rule.

The problem I see laying out 45" with a small 12 inch rule by stacking or chain measuring is the errors introduced with each step. Add to this, the confusion in avoiding the first inch or half inch of the shorter rule. It is just too easy to make a mistake.

Another one of your considerations should be thermal expansion. Forty Five inches of steel will change in length 1/16 with 20 F

Thank you for the post. It is these kinds of problems that challenge our minds. I am looking forward seeing other solutions.

Scottike
12-23-2011, 03:32 PM
Just a thought here, .788 in = 20 mm. Perhaps you should be measuring in metric instead of imperial. Are the rest of your diminsions converted from metric as well? It might be easier and more accurite than trying to "guesstimate" it's fractional equivelent. It shouldn't be too difficult to find a tape measure that has both metric and imperial scales.
(I always seem to find them when I don't want them)

LKeithR
12-23-2011, 03:58 PM
I once needed to machine a couple dozen 60" + shafts to an exact--and consistent--dimension. Having no means of accurately measuring that length I took a piece of cold finish flat, set it up in the mill and machined one end square. From there I machined a precise 1" slot with it inner edge exactly 18" (I have an 18" vernier) from the end. I stepped over another 18" and milled a second slot. Doing this a third time gave me a slot with its outer edge exactly 57" from the end. From this edge I machined the end of the bar to the exact length I needed.

To use the gauge I faced both ends of each bar--leaving them about 1/16" long--and then, using a stop and a depth mic, I determined how much to shorten each in the final operation.

Using this method does raise the possibility of having incremental errors creep in but I'm sure--given the care I took when machining the bar--that I was much closer than I'd ever get with a tape measure. And of course this method ensures much greater consistency, something which was--in my case at least--as important as the actual final dimension...

lbhsbz
12-23-2011, 04:51 PM
Pick a division of that final number that you can accuratly measure...like 15.262...use your mill to drill 2 holes in a piece of stock, exactly 15.262" apart. Size the holes to accept a couple of transfer punches with a light press fit...stick a transfer punch in each hole..this is your tool.

We'll call one end of the tool "A" and the other end "B".

lightly center punch the location of the first hole in your workpiece, set the transfer punch tip in end "A" in your dent. Apply a bit of pressure to punch "B" to make a little dent, then pivot your tool around "B" to put "A" now 30.524" away from your first hole. Repeat the procedure and "B" should be exactly 45.788" away from your first hole.

Use a long straight edge as a guide or put marks at your intevals in order to keep all 4 dots in a straight line, and your distance should be fairly accurate.

Fasttrack
12-23-2011, 04:53 PM
There might be someone who can layout to .010 but I've never seen it done. Basically there is no such thing as layout to those tolerances. If really needs to be to .002 then you need a machine that can measure it.

What? Have you been watching fabricators layout with soap stone and tape measures? What is your definition of "layout"?

darryl
12-23-2011, 09:21 PM
I suppose another question might be how are you going to deal with the layout- are these marks which you center-punch, then drill? That would introduce it's own possible errors. And how would you determine whether the marks are the right distance apart- what measuring device would you use?

If you can trust the leadscrew on your mill for accuracy over length, here's a method of drilling some accurately spaced pilot holes in a piece of flat bar. Mount a temporary fence parallel to a t-slot. Also mount a piece of scrap near one end of your table. Have some more of that scrap ready to use as shims.

Run the table over to the mounted piece of scrap and set up to drill. Index the table for a starting point, then lay the flat bar along the fence on top of the scrap piece, and support the rest with some shims. Carefully drill a hole near the end of the flat bar and then into the scrap piece. Insert a tight fitting pin, then run the table over by exactly ten inches. Place a shim directly under the flat bar there, secure the table at that point, and drill another hole in the bar. From there you move the bar over, using the newest hole and the pin to 'index' the flat bar. Two more moves and you have the outside holes exactly 40 inches apart. From there you can leave the bar indexed where it is and move the table over another 5.78246 inches, or whatever it was you needed. Drill the last hole. All the while you are keeping the flat bar up against the temporary fence.

