Plotter as digitizer

I bet you could reproduce that drawing in CAD in less time than it takes you to digitize with the plotter- no contest.

Yup. It will measure 'long' for 'outside' measurements.

• For me that usually means a small overlap if I cut very close to the line. This is good.

• The scribed line will echo any angular displacement in the joint. Sometimes a 90 degree cut will not rhyme very well against a twisted beam. :)

(...)

Yes.

Trivial. You can move and scale the bitmap in Rhino3D to match any reference dimension on your drawing. Ferinstance, I have a 70' long fence section on my property. I draw a "70'" long line in the correct orientation on my drawing and move (and scale) the 'plan' bitmap until the image of the fence section equals my "70'" reference line.

Now I know my reference drawing and my CAD package agree what a 'foot' is, to within an acceptable error.

(...)

That is better, but still not terribly fast, accurate and easy, IMHO.

The error using the scanner at 600 dpi is about +- 0.003". Multiply that by the reduction scale to reveal the real- life error. Could you hold +-0.003" using the plotter digitizer jig? I am sure that I couldn't, though it is easy to do, all day using a scanned bitmap in Rhino3D.

:)

I concur, especially if you have a second-level data extraction process that creates a Bill Of Materials.

(...)

Do you have a smaller digitizer tablet available? It would be highly interesting to do a 'trial run' using hardware on hand.

I can move a sandpile with only a teaspoon, but I can do it much faster and more easily with a shovel. Discovering that early in the process can save me a lot of time and sweat.

:)

--Winston

(...)

This assumes that everything on the source drawing is dimensioned completely and there is no need to scale the drawing to extract the missing data.

I don't believe that is the case here.

--Winston

Hard to be much closer than the "local auction" (probably

6 miles, as the crow flies) :>

I've already got three large-format plotters. The Calcomp *tablet* would allow me to "not solve" the "plotter as digitizer" problem.

Hmmm... I thought all HP's were serial/HPIB (?). I don't know how you could provide the digitizer functionality using a (conventional) parallel port...

I wasn't suggesting the use of a parallel port - just mentioning that my plotter had one. It's an HP C3171A, better known as the HP DraftPro Plus.

But you are indeed correct - I did have it wired for serial (it also has a serial port). And I use the WinLINE driver for XP.

-mpm

Correct. They represent the architects *intent*. They don't represent the inevitable screw ups!

In our case, the plans aren't even for this *exact* house but, rather, for another "identical floorplan" (different elevation) in the neighborhood. E.g., they don't show the living room as sunken, don't show the "frontier style" roof, don't show the fireplace, etc.

There are also obvious errors *in* the original drawings -- like the RO for the kitchen window extending *below* the height of the adjoining countertop! (oops!)

Yes, and dimensions alone won't help you decide *exactly* how many 8, 10, 12 or 16 ft sheets you'll need. It's an

*estimating* tool.

Until I saw the "professional" plans (i.e., more than just my

*sketches* of the floorplan), I had never considered how significant *where* I laid the first tile would be! Having seen it, now, I will spend a good deal more time evaluating potential layouts before placing that tile as it has a significant impact on the final appearance of the tile in total.

This is *much* easier to do with a scaled CAD rendering than "sketches"!

Yes, I've learned many of these lessons the hard way :< I installed blocking in the kitchen walls off of which to hang the new cabinets. Nice square cuts with the chop saw. Yet, few of them "fit" properly due to the twists and "non-square" placements of the studs.

OK.

Recall these are just *walls*. Walls that I am intimately familiar with! :> Find a corner. "ENTER" it. Slew the pen along the wall to the next "corner". ENTER that. Lather, rinse, repeat.

You're just walking along the surfaces of the house interior (then exterior) so its really just a couple of bizarre polygons.

Once all the points are ENTERed, go back and figure out if you missed any "features". E.g., we have one partial wall that separates the living room from dining room. Not load bearing but, rather, present for "artistic purposes" to break the sight line (we visited another house with similar floorplan where this wall had been removed and the "feel" of the house was decidedly less friendly/intimate).

I think the bigger problem would be finding a symbol library that was consistent with the symbols that I used in the original schematics. Note that it not only has to draw the components the same way but, also, the relative placements of the various connections to those components must be correct/consistent.

