Shop OBEX P1 Docs P2 Docs Learn Events
QuadWalker walking backwards... — Parallax Forums

QuadWalker walking backwards...

jcpolejcpole Posts: 92
edited 2007-01-01 02:38 in Robotics
Hi...

I just finished building a QuadWalker kit, and the thing seems to be completely backwards.· The gaits for forward move backwards, and the right and left directions also seem to be reversed.

I have quadruple-checked the circuitry, and it is correct.

Any ideas?

Thanks...

Jamie


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--
Jamie C. Pole
Principal Consultant
J.C. Pole & Associates, Inc.
http://www.jcpa.com/

Comments

  • PARPAR Posts: 285
    edited 2006-12-22 10:00
    jcpole said...
    I just finished building a QuadWalker kit, and the thing seems to be completely backwards.· The gaits for forward move backwards, and the right and left directions also seem to be reversed.

    I have quadruple-checked the circuitry, and it is correct.

    Any ideas?
    Maybe posting photos would help, ·of the built robot's legs (servos as installed), and labelling each servo's wire bundle and posting a photo of the labelled, plugged-in wiring (to show which servo is plugged into which PSC socket).

    And, then post the program listing you actually downloaded and ran, and describe the movement actions taken by the robot under that program's control.

    Someone with the same Crawler might be able to help more easily then.

    PAR
  • dandreaedandreae Posts: 1,375
    edited 2006-12-22 19:30
    Can you list the program that you are using as a .bs2 file?

    Thanks,

    Dave

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Dave Andreae

    Parallax Tech Support·
  • jcpolejcpole Posts: 92
    edited 2006-12-23 06:47
    Here is the program I'm using.· It was downloaded directly from CrustCrawler.

    Thanks...

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
    bs2
    bs2
    QuadBase.bs2
    25K
  • PARPAR Posts: 285
    edited 2006-12-24 18:54
    jcpole said...
    Here is the program I'm using.· It was downloaded directly from CrustCrawler.
    I looked at the quadcrawler list (main and archive, http://www.crustcrawler.com/code/index.php?prod=20) of code, but didn't find "QuadBase.bs2".

    Exactly which program did you download? (URL, filename)

    BTW, are the values in your copy of the program (the one you attached on 12/22) the same as the ones in the source (i.e., the Crustcrawler site) of the download, or did you have to change any of them when calibrating your crawler?

    PAR
  • jcpolejcpole Posts: 92
    edited 2006-12-24 19:33
    The program came directly from the CrustCrawler site - http://www.crustcrawler.com/products/quadcrawler.php?prod=2

    It's the first program on the list, under "More Project Code".

    As far as changing values, I didn't have to change anything - miraculously, all of the servos centered without modification. That's definitely a first for me. smile.gif

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
  • AImanAIman Posts: 531
    edited 2006-12-28 17:47
    My Spider III started going backwards. Turned out that my motor was connected backwards.

    If you haven't checked that I would. Simply but easily missed.
  • jcpolejcpole Posts: 92
    edited 2006-12-28 17:59
    Thanks - I did check that, and it looked correct. The leg motion and gait are correct - just backwards. At this point, I'm going to tear down all of the servo wiring and start from scratch.

    Honestly, I'm a little discouraged about the lack of support Parallax has provided. I know they are closed until 01/02/07, but my original post was from 12/21, and so far I've only had a request for more information from Parallax.

    :-(

    I'm going to try contacting CrustCrawler directly to see if they will provide support.

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
  • WarrlokWarrlok Posts: 77
    edited 2006-12-28 20:11
    u could try some other codes,,,or write ur own or try 1 servo at time ,,,wohhh just had a brain fart,,,, swap the servo,s leads that go to the left side with the ones tht go to the right ...I think thts ur problem.

    hope that helps
  • John R.John R. Posts: 1,376
    edited 2006-12-28 20:20
    I would concurr with Warrlok. If the walking is working, but backwards, it is most likely that the servo connections are somehow swapped, probalby right side for left side.

