Journal97 oct

From Simreal

---

October 1997

---

Thu, 2 Oct 97

Fabricated and attached the mounting brackets for the leg joint potentiometers. Just hacked them out, welded them into place - totally cheezy, but should do the job okay. I used one of my pots for alignment while welding - I think I'll need to replace it now ;-}

I also fabricated all of the brackets for the cylinder tail end mounts, and attached two of them (the Thigh and Leg joint mounts). Before I can attach the Hip joint mount, I need to do some more fill-in on the body, and before that, I need to design more of the body.

To get the cylinder to pivot inside my mounting brackets, I had to chop off the corners of the brackets and make them more round. Okay, can't clamp it into the cutoff so I get the hacksaw. Whack, whack, whack, bam! Oops! At the end of a cut, I tend to smash the nose of the hacksaw into the table (I get a bit, um, exuberant). This time, the whole thing popped apart, and I lost the little thingy that holds the blade in place. Darn. I turn another 1/8" off my grinding wheel into dust making everything round.

Yesterday I got a nice set of catalogs from my Clippard distributor. Those valves are 3-way... and two of them gives me control of my piston. However, I didn't realize at the time that the center position will be open - not closed like in my current 4-way 3-position valve. So the joint will float when both valves are off. Then, I will waste a bunch of air pulsing the leg back into position. Hmm. But I want those valves. Aha! If I keep them both on all the time (or get normally open valves), the piston will be locked into place under pressure. Of course there will be a slow drift, due to the difference in area between the tailed and roadside of the piston, but that will be less work to compensate for that a floating piston. I order a set, and they should be in for next weekend.

Next (hopefully tomorrow), I need to build the cylinder tipped mounts into the legs. This shouldn't be overly difficult. Then, attach the cylinders, valve, and presto! Testing time! I want to have something to show-off this Saturday evening, at our monthly first-Saturday party.

Oh yeah - Tomorrow is also Melissa's birthday. I can't say how old, but she is still the prettiest girl in town. So I don't get to stay up all night tonight. Working on the robot, that is.

And finally - we are into October. It is time to build the Boris web pages. Next week they will be published, and I will be committed (drum roll please), or at least should be.

Fri, Oct 3

I managed to finish the cylinder attachments for the Leg and Thigh. I also started the web pages. I'm not going to get overly fancy here, just get the messages across. Had a nice birthday dinner at the Sushi place. Yum!

Mon, Oct 6

Life has been good. Got my 10 BotBoard2 PC boards in. My MC68C811E2FN chips shipped Friday (and, boy did I get a good price). My second prototype leg is assembled with 2 pistons attached, ready for testing. And I am roughing out my test program so I might be ready to run them tonight, on my one sad little valve. I'm going to push it to the limits this time. I also added flow limiters to the exhausts, so I can slow it down to make for easier testing at first.

I also published my Boris web page (which, of course, should be obvious, if you are reading this now).

Tue, Oct 7

Well, the response to the Boris web page (all five of them) is unanimous - I have to get my rear in gear and put up some more pictures! Soon I will shuffle out to the store and get a stack of film, and create some nice photos to spackle around the text. By the time you read this, there will be photos all over the place.

After dinner tonight I finished my test program ... a really simple number that tests the speed response of the (slow Parker) valve I am using. For each speed of interest, I cycle the cylinder down, then up, then on to the next speed.

Interestingly enough, the valve starts showing a response at about 9ms - three times faster than spec! I wonder what my fast valves will do when they come in! Also interesting - the flow limiting valves were totally unnecessary. Oh well, small expense.

I am worried however - the leg joint has sufficient strength (nice attachment point, nice leverage ratios), but the thigh action just seems too weak. I will upgrade that cylinder to a 2î or 2.5î diameter and really make it sweat. That will, of course, slow it down a lot... but the ratio amplifier and range of motion is such that I think that will be okay. I will, of course, let you know when I find out!

Finally, I hooked up the potentiometer at the knee joint to the analog input port of the PIC. It tracked fairly nicely, though there was incredible backlash (about 5 degrees motion before it noticed the reversal). I rather expected that, I guess, with my wimpy pin-in-slot arrangement holding the pot angle. I will need to have a positive action set screw. Of course, then I have to worry about slop in the joint - with a pin-in-slot, the leg could slide back and forth a bit, and would just slide in the slot. With a set screw, lateral motion would make a mess of my pot.... Unless it had a more flexible mount. Or the joint was rigid.

It looks like this second test leg is, so far, a rousing success! All elements appear to be strong, all motion is sufficient, and the strength is approaching the correct values. Soon, I delve into the final designs.

My friend in California says that, if I visit him, he can help me machine stuff. Simple 2-axis aluminum plate types of things, but I need to keep that in mind as I do the final design.

On the research end of things, I found some nice accelerometers from Linear Devices... not terribly expensive, and they appear to be easy to use. I am going to have to provide acceleration (and tip) feedback for Boris, as part of the feedback for when he learns to walk. Maximize forward acceleration, minimize vertical wobble, minimize vibration. Who knows what other exciting feedback they can provide while in combat! Now to decide when to spend the $60 for the sample pack with boards and doco...

