Pi Robot 2 - Take 2
Patrick Goebel
Now that I have played with the (Maxwell inspired) Pi Robot 2 for a month, I've made some modifications to the design to address some of the shortcomings I have discovered:
* center of gravity is too far forward--too easy to tip the robot over forward
* 12" round plates make for a tight fit for all the components and wiring terminals
* no accommodation for the linear actuator
* battery not easily accessible to swap out if using a second backup battery
I've attached a picture of the new design (without the arm). The new plates are 14" long by 13" wide with rounded edges. The wheels are further forward and more directly beneath the vertical column. The battery sits toward the back to offset the front-end weight and to provide easier access to swap out with a backup battery.
I also have a few questions for Fergs:
* Is your T-slot attached at the bottom of your base as well as on the top plate? How stable is the column in terms of wobbling when the robot is moving?
* You built Maxwell's base as a closed box--how did you join the sides, top and bottom?
* Maxwell's main box sits close to the ground but I think you said you have a fairly large rear caster attached to an upward sloping back end. How did you make that attachment? I'd like to be able to use the Parallax caster wheel but I don't have the 3" clearance even with the motors mounted on the bottom. So I am wondering how did your mounting.
* If you look at the second attached image, how much clearance do I need indicated by the red arrows to allow the bracket holding the arm to slide through the gap in the plate?
I can't help think that I have made this way too complicated and every time I look at Maxwell, it seems infinitely simpler. I don't think Maxwell's interior has enough room to hold everything I need, including battery, large voltage regular, power terminals, USB hub, Serializer and ArbotiX. And it *is* nice having the laptop mounted higher up as it is easy to flip it open and do some quick coding while kneeling or sitting on the floor.
In any event, any suggestions are most welcome!
--patrick
Michael Ferguson
I've attached a picture of the new design (without the arm). The new plates are 14" long by 13" wide with rounded edges. The wheels are further forward and more directly beneath the vertical column. The battery sits toward the back to offset the front-end weight and to provide easier access to swap out with a backup battery.
I also have a few questions for Fergs:
* Is your T-slot attached at the bottom of your base as well as on the top plate? How stable is the column in terms of wobbling when the robot is moving?
The T-slot goes through the top deck into a receptacle on the bottom deck. It is attached on both sides on the top deck. Since the base is effectively a torsion box, there isn't much flex.
* You built Maxwell's base as a closed box--how did you join the sides, top and bottom?
I use 4-40 corner brackets and a tab&slot construction. The picture in http://www.showusyoursensors.com/2011/03/look-inside-maxwell.html shows one of these brackets just below the left-side drive motor.
* Maxwell's main box sits close to the ground but I think you said you have a fairly large rear caster attached to an upward sloping back end. How did you make that attachment? I'd like to be able to use the Parallax caster wheel but I don't have the 3" clearance even with the motors mounted on the bottom. So I am wondering how did your mounting.
The 'box' is only about 8" deep, the sides extend beyond it and the top deck is unobstructed below. See the second picture here: http://www.showusyoursensors.com/2011/02/introducing-maxwell.html
The caster is a regular old 3" swivel caster, works just fine, has some 1/4"-20 standoffs above it to mate with the upper deck.
* If you look at the second attached image, how much clearance do I need indicated by the red arrows to allow the bracket holding the arm to slide through the gap in the plate?
This entirely depends on how wide the arm mount bracket it is -- which also depends on the servos chosen. On Maxwell, this assembly is about 2.5" wide.
I can't help think that I have made this way too complicated and every time I look at Maxwell, it seems infinitely simpler. I don't think Maxwell's interior has enough room to hold everything I need, including battery, large voltage regular, power terminals, USB hub, Serializer and ArbotiX. And it *is* nice having the laptop mounted higher up as it is easy to flip it open and do some quick coding while kneeling or sitting on the floor.
There probably is enough room, see the image linked above. Of course, Maxwell is also 16x16", which is a bit larger.
In any event, any suggestions are most welcome!
--patrick
--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To post to this group, send email to hbrob...@googlegroups.com.
To unsubscribe from this group, send email to hbrobotics+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/hbrobotics?hl=en.
Michael
Here are a couple of other options for the box joinery:
http://www.servocity.com/html/attachment_blocks.html
http://www.kitbots.com/searchquick-submit.sc?keywords=nutstrip
Regards,
Michael Sprague
Wayne C. Gramlich
The major problem with the multi level pancake design is that it
is hard to access the lower levels. In order to get at a lower
level, you need to take off all the levels on top. The cables
are rarely long enough unless you have connectors between each
level. When you add the connectors, very frequently, they add
a potential failure location. Please think long and hard about
how easy/hard it is to futz on the robot.
Seriously consider putting a frame on the bottom of your robot.
