Wednesday, April 23, 2014

161> Scribblings from the Tundra

First of all, thanks for the support in regards to my last post. With everybody seemingly stable, I decided to let my sister “take the wheel” for a couple of weeks so I can unpack my New Hampshire cabin for the summer construction projects, and see how it has fared through the worst winter in recent memory.   Unfortunately, I arrived to find snow drifts of up to 4 ft., with 12 hours of shoveling just to get my car to the road. The picture above is my “runabout” Toyota, which I only use on the land, down by my cabin.  (Boards on roof with tarp over that, if you are wondering what you are looking at) And that was is April 13th! Then, on April 15th, we got 3 more inches! Now it seems my dad has had another stroke so I am going back to Houston tomorrow. Not that I am getting much done here, although the snow is only patchy now…

I said I would post on the progress of my 1/6 scale model, so here is how it stands, the picture taken just before I left.

This really shows how different this bogie design is than one where wheel-wear is not considered as a primary consideration, usually because of slow anticipated speeds and/or more limited trip distances. Here the diameters/widths are all proportional to service demands. It is not easy to see, but there are four black wheels, in two counter-rotating pairs. The small bottom wheels only engage during switching.

I mention this because this is NOT how to make a small model for running around a track to demonstrate PRT. Designs adapted to the bench top would be much easier, and would look a lot more logical. This is true “to scale”, where the full sized wheel-motors would be about the same size as the wheels of a compact car. Our personal vehicles, I would point out, are real work-horses, when you think about it. Would you want any less for a commercial system? You re-invent the wheel at your own peril!

Unfortunately, the electronics have been really slow going. Designing the circuits forces you to finalize your logic in excruciating detail. How many digital inputs from the track might you want? How many sensors will the bogie have? How and what will force a program to branch? Will future iterations of motors have other wiring needs?  Another problem is that hand-made circuits can’t be shrunk to fit a scale model, and I want to make the electronics modular, so that PC boards can be swapped out for later iterations.  These boards need to be able to physically attach to the frame and make all of the proper electrical connections with each other. Finally, the motor winding discs have proved very tricky to produce. I have killed fully 50% of my attempts. Pictured below are 6 components to prove my point, all destined for the junk pile. Each represents dozens of hours of work.

There is, at least, some upside to such failures. Your attempts get better and better. Along those lines, I thought I might share one other reject that I must say I’m rather proud of.

I’ve been looking around online for the latest techniques for making circuit boards at home. I used to photo-etch, sensitizing my own boards, but my chemicals went bad and are now illegal, leaving me with a lifetime supply of double-sided copper-clad board and no way to sensitize it. Luckily, I stumbled across the “Laser printer iron-on transfer method,” which various people have demonstrated on YouTube, Instructables, etc. Unfortunately my printer is a real cheapo, and the faint, line-ridden output isn’t what these online instructors had in mind, even for the (generally) large, crude boards that they show. One of these days I’ll have to reveal my secret techniques for getting such fine detail! In any case, it shows that all of that time wasn’t completely wasted. Learn by doing!

By the way, since an effective PRT demonstration should have more than one vehicle, I am keeping an eye on ways to mass produce these components. Circuit boards are fairly cheap to have made in quantity, once you know, for sure, that the circuit paths are what you really want. You could waste a whole lot of time and money in the meantime, though, discovering what is wrong with your designs! They still need to be prototyped in the finished form and scale to make them “plug & play.”  

So here is where I stand, after all of my misguided efforts. I have now, reluctantly, decided to use 5 Arduino microprocessors per vehicle; one per wheel with a fifth as a controller/clock. In theory I could run all four wheels with a single microprocessor and still have a couple of IO pins to spare, but that limits improvements to the motor drivers, along with other issues. The circuit board above was to run two wheels with one microprocessor, (for 2 or maybe 3 microprocessors per vehicle) since wheels always work in pairs anyway. But the whole left side of the board is for two buffer chips that would have directed that output, and they take as much space as the microprocessor board they would replace. So in the end I am lavishing the system with microprocessors galore. This goes against my grain, as I am sort of fanatical about keeping designs concise. This will, however, do more than ensure that we have plenty of spare IO pins. It will give us four identical driver boards, and a separate microprocessor only for functions like routing and sensing. It is a more physically logical and symmetrical architecture, and it neatly separates the software code into higher and lower functions. Below is the artwork, almost complete, for such a driver, shrunk as small as I can make it. Never mind the “floating”” components… It’s just a screen-shot… This is what I have been working on at night by the wood stove!

Unfortunately, wrapping this iteration up is showing me that there are a number of connection pins that simply won’t fit where I want them. Period. The board cannot be made any bigger and still fit, and the density of the circuit traces is as tight as I can go. So, believe it or not, it’s back to the drawing board yet again! I need to explore modifying this design into a piggy-back design, since I do have room if I build upward. Well, I said this would take a long time, didn’t I?

Here. Enjoy. This is my stream after the mid-April snow. We’ll see you next time.