    As far as Parallax support, realize that while backwards, you are functioning, so it isn't a "life or death" support issue, and your posting of information after their request was getting close to the time they were getting ready for the holiday. I would also add that the request for additional information was an appropriate request on their part.

    If you really felt you needed "Parallax" support, you could have (should have?) also directly contacted their support team, via phone or e-mail. This site is not the "official" support site, and they are not "obligated" to solve your issue here.

    I wouldn't begrudge them on this issue.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    John R.

    8 + 8 = 10
  • jcpolejcpole Posts: 92
    edited 2006-12-28 20:38
    Wow - a request for help turns into an indictment. For your reference, here is the procedure Parallax posts for obtaining support:

    "To obtain technical support on Parallax products use the following resources in the order shown below. We are pleased to answer questions by e-mail and phone, but encourage you to first use available on-line information.

    Discussion Forums
    Knowledge Base
    Protect your Stamp
    Parallax Downloads
    E-mail Support
    Phone Support
    Authorized Consultants"

    The operative part of that statement is "in the order shown below". You can read this for yourself at http://www.parallax.com/html_pages/tech/index.asp

    I followed the published procedure, and have not received support.

    Please do not talk to me like I am a child.

    As far as something being "life or death", unless this QuadCrawler somehow became critical to my well-being, I don't recall ever having suggested that I was going to die if it didn't get fixed. I'm also not saying that a request for further information was inappropriate - I was merely stating that their entire support effort to date consisted of asking for more information.

    A posting like the one you just made adds no value to the conversation, and only serves to raise the frustration level associated with the problem. I did not ask for help using Parallax support, and I most certainly did not ask for your critique as to the procedure I used to request support.

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
  • PARPAR Posts: 285
    edited 2006-12-28 21:02
    jcpole said...
    I followed the published procedure, and have not received support.
    Your original posting was Dec 21 at 6 in the evening. Parallax staff replied on Dec 22, at 11:30 in the morning·--the first day that Parallax was closed for the holidays.

    Not only is that prompt service from Parallax, but it was DURING their holiday period.

    Earlier that same day (2am) I suggested to you:

    "Maybe posting photos would help, ·of the built robot's legs (servos as installed), and labelling each servo's wire bundle and posting a photo of the labelled, plugged-in wiring (to show which servo is plugged into which PSC socket)."

    While this may seem to be a bit of overkill (on my part, to request such information), you stated that your circuitry was correct, and at the time you had not provided your source code either.

    For those of us who do not have a Crawler, we can not run your source code, and so can not EASILY confirm that it is correct (we could tediously read the entire code, envision the mechanical linkages of the Crawler, envision the possible wiring (both correct and multitudes of combinations of incorrect wirings, but frankly that's expecting a lot from volunteer forum participants).

    It seems (seemed to me anyway) that a few photos of the mechanical and wiring linkages posted to the forum might be EASIER for some volunteer to inspect and comment on. Photos of hard-to-see connections are not ideal, but it was a shot in the dark. Since you did not post any photos in response to the request, we're left with only your source code and your word that your Crawler is correctly constructed and wired. (BTW, comparing your source code to the original source code compares equal, so we might conclude that either the original code didn't work, or that your wiring or mechanical linkages are wrong; or, perhaps it's a question of which end of your Crawler is the "front"?)

    Parallax no doubt has a Crawler handy and could run your posted code on it --once they return from holiday vacation.

    PAR
    ·
  • jcpolejcpole Posts: 92
    edited 2006-12-28 21:33
    Okay, PAR - your point is well-taken.

    For my part, I had it in my head that Parallax closed for the season on the 12/25, not 12/22. The web site clearly says that the period begins on 12/22, so that is my mistake. I don't know where I got a date of 12/25 from - that was Christmas Day.

    I have never disparaged Parallax support - I've had great conversations with Jon, Chris, and Beau in the past. All I said was that I was disappointed. Understanding now that they were closed when I received my first reponse, that clearly resets my "expectation" of when I should have received a response. Thanks for pointing this mistake out

    As far as photographing the wiring, this is not practical because the wiring crosses from the top of the upper deck to the bottom of the upper deck. There would be no way for you to discern which servos were connected to which wires. I'm rewiring all of the servos, and I will post again if that does not solve the problem.

