Custom built Phoenix-like Hexapod

Lynxmotion Legged Robots
41

Hi pyrrhicpk, hear that your phoenix is standing up and can move. Congradulation!! I’m for several weeks, now only going to start on my Phoenix. I have a question to ask you. How you actually export the PEP out and import into VSEQ ? I’m using Excel 2000 and when I was trying to save as CSV file, I only can find CSV (Comma Delimiter), I can’t get Semicolon seperator. I’m using Excel 2000 Premium. Please help! Thanks!

42

You can change the seperator in the regional settings of your pc.

Start->Control panel->Regional and language options.
Then click “customize this format” on the “Formats” tab.

Xan

43

Yeah, thats the right way to customize the separator in windows. otherwise, before this knowledge I had been saving the CVS by opening it in notepad and replacing all the commas with semicolon.

Thanks for the compliments, but so far I have just made some body rotations and some stand-up/curl positions. Meanwhile I found the challenging task of determining the FK and IK, which now I have set as my very next milestone. I agree, its cumbersome, and re-working an already worked theory is not suggested generally but by the time you understand accurately what’s going on behind the digital chips as the bot walks, it gives even a better feeling :slight_smile: .For the matter of action and plug&play, zenta’s PEP suffice all needs :slight_smile:.

Anyways, looking forward to see some updates from you.
Regards,
Umar

44

Hi zenta,

I worked through your tutorial about 12-Steps-Ripple Gait and below you may find what I achieved :slight_smile:

youtube.com/watch?v=4LaVqtc-qqo
**
from a different angle**

youtube.com/watch?v=mZKNnkkEOfE

As you can notice in the video, the gait is not running as smooth as we see in your videos or in Xan’s. I tweaked around some settings in the PEP,but still there is an abrupt motion of legs which is apparent in the video.By abrupt I mean, its spontaneous in comparison to the fluid like motion you and Xan have achieved in the gait.

**So, where should I start from now? would reducing the step size result in better flow? **

Moreover, If I want to loop the gait for a number of times so that the bot walks straight for a while before stepping down, what should I do?

Here is the PEP sheet which I imported. The step where 12-step ripple gait ends, I have added the “All1500” position to stabilize the bot before steppng down. Is that causing the jerk?

img232.imageshack.us/img232/3560/70072758lh8.jpg

Another question is that, whenever I open the Sequencer in VSEQ, a dialog box pops-up displaying that “No servo locations are stored in SSC-32, Initialise the home position and enable all servos from the ambulance button”, the message is something like that. I have stored the home position in the CFG file which I load as soon as VSEQ asks at the initial startup. And is there some way to enable all the servos at once? by that I mean a single click to enable all the servos in a project than selecting each servo and enabling it by checking the enable servo checkbox.

Zenta, Xan, your guidance is required in particular about making the gait smooth and about anything suspicious you suspect from the video.

I appreciate your time and cooperation.
Thanks,
Umar

45

Hi Umar,

It looks like the single 12 step ripple gait are correct. If your servos are correctly calibrated in the ssc32 configuration it should be ok. I just looks not so smooth because of what happens before and after the 12 step ripple gait.

Try to run only the 12 step ripple gait in a loop with Visual SEQ, use the advanced player in SEQ.
Another thing is that you should start and end the gait with a transition from the all 1500 to the first step position in the 12 step ripple gait. I think I gave Harry Mueller an explanation of how the transition work, I’ll see if I can copy it into a post later.

46

Hi Umar,

Here it is:

The transition gait is not a walking gait but a method I use to move all legs from one pose to another (therefore the name transition).

I use the transition part a lot when I create more advanced moves. You can also do this manual by moving one leg at the time, but this function can be a real timesaver.

Example:

  1. Start with the standard pose when Phoenix have all coxa’s set to 0 deg.

  2. Select transition gait, and the textbox show:â€

47

Hi,

Zenta, this was awesome. it makes much sense now.
thanks a lot for such an explanatory guidance. Concluding, if I got it right then there should be a transition part at the begining and at the end of every gait?i’ll definietly see how it runs after integrating it within the gait routines.

