Let's Make Robots!

Houston (WIP)

Now? Get's in the way
robot_cost_and_sources_sheet.xls18 KB

There's a science fiction novel I have, (About a robot of course.) Where a secondary character attempts to use some time management software to schedule a minor task he's about to embark on. When the computer informs him he can do it in (as i recall) eleven years, he gets insistent and demands it sets this task to a higher priority.  SO it lists for him all the other high priority tasks that MUST be done first--Repair of structure that's about to collapse, requiring  tools that are in unavailable and, parts that have to be made, etc. In the end,  he gets so frustrated and angry that he beats his newly adopted robot almost into scrap metal. (Well, there WERE other extenuating circumstances, when the robot counted to eleven. But I digress.)

I know just how he felt.

Houston started back in, October. I think. When I ordered a bunch of stuff from YourDuino, (Before they became the LMR favored supplier.)  For the life of me, I can't think of anything from that order that's actually gone into Houston, or will.

And that's been the story of this project. I get an idea, put some money aside for parts. Order the parts and while waiting for that proverbial slow boat, I do some research and testing. By the time the parts arrive, I decide something else would do a better job, or I'll find a better way to accomplish my goals and begin again. And then as I make progress, I get to a point where I can't go forward--Indeed at times I couldn't even start--until I get ahold of something I haven't even ordered yet. And trust me--any time I jumped the gun I ended up starting all over again.

Here's an example.

This is a scaled drawing of Houston's base.  I made it so I could determine the spacing of parts I had, and other parts I intended to order.  Above this  deck sits a single board computer all alone on a deck of its own.  In the center of this deck is an 8.4v volt battery pack sitting inside a Rover 5.  In the foreground is a 4 output DC-DC converter that would have been the perfect power supply had I bothered to think for a minute, calculate a little ohm's law, and determine that a computer, Rover 5 and a dozen or so servos are going to need considerably more than a 15w power supply.

Also, do you see a hard drive anywhere?  No it's not on the (unseen) upper deck.  The only thing there is the SBC.  My Houston blog has more details.

Here's how it turned out when it came time to put things together.

The 3 blue boxy things are three DC to DC converters that act as (from left to right) a 5 volt power supply, primarily for the servos.  A 7.2v supply for the Rover 5's motors, and an 18v supply (I thought it would be 12v like all the OTHER SBC's I'd ever bought.) for the computer.  The batteries in the middle were for sizing of a 14.4v 5a battery pack I'm currently assembling out of individual nimh c cells.  (Another tale of woe I'll get to in a future post.  Me and my bright ideas.)

I ended up using an Arduino 2650, that I already had, instead of the Redback Spider controller I would have had to wait (and pay) for.  Sorry Oddbot.

I also managed to squeeze in the  notebook sized hard drive I forgot, but you may notice the red board in the background isn't on the deck any more.  (Challenge--What is it?  I've never mentioned it or what it actually does)  I might still use it, but if I do, it will have to be mounted higher up somewhere.  Probably a good idea to keep it out of the way.

There IS a minor improvement visible.  I took the center section that I cut out, cut a slot on one edge for the motor wires and. . .well, as cool as Rover 5's are, their insides tend to be a scary mess of wires.  Except for mine.

Anyway, I have this thing about posting too early, and having dead projects staring at me from my monitor.  But then, I realized that I HAVE seen a number of projects here than started out with a drawing.  And I've watched ChrisTheCarpenter bulid, and then REBUILD his magnificent CNC machine--chronicling all his mistakes, literally over the course of days. (Never mind how intimidating that time frame was.  heh.)

I've got lots more to talk about.  Building a battery pack.  Figuring out what the hell you actually DO with an Xtion Live Pro, servos and heavy arms, juggling Arduino pins, resource hungry motor driver boards and too many servos, and. . .

Well. . .it will be there.  With pictures and everything.  Maybe even some video. Trust me.

Oh by the way, did I mention I'm realizing a life long dream and having a ball?

EDIT 3/16/12  I put together a visual chart of Houston's hardware configuration.  Comments welcome.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I know how this goes.  On project after project you will see something someone else build and its looks GREAT!  Brilliant!  Cool.  It looks so easy like it took no effort at all, but that is almost never the truth.  In many cases you are looking at the 3rd or 4th try and this is first one that worked at all.

I have been here more times than I want to count.  When you have something that works, making small improvements often works, but when you go into big new changes, you have to hit ...every speed bump... in the road to get past them.

The trick is to realize that everyone does this and not get discouraged.  Also it should be fun.  Slow and steady will get you there while keeping your expectations for the current version reasonable.

This actually looks like a very interesting and challenging task.  Its quite tall and looks like it could do some very cool things.

        Good luck with it,


Thanks.  kingart3.  That's just the kind of "LMR love" that will get this project completed.

It finally occurred to me that a WIP is a better post than a polished robot.  Why follow something that's done?  The technology to build the robot I've always wanted is just about in reach.  And I've finally managed to gather pretty much everything I need to get it going.

Next is the boring task of figuring out exactly which Arduino pins will accomodate all those servos and motors.  Then I can put the SBC back on it and figure out how to make it DO something.