As the payload starts to increase, the plastic will start to get
floppy (technical term.) Only the bottom level needs to be framed.
T-slot aluminum extrusions are fairly cheap and easy to work with.
I do not understand the geometry of the linear actuator, but I
would attach it to the frame somehow. Think about putting the
lead acid battery on its own tray so you can easily remove it.
My robot has a handle on the battery tray.
Horizontal pancakes are not the only geometry to consider. Think
about have the bottom plate be horizontal and the other units mounted
horizontal. The Laptop can be on a top most tray which can be
horizontal again. Again, think outside the box, err, pancake,
err, you know what I mean.
Also, consider putting the lidar scanner on a platform that sticks
out a little. Likewise think about letting the vertical column
stick out in front a little.
Regards,
-Wayne
On 02/21/2012 11:46 AM, Patrick Goebel wrote:
> Hello HBRC'ers,
>
> Now that I have played with the (Maxwell inspired) Pi Robot 2 for a
> month, I've made some modifications to the design to address some of the
> shortcomings I have discovered:
>
> * center of gravity is too far forward--too easy to tip the robot over
> forward
>
> * 12" round plates make for a tight fit for all the components and
> wiring terminals
>
> * no accommodation for the linear actuator
>
> * battery not easily accessible to swap out if using a second backup battery
>
> I've attached a picture of the new design (without the arm). The new
> plates are 14" long by 13" wide with rounded edges. The wheels are
> further forward and more directly beneath the vertical column. The
> battery sits toward the back to offset the front-end weight and to
> provide easier access to swap out with a backup battery.
[snip Fergs questions]
-Wayne
Wayne C. Gramlich
> Patrick:
[snippage]
> Horizontal pancakes are not the only geometry to consider. Think
> about have the bottom plate be horizontal and the other units mounted
> horizontal. The Laptop can be on a top most tray which can be
> horizontal again. Again, think outside the box, err, pancake,
> err, you know what I mean.
Think about having three vertical mounting surfaces (left, right, and
back) mounted to the base using simple hinges. To access the guts of
the robot, just swing the sides down and it opens up like a flower
(or a Mars rover base.) When you are done futzing, close it up and
let it go.
Regards,
-Wayne
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On Behalf Of Patrick Goebel
Sent: Tuesday, February 21, 2012 11:47 AM
To: hbrob...@googlegroups.com
Subject: [HBRobotics] Pi Robot 2 - Take 2
Patrick Goebel
Now that I understand more about Maxwell's details, I'm starting to think that building another Maxwell--perhaps with a slightly smaller base--is probably the best way for me to go. Wayne is right that the stacked plates make it a pain to get at anything to fiddle with. I'm wondering if Maxwell's top plate could be modified to have a hinged cut across the width near the base of the T-slot so that it could be flipped up to give access to most of the components inside?
I really like the idea of a sliding battery tray and I was Googling around for such a beast a few weeks ago but couldn't find one--are they commercially made or is a DIY item? Of course, if the flip up top described above could work, a sliding tray would not be needed.
As for putting the lidar and vertical column further forward: the main reason I have the lidar tucked away is to protect it from being kicked or bumped if the ROS obstacle avoidance fails. And I think the T-slot needs to be as much over the wheels as possible to help the balance issue. Mike's lidar placement (seen in this image) seems pretty ideal--it's a little exposed but the front of the base protects it from collisions.
I also noticed that Mike has the ArbotiX mounted vertically on the inside of the front panel so one could save horizontal space for additional microcontrollers (such as my Serializer) by mounting it the same way. Ditto for the USB hub.
--patrick
Patrick Goebel
I'm guessing you saw the links to the CAD drawings on this page (halfway down on the right under "Downloads and Resources")? I don't have the ability to open any of those file types but perhaps you do.
Yeah, drive wheels up front with a single tail dragger seems to be a good option. The two reasons I like to tuck the drive wheels in a bit is (a) hopefully part of the base will bump something before a wheel does so the wheel won't get stuck and (b) if I want to mount a front bumper at some point, it would be easier for the bumper to shield the wheels if they don't stick out from the base.
--patrick
Wayne C. Gramlich
> Thanks Mike, Wayne and Michael,
>
> Now that I understand more about Maxwell's details, I'm starting to
> think that building another Maxwell--perhaps with a slightly smaller
> base--is probably the best way for me to go. Wayne is right that the
> stacked plates make it a pain to get at anything to fiddle with. I'm
> wondering if Maxwell's top plate could be modified to have a hinged cut
> across the width near the base of the T-slot so that it could be flipped
> up to give access to most of the components inside?
Hinges are your friend. Apply them liberally.
> I really like the idea of a sliding battery tray and I was Googling
> around for such a beast a few weeks ago but couldn't find one--are they
> commercially made or is a DIY item? Of course, if the flip up top
> described above could work, a sliding tray would not be needed.
