I have a collection of drawings of historic steam equipment that I would like to have digitized. I have heard that you can take a drawing and digitize it into a format that can work on a CAD program and then manupulate the drawing to create isometrics, various different views etc.

Anyone know more about the process and could offer me some advice?

The Tod Engine
todengine@woh.rr.com

> I have a collection of drawings of historic
> steam equipment that I would like to have
> digitized. I have heard that you can take a
> drawing and digitize it into a format that
> can work on a CAD program and then
> manupulate the drawing to create isometrics,
> various different views etc.

> Anyone know more about the process and could
> offer me some advice?

Hi Rick:

Congratulations on your progress to date on such a massive project. Also, my condolences on anything else you might want to do with your spare time!

You are correct. You can scan original drawings into a CAD format and then manipulate them. But, this is not a very good way to take advantage of the real power of CAD systems because the original scaled hand drawings don't offer enough accuracy to allow the CAD system to "know" where everything is on the drawing. Therefore, scanned hand drawings are only "rough pictures" of the original engine.

It is very important to recognize that the original engineering drawings for an engine like the Tod machine are merely guides for the basic form and function of the machine. The finish and fit of all the parts were part of the shop trades and were not dictated by the engineering department. Therefore all the fitting and tolerances were generated by skilled craftsmen in the machine shop and on the erecting floor. I would wager that if you study the drawings very carefully you will find no tolerances on most dimensions and no finish callouts on machined surfaces like you would expect to see on modern mechanical drawings.

Engines like the Tod engine were built to order -- they were not mass produced. Therefore, they were erected and fitted by hand as individual machines. They were manufactured and assembled in one shop complex and all the parts were fitted and assembled under the watchful eyes of skilled, and experienced erecting shop foremen. This has major implications for anyone trying to digitize the original drawings or the engine itself.

Basically, the old assembly and erecting skills are gone for this type of machinery. If you re-erect the engine for operation you will have a difficult time unless you establish your reference points and stack up your tolerances on related parts. For example, if you do not place the main shaft square and at the proper distance from the cylinders, or you do not establish the proper length of the connecting rod you can easily break out a cylinder head. These parts are all adjustable so the overall tolerance stack up between all the wedges, shims and main parts is quite critical even though the individual tolerances are fairly loose.

The real power of digitizing the engine on CAD drawings is the ability to know where everything is within about 16 decimal places. This is essentially perfect accuracy and allows the draftman to assimilate the old shop fitting skills on the drawings because you can "play" with all the parts until you get the relationships correct so you don't do things like blow out a cylinder head.

This means carefully re-drawing the engineering drawings on a CAD system from scratch and carefully checking the drawing dimensions against the real parts to make sure something hasn't changed since the original drawings were produced. This then is essentially a set of "as built" drawings which represent what you actually have now.

These drawings should represent the "perfect engine" with no allowance for fits and tolerances. For example, there should be no clearance between bearings and journals because you don't know what condition those parts are in now. When you have the whole engine digitized you can begin to add standard clearances (and press fits) where they belong. Once you assign the fits you can begin to check the real parts to see if they match up. Finally, you can stack these fits up to see if you are going to have a problem with interference somewhere (like a piston striking a cylinder head).

The old timers had the whole engine laid out in their minds so they knew where their reference planes and centerlines were. They established these points and made all the various parts comply with their spatial locations relative to the important references. This takes years of experience to do such a thing effectively without making a lot of mistakes that can break up irreplaceable machinery. However, with a CAD drawing that has been developed with impeccable geometry you don't need the engine lay out in your mind and you don't have to figure out how to measure everything twice to establish your relationships. This is accomplished by drawing everything on a one-to-one scale and making sure all objects are accurately developed with their true dimensions. For example, what if you draw a through hole somewhere that is supposed to be 1 1/2" in diameter, but you goof and draw it in as a 1" hole? You can't just change the dimension, you have to go back and actually draw the right size hole. You may recognize the dimension, but the computer will still think it has a 1" hole. Too many mistakes like this means you can't trust the geometry on your drawing and it will be worthless as a layout assembly and erecting tool. If your drawing has impeccable geometry though you will know a lot more than the old erecting shop foremen did and you can carefully "erect" your engine in a virtual world before you begin to heft the real parts around. Your main challenge will be accurately measuring real parts so you can enter them accurately into the CAD drawings.