    I do appreciate your response, and I should have replied to it directly.

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
  • John R.John R. Posts: 1,376
    edited 2006-12-28 21:48
    Jamie:

    You may also want to check that the linkages on the "lift" servos are installed correctly. I'm not sure of the configuration on the QuadCrawler or other 2 DOF crawlers, but if that is istalled off by 180 degrees, the lift and lower would be reversed, and this could also cause the crawler to move backwards.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    John R.

    8 + 8 = 10
  • jcpolejcpole Posts: 92
    edited 2006-12-29 14:58
    Yeah - I had that problem with the first leg that I built, but it was pretty easy to fix once I figured out how the "dog bone" connector was supposed to be constructed. The documentation shows a gap in the shaft, but the individual parts are slightly longer than the diagram shows, so allowing a gap in the dog bone connector doesn't work. That dog bone connector is a really fascinating piece of engineering (Although I freely admit that I'm easily impressed by things like that). Once the dog bone connector had the correct length, getting the proper raise/lower motion was just a function of making sure that the servo was centered properly. I found that with a gap in the dog bone connector, each raise/lower cycle actually produced 2 mini-movements (that is a highly technical term) each time, instead of a full raise/lower.

    I'm going to re-wire the servos in the QuadCrawler today, so we'll see if that fixes the problem.

    Thanks...

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
  • ZactlyZactly Posts: 7
    edited 2006-12-29 15:37
    If servos aren't directly driving the legs do the gears rotate correctly?

    If the servos are directly driving the legs then simply reverse your pulse sequence in your code and you probably will have your problem solved.

    Example
    instead of
    counter = 1 to 500 step 5
    do
    Counter = 500 to 1 step 5
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2007-01-01 02:04
    Jamie,
    ·
    ·· I apologize your issue wasn’t handled in a timely manner. ·As you can imagine with Parallax being closed for the holidays we’re a little slower getting to all the posts and e-mails. ·In any event the fact that your robot is walking seems to indicate that there’s no hardware or technical issue. ·Either there is a problem in the code or the servos are connected out of order.
    ·
    ·· We often ask for the code you’re using because sometimes the code is typed in and one wrong character can make all the difference. ·I have never seen this piece of code before myself, but a quick glance at it suggests that since it is designed to create and test walking gaits, and since it contains a backward gait, perhaps somehow that is getting selected. ·I would also triple-check my servo connections.· I hope this helps.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Tech Support
  • jcpolejcpole Posts: 92
    edited 2007-01-01 02:32
    No worries Chris...

    Somehow, I had it in my head that you guys were open until the 25th. No idea where that one came from, but it was definitely my mistake...

    As far as the QuadCrawler, it's still walking backwards, but I substituted the backward gait values, and the thing walks normally.

    I'll figure it out eventually, but it's working fine now. The servos were correct before I tore it down and rebuilt it, so I was a little relieved that it didn't start working correctly. smile.gif On the plus side, I did find that the vertical servos in each leg were not completely centered, so the rebuild actually produced a better gait than the first iteration.

    I do have one observation that you might want to be aware of moving forward - the PSC that came with my QuadCrawler had the jumper installed to select the high channels instead of the low ones. Unless I missed something, the build documentation doesn't mention this jumper. As a result, my first build didn't move at all. I figured it out when I put one of my known-good PSCs into the QuadCrawler and it started working (backwards, but moving). If there are any plans to update the documentation, it might be a good idea to mention that the jumper needs to be removed into order to select the low channels. It might save you guys some support calls.

    I hope you had a great holiday...

    Jamie

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --
    Jamie C. Pole
    Principal Consultant
    J.C. Pole & Associates, Inc.
    http://www.jcpa.com/
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2007-01-01 02:38
    Jamie,
    ·
    ·· Thanks for the note…I will mention this when we re-open. ·Take care.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Tech Support
Sign In or Register to comment.