Requesting Recommendations for 6-axis Desktop Robot Arm

138 views
Skip to first unread message

Joseph Heady

unread,
Jul 9, 2021, 5:09:03 PM7/9/21
to hbrob...@googlegroups.com
Hello everyone


I am working on a project that requires that I move an object linearly along x/y/z, axially about x/y/z, in a figure eight motion, and similar. I am reaching out to see if anyone has any recommendations for a desktop 6-axis robot (could also possibly get by with a 5 or 4 axis robot!)

The robot should match the following specifications: 
  • Reach of >=500mm
  • Payload capacity of >500g
  • Repeatability <0.1mm
  • Able to program using python
  • Desktop size
  • Price <$5000
  • Qty needed: 10 (one initially for testing)
  • 6 axis ideally, also interested in 5 and 3 axis robots as well
Your input would be greatly appreciated.


Thank you,

Joseph Heady
HBRC | Programs Director

Marco Walther

unread,
Jul 9, 2021, 6:59:30 PM7/9/21
to hbrob...@googlegroups.com
On 7/9/21 2:08 PM, Joseph Heady wrote:
> Hello everyone
>
>
> I am working on a project that requires that I move an object linearly
> along x/y/z, axially about x/y/z, in a figure eight motion, and similar.
> I am reaching out to see if anyone has any recommendations for a desktop
> 6-axis robot (could also possibly get by with a 5 or 4 axis robot!)
>
> The robot should match the following specifications:
>
> * Reach of >=500mm
> * Payload capacity of >500g
> * Repeatability <0.1mm
> * Able to program using python
> * Desktop size
> * Price <$5000
I don't think, a classic robot arm can fulfill all that at that price
point. Maybe at 10x?!

https://www.youtube.com/watch?v=2cyBPkZRIUc This can hold < 15g, reach
is something like 400mm, repeatability >= 25mm, price around $