Some of you might remember me from a while back. I built an SES arm but went too cheap on the servos, getting all 422s. I recently had some cash to paly with (a RARE occurance) and rather than order new servos for the arm, I ordered the brackets to make a brat using my 422s, which I heard will do just fine for the brat.
question is, how good is just fine? What kind of performance will I see off the 422s? I assume they will be able to handle a basic gait, but will they be able to handle the one leg stand, the standup from back or belly, etc?
I think all of those neat tricks were done with 422ās so that Jim could show us how a basic BRAT kit can perform all those neat show-off moves So 422ās will suffice.
Great! Iāll upgrade the little bugger as I can, since he is a great stepping stone to my ultimate hex project.
Now I just gotta figure out the best way to bolt on my little gripper from my arm project to the front of the brat. I figure I can have him lean backwards isntead of forwards to balance the weight of whatever heās carrying.
I just got my BRAT last week, I donāt have the sequencer but I have been working on my own program to control the BRAT. I found no problems at all doing a one legged lean, then adding in a fast kick. I also have him leaning way back on his knees and he stands back up no problem.
I think your only issue when you add the gripper and weight that your picking up. Anyway welcome to the BRAT Brigade looking forward to picks of a Gripper BRAT.
Well Iāve had a few thoughts as to how to attach the thing, but not the parts to try it out. Is it considered sacriligious to modify lynxmotion parts? Iām thinking the easiest way would be to cut the back plate of the LG (where the wrist attachment is) clean off, and slide a universal servo bracket in there. Id lose the back end of the servo clip, so Iād be depending on the strength of the servo horn screw to hold the gripper onto the bot, maybe some double sided stickyness ala the passive hinge. But for this kind of light application, I think I can live with that until I find a better way.
Also, on the brat itself, is there enough clearance between the legs to attach an L bracket to the U channel so it comes down between the 2 universal brackets that make up the hips? Please, no anatomical jokesā¦
OK lets give this a whirlā¦ http://i187.photobucket.com/albums/x3/KamenBots/hacked_grip.jpg
Heres my horrible mutilation of the back end of the gripper. I used a hacksaw blade to shave off the wrist mount and the left side tabs which I presume are for the LGA-KT (I dont have it), as well as clipped the corner and cut a notch to accomodate the bracket arms that support the servo. The ākeyholeā is to accomodate the topmost screw in the SES hole pattern on the bracket.
EDIT: After some tinkering I managed to mount the bracket upside down. This puts the SES hole pattern directly behind the gripper which could make lining it up on your bots easier. Those top holes could make good mounts for a cam! http://i187.photobucket.com/albums/x3/KamenBots/invert_mount.jpg
Please, no remarks about my spaghetti cabling. Due to unforseen circumstances, I do not currently have a WH-01 to hook the battery up, so Iām tethered for power now. That works out nicely since Iām also tethered for control! (RF was beyond this months budget)
I also included a shot of the back because of the way I mounted the SSC32. Following the assembly guide, the electronics carrier came down too far and blocked the backward range of the hips. The pics on the guide arent too clear so it might have just been me mounting it wrong, but either way mounting it horizontally works nicely. Now, since I dont have the SEQ Iām having to hammer out the sequences in lynxterm by error, error, error, and trial. Once I learn to drive this thing there might be some videos coming.
Oh yea, dont try this with with 422 servos kids, with the extra weight of the gripper it cant pick itself up when I make it fall down.
Some updates on gripbrat. Mounting the SSC32 horizontally is a good idea for range of motion on the hips, but a bad idea if you are using a serial cable, the weight of the cable drags the bot to one side. I was wrong about the 422s being too weak to let GB stand up, its just a matter of using them together. I have gotten him to take a single step without falling over, but the second step does him in.
My problems are the uneven weight distribution from the cable, and my lack of being able to send command strings at the right time to make the motions fluid. I assume the SEQ software will solve that one whenever I scrape up enough coin to buy it. For the cable weight, I plan to remount the board so the cable comes down like a tail as a temporary fix until I can afford a bluesmirf, and once my wiring harness gets here Iāll be able to get off the power leash. Probly going to pull the gripper too as it makes him a bit front heavy, and try it again once I can drive a stock brat.
Out of curiosity (and sheer frustration) has anyone managed to get a BRAT walking off lynxterm (or any other telnet app) alone? If so, Iād love to see your command strings
you could try making your own rs-232 cable from some smaller wire as you only need 3 conductors and you need to be more concerned about flexibility than current carrying capability. I was thinking that ribbon cable is usually 28 AWG and it is usually pretty easy to strip any particular number of conductors off of a length of it. You may find that dragging a tail on the floor could just as easily add or remove stability depending on (I am guessing) the center of gravity of the BRAT relative to where the wire exits the structure (the connector). There may be some value in controlling how the wire comes down to the floor, perhaps using some nylon or PET sleeving over it to still allow it to be slippery while limiting just how much it can flex. The analog in my mind is like a lizard tail where it is fairly flexible within a range of movement but it canāt fold back over on itself without a significant radius. It will be interesting to hear how this works out for you.
Way expensive, but you could use the below head phone extension cables from Radio Shack for the serial connection. For two way communication (tx, rx, ground) use the three conductor. The two conductor would be for tx and ground only. The two conductor is very thin light wire which would have minimal impact on small bots.
42-2562 three conductor 20ā $8.99
33-176 two conductor 20ā $7.99
Zoomkat, I plan on making my own serial cable as soon as I get the tools to do so, but your suggestion on using an audio patch cable was just brilliant. All I have to make are DB9 to 1/8 audio adapters to pull that off nicely.
I found another problem entirely unrelated to the bots design. I didnt want to risk these expensive parts falling from my desk so I kept the bot on the floor while working with it. My floor is carpeted. Yes, it took me this long to realize the problem. Now that Iāve gotten GB on a wooden board he is performing much better. Cable drag still keeps him from walking straight, but at least he walks now!
If you go to the DollarTree store (where every thing is just $1), you can usually find straight 15ā 4 conductor telephone extension cords and telephone splitters (female RJ11). With 4 conductors, you can send power (although not a lot), serial commands, video, and have a common ground. Some body needs to do a show nā tell in the electronics section on how to make your own serial cables.
So 422ās will suffice.
Iām thinking the easiest way would be to cut the back plate of the LG (where the wrist attachment is) clean off, and slide a universal servo bracket in there. Id lose the back end of the servo clip, so Iād be depending on the strength of the servo horn screw to hold the gripper onto the bot, maybe some double sided stickyness ala the passive hinge. But for this kind of light application, I think I can live with that until I find a better way.
Now that Iāve gotten GB on a wooden board he is performing much better. Cable drag still keeps him from walking straight, but at least he walks now!