The end result should be a bar with end holes the exact distance apart that you need, within the cumulative error of the sloppiness of the pins fit. When you use it to pilot the holes you need, you'll be eliminating the error that could creep in due to marking, punching, and drilling wander.

Knowing that the temperature of the bar will affect the distances, you would take care to keep this flat bar at mill table temperature as you drill it and as you use it.

I'll call this a pilot hole jig. It would be especially handy where you need to mark out more than one placement of this exact distance. As long as your leadscrew doesn't have accumulating error, you should be able to use this jig to drill two holes within about 5 thou of exact length- maybe as close as 3 thou.

tdmidget
12-23-2011, 09:53 PM
What? Have you been watching fabricators layout with soap stone and tape measures? What is your definition of "layout"?

Well it certainly would not be with a tape measure where the lines would be wider than the tolerance we are talking about. Go back and read this thread. You should all be ashamed. The OP talks about +/_ .002 and you are talking tape measure. Some have immediately assumed 1/16" is is OK . And we have one who thinks it was called out as 45 inches and 20mm. No you can not lay out to .010 with instruments that have less resolution than that.

RussZHC
12-23-2011, 10:08 PM
Watching this thread with interest as well.

Playing devil's advocate, how did you decide on such a given dimension?
The point I guess is, if you don't have a method of determining to such exactness over a distance greater than an instrument you have to measure with, why does it need to be that number?

What I am saying is your statement about the mating pieces may in fact end up being what you have to work with in the situation.
Do you need pieces 45.788" or will 46" or 45"work?
If you need say 4 pieces and two ends to make a box structure, if you can make those four pieces the "same" (or 2x two "same" pieces) and two matched ends, you can make that box structure. And if you can do that once, you can do it repeated times with other boxes to make units to form the entire machine (mind's eye views this as being one of those CNC "gantry" type machines).
The entire thing may not end up with the precision you state over a given dimension but provided you make it large enough for the work pieces you want to use it on/with, does it matter?

Just saying, typically, for me, the restriction is the size of a 4x8 plywood sheet...you can make things the "same" going old school with arcs, points and edges square to one another...square or a given angle can be more important than precise linear distance (and the square and angles, to some degree, you can arrive at with mathematics...and in conjunction with the accurate measuring devices you do have, I suspect a fair bit can be accomplished)

Edit: I wondered about this in my own situation, I have micrometers that are, to me, silly accurate, up to about 3", my small lathe is about 24" capacity center to center, how do I measure that to the same degree? My answer is, I don't/can't.

Edit: to correct measurement so as to match OP

strokersix
12-23-2011, 10:15 PM
As mentioned above, this sounds like an ideal situation for match drilling. It has to match but can be off 1/16 inch.

Scottike
12-23-2011, 10:16 PM
and we have one who thinks it was called out as 45 inches and 20mm

I did not say that the diminsion was 45" +20 mm. What I said was that .788" = 20 mm. It's such an odd diminision in the imperial scale that I thought perhaps the OP was working from a set of plans that may have originally been diminsioned in metric, hence the odd decimal measurements. I suggested that if that was the case, perhaps he should use the metric scale for his project rather than convert all the diminsions to imperial. By the way 45.788" = 114.5 cm, would it not be unreasonable to assume that metric would be the scale to use if the other diminsions in the project were just as odd in the imperial scale?
edit: Presuming that even if it was just the parts the OP is using have existing mount points that are based on the metric scale, why convert everything to imperial? just go with it.

Fasttrack
12-23-2011, 11:24 PM
Go back and read this thread. You should all be ashamed. The OP talks about +/_ .002 and you are talking tape measure. Some have immediately assumed 1/16" is is OK .

LOL - You're on a "you should be ashamed" kick, lately. Actually, if YOU go back and read the thread, you'll see this:


The dimension in question can vary by a 1/16", but I have to be able to read the dimensions, as other parts must mate up, such as bolt holes.