I love the grand scope of your project even though I am not a hundred percent sure what it is going to be when you are finished.  I assume some sort of cool, servo armed rolling around super robot that will scratch your back and clean out the litter box.  Your comments about ordering parts that you didn't end up using or researching better methods for solving problems than the ones you had originally envisioned really spoke to my own experience trying to get my first robot project for this site completed (I am starting over after ruining my first picaxe project board, at this point I am pot committed with picaxe)

Keep at it!  Don't let the robot win!  I would love to see the finished product or even some more build diary entries.


Thank you for your encouraging comments. 

It's not that my project is beating me--It's just that between accumulating the parts and some of the failures and false starts--it's taking so long.  And actually--I've got all the important stuff ready to go.

I've started robot projects a number of times in the past.  Decades.  Then something or other happens.  Money dries up.  Time dries up.  I come to the limit of what I understand.

I've got all the parts. Except for some cheap unimportant "arty" stuff.  And the vote is still out on a stack of batteries I savaged.  The ironic thing about the batteries--Building battery packs is something I used to do fairly regularly.  (Rechargeable tabbed D cells used for 1980's point of sale battery back ups.)  This time around, I started out REALLY carefully.  Made a few simple mistakes, which ballooned.  Ended up with a bad cell, and some shredded plastic.  Not a total loss really, it just irks me.

So far, this robot is a victory.  It's going to roll around and listen and respond and find it's way around the house better than I imagined it would.  (Now I'm jinxing myself.)  The arms probably won't be strong enough to actually DO anything--Actually, back scratching might not be a bad task for them.

I might be grumbling a little, but maybe that's because I feel it's close but still just a little beyond my grasp.

But I'll get there.


Wow, nice project! I hope the Rover 5 is sturdy enough to hold everything you want to put on it. I had like 3 versions of my butler robot, functional enough but not programmed since I suck at that chapter. I can program microcontrollers but not PCs. Not that it's impossible for me to learn, but I kind of have little time here and there and I tend to forget what I did last time and I have to use part of the available time to remenber where I was and what I was thinking. I see that you're on the right track here, skipping a few of my steps so I think you'll have a better chance of success than I had. So good luck, work hard and keep us posted! I should do the same and post a WIP of my current version of the robot, even if I am not working on it atm.

Uh. . .yeah. . .You hit a nerve there.  I'm not sure that the Rover 5 can support everything I've put on it.  Just about half the weight is batteries.  (Which I've found is typical of a robot project.)  I'm pretty sure it will move.  And I don't have any treads to throw.  I'm pretty sure it will move with no problem, but I'm not sure how well it will hold up.

As for software. . .I've got a few ideas in mind.  I've been flirting with ROS for a long time now, and most components were chosen with ROS in mind. (Such as the Xtion)  But I'm using cheap servos, and a base no one else is using.  (I'd like to see a Rover 5, D4CMC, and Arduino stack take it's place in the ROS repositories.

That being said. . .I talked to GroG about his myrobotlab.  It has a lot of the same features as ROS, but with a milder learning curve.  But we've had some issues getting the new (unreleased) version installed, and circumstances have made him unavailable for the time being.

I've not been on this site that long but it took me a long time to build something very simple (before I stumbled across LMR).

I always set myself a simple attainable goal but keep in my head the long goal so you can make things generic that need to be. However, I'm sure it is probably faster to make something bespoke and then rebuild it on version2 as you learn a lot from version1.

Version2 is then version1 + another small attainable goal.

Time is the thief. Everyone wants it and having enough to go round is problem. If this working marlarky didnt get in the way, there would be plenty to go around.

Now those rich victorian gentlemen who could tinker all day without ever really working had it right ... :-)

You know, you made a good point about simple goals.  Currently there isn't much on Houston that's especially difficult.  (Future plans could be a different matter.)  It's all been done here on LMR.  There are plenty of arduino projects, arm (as in limb) projecs.  Rover 5/D4CMC projects.  What I have here is a bunch of simple things that I keep jumping around between.

The best thing I could do right now is get everything plugged (permanently) into the Arduino.  Even if I have to wait a couple more weeks to get new batteries, I could finish wiring my power supply and get into the real developing.

I got a good taste of my "own advice" today while trying to load my PICAXE 28x1 on my developement board.  I am a computer programmer, but not an electronics guy so when I got the header pins soldered in the board last night, I thought I was home free.  I didn't take into account installing the driver for the PICAXE USB cable under Windows 7.  What a PAIN!

I am guessing it took 2 or 3 hours.  My first mistake is I used their auto installer which loaded the unsigned driver.  It took a while to realize that an unsigned driver was the problem, then I went off trying to figure out how to force Windows 7 to accept an unsigned driver.  A few hours later, I check the PICAXE web site again and discovered the SIGNED DRIVER and I downloaded that.

Now I am going to win right?  Uh...no....

I loaded the driver, but you have to go back and also load the virtual COM port driver. (Basically you load drivers twice to make the cable work).  As it turns out this is covered in their documentation, which I read closely then ignored after having the first problem.

After I loaded the driver for the virtual COM port suprise it worked.

I then wired up the "blinking LED" example and it now works great.

Good news, I win (tonight), but what a wrestling match that was.

Yeah but. . .Doesn't it feel good when you make it over that hill?