I'm not sure what my next immediate tasks are for Boris - I'll think about that as I photograph everything and tidy up my journal. I need a bit of a break, anyway. When the final design is in place, then the big work begins. I'll start by building all the frameworks, while saving for the big piston, valve, and compressor expenses. I need to build the new BotBoard2 computers sometime, too. <sigh>. Feel free to donate money to the cause! I'll even give you a T-shirt when I make some! Full recognition, that's my motto.

Thanks for reading!

Thu, Oct 9

Shot a bunch of pictures for the pages... scattered all through the journal! Darn, now you have to read everything again. Since I don't like slow web pages, I'll thumbnail everything into unreadable specks, and link to larger images. Yeah.

Talked to my Clippard rep - the manufacturer doesn't stock the valves! I have to wait while they make the darn things... another two weeks before I get my hands on them. I guess I'll have to do other tests first. Need to try a larger piston (that will cost me a chunk), need to fix the awful motion sensing on the leg, need to start writing good control programs.

Heard from a friend yesterday (Hi Greg!)... he has a friend (Walter) that is a retired engineer or something. Has lots of metal experience, access to tools, and....<drum roll>... is interested in my project! Yes! Score! This third (and final) re-design is going to be great.

Halloween is coming!

I had such great plans, too. I love Halloween. But Boris has taken all of my resources. But I have a reduced plan now, and that should be sufficient.

Every year, we have a rampaging hoard of pumpkin smashers come through. I don't like that. The kids work hard on their pumpkins, and I find it unspeakably rude that these rats trash them.

Don't think I'll have any problems next year though.

This year, well.... I have a microcontroller, some SCRs, some this, some that.... I saw some great fire alarm bells at the surplus place earlier. I can get a nice electrical water valve... hose is right there... lights... switches.... Pumpkins.

Lift a pumpkin, get an alarm, sprayed, and well illuminated. Heh heh. Caught in my trap...

Oh, and that corner of the deck is where we have the trash cans. I think a trash monster that jumps out (heck, I have all this pneumatic gear....) on my signal would be handy, too. Automatic for the pumpkin trashers, but on manual too... just ëcause I can.

See Phantasmechanics for more seasonal ideas...

Sun, Oct 12

Friday, I had the pictures scanned, and madly started processing them to put into the page... I told my Air-Oil friend that I would have them in that night . But it was not to be. That night, we packed it all up, and went to Train Mountain (Hi Quentin!) Yow!

Imagine this.

Over two thousand acres at about 4,000 feet up, in the Southern Oregon mountains (whatever). Near Klamath Falls. Beautiful.

On that land, lives an amazing person who has a dream -- to create the most magnificent 1/8th scale train system in the world. And so far, it is exactly that. Now, I thought my hobby was involved, but these train people have got a corner on obsession. They have the most incredible models! This is a members only place, but they have train meets all the time, and give rides, and share stories, and generally do what all other people with amazing hobbies do.

But when they do it on Train Mountain, they do it in style. Quentin has about 10 miles of track laid already, with two tunnels, several (scale) villages, bridges, it just goes on. The large loop takes two hours to run. The small loop goes in, oh, 45 minutes. And you thought that HO gage train in the basement was taking over...

But this is just the start. He has a map showing how he wants to expand the rail ... over the entire 2,000 acres!

Humbling.

But I am back now, to post my pictures. Oh, be sure to let me know if my journal pages start getting too slow to load -- I can break them up buy weeks, or something. I have a few extra pictures here. The roto-tiller represents all those household tasks I'm busily not doing. The black dog is "Bear" -- an amazing whippet/pit bull that I inherited from my jewelry instructor (Hi Leo!). The big white dog is "River", and was found by a friend by (you guessed it) by the river. The little brown pup (but only little for the short term) is "Sham", short for "Shamrock" and isn't ours, but is visiting.

Thu, Oct 16

It is amazing how the time runs off when you don't watch it carefully... how does that theme go? ìLike the grains of sand...î

I am near the end of taking the week off from Boris - and working on the Halloween project instead! I even had to put my $100 jewelry money into it.

Let's see... where am I? Monday I didn't do much of anything, but I probably thought about the Halloween project. The trash can jumper, the alarm, lights, water... and synchronization through the MicroBench.

Tuesday, I built a handy trash can pop-up mechanism using scraps of steel and a ìspareî cylinder. Essentially a parallelogram with one long side attached to the garbage can (actually, press fit into the can using big metal hoops as clips), and the opposite (parallel) side tall the other direction attached to the lid. I know, I know ... PICTURES! When I develop them, I'll put them in.

Wednesday, I went to Radio Shack and got a bunch of discretes, a strobe, and some this-and-that. Spent my evening building the control panel.

