Journal98 apr
From Simreal
April 1998
Tue, 7 April 98
Well, I didn't quite do what I expected last week -- but I made good progress nonetheless! I was thinking originally about welding together only part of the body, for local testing. But I remembered the lessons I learned from the leg framework. I'll try to articulate those here (I glossed over it last week).
If you recall the form of the leg frames -- there are lengthwise beams connected with a criss-cross of supporting beams. The first leg, I took a lengthwise beam and welded the supports to it. I did that for all four beams. Then I took each sub-assembly and clamped them together into the full leg frame and welded that shut. However, it was quite a challenge to get it all to clamp cleanly; little errors and differences were frozen into place.
The second leg section wasn't a full tube, but a half (the lower leg). For that one, I took two of the lengthwise beams, and the two end supports, and clamped it all at once into a parallelagram, and then welded it that way. Then, I could weld in the internal supports. The second side was added similarly. This assembly order seemed to help distribute errors around the perimeter smoothly, and it went together much easier.
I figured, if I made the body up out of several independent pieces, it might be a terrible struggle to get them assembled into a clean unit. So I ran out and bought four more adjustable clamps (so I have eight now, one for each corner of the body's octagons). Then, I took all eight body sections and clamped them into a circle, and welded. Did this for the two main body hoops, and again for the internal body hoop. Finally, I spent most of the rest of Saturday finishing up the welds -- there are four sides to each tube, and each one gets its own pass.
Another side-lesson -- you can never have too many adjustable-angle clamps. I used all eight again when I assembled the body hoops, and it wouldn't have hurt to have more!
I am also slowed by being under the weather this weekend (and Monday, and today, and at this rate maybe tomorrow too). But I struggled valiantly, and now I have the two body rings connected by the sixteen vertical supports. I haven't added the inner body ring yet, since I still have to drill the cylinder support-bracket holes into it. The body looks really cool, sitting there in my shop. I'm struck by the proportions of the system -- Boris will look far more crab-like than spider-like. I guess that's okay ;-}
I had to cut one set of body pieces twice -- the upper and lower braces on the part breakdown for the body had one of the end angles backwards. It shows those pieces as a trapezoid, where it is actually a parallellegram! Argh! I also discoverd that once you weld some pieces together, the edges get all indistinct. Exactly how far is 1" in from the left of that piece now!? So -- do all of your measuring and marking before you weld anything. Drilling, too... I need to drill the cylinder mounting holes on the center ring, and not only was it difficult making the placement measurements, it will be much harder to drill the holes. I'll have to use my hand-drill or rig something really awkward with the drill press.
So -- my todo list this time. Drill the inner support ring, and attach it to the main body. Create some cylinder brackets. Create the sensor mountings. Attach the joint flanges and sensor mountings to the frame. Assembly the frame. Attach the cylinders.
I don't think I'll get all of that done this week! But who knows...
I got Boris' "eye" in the mail yesterday. It is a Vision 5300 sensor on their VV300 Sensor Card. It has a fairly low resolution (164x124), but I expect to be driving that system from a PIC MCU, and they don't have very much oomph. I don't really know how I am going to manage vision on Boris yet... maybe active vision. Maybe I'll buy another eye and go for stereo vision. Maybe I'll manage depth perception through parallex like lizards do. Maybe I'll just make a system that detects motion and orients to it. I'm going to hook it to my PC for initial testing, and see what I can brew up in my spare time (hah! Proof of my insanity! "Spare time"? Who is he kidding...) The Vision card is nice, though -- it outputs digital data in either parallel or serial format. It should be extremely easy to interface, if I can ever squeeze the technical data out of my supplier.
I also got my additional pneumatic cylinders yesterday... though the valves and valve manifold are still on order. The 3" cylinder is ridiculously huge; I'm going to try to not use it. I'm thinking the 2" and 2.5" cylinders will be best for everything.
Wed, 15 April 98
Tax Day! Woo-hoo! Pay those puppies! *I* get a *REFUND*. Neener-neener-neener!
This weekend was great, though some sneaky person snuck in a major Holiday on Sunday -- Easter sucked the juicy center out of my robot time, with six hours of festivities. But it was great.
