Problems with Femisapien

17 posts / 0 new
Last post
kinakono
kinakono's picture
Problems with Femisapien

1)  she dont walk straight, esp with rechargeable batteries.  while on alkaline, it's much straight, but the turning is not accurate either.

2)  no wheels on her foot like the mini, she dont lift her legs too, any surface other than flat will immobalize her.

3)  i opened her up n found that her mic is way inside her chest plate, sound wave is harder to reach, any wind near the hole will be activated as sound wave.

roschler
roschler's picture

kinakono,

Try Juice Alkaline rechargeable batteries. Search for "Juice Alkaline rechargeable" on eBay. That's what I use in all my WowWee robots that need AA batteries.

-- roschler

kinakono
kinakono's picture

I'm pretty sure it wont walk straight even on Juice alkaline, or perhaps I got a defective unit?

roschler
roschler's picture

kinakono,

I wouldn't expect it to work perfectly straight, but you said it did better with Alkalines so I told you about the rechargeable alkalines.

-- roschler

kinakono
kinakono's picture

actually i knew about the alkaline rechargeable before you mentioned it. but i was pointing more to the problems of femi rather than batteries alternatives. i am thinking of hotwiring 2 additional 1.2v rechargeable to give the motor more juice and see if it works better, hope the fuse dont blow. anyhow, thanks for mentioning the Juice alkaline.

Shai Alyt NEMO
Shai Alyt NEMO's picture

kinakono said:

 1)  she dont walk straight, esp with rechargeable batteries.  while on alkaline, it's much straight, but the turning is not accurate either.

Personally, I did not tried to run her on rechargeables yet (but it is in my plans). I think, the lower voltage of rechargeables is causing the insufficient motor power, making her movements less predictable. And for all Robosapiens' walking systems - they are just toy-grade ones, not designed to provide any serious level of steering accuracy, especially on more complex (then a tabletop) surfaces. But overall - she is not a jungle explorer, she is a parlor performer.

Here you can see in action the prototype of a really impressive walking robot (but shurely of a different price level Wink ):

http://www.youtube.com/watch?v=W1czBcnX1Ww

 2)  no wheels on her foot like the mini, she dont lift her legs too, any surface other than flat will immobalize her.

It is a well-known shortcoming of her, which is a result of her walking system construction (the price for a more realistic (then that of RSv1) legs). For now I did not tried to somehow improve her in this sense, but for the future, I think something similar to the traditional japanese "geta" or "pokkuri" may allow her to walk better.

 

 

 3)  i opened her up n found that her mic is way inside her chest plate, sound wave is harder to reach, any wind near the hole will be activated as sound wave.

Sorry, I did not competely understand your thought.

 As for me, her audio-sensitivity straight "out-of-the-box" is too low. But this is curable (if you are not afraid of some hadicraft and self-making of a simple electronic devices). See "Robots hacks and mods" section. The final schematic for an audio amplifier for Femi will come soon (maybe even at today's evening).

kinakono
kinakono's picture

i just found out that diff surface makes a difference.  and i dont think making her shoes will make any difference.  she rely on the friction of the soles and swaying of her body to move.  making her skates would be better. 

rechargeables r mentioned in the manual, i suppose they should mention rechargeable r not recommended if it dont work too well, or make a chip to draw more power from rechargeables.  i feel that the leg motors r not powerful enough too.  as compared to the hands, i feel that the motor power r the same, which should not be the case.

if you open up her chest plate, u'll find that the mic is burried deep inside.  usually for mics, they r next to the inside of the plate, but for femi, it's on the chip inside.  this might be the cause of audio pickup too soft.  amplifiers might not be necessary if you could strip out the mic n paste it in the inside of her chest plate.

 

kinakono
kinakono's picture

some more problems:

4) screw holes are not covered, it makes her back very ugly. maybe if i got time, i'll find some pugs to cover them up n do some paint job.

GWJax
GWJax's picture

kinakono said:
some more problems:
4) screw holes are not covered, it makes her back very ugly. maybe if i got time, i'll find some pugs to cover them up n do some paint job.

If you want to make your own plugs for the holes I would use ShapeLock Thermo Plastic. This stuff is great for this kind of stuff and also making your own body parts. you can find it at www.shapelock.com if your intrested in it.

 

Jax

Shai Alyt NEMO
Shai Alyt NEMO's picture

kinakono said: ...  and i dont think making her shoes will make any difference.  she rely on the friction of the soles and swaying of her body to move.

Exactly! The friction of her soles! It can be significantly decreased by the shoes on high supports and it may (i'm not shure, but it looks possible) help to conserve her motor power. BTW, did you think that rubber patches on her soles are mere decoration? I think replacing them with significantly more thick (and maybe hard plastic) ones may improve the situation with "carpet walking". BTW, examine the feet of RSM. What are that (rubber?) half-spheres on his soles Wink

kinakono said: ...making her skates would be better.

