My first Bioloid

[PeterW]

I am new to humanoid robots and although I have a good deal of experience in other fields of hobby robotics I do feel some what out of my depths in this field. So what I am proposing to do is to put my ideas for modifications down here in detail before carrying them out so that the more experienced and helpful of you out there could comment on them before I make any expensive mistakes. As I have mentioned on this forum to start with I am proposing to replace the 2000mah batteries in the CM5 unit with two 1000mah LiPo's. (one in each foot). I further propose to fit four 0.2" FSR pressure sensors in each foot. A PCB again in each foot with an Atmega44 will monitor the pressure sensors and the batteries and hopfully comminicate with the CM5 over the AX12 bus. I havent decided if I should switch batteries one at a time or regulate power from both. The batteries are some older ones I had that are rated at 10C so thats 10Amps from each. Here is a sketch of the mod so far.
http://i109.photobucket.com/albums/n66/ ... idFoot.jpg
Any comment even negative ones will be well received.

[JonHylands]

I would advise against putting batteries in the foot. It makes for good stability while the robot is standing still, but makes walking much harder.

On the pressure sensor side of things, you're looking to do what I've done:

http://www.huvrobotics.com/shop/index.p ... roductId=4

I've had a really hard time making boards successfully with the MLF-32 footprint. I have successfully made 6 of them over the holidays, and hopefully I'll be able to make some available from my store soon. The source code (which runs on an ATmega168) is freely available if you want to use it with your own boards.

- Jon

[PeterW]

Are the problems with walking associated with the extra weight in the foot, in this case 65gms, being more off centre and effecting the balance or just the fact that the leg has more mass and is therefore more difficult to control. Also presumably the increase in height of 8mm won't help.
On the software side a look at the code would be very useful as I have yet to use the serial comms on the Atmega range of processors.
Either way thanks for the input this is exactly the sort of help I need. I might try some experiments with just adding weight into the feet to see how much I can get away with before I start modifying the robot permanently.

[JonHylands]

I'll point you to a thread from a year ago on RoboSavvy here:

http://robosavvy.com/forum/viewtopic.ph ... c&start=45

read the fourth post down on that page, by BaurMech...

"Yeah, I noticed how much strain is put on the hip servos when Rook's Pawn (v1) lifted his leg. His movements had to be dramatically slowed down or he'd loose his balance on one leg. The extra weight is nice when he was standing in place, but kinda' defetes the purpose of having a walking bi-ped robot. Wink

I'm looking for that happy medium between stability and agility."

- Jon

[PeterW]

I have just done a quick test using the standard walking sequence that comes in the standard Humanoid test program. First it walks foward then backs up and turns when you trigger the front sensor. I did this first with the standard robot and then with a 1000mah battery pack blue tacked to each foot. There appears to be very little in it to my untrained eye see below.
http://s109.photobucket.com/albums/n66/ ... ioloid.flv

[JonHylands]

Yeah, if all you're looking to use are a couple 1000 mAh batteries, you'll probably be okay. However, you're not going to get much more in the way of run-time out of 2000 mAh, whether it be lithium or NiMh. You'll get some, because (assuming you're using 3s cells) your voltage will be higher for longer. But realistically, you can't let a 3s battery go down to 7.5 volts anyways, so you've still only got about 2 or 3 volts drop, from fully charged to fully-discharged.

- Jon

[PeterW]

I wasn't really looking for longer run times as I assumed that the recent doubling of the Bioloid pack from 1000 to 2000mah was probably enough especially as you only have to lift 65gms of battery instead of 250gms. That comes from my antweight robots where I never carry a battery bigger than needed for a couple of fights due to the 150gm weight limit. I just make the batteries quickly exchangeable and have about five sets so they can recharge while the robot is in use. In this case with two packs I might have to cut that down to two spare packs on cost grounds. Any way it is starting to look like this mod is worth a try so I will keep you informed on any progress.

[PeterW]

Just been on the trossenrobotics site for the FSR400's and 10 of the sensors are $57 dollars which seem quite good but delivery is nearly $70. Is there an online distributor in this country or an international one with more reasonable postage charge.

[PeterW]

I seem to have solved the problem Steadlands.com in the UK sell them. They are not available to buy directly on line but I emailed them last night got a quote today and have ordered them by phone this morning and they are due tomorrow.
10 FSR400 £30 + P&P £7.50 + vat £6.56 total £44.06. That’s compared to over £60 from the USA plus possible import duty / vat and a much longer delivery. They were extremely helpful and are quite happy to deal with individuals and small quantities.

[PeterW]

The FSR400 sensors came today less than 24 hours after ordering. I haven’t tried them yet but I have rethought the new feet brackets. They now not only have space for the batteries and the board but also allow the pcb to extend to the front and back of the foot. Not sure about the back but the front may be going to have IR send receive for obstacle / stair height detection.

[PeterW]

Starting to make some mechanical progress.
Bracket design and PCB outline finalised.
http://i109.photobucket.com/albums/n66/ ... dFoot3.jpg
Brackets made
http://i109.photobucket.com/albums/n66/ ... CF0732.jpg
Brackets fitted
http://i109.photobucket.com/albums/n66/ ... CF0738.jpg
Loads of space now for the sensors PCB and batteries and only 8mm increase in height.

[JonHylands]

Wow, that's really nice. It looks like a lot more than 8mm...

I'll give you one piece of advice WRT the PCB - don't use MLF package chips, unless you're an absolute expert at surface mount stuff.

- Jon

[PeterW]

Jon it does look more but I think that is because there is no bracket through the middle of the foot.
I always use the TQFP package as they are not only easy to solder but you can inspect the joints after.
You mentioned that the software for you pressure monitor board was freely available could you point me in the right direction as it might help with my first attempt at serial comms on the Atmega.
Just found some cheap 1000mah 15C 3 cell packs for $12.10(USD) at
http://www.unitedhobbies.com/unitedhobb ... oduct=5470

[JonHylands]

The code can be found here:

http://svn.hylands.org/avr/bioloid-foot/

Note that it uses a bunch of other files on the repository...

- Jon

[PeterW]

Thanks Jon I can see I might have to brush up on my C as it has been a few years since I used it. All my microcontroller code for both MiniSumo and MicroMouse have been in assembler.