Engineering and Computers

[sciguy125]

After several experiences over the last few days, I've done some pondering. In the Digital Age, we have computers that help us every day. We can communicate with people around the globe. We can perform millions of tedious calculations in a matter of seconds. We can design and test an airplane before it's even built. So, my question is, how did people do any hardcore engineering without computers?

After having a short lecture about the operation and use of analog vs digital oscilloscopes yesterday, I was in awe of the added features in a digital scope. I'd never used a digital one before and there were so many time-saving gadets on it. For one thing, it gives numerical readouts of delays and various voltage settings. The thing that I thought was most interesting was the fact that it will freeze the display and shows the last displayable waveform if the trigger level is never reached. Try getting that from an analog scope. There's even a print feature to send a picture of the display, along with some info about settings, to a networked printer. (Although, my scope was one of the few that wasn't connected properly, so I couldn't print ) Apparently, there were even more useful things that were disabled because they want us to do it the old fashioned way.

Also take a look at CAD. You design something in 3-D, put all the pieces together and see that it all fits. Making 3-D drawings by hand is next to impossible. If you want to test fit everything, you'd either have to do some very careful number crunching, or take the time to build models. On the computer, you can start out in 3-D, do whatever you want with it, get the mass, find the volume... You're even a few clicks away from a CNC program if you have the right utilities. Wow, computers even eliminate the manual labor involved in making the part in the real world.

And now, how I spent a few hours of my Wednesday night. In our lab reports, we have to have hand drawn graphs. While Excel could easily take all my data, produce a neat little graph for me, and even fit a polynomial if needed, I have to do it by hand. On the last report, I lost some points because I should have plotted it on a logarithmic scale rather than the linar that my engineering paper has. Excel could have easily done this, but again, I had to do it by hand. Do you have any idea how hard it is to find logarithmic graph paper? Google turned up a bunch of stuff, but nothing that fit my exact needs in terms of scaling and borders. Eventually, I just settled for one that was close enough to what I wanted that I could work with it. During this process, I also had to figure out how to use the stuff. Plotting on semi-log paper is very strange to say the least. My graphs didn't turn out as pretty as I wanted them to be, but I guess they were good enough. Again, rather than having to go through this whole ordeal, I could have simply fed numbers to Excel and ordered it to spit out the exact graph I wanted.

I have more examples and I'd love to share them, but alas, my ranting time is up. Now, if you'll excuse me, I have to follow the black pavment road to see the wonderful wizard of differential equations.

[sanddrag]

Log Log or semi-log graph paper was pretty easy to find for me. Just print it out.

For my Physics and ME courses we have done a fair amount of graphing including graphing on log log and semilog paper.

For everyone: Did you know excel can plot on a logrithmic scale? Also, did you know you can do a best fit line? I never knew these things before so that's why I'm telling.

For computers aiding in stuff, you're 100% correct. I couldn't imaging designing many things without one.

On a personal note, I've got a nice little Excel Spreadsheet where you input vecors' magnitudes and directions and it breaks them into x and y components, and finds the resultant magnitude and angle. Cool huh? Excel is great. However, we still have to write out everything on our homework like Fx=Fcosθ and so on and soforth.

[Tristan Lall]

Originally Posted by sanddrag

For everyone: Did you know excel can plot on a logrithmic scale? Also, did you know you can do a best fit line? I never knew these things before so that's why I'm telling.

I've got an added challenge: get Excel to print out a graph in (what I think is) y —> log(log(y)) format. I've got set of data for kinematic viscosity vs. temperature for 5W-30 and Mercon SP oils, which I wanted to put in Excel—but Excel seems only capable of doing conventional log(y) graphs, and I haven't bothered to check if Mathcad (which I don't really like) can handle it.

Since this description likely makes no sense to someone who's never seen this graph before, I'll try to clarify: instead of having the major divisions of 10ⁱ, for all integers i, each displayed at a fixed distance apart on the y-axis (such as would be the case with log(y) graph paper), this graph displays them at ever-decreasing distances apart, in the same proportions as the logarithmic spacing (so a big segment for 10¹, and decreasing segments for 10², 10³, 10⁴,...).

Why all the thought about how to do this? Well, theoretically, kinematic viscosity should appear mostly linear on this y —> log(log(y)) plot, rather than logarithmic or worse. I wanted to see how well a list of measured data points in a chart compared to the (dubiously accurate) graph which I already had.

[sciguy125]

Originally Posted by sanddrag

Log Log or semi-log graph paper was pretty easy to find for me. Just print it out.

That's what I did, but most of the ones I found had little problems. I didn't like the orientation of the numbers, the margins weren't the correct size, lines were too light or too thick, strange non-round number of squares... One of my graphs needed 3-cycle paper, the other need 4-cycle. Just a bunch of little things that made it hard to find something that would make a clean graph to put in my otherwise neat, computer generated report.

Originally Posted by sanddrag

For everyone: Did you know excel can plot on a logrithmic scale? Also, did you know you can do a best fit line? I never knew these things before so that's why I'm telling.

I actually learned about that stuff in class. The first half of my Intro to Engineering class was about Excel. You never learn to truely loath a program until you're forced to spend 3 hours at a time with it.

The other example I wanted to bring up before I was rudely interupted by the beckon of class was a project I recently did. For Thermodynamics, we had to analyze certain aspects of a steam turbine. Needless to say, this takes an ungodly number of calculations. In addition to that, we had to perform the analysis for 6 different cases. After figuring out exactly what needed to be done, it probably would have taken about half an hour to manually (calculator allowed, but no computer) do the calculations for each case. My group got smart and decided to make Excel do the grunt work for us. Another group wasn't so smart. I suppose they are very adept at operating their caculators now. After finally analyzing the numbers we got, we realized there was something seriously wrong. You normally don't expect a turbine to have an isentropic efficiency of 0.45%. After about an hour and a half of triple checking all our formulas, we discovered that one of our conversion factors was wrong. With Excel, you change the number and everything fixes itself. By hand, you start over and sleep later.

[mechanicalbrain]

I think a majority of engineering involved lots of mathematics and practice, practice, practice. I think one of the largest hurdles facing computers is a lack of compatibility. I'm waiting to see but as a regular reader of scientific websites i see some very interesting product coming out or being developed that will probably be on par with the invention of the hard drive or mouse (Despite being rather young I remember using a keyboard to scroll through a black and white menu and yes the mouse is quite an invention). Brain controlled computers, self configuring processors, optical memory cool stuff on the horizon.....

[sanddrag]

sciguy what college do you go to? The stuff you do sounds quite similar to what we do over here at CPP.

Also, Excel is quite a little superstar on the PocketPC (Windows Mobile 2003) platform too. No graphs but still TONS of operations (I guess just like the big brother on XP) and borders and fills and everything. It is a great thing to be able to have everywhere you go.

Last, does anyone know of a good graphing utility for PocketPC?

[JonBell]

Originally Posted by sanddrag

Last, does anyone know of a good graphing utility for PocketPC?

Me and my TI-89ti are doing just fine (It can do pretty much anything that excel will do, albeit at 10 (?) mhz)

[sciguy125]

Originally Posted by sanddrag

sciguy what college do you go to? The stuff you do sounds quite similar to what we do over here at CPP.

I go to San Jose State. But, I suspect that most schools in California will have very simmilar policies/curriculum.