Brilliant idea! But I fear, on skates she will not be able to walk at all (nearly zero sole friction!). At least, her skates' wheels must be turnable in only one direction. And it may present an additional trouble for her "turning efforts" as it is indeed a long and exhausting struggle for her to turn. In a long time passed childhood I have owned such a "robot" toy. It was powered by a spring, like a mechanical clock (no worry about batteries!) and "walked" forward, moving it's legs back-and-forth on such a "unidirectional skates" in it's soles.

kinakono said: ...i feel that the leg motors r not powerful enough too.  as compared to the hands, i feel that the motor power r the same

Indeed, they are the same. But gearboxes are different.

kinakono said: ...if you open up her chest plate, u'll find that the mic is burried deep inside.  usually for mics, they r next to the inside of the plate, but for femi, it's on the chip inside.  this might be the cause of audio pickup too soft.  amplifiers might not be necessary if you could strip out the mic n paste it in the inside of her chest plate.

To desolder the mic. from the board and move it closer to the openining in Femi's chest is indeed very easy, but improvement of her "hearing" will be nearly undistinguishable. I've tried this before thinking of an amplifier. It's pity that WowWee engineers make her audio sensor (and of all their products, in fact) so frustrating "deaf". Maybe, they expected her to be operated only by a bad-behaving always shouting aloud children? I think, amplifier is the only real (and now, practically proven) solution...

What about hiding the screws? Their presence did not offend me at all. She is a mechanical robot, after all, not an organic cyborg.

kinakono
kinakono's picture

Shai Alyt NEMO,

making the soles taller will indeed help her cross "jungle terrain" carpet. however take note that the soles must have friction. i would expect those rubber pegs of RSM to have better performance than hard plastic. remember, she's swaying her weight to walk.

if u've played mini femi, u'll notice she has wheels. the wheels move in both directions too. this will need a test to see if my theroy is right. but i've not found suitable parts yet. most likely i'll stip those racing cars for parts.

one idea i've thought of is to change her leg motor to more powerful ones. but i'm no engineer, so no idea how to start.

i'll try if without the chest plate, the mic works better. well, this definately counts as violating a girl. i hope i dont get sued. lol

Shai Alyt NEMO
Shai Alyt NEMO's picture

kinakono said: i'll try if without the chest plate, the mic works better. well, this definately counts as violating a girl. i hope i dont get sued. lol

And to remove her breastplate, you have first to take off her "panties"! Surprised

redstorm
redstorm's picture

femi is always falling, when i trying to issue a "Counter-clockwise step turn." command ([1], 14). while the opposite command "Clockwise step turn." ([1], 12) is ok. is anyone else have noticed this?
ps. im using novii remote from ppc, and these codes from "green" area (so the problem with command "step >>")

[1] http://www.pollensoftware.com/femisapien/femisapienbuttons.htm

Shai Alyt NEMO
Shai Alyt NEMO's picture

I did not noticed such a strange behavior.

Just an idea:
Try to observe if her "Full Lean Left" and "Full Lean Right" movements are symmetrical. Maybe your specimen of Femi is somehow inaccurately adjusted...

redstorm
redstorm's picture

i opened her up and look inside.. at first look mechanics seems to work symmetrically for both of legs in case of 2 this commands, but maybe i need to take a futher look. i guess, for this time, it may be caused by some nonsymmetrical hips movement, which is connected to her foot.

FreddyA
FreddyA's picture

Check that her feet position look the same. I had one foot on my Femisapien that would not move back completely to the same position as the other foot. I found that her tendon had moved out of its place. I thought first that the wire had slipped loose a bit, but it was actually the outer metal tube had jumped out of it socket making the wire pull a mm more than it should. I hope i explained it well enough.

redstorm
redstorm's picture

im currently trying to figure out the problem .. already found metal tube out of socket, and glued it. but this didnt solve a problem. also, while dissassembling, found arm position sensor wires that was damaged a little by arm moving mechanics. fixed, placing a wire away from moving parts.. now i notice, that her right leg moves backward not so free as like left leg. so i will try later to dissassemble the right leg motor assembly and try to fix it. by the way, i starting to like her even more after i looked inside by myself...
ADD: while further investigating into the problem, i noticed, that while femi is on in learning mode problem leg moves just like the another! but, when she is off, the same as was described above. strange problem. maybe, it is even in electronic part of her, not mechanical.
ADD: also, noticed that led indicator on arms is on while moving legs forward and backward on left leg, and only when forward on the right! indeed, some electronics problem. however, i noticed, that femi improoved a little her walks on carpet after my small fixes and adjustments to mechanics )))... i just guess now, it is some diode failed or so on. hope, i can fix it...
ADD: noticed also, that the left arm also enters this mode. so, when femi is off, if i try to move her right leg, or left arm for some time - it enters the "forward move led indicating, backward - not" mode. i can't guess, is this normal, or not. (im only starting to understand the functionality realisation, so im on a 'blind guess' way now.) can anyone tell, if the leds indicating on a backward move of the right leg or left arm while moving them for some time when femi is off?