As StrokerSix says, the accuracy doesn't matter, just matching features. Regardless, there were several methods presented that made no use of a tape measure. Your post indicated that it was somehow impossible to layout a distance of 48" accurately without the aid of machines to which I asked what your definition of "layout" is. And yes! It would be impossible using a tape measure and soap stone, as I said. :rolleyes:

Mcgyver
12-24-2011, 12:25 AM
The dimension in question can vary by a 1/16", but I have to be able to read the dimensions, as other parts must mate up, such as bolt holes.

transfer punches?

Scottike
12-24-2011, 01:14 AM
let's get realistic here.
45.788" (+/-.002") between holes?, even Boeing doesn't have those kinds of tolerances. (well, they might in a special circumstance) Much less the average home shop machinist.
And to apply those specs to a piece of particle board? I hate to say it, but the OP needs to apply some realistic tolerances. Just drilling a hole on a centerline of +/- .002 in the material would be difficult, if not impossible in a home shop enviroment. Not that it can't be done, given the right tools, but how much do you want to spend? And then get two holes 45.788" apart (+/-couple thou?). the diminsion of 1/16" is easy, a 32nd or 64th is doable with care, +/- .010 should be able to be done with the right tools and care.
but at what temp?

LKeithR
12-24-2011, 02:38 AM
If you can trust the leadscrew on your mill for accuracy over length, here's a method of drilling some accurately spaced pilot holes in a piece of flat bar. Mount a temporary fence parallel to a t-slot. Also mount a piece of scrap near one end of your table. Have some more of that scrap ready to use as shims.

Run the table over to the mounted piece of scrap and set up to drill. Index the table for a starting point, then lay the flat bar along the fence on top of the scrap piece, and support the rest with some shims. Carefully drill a hole near the end of the flat bar and then into the scrap piece. Insert a tight fitting pin, then run the table over by exactly ten inches. Place a shim directly under the flat bar there, secure the table at that point, and drill another hole in the bar. From there you move the bar over, using the newest hole and the pin to 'index' the flat bar. Two more moves and you have the outside holes exactly 40 inches apart. From there you can leave the bar indexed where it is and move the table over another 5.78246 inches, or whatever it was you needed. Drill the last hole. All the while you are keeping the flat bar up against the temporary fence.

The end result should be a bar with end holes the exact distance apart that you need, within the cumulative error of the sloppiness of the pins fit. When you use it to pilot the holes you need, you'll be eliminating the error that could creep in due to marking, punching, and drilling wander.


I'll call this a pilot hole jig. It would be especially handy where you need to mark out more than one placement of this exact distance. As long as your leadscrew doesn't have accumulating error, you should be able to use this jig to drill two holes within about 5 thou of exact length- maybe as close as 3 thou.

Your idea is similar to mine (Post #13). I like yours as a means to position holes; mine was intended to measure a length. You could improve the accuracy if you had a DRO on your mill--as I did in the case I mentioned...

Hopper
12-24-2011, 02:40 AM
How about one of those new-fangled laser measuring thingies. I know not much about them, other than having seen them in the hardware store. There must be versions accurate to .001"?

vincemulhollon
12-26-2011, 04:20 PM
but at what temp?

Ah don't worry about that. His spec is for .002/45.788 * 1e6 = 44 ppm.

He's working MDF. That varies by something ridiculous from bone dry in the desert to almost moldy in the summer tropics, like 1 to 2 percent in expansion which is like 10000 to 20000 ppm.

Its vital that his sticks of MDF be stored in the same environment for a couple weeks with great air circulation on all sides before cutting to "length".

I can't get a straight answer on thermal expansion of MDF but "plain ole wood" as an engineering rule of thumb expands about two to four dozen ppm per degree F, so he's only got to hold the temperature of his sticks to within one degree or so, which admittedly in a garage would be impossible, but in a basement away from sunlit windows is not much of a problem, at least compared to the humidity problem.