I have a couple of 12x12 and 12x18 tablets on hand. These will work well in AutoCAD, for example, -- but, only for drawings of those sizes (I regularly use the 12x12 with autocad for this sort of thing).

The Calcomp tablet in question is at least D size (this is what would make fetching it home difficult) so it would allow me to just mount the drawings and go to it! It would take up a sh*tload of room, here, but could double as a drafting table when not in use as a digitizer.

Just cover it up with *dirt* and buy *new* sand for the "relocated pile"! :>

Exactly. And, any data that you would have to extract with, for example, a *scale* ("ruler") would be suspect. You would want to make all your measurements relative to a single (arbitrary) "reference point" on the page lest you don't start *accumulating* errors:

"Hmmm... I *know* this wall lines up perfectly with this other one -- so why don't they appear that way in the CAD drawing?"

"Because you measure the length of this *part* of this wall... then added the length of this *other* wall... then made an 89 degree turn before you added this final piece..."

Note that The Great Ruler Conspiracy has, as its stated goal, the introduction of anarchy through ambguity and imprecision! Be ever vigilant!!

I had thought of attempting something similar with my larger (12x18) tablet:

- Cut (or even *pan*) each sheet so that a "new portion" was overlaying the active area of the tablet.

- digitize key points in that portion SUCH THAT there are some shared points with the previous/adjacent portion(s)

- pan and rotate the individual portions until they "snap" together

I think this has more opportunities for error. And, if you opt to cut the originals, you deprive yourself of the ability to "try again" if you screw up the first time!

(...)

If the resulting accuracy is sufficient to the task, you are all set, (except for the space taken up by the digitizer). :)

The best solution would be to use a laser scanner to capture the existing structure. :)

(...)

Luckily, most systems allow you to create your own library components.

(...)

This is excellent. (If you need a drafting table.)

That's it! Smarter, not harder.

--Winston

(...)

Roger that!

--Winston

Note that the plotter(s) is already here so that space is already "taken up"

I'd like to make the problem *simpler*, not *harder*! :>

Yes, but some of the drawings were done "free-hand" (others were done with self-adhesive "templates") so the resulting symbols might not be easily created one-to-one in some packages. E.g., getting the text to fit in a shape that has been defined without concern for that text (font, glyph spacing)

We already have *one* so this would just be a *replacement* (although I may be overridden in that decision :-/ )

That's the beauty of creating your own components. You can change things around to suit.

(...)

You can be a hero!

Lose the existing table and have the scans done outside. :)

--Winston

But if the goal is to mimic the existing drawings (to eliminate the need for figuring out a *new* routing scheme for all the individual signal paths), then you want to be able to re-create *exactly* what was there before.

[One of the designs is a CPU built of discrete tonka toy logic -- *lots* of (named) signals!]

No, the existing table isn't *mine* -- so, I'd quickly find myself facing many scowls! :-/ The trick is to convince that the *tablet* would make a FAR BETTER drawing ("drawing" since "drafting" has special connotations) table! ;-)

And, of course, the cosmetic appearance is something that will have to be evaluated (something I can't do from photos) to ensure that it meets the "standards for acceptability"

[Never could really understand this sort of thing: "It *works*, doesn't it? What *more* do you want??"]

:(

(...)

Sounds like an opportunity for creative reinterpretation. :)

That is where Eagle EDA is a much better answer than a conventional CAD program.

(...)

Uh oh. Dangerous water, here.

--Winston

If I ever have to put up cabinets again, I'm going to put 3/4" plywood on the wall to hold them.

--
You can't have a sense of humor, if you have no sense.

Ever put up 4' * 12' sheets by yourself? That was a 'fun' job.

--
You can't have a sense of humor, if you have no sense.

I'm debating the rent vs. buy issue for a lift. The rental units (Panellift brand) are considerably nicer (in terms of build quality). OTOH, buying something for $200 gives you "all the time in the world" to finish hanging!

I'm not too worried about the walls. I'll hang them vertical (more taping but I won't have to worry about butt joints).

The ceiling, OTOH, will be a pisser, regardless. Even 4x8 sheets will quickly show my age :-/

I used 2x6's at the top and bottom ends of the uppers. And, at the top of the base cabinets.

I figure it gives a big enough "target" to ensure the screws "find wood". A couple of the uppers are very narrow and there are no studs behind them. I don't want to risk having to hang them *off* their neighbors.