I suspected that a few of the servos were not aligned properly, but when I tried to offset them from the SSC-32 registers, I recieved some garbage values in the offset and width columns. See the screen capture

img91.imageshack.us/img91/1853/ssc32vh2.jpg

I tried reading the values again and again,and every time the values kept returning either empty or garbage values. Moreover, the offset and width sliders were non-responsive. Any thoughts?

I would like to interject my previous question again to get some confident confirmation about the servos sound I am suspecting. There is always a theethering sound in some servos in loading condition. Is this a normal behaviour? Or does this has to do something with the battery condition or if the weight is greater than the maximum torque, servos can provide?

Thanks,
Umar

48

Hi everyone,

Has anyone got a higher resolution of these pictures? I dont remember from where I copied these pictures, but the resolution is too low to read the details.

img139.imageshack.us/img139/1560/19a9mk1.jpg

img137.imageshack.us/img137/2383/42b8yv8.jpg

Thanks,
Umar

49

Thank! I’ve got it done! Now trying to create some movements!

50

Hi,

Zenta, ripple gaits are working fine. I figured out that the 6-step ripple gait runs more smoothly on my bot than the 12-step ripple gait. I have no conceptual thought on this :slight_smile:

So, now how to turn the bot in either direction? I tried rotating about Y-axis but it just turn the body keeping the legs fixed at their places.

Thanks in advance :slight_smile:
Umar

51

Hi, pyrrhicpk

Can you send me the CSV sheet for 12 step ripple gait ? I need to do a compare because my bot won’t move like what you had shown in Youtube. Thanks!

52

Hi,

Yeah I have made a lot of progress in using PEP, and by now my gaits have evolved to much better accuracy. Soon i’ll upload a new video displaying much better results I achieved using PEP. It’s really a lot of fun with ease to work and play with PEP.

As zenta said, “calibration” and bot’s “body dimensions” are imperative factors in determining a smooth accurate gait, so double check your PEP setup sheet.

PM me your email address, i’ll send you the ripple6s and ripple12s gaits I created.

Regards,
umar

53

Hello,

Anybody got these?

54

Hi, I think Alan could help you. He have a great collection on a yahoo group. Link to picture : tech.ph.groups.yahoo.com/group/HexapodRobotIK/photos/view/c0ca?b=1&m=f&o=0

And the document : biorobots.cwru.edu/publications/Thesis05_Lewinger_BILL-Ant-p.pdf

55

Thanks a lot zenta for those links :slight_smile:

I really appreciate your cooperation and assisstance.
Best regards,
umar

56

Hi, Umar,

Thanks for the CSV file.

Here I have 2 videos, need to ask you for some advise.
The 1st one is regarding the Visual Sequencer. When I import the CSV file inside I only see 6 motor blocks and I have to manually open and enable it. It suppose to have 18 servo, right ?

youtube.com/watch?v=2sM3Vc8fevc

The 2nd one is after I run your CSV file. Seems like there is 3 legs got problems. Sorry for disturbing you, I need your help and advise. Thanks!

youtube.com/watch?v=1s8TptRLSSE

Regards,
Sam

57

Hi Sam,

Looking at your video tells me that the problem is in the direction. The calculations seems to be right because they move at the right time, but in the wrong direction. I think you might need to add some “-“ signs on the angle values of the servos which are moving in the wrong direction.

But I have to say that I’ve never used the Sequencer to drive the SSC so it could be something else… :wink:

Xan

58

Hi Sam,

You have to create a project in the SEQ Sequencer window like this:
http://i214.photobucket.com/albums/cc6/ZentaOlbaid/SEQSequencer.jpg

EDIT: After creating this project you have to export it for creating a .shp file. (As I explained in the PEP manual). Then import your PEP .csv file with the same filename as the .shp file.