Over a realistic range of basement humidity and temperature, the humidity is going to be about a hundred times harder of a bear to tame than the temperature. I suppose it depends where you live...

Aluminum and friends have about ten ppm thermal expansion. I'm told if you bolt an aluminum flat very tightly to an oak panel with SS hardware you have created something like a bimetallic strip thermostat which will visibly flex with temperature changes outdoors, like you can make an artsy craftsy outdoor thermometer, although I've not seen it myself. Obviously you have to seal the wood or you've created a humidity meter instead.

Ian B
12-27-2011, 04:09 AM
The OP isn't using MDF, he's using MDO, about which he says "The material is plywood, actually MDO, a resin impregnated material, very stable dimensionally and strong". MDO is plywood with a phenolic coating, so reasonably moisture resistant. This stuff: http://www.ehow.com/about_6466437_mdo-plywood_.html

It sounds like he's using a sheet of it as the baseplate for a CNC router, and will probably be bolting the bed rail supports to it. Absolute length is probably less important than alignment and squareness. A sheet by itself wouldn't eb that rigid, but for all I know he's making a box section out of MDO, which could be very rigid.

If that's the case, then he could probably do worse than to measure in from the edges of the sheet - plywood seems to come cut very square from the factory. This will let him determine where to fix the components. Then, as McGyver said, transfer punches for the actual holes.

Ian

darryl
12-27-2011, 04:55 AM
I would not count on any sheet goods being square these days-

mototed
12-27-2011, 09:03 AM
Divider/caliper? Spread points out to 11.477, lay on a straight edge guide on work piece. swing points 4 times.
Never checked on how close this would be.

darryl
12-27-2011, 04:16 PM
Sometimes I need to precisely dimension something that may be on the long side, then precisely divide that. I've looked for various pre-made things that would help in this. One thing I came up with is shelf bracket material that you would find in a cupboard. These are vertical metal strips that you clip little brackets into to support a shelf. The strips have evenly spaced slots at 1/2 inch spacing. 90 slots would be 45 inches, etc. I've checked some of this stuff at work, and found the spacing of the slots to be quite accurate, and over length as well, checking with a tape measure. The slots are 1/8 inch wide and just over 1/4 inch long. You could custom make a pair of brackets that would be a close fit into the slots, and each would have a pilot hole in it. This gives you a method of setting pilot holes any multiple of 1/2 inch apart.

This jig could be quite handy for pilot hole drilling, especially where you would be repeating the same measurement more than once, and where the layout is relatively large such as over a 4x8 sheet. It would be handy for squaring up a layout as well- using the 3-4-5 method and some flat bar, you could make up a triangle that is pinned together at the corners. You could set one pin into a pilot hole at 'zero, zero, say, then the other pins would be rapped with a hammer to leave a pair of dimples, precisely laying out a 90 degree corner.

The jig would also be good as a compass, where you might want to draw out an arc where the radius might be measured in feet or so. This slotted material comes in full lengths, as in 8 ft, or possibly 10 ft.

Just another idea.

boslab
12-27-2011, 07:40 PM
You needa straightedge bigger than what yourmarking any, so buy a 4 footer, if too expensive and you already have a straightedge then blue up, scribe a line longer than required, divide the required dimension up to a size that you can accomodate with your biggest dividors [easy to make project!] and step off scribing at each step, its not hard to get within thous never mind 1/16", thats the old way to do it, time proven, if you repeat the stepping 1/2 a dozen times you get a range of endpoints, with a magnifying glass punch with a DOTTING [60 degree point] punchat the middle of the group of scribed lines so you get the average of the 6 markings so that cuts down on the errors too.
I admit its not perfect but unless you have a big vernier or stick mike its difficult, you can just drop in to your local college and use thiers, they usually dont mind [lube required for dry throats though!] its worth getting to know them anyway as they are a mine of information left untapped by most home shoppers.
good luck an merry Christmas
mark