All of the cylinder brackets, etc, etc are made, and most of the leg/body joints are in place... I am --><-- THAT CLOSE to having the leg on the body. It would be there if I hadn't fried the &#&@! out of my hand.
Yup, there I am, making great progress... just welded on a couple of joint flanges. BTW, I thought of a few parts that, if I can get my friend in California to machine them for me, will make my life much easier for the other legs. Anyway, after the flanges I did a little touch-up weld on another part of the leg, where I had cut off a small projection. I forgot I did that touch-up, so after I turned away and (of course) removed my glove, I decided to quench the hot parts. So I picked up the leg section by a beam that was far away from the hot parts... and happened to be right where I had done the touch-up welding.
So... burning skin smells just like burning hair.
Tastes bad, too (what is that reflex to stick a hurt body part into the mouth, anyway?). My mouth didn't make it feel better (duh), so I stuck it into the (really refreshingly cold) quench water. Ahhhhhh. A few minutes of that, I decided I really wanted to make the leg cold, too, so I quenched the flanges. Then the touch-up. After several minutes of sitting, it still steamed.
So... here I am, my hand stuck in cold water. I transfer it to a vinegar-soaked wash cloth with frost and ice from my big freezer in the garage. I keep refreshing the vinegar and ice... cold is good, numbs the screaming nerves. I don't know why you use vineger, but my Mother in Law said so. Looking at the burn, there is a huge line running down the pinky-side of my palm from right below the joint, to right above the ball of the palm on that side, and then a sharp turn toward the middle and a broad sectoin going up... like a big check-mark. Later, I start adding a particular salve called Sooth-Rite to the vinegar and ice mixture. Had to keep it on Ice until I fell asleep that night, and I wasn't sure if I was going to get to sleep, because it hurt so much when the cold was off it.
From the pain, and the look of the burnt skin at the time, I was absolutely sure I was going to blister and be in agony for days. What happened was, I woke up the next day with a big, smooth check-mark branded into my hand, and almost no pain. Wow. Was it the vinegar? The Sooth-Rite? The five or six hours of being nearly frozen? Or was I just damn lucky?
I don't know.
But I think I'll stop picking up stuff without gloves, no matter how cold I think it is. Important safety tip.
Anyway, I will have pictures of the entire thing to put up on the Web site next week if I can get my friend with the digital camera out this weekend...
Of course, I have one more difficult decision to make -- should I leave Boris a gleaming brushed-steel color (which, BTW, will not show damage in combat, but which will eventually rust), or should I paint Boris a combination of Black and Safety Yellow and Red? Hmmm... I need to paint it before I insert the bearings.
Wed, 22 April 98
Well, I had my friend take pictures of the full leg/body assembly with his digital camera -- now we are just waiting for him to e-mail me the results! I may do a mid-week partial update when I get them...
Since this is a test leg, I decided to paint it and see how that works for me. It is a gleaming black, which is quite attractive. If I go with color, I will later paint the cylinder bodies red, and use red tubes and other accessories. It should make for quite the nice effect.
My hand is much better, now... it only hurt that one night in fact! After a couple of days, it blistered up, but no pain. Now, it just itches from time to time, and the blistered area is slowing shrinking.
There are a number of modifications on the leg from when I did the technical drawings -- I'll have to update my blueprints eventually. Everything fit together nearly perfectly... one slight oversize in the hip joint (I'll just add a washer), but the two leg sections, for being full of complex measurements, fit perfectly. Oh, and the cylinder inside the body that actuates the hips is about .5" too high, so I'll need to fiddle with the hip joint a bit.
I finished the body assembly. I wasn't sure how it was going to work, or even how I would clamp it to weld it together (it turns out it took me a couple of days just to figure this out). But it is strong. I can stand in it (with my 180 pounds) and bounce up and down on the inner ring, with no visible flexing. And it is light. I am very proud of this guy. I just hope I can find a clever way to mount the compressor in it when I get it! I'm also a bit worried about the strength of the leg attachment points... we'll see.
