Finally received my new Bioloids: HOWnotTO build a Bioloid

[StuartL]

Yesterday my Bioloids turned up (yes, two of them). I wasted no time at all in assembling the humanoid straight away (anyone say "Steep learning curve"?).

This post is intended for new users to learn from my mistakes.

I had some minor wiring problems in the build as I didn't initially realise that there were both 18cm and 20cm data bus cables and put the 18cm in the ankles, requiring a rebuild of the ankles when I ran out of 18cm cables later in the build.

This frustration aside and taking half the day putting those damn nuts in place (please Robotis: Preinstall those nuts. Glue them in. Don't make me put any more in!) I had a humanoid built in about 2.5 hours.

Obviously I couldn't wait to try it and installed the software, updated the Bioloid firmware, loaded the test routine, couldn't figure out why the LEDs weren't lighting, couldn't figure out why it wasn't assuming the test position but thought "Hell, I'll just put the demo code on it and see what happens".

Of course I know NOW why the test code didn't work. I put the demo code on and pressed the start button. I blew a fuse. I had been warned about the blowing fuses problem by Royce at RoboGames so I fortunately didn't spend the next 3 hours trying to figure out what was wrong. Instead I popped the top off and put the supplied spare fuse in. The console had reported loads of "Invalid page" errors which confused me for a minute.

I guessed what was up and scanned the manual. The manual says loading the motion data is optional. Of course what it means (in hindsight) is that the motion data isn't overwritten by the BCP but plain doesn't exist on a new configuration. I uploaded the motion data and while the test program still didn't attain the crucifix position the LEDs were now lighting.

I loaded the motion editor and verified that the motor positions were being read correctly.

I then gritted my teeth and reloaded the BCP and ran it. It assumed the standing position! The clap test worked and so finally, at 12:30am, I went to bed.

Summary:
- Make sure you measure all your cables before plugging them in!
- Make sure you verify all motor positions before tightening those screws (this one didn't catch me but would be a BIG problem if you didn't get it right).
- Make sure you load the motion program before you run your BCP. If you don't BAD things will happen.
- Buy lots of fuses. 20x5mm 250v 5A. Buy loads, they're cheap.

Things I didn't like:
- The absence of pre-installed/glued nuts. This seriously extends build time.
- The absence of automatic torque overload control preventing fuse overloads.
- The size of the holes for the screwdriver to fit the CM-5 into the case. None of the screwdrivers I have with me in SF fit in these holes and still reach the screw heads. You need a screwdriver with a very long shaft.
- The zig-100 module is solder on, not plug in. I will have another look at this later next week and see if I can solder on a socket for it to fit in.

Things I did like:
- The flexibility of the kit.
- Torque feedback capability.
- The clarity of the instruction manual build instructions.
- Future build possibilities (e.g. double knees, grippers etc).

[cdraptor]

Hey StaurtL or anyone that has a Bioloid. I found an issue and wanted to see if anyone else has run into this. The S1 screws are sized to attach frame pieces to the Servos - the size of the ensures they do not go all the way through into the Servo casing. Well on 2 of my servos so far ID3/ID7 I ran into the issue where the S1 is not catching in the hole, it's as though the threads are set a little too far back so it never catches, but sets flush in the frame, of course it slides right back out, so I have 2 servers with just 3 screws connecting the frame.

At this point I put together the Walking Droid, now doing the Battle Droid which is basically the walking Droid with the arms. Next I am looking to do the King Spider. So I haven't gone through any more than servos 1 - 8 at this point.

This happen to anyone else, the 3 hold pretty solid, but I am wondering if I should look at getting those servos replaced??

[Dewey]

cdraptor

The problem with S1 screws not 'catching' the nut built into the AX12 head is easy fixed.

Take the AX12 and a long screw. By hand gentle screw the screw into the offending nut on the AX12 head until it 'bites' on. Now take a pair of pliers and gently pull the screw away from the head, the aim is to pull the nut back into its seating in the head ( If you lever the pliers against the body of the head as you do this you are not putting any load on the AX12 bearings). Remove the long screw and now an S1 screw should thread on OK during subsequent assembly.

Dewey

[JonHylands]

Dewey has it right, but to make sure you know what is happening, here's a picture of the back of the servo plate:



What happens is the captive nuts slide out of their slots, towards the main body of the servo, and then the standard screws can't reach anymore.

- Jon

[billyzelsnack]

Those nuts bugged me at first, but after awhile you'll get the hang of them. The trick is really to have the right size screwdriver to help put them in.

[cdraptor]

Awesome guys thanks for the advice, I follow what your saying and I'll try it out tonight. Yeah I found having some good screwdrivers - I have a precision kit with some length to it. Also tweezers and needlenose pliers are excellent.

[cdraptor]

Thanks for the advice, I tried it out tonight, worked perfectly. Jon something you put on the bioloid wiki. Anyway I whipped together the BattleDroid very quickly, once you get the hang of it, it's pretty easy, worst thing is the running the wires, but I now fully understand the diagrams and they are very good. First build attempt on check assembly and the Walking Droid was dead on. So I'm going to spend some time making a bunch of motions. I'll load them up to Robosavvy.

[Kess]

Glad I found this advice as I'm literally just about to start the full Humanoid build.

I've only tried one of the beginner projects so far and I agree with most of the negative comments, especially the lack of pre-installed/glued nuts and the screwdriver problem (none of mine will quite reach the bottom of the CM-5 holes either). Given the cost of the kit it's a shame they didn't throw in a screwdriver.

I'll add a few other things I've found disappointing:

- The CM-5 unit looks awful - cheap drab grey plastic. It really detracts from the quality of the rest of the kit.

- A printed copy of the full User Guide would've been appreciated.

- The QuickStart guide has such a stiff spine that it won't stay open on the table while I'm trying to follow the instructions! I've ended up printing the relevant pages from the CD. A ring-bound manual would've been better.

Hopefully the positives will outweigh these negatives.

Kess

[jerome]

Good work StuartL. I though I would post my experience after assembling my Bioloid but never did it ...

Thanks also Dewey for the screws not "catching". I though those holes were not threaded! However the screws become loose on my kit. I have to tighten them from time to time.

My personnal impressions on the kit:
+ good building instructions
- the building of the humanoid is easy, but ends with two critical operations that could ruin your work and force you to move several steps backward: the taped nuts, and the assembly of the CM-5 that requires a long and magnetic screwdriver
-- the wiring. I never got it right. The wiring on the Robonova looks much cleaner although it is supposed to have many more wires. My humanoid just leaves with all those loose and inaesthetic wires.


Overall, very satisfied with the Bioloid, especially because there is a lot of infos on it and it is pretty flexible.

Jerome.

[jerome]

And also ... the Bioloid is heavy enough like this, I do not thing they should preinstall nuts everywhere! It is not that bad after all, you do not have to hold the nut with one finger while trying to fit the screw with the other hand.

Maybe they could optimize a bit the design so that the nuts are easier to insert ...