I'm not aware of any off the shelf tray, other than for car batteries
which are too big for this application. The goal is to be able to swap
batteries. Charge one while the other is in use. My battery is held
in by 4 screws. Pop the 4 screws out and the entire battery and power
switch assembly lifts out. As you know, I have all screws mounted on
1" x 1" centers that is compatible with the peg board that can be
purchased at Home Depot/Lowes.
> As for putting the lidar and vertical column further forward: the main
> reason I have the lidar tucked away is to protect it from being kicked
> or bumped if the ROS obstacle avoidance fails. And I think the T-slot
> needs to be as much over the wheels as possible to help the balance
> issue. Mike's lidar placement (seen in this image
> <http://2.bp.blogspot.com/-JZUcuEQ7DHI/TYgZPuUt31I/AAAAAAAAAOM/SoQngH30-0E/s400/maxwell1.jpg>)
> seems pretty ideal--it's a little exposed but the front of the base
> protects it from collisions.
There is no ideal place for the arm. The further back you place it,
the less chance there is to tip it forward, but the harder it is
for the fore arm to reach things. Basic physics gives the torque
of the battery vs. the torque of the robot arm. When the mass in
the robot arm causes the robot arm torque to exceed the battery
torque, your robot tips over. Gravity always wins. ;-)
> I also noticed that Mike has the ArbotiX mounted vertically on the
> inside of the front panel so one could save horizontal space for
> additional microcontrollers (such as my Serializer) by mounting it the
> same way. Ditto for the USB hub.
Keep mashing the ideas around in your head. The next revision should
be quite interesting.
Regards,
-Wayne
Patrick Goebel
> My battery is held in by 4 screws. Pop the 4 screws out and the
> entire battery and power
> switch assembly lifts out. As you know, I have all screws mounted on
> 1" x 1" centers that is compatible with the peg board that can be
> purchased at Home Depot/Lowes.
Hi Wayne,
Are you using a block-style lead acid battery? If so, how are you
screwing it down? Or do I recall that you are using the elongated
battery packs?
--patrick
Wayne C. Gramlich
I using a gel cell lead acid battery. Using an aluminum plate,
a couple of long screws, and a short piece of aluminum plate,
I built a bracket that held the short aluminum plate across
the battery. A couple of wing nuts completed the process.
I'm currently using two batteries -- a lead acid for the motors
and an 7.2V RC battery pack for the logic. On the next iteration,
I'll probably replace the RC battery pack with a high quality
DC/DC switching converter. The ground leads of the logic and
motor buses are only connected in one location close to the
batteries. There is no electrical connections between the
two buses anywhere else, thus, there can be no group loops
(no loops => no ground loops.)
Regards,
-Wayne
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
To: hbrob...@googlegroups.com
Subject: Re: [HBRobotics] Pi Robot 2 - Take 2
Patrick Goebel
> Patrick:
>
> I using a gel cell lead acid battery. Using an aluminum plate,
> a couple of long screws, and a short piece of aluminum plate,
> I built a bracket that held the short aluminum plate across
> the battery. A couple of wing nuts completed the process.
>
Ah, I see. Thanks for the tip!
--patrick
James M. Geidl
Patrick Goebel
--patrick
Alan
Get some PVC or Delrin ½” sheet, should be easy enough to do!
Clamp the castor down to the sheet, mark the contour and holes.
Cut it out, drill it.
Get long bolts of appropriate size, and you’re good to go!
Bring your castor, (drawings?) Maybe I can help.
Tonight is SIG, I can’t come, but I’ll bet someone there with a membership could just about knock it out for you!
Alan KM6VV
On Behalf Of Patrick Goebel
Yeah, it's really close. If you can explain how you might make a 1/2" shim that the Pololu brackets could mount to, I'd love to hear it--I'm somewhat mechanically challenged when it comes to that kind of thing.
--patrick
Patrick Goebel
--patrick
James M. Geidl
KM6VV
Are you sure you want a 1/2" thick spacer? Wood would be fine, I like metal work.
They don't specify the radius of the corners, or the hole sizes, but I see #6 mounting hardware, so we know what hole size to make (6-32 clearance).
1/32" off in height won't matter anyway. Or just use 1/32" sheet metal as a "shim" (1/2" is a little thick for a shim...). ;>)
Let me know if you wan one!
Alan KM6VV
On 2/22/2012 8:13 PM, James M. Geidl wrote:
Patrick,
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On Behalf Of KM6VV
Sent: Wednesday, February 22, 2012 8:34 PM
Patrick Goebel
(Note: I changed the subject line to better reflect our current discussion.)
This sounds like a great plan. I'll wait to see how James does with his caster and test shims then follow your guys' lead.