I think the three legs on the one side need to be reversed. Remember to calibrate your hexapod in the SEQ SSC32 configuration. Here are the pics of all mine 4 banks:

Bank 1:
http://i214.photobucket.com/albums/cc6/ZentaOlbaid/SEQSSC32configBank1.jpg

Bank2:
http://i214.photobucket.com/albums/cc6/ZentaOlbaid/SEQSSC32configBank2.jpg

Bank3:
http://i214.photobucket.com/albums/cc6/ZentaOlbaid/SEQSSC32configBank3.jpg

Bank4:
http://i214.photobucket.com/albums/cc6/ZentaOlbaid/SEQSSC32configBank4.jpg

Since no servos are 100% equal your values won’t be exactly as mine, but the Min,Max deg values should be calibrate as mine. Depending on how you placed batteries etc the coxa Min. Max deg values might be different. Use a protractor to measure the angle deg.

Hope this was helpful!

59

Hi Sam,

The problem is exactly due to the direction of those 3 servos. They have to be moved in reverse direction. So reverse them in SSC-32 Setup in the Sequencer. Zenta has given you a wonderfula visual presentation on which servos are to be reversed. So I hope it might solve your problem.

As zenta said, first create a generic project in SEQ. You can have the following arrangement of servos as I have.

img219.imageshack.us/img219/6214/vseqfk5.jpg

After arranging the servos, read this

img201.imageshack.us/img201/5861/pepqb5.jpg

Once you get a SHP file, place it at the same location as that of your CSV file with the “SAME NAME”.

Sam, I would like you to take careful note of what zenta explained here, since the calibration is the imperative factor is determining an accurate gait. The more accurate your servos are calibrated, the more precise the legs of your bot reach their destination points.

If you ask me, then I didnt rely on just calibrating the All1500 position but “All-Legs on Ground” position as well. I just wanted to make sure that when my bot sits on ground, the Tibia’s tips hit the ground at the same coordinate locations as determined by PEP. So, in a way you can double check any position within a gait against PEP results. You may need to setup a white paper sheet, project the servo’s axis lines and mark the coordinate points obtained from PEP.Then run the sequence and see whether your bot hits those points accurately or not, and if not then what is the percentage of error. I hope that may aid to your calibration thoughts. Thats the way I did :slight_smile:

You are welcome,
Umar

60

Thanks to Zenta, Xan and Umar,

I’ve solve the sequencer problem.

But for the movement, here is the video after I reversed the angle direct from the CSV file. Please advice.

youtube.com/watch?v=S8b4kCRuzuM

Umar,

from what Zenta had teach and explain on the previous for the steps to generate the 12 step gait. I’ve try to generate my self but I cannot get the result. At least the one you send me can see the 3 legs move in correct direction but the one I generate is totally won’t function, all the 6 legs gone crazy.

For the ssc32 config

Bank 1
http://img401.imageshack.us/img401/1122/bank1tp7.jpg
By visible at 2008-09-26

Bank 2
img525.imageshack.us/img525/1614/bank2we4.jpg
By Visible

Bank 3
http://img242.imageshack.us/img242/6392/bank3sx9.jpg
By visible at 2008-09-26

Bank 4
http://img242.imageshack.us/img242/7430/bank4fq6.jpg
By visible at 2008-09-26

This is the result I get

http://img401.imageshack.us/img401/7516/27092008zd2.jpg
By visible, shot with E65 at 2008-09-26

http://img88.imageshack.us/img88/4433/27092008001bg5.jpg
By visible, shot with E65 at 2008-09-26

I have 2 set of Phoenix here, one is I’ve make myself, that I post earlier “DarkSkitter” and this one is which I bought from Lynxmotion distributor Roboshop. The reason is because firstly I’m not understand how to make it move like what I had seen so decide to follow the Lynxmotion then if I can make it and continue with the one I make.
Thanks!

Regards,
Sam