As you have probably guessed by now all this cyber work outlined above is just as massive a project as working on the real engine. But, if you can get it done you will know so much about your engine you will save untold amounts of time, money and grief when you are putting it back together. Furthermore, you will begin to understand the engineering basis and limitations associated with large, high torque, slow speed engines -- and that is perhaps the most important reason for saving the whole thing in the first place.

I worked at the Skinner Engine Company in Erie, Pennsylvania for several years and learned many of these things in their historic old shop. While at Erie I read about and saw pictures of another big old rolling mill engine in the Republic Steel mill in Cleveland. It was called the Mesta engine. You may have heard of it. Industrial Archeaology like your efforts is very important and I applaud your work.

I have used AutoCAD for many years and prefer it because its .dwg format is almost universal and it is easy to find draftsmen that are fluent with it. AutoCAD LT is a 2 dimensional program that is relatively inexpensive and easy to use. I am a long way away in Seattle and am fairly busy with my mechanical engineering consulting business, but can help you out in a limited way if you are interested.

Bill Petitjean



petitinc@nwlink.com

Rick:

I can add a little to what Bill had to say. It's an easy process to scan engineering drawings and get them into electronic format, but it's considerably more difficult to convert them to CAD, tolerances and accuracy aside.

My office purchased a large drawing scanner a couple of years ago along with a "raster to vector" conversion program. Scanned images are "raster" files- a collection of electronic dots on a page which comprise a picture. Autocad and other CAD programs use "vectors"; i.e.- a line is defined by a length and a direction. The conversion program converts one into the other and has hundreds of settings which can be adjusted to manipulate the conversion process.

From what I've seen, it doesn't work very well. The program has to "decide" whether a circle is a circle, or an ellipse, or a slightly curved line, etc. It has to decide if the minor gap in a line is intentional or if it's merely a defect in the old print. It has to decide if a bunch of random squiggly lines are text or just random squiggly lines. In the few conversions I've tried, it would take less time to re-draw the drawing from scratch in Autocad than it would to scan the drawing, convert it from raster to vector, and then manually clean up all the mistakes the program made. It may be that the software we purchased isn't the best, but I wouldn't be suprised if this is typical of most raster-to-vector conversion programs.

Scanning your drawings may be a good way to preserve them and make them portable, and to make later reproduction easier, but converting them to CAD is a whole 'nother ballgame.

Good Steaming,
Hugh Odom

> I have a collection of drawings of historic
> steam equipment that I would like to have
> digitized. I have heard that you can take a
> drawing and digitize it into a format that
> can work on a CAD program and then
> manupulate the drawing to create isometrics,
> various different views etc.

> Anyone know more about the process and could
> offer me some advice?


The Ultimate Steam Page
whodom@awod.com

Bill and Hugh,

Thanks so much for your comments. Our initial priority is to scan all of the prints (about 90 of this particular engine) to a TIFF format so that the original drawings can be donated to a local historical society that has a climate controlled arcival storage area. With the drawings digitized then we can have copies made at will without undergoing the generational problem of photocopying originals.

A second phase of the project is to redraw a select few of the drawings to a CAD format, so that we can have a general arrangement drawing of the engine that would be able to show all angles, three dimensional views and the like.

Bill you are right that these were custom machines and hand fitting played a major part in their erection. There were two of these engines built, practically identical to each other. Each part was labeled for a specific engine, and were not generally interchangeable.

The project that you suggest of completely redrawing the engine in CAD including all tolerances, dimensions etc. is something that I had not thought of but something that I feel can be accomplished, and should. We plan to re-erect the engine next year at tne new site, not for operation but just static display. We do not have the funds to construct a proper foundation so that we can set the engine up for operation, but we will be able to loosely assemble it for display. Perhaps in twenty years or so once all of the loans are paid off we can construct a proper foundation and set up the engine for operation. Since I am 28 now, I may just live long enough to see her run under steam once before I go to the great steel mill in the sky. :-)

I am working with someone who is quie talented at producing animated GIFs, and who sent me two test versions that he has been working on of this engine in operation. They are quite detailed, and it is my hope that we could expand the GIFs to include animation of the rolling mill drive train and actually follow a piece of steel as it is shaped by the rolls. That would probably be a tremendously huge file, but it would also be the most effective method of visually demonstrating what it is that the Tod engine did.

The Tod Engine
todengine@woh.rr.com