BTW, for anyone else interested in the Parallax caster wheel, it is $39.99 from Parallax itself but as much as $79.99 elsewhere.
--patrick
Alan
James, Patrick,
Sounds good. I have enough info to make a spacer, just need thickness wanted. Next Wednesday is the club meeting, probably wouldn’t be too hard to make up a part or two before meeting if either/both of you guys want them. Or it can wait. I’m working on my latest version of jaws, now I have a “lift” servo! I changed to a pair of mini servos, HS-5245MG. Same jaw system will work on my bigger “MicroMoose” ‘bots.
Alan KM6VV
On Behalf Of James M. Geidl
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
From: hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com] On Behalf Of Alan
Sent: Thursday, February 23, 2012 10:46 AM
To: hbrob...@googlegroups.com
Subject: RE: [HBRobotics] Pi Robot 2 - Take 2
Alan
Oh I get it. I had it backwards. The motors need the spacers. Still, relatively simple matter to cut a spacer.
Let me know if I can help.
The triangle shape of the caster base would have been more fun!
Alan KM6VV
On Behalf Of James M. Geidl
Alan,
It isn't a spacer for the caster that will be needed. The caster is 4" tall and the highest configuration for the motors is about 3 9/16" All that will be needed are a couple of rectangular spacers around 1/2" thick to fit under the motor bracket or make new motor brackets that put the motor further away from the base. But, I still want to get the caster in my hands before I do anything about spacers or shims.
James M. Geidl
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
Sent: Thursday, February 23, 2012 9:47 AM
To: hbrob...@googlegroups.com
Subject: [HBRobotics] Parallax caster and shims for Pololu motor mounts
Patrick Goebel
--patrick
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
To: hbrob...@googlegroups.com
Subject: Re: [HBRobotics] Parallax caster and shims for Pololu motor mounts
Alan
Yup! You’ve got it!
Alan KM6VV
On Behalf Of Patrick Goebel
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
To: hbrob...@googlegroups.com
Subject: Re: [HBRobotics] Parallax caster and shims for Pololu motor mounts
Patrick Goebel
--patrick
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
To: hbrob...@googlegroups.com
Subject: Re: [HBRobotics] Parallax caster and shims for Pololu motor mounts
Patrick Goebel
--patrick
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
Patrick Goebel
--patrick
James M. Geidl
James M. Geidl, K6JMG
D.B. Cooper, you have a
message.
Wayne C. Gramlich
tonight? Such artwork should shared. ;-)
-Wayne
On 02/29/2012 03:09 PM, James M. Geidl wrote:
> That's a perfect way to describe it. It's almost too pretty to use.
>
> James M. Geidl, K6JMG
> /D.B. Cooper, you have a message./
>
>
> ------------------------------------------------------------------------
> *From:* hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com]
> *On Behalf Of *Patrick Goebel
> *Sent:* Wednesday, February 29, 2012 2:01 PM
> *To:* hbrob...@googlegroups.com
> *Subject:* Re: [HBRobotics] Parallax caster and shims for Pololu motor
Patrick Goebel
Alas, I can't make tonight's meeting otherwise I'd would have been happy
to. Sorry!
--patrick
Arthur Wait
Wayne C. Gramlich
-Wayne
jge...@sbcglobal.net
Jim, K6JMG
Sent from my HTC Inspire™ 4G on AT&T
----- Reply message -----
From: "Wayne C. Gramlich" <wayne.gra...@gmail.com>
To: <hbrob...@googlegroups.com>
Cc: "Wayne C. Gramlich" <wayne.gra...@gmail.com>
Subject: [HBRobotics] Parallax caster and shims for Pololu motor mounts
Date: Wed, Feb 29, 2012 3:25 pm
Is there any chance one of you could bring it to the meeting
tonight? Such artwork should shared. ;-)
On 02/29/2012 03:09 PM, James M. Geidl wrote:
> That's a perfect way to describe it. It's almost too pretty to use.
>
> James M. Geidl, K6JMG
> /D.B. Cooper, you have a message./
>
>
> ------------------------------------------------------------------------
> *From:* hbrob...@googlegroups.com [mailto:hbrob...@googlegroups.com]
> *On Behalf Of *Patrick Goebel
> *Sent:* Wednesday, February 29, 2012 2:01 PM
> *To:* hbrob...@googlegroups.com
> *Subject:* Re: [HBRobotics] Parallax caster and shims for Pololu motor
> mounts
>
> That was fast--just got it in the mail (Parallax must be down the street
> from me.) This thing is a work of art. Never thought I would get that
> excited by a caster wheel. More importantly, I think the pneumatic tires
> will help get over certain edges without catching--my intuition is that
> the tires will dent around the edge, grip and keep rolling. The hard
> plastic one I have now tends to catch and stop rolling.
>
> --patrick