I originally controlled the end leg section from a cylinder on top of the thigh section -- actuating a strut above the joint. This just wasn't working for me -- the motion was off, and the larger cylinders I am using interfere to readily with the support structures. So I moved that cylinder to underneat the thigh section... it is way more exposed to damage, but should give me better strength and range of motion. It also lets me use a longer cylinder, for more motion. I tried a 6" stroke, but wasn't happy with it. So I placed the bracket for a 9" stroke cylinder. I couldn't get a 9" stroke at my supplier, though, so I picked up the 10" stroke cylinder they had on hand. It is too long, so I ordered an 8" stroke cylinder instead... which should work very well. I'm going to see if I can arrange to get cylinders with the air holes at a different orientation, to help isolate those connections. Then, I'll just have to keep opponents from getting into the soft underside of my legs... if I just keep them pointed at the bad guy, it should be okay.
The thigh cylinder has always been kept inside the thigh framework (as opposed to the end section cylinder, which has always been external). Using the 2.5" cylinder made for some problems with interference, though. Since I moved the end cylinder (as described in the previous paragraph) down, I no longer needed the tetrahedron on the thigh next to the hip, so I removed it. Not only will that save me several pieces during assembly, it opens up the thigh a lot more, giving me greater flexibility in the size and placement of the thigh's cylinder. I may adjust the cylinder mounting position on the hip to get better mechanical advantage from this, and I could even upgrade to the 3" cylinder, or go for a longer stroke, now that interference issues are gone.
The hip cylinder (which attaches inside the body) was designed originally to be a 1.5" cylinder; this was before I decided to upgrade everything. The 2" cylinder just doesn't fit right... it is slightly longer, which changes the range of motion of the hip, and it interferes slightly with the body framework supports. Also, where I've attached the cylinder to the hip doesn't allow me full swing in towards the cylinder -- it loses mechanical advantage before it is done. So I am going to move the cylinder attachment point from the inside face of the hip joint, off to the end. This will move the attachment about 2" out from the center, which will both give me better pull strength, AND it may allow the use of the 2" cylinder without compromising my range of motion much.
All in all, I am very happy about the mechanical assembly so far. I need to run out and get a few pins to fasten the cylinders to my brackets, but then this assembly is complete. The next step is to wire the various sensors and run the cylinder air lines, and mount the (really cool) valve manifold. Then, I spend what is left of this month and probably all of May doing low-level microcontroller programming and system tests on this leg and its sensors. Once I am happy that I can control a single leg system, plus make intelligent use of the feedback I can get from the pressure and angle sensors, I will be ready to run the big system -- all six legs.
So, end of May and through June I will be building a full set of six legs. I have a friend in California who owns a machine shop, and he said he will make some of my brackets for me... this will be excellent, and will save me days of work. All I will have to do is weld up the framework then. Oh, and somehow, I keep hoping that my sponsor or someone will drop $4,000 in my lap so I can buy the complete pneumatic system. I want July to be available for the high-level programming and testing the system controls etc.
Of course, this may still be impossible. I still need to work out vision for Boris, if at all possible. I need to get a high-level computer to do the system work (unless I go with a high-end PIC or Atmel MCU for the main brains, too). I need to develop and get etched circuit boards for all of the electronics (otherwise, I will spend way more time wiring and debugging the circuits for the six legs, etc). There seems more work than can possibly fit in the next three months... but who know, maybe abject panic will put some speed into my efforts!
Mon, 27 April 98
Whoa... this was a quiet week. Didn't do a darn thing. Mostly I slept.
Well. I worked on the high-level software some (you know, that stuff I'm doing for that other company). Re-worked it so it is highly modular, allowing an amazing amount of user-added Brain modules. We are working on a Web Page for that company now... when it is ready, I suppose I'll point to it ;-}
There are a few minor parts on order for Boris... some pneumatic fittings that I overlooked. An 8" stroke cylinder for the leg. Probably get it all in this week.
On a personal note -- it looks like we will have our house and land sold by the end of this month. Yippee! We're moving! To Texas! AAAIIIIGGGHHH!!!! <ahem> <er> ... well, Austin, which is kind of Texas, but I hear it is very nice. I expect the move to be around end-June or so. By then, most of Boris should be either built or on the verge of built.
Haven't heard back from my Machine Shop friend yet, but that is expected -- he is moving into a larger building even as I write this. I guess I know more from him in a week or two.
Well, I'll keep it short. Hope there is more to write next week!