Today I programmed the system. I can sequence the four types of actions, and their on and off times. I can manually trigger any given action, and the pumpkin-lift-switch runs the automatic sequence. I can also independently disable any given action from the sequence.

Tomorrow, it looks like I am off to Train Mountain again. Vacation time! But the next step in Halloween is to finish the electronics to control the air and water valves, and to see if I have the parts to control the 110 volt strobe. Then, dry runs... then Halloween proper!

I think I'll spend some time this weekend reviewing the Boris designs.

Tue, Oct 21

Ahhh... those Train Mountain visits are very nice and relaxing, but they are putting terrible delays into my projects. It takes at least a day to wind up again after getting home.

Tonight I wired up half of the power control card, and did full tests of the Jumper. Worked like a charm. I'll eventually photograph everything and post pictures.

I went to test my water valve, but the hoses didn't fit the thread style. Tomorrow, another trip to Jerry's. <sigh>

I'm going to forgo the light - I don't want to mess with the 110 volt interface this year, and I'm using the light channel to get better control of the down action on the Jumper.

On the Boris front - I am running all the issues over in my mind. I want nicer joint attachments. No resolutions yet.

Wed, Oct 22

Score! Our friends at UPS gave me a visit today, and delivered unto me... Clippard valves! Rummage, rummage... shuffle, shuffle... argh! Those weenies gave me the wrong part ... a package of little brass thingies instead of the valve manifold. A couple of phone calls later, and it is all better. We'll be doing a swap through the mail, so I'll have my manifold soon.

Hmm... I'll have to detach the PIC board from the halloween system to test the new valves.

Speaking of Halloween, I soldered the other two power drivers in, finished wiring the control panel, and stuffed it all together. A cold run shows the computer is working, and all the buttons and lights. I'll do a hot run soon, with all the valves and stuff hooked up. <chortle>

After heavy thinking on Boris' framework - I think I'll keep my current design. I just can't find an improvement on the joint work . I'll just do some fine tuning instead of radical re-design. For instance, on the potentiometer, I'll make a better bracket arrangement. And I'll use some nice parts I found to join the Pot to the Shaft (more details when I pick them up from the store).

Thu, Oct 23

Didn't get to do much fun today - made about 40 pieces of jewelry for the shop, to catch up a bit! Gotta work to pay for the play.

Picked up the shaft coupler today. Cute little things. Three parts. First part has a set screw to attach to the potentiometer shaft, and two fingers on the other side. Second part is the same, but larger hole to fit my axle. Third part is called a ìspiderî (I like it already!) and goes between the first two. Small, light, simple. Cheap (mostly). I am really beginning to like my supplier of odd parts, Applied Industrial Technologies (previously, Bearings, Inc.).

Mon, Oct 27

Okay, Boris Fans - still waiting on that manifold. Called Air Components etc. today, and they said the factory would take two weeks to get it to me... for a part I should just make in my garage with a chunk of aluminum and a drill press. I don't think so. So they are going to try to get it from a different store.

Halloween countdown. The entire system is set up right now. And it works.

Had a scare yesterday - the pumpkin switch wasn't working... when triggered, it caused the whole thing to lock up! Still does, sometimes. It all worked on short wire, but run the 100 feet of cable through the kitchen window, and the switch goes nuts. Added a series resistor to it, and it seems much happier now. Don't know why, I was just running on a hunch.

I'll take some pictures of it all on Halloween, in context before the light fails. That means, it will take a week or so after this Halloween post before you can see it.

Thu, Oct 30

I just ended a mind numbing sequence of finding interesting Robotics and AI papers in my information links, downloading them, printing, then reading (or scanning). There is a 4 inch thick stack on my desk that I waded through. Now I need to buy a new toner cartridge...

This orgy of information download was precipitated by my finally getting a postscript viewing-and-printing system on my machine. I had tried Ghostscript, from GNU, before - but always had trouble getting it to do anything (does that make me a weeny?). This time, I also found Ghostview (GSView), which makes it all better.

I have an old paper I wrote on giving computer adventure game characters personality. Now, I am going to build on that and work this new information into a second paper on robots, learning, AI, and personality... which will be the background for Boris's brain development. Whether I get all of it in for Robot Wars or not... well, either way I intend to hang out in this robot world for a long time. Maybe I'll post those papers here when I get them cleaned up... a lot of it is blue-sky thinking, but it might help other people out some too.

Called Air Products again - I'll have the manifold on Monday! Ohboyohboyohboy...

Tomorrow is Halloween... all of the gear is sitting in a clump in the kitchen, awaiting deployment. Trick-or-treaters are working on their costumes, dreaming about their makeup, salivating at the thought of all that candy just waiting for them... waiting, like the monster in the trash can is waiting for all those tender, juicy trick-or-treaters to wander past its lair. My vote this year: TRICK!

Well, I'll hand out candy too... to the brave ones.

Fri, Oct 31

It's Halloween.

The trap is set.

The ìtrashî is out.

The control panel is blinking.

It all still works.

Now we wait...