Pressure Sensors

[PeterW]

I know I am probably getting ahead of my self as I don't get my Bioloid until Christmas but it seems to me that four pressure sensors on each foot would be very useful in helping the robot balance. In the free copy of Robot Magazine that came with the bioloid from Robbosavvy I see Trossenrobotics offer some that look just about right.
http://www.trossenrobotics.com/store/c/ ... nsors.aspx
Has anyone tried these, do you know of any others or am I on the wrong track altogether.

[JonHylands]

http://www.bioloid.info/tiki/tiki-index ... re+Sensors

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

I've been having trouble building this board - specifically, getting the micro-controller to reflow correctly to the board. I'm going to give it another try over the holidays now, and hopefully will get some more boards in stock...

- Jon

[NovaOne]

Peter, are you thinking of using flexiforce sensors or interlink FSR's.

I'm not having much luck with FSR's, at the moment. I'm hoping someone will point me in the right direction with them..?

I wish I had seen the flexiforce sensors on trossen's site earlier because, it clearly says that they are more accurate.

[PeterW]

Looks like I must be on the right track.
I have actually been on your site before looking at your rather smart minisumo but I never scrolled down far enough to see the bioloid. I have never tried reflow soldering I just place a thin deposit of solder paste along the pads on each side of the chip. I then carefully place the chip and when aligned touch two opposite corner pins with a fine soldering iron. After a final check of the alignment I place the iron between each pair of pins in turn until the solder flows out of the gaps and around the pins and pads. At the moment I have gone upto the Atmega48/168 (32pins) but my next mouse board will have Atmega128 (64pins) and I don't think I shall have a problem with them. Here is a close up of a 48 on my wallfollower mouse the whole board is hand soldered.
http://i109.photobucket.com/albums/n66/ ... CF0710.jpg
It could do with the flux being cleaned off but the joints seem solid enough.

[PeterW]

Sorry not trying to ignore you NovaOne you must have posted while I was uploading the picture. I wasn't quite sure I haven't gone into it in great detail yet but I had assumed the 0-25lb flexiforce but was hoping on cost grounds that the FSR's would be good enough. Mind you as I mantioned above I still haven't got the robot yet so I am rather jumping the gun.

[NovaOne]

Peter,
I'm waiting to see what results Bren gets with the flexiforce sensors he has on order.

PS
Your maze robot looks familiar, I think we met last month at ROBOtic'07?

[JonHylands]

Looks like I must be on the right track.
I have actually been on your site before looking at your rather smart minisumo but I never scrolled down far enough to see the bioloid. I have never tried reflow soldering I just place a thin deposit of solder paste along the pads on each side of the chip. I then carefully place the chip and when aligned touch two opposite corner pins with a fine soldering iron. After a final check of the alignment I place the iron between each pair of pins in turn until the solder flows out of the gaps and around the pins and pads.

That's actually quite easy to do with chips that have the pins sticking out. Unfortunately, the MLF-32 I chose for that board (and my IMU) has the pins tucked under the board, and you can't really solder them by hand.

I may end up redesigning the board to use the 32 TQFP package rather than the MLF-32...

- Jon

[JonHylands]

I'm not having much luck with FSR's, at the moment. I'm hoping someone will point me in the right direction with them..?

NovaOne,

What problems do you have with FSR's?

- Jon

[NovaOne]

Its seems that I either can't actuate them repeatably,or they are not accurate enough?
Its all detailed in this thread:
http://robosavvy.com/forum/viewtopic.php?t=1991

I may be wrong, but I have come to the conclusion (from the data sheet) that i need a precision elastomer "puck" fixing "perfectly" over the sensing area of the FSR, and even them i can only expect to get a ± 5% accuracy? assuming this is repeatable, ± 5% maybe OK for accuracy...but its early days?

Do you know what the best material to use for the elastomer might be?
The backing is self adhesive, suggesting the sensing area may be exposed to damage facing downwards? Also any elastomer would increase the friction between the foot and the walking surface, hence affecting "walking" gaits that require slipping?
It might make sense if the recess on the foot of the bioloid was slightly domed in the centre?

[PeterW]

I hadn't noticed you were using that package Jon as you say the TQFP I use are much easier to solder. It looks as though a lot of work is already going on in this area I hope once I get sorted and up to speed I might be able to contribute something.
NovaOne I am not sure but I think we had a discussion about the optical edge sensors used on my minisumo as you were looking for some to detect the edge of steps in the feet of you humanoid. The Maplin ones are now listed again but out of stock so you might be able to get some soon.
http://www.maplin.co.uk/Module.aspx?Mod ... &doy=24m12
Can anyone tell me what are the connectors used for the BUS and where can you buy them.

[JonHylands]

Can anyone tell me what are the connectors used for the BUS and where can you buy them.

http://www.bioloid.info/tiki/tiki-index ... AX-12+page

I'm pretty sure Maplin carries them as well. The main, straight 3-pin header you find on the servos is Molex part # 22-03-5035. The 3-pin plug you find on each end of the wires is Molex part # 50-37-5033. A 100-pack of the crimp connectors for the plugs is 08-70-1039.

On another note, I bought a crimp tool from Digikey, and I can't believe I've spent so many years making up cables, and soldering the crimp terminals in place. The crimp tool makes it soooooo much easier...

- Jon

[PeterW]

Thanks Jon using your part No's I have found them at Farnell
http://uk.farnell.com/jsp/Connectors/PC ... tid=142126
http://uk.farnell.com/jsp/Connectors/In ... tid=142151
http://uk.farnell.com/jsp/Connectors/Co ... tid=142163

[NovaOne]

Thanks Peter, I'll ask about the maplin ones, although I've been looking at these RS devices recently:: 497-5532

What do you think?

[PeterW]

That looks a better choice although that version is the non ROHS version which is now obslolete. The SFH9201, SFH9202 and SFH9240 are the new versions and are even cheaper. These are surface mount devices and are orders of magnitude faster than the maplin ones. I think I will invest in some when I next order to have a play. Happy Christmas by the way. I can't open the bioloid until the rest of the family arrive.

[PeterW]

As well as fitting pressure sensors in the feet one of the other thing I was considering was upgrading to LiPo batteries which I use in all my other robots. Having seen Meltdowns modified feet
http://robosavvy.com/forum/viewtopic.php?t=2024
I recon I might be able to kill two birds with one stone and get both the sensors and the LiPo's in the existing feet by fabricating new feet brackets. I might have to gain a little height but not only do you gain in stability but you have 250gms less to lift when standing and only ever have to lift half of the battery weight at any one time. That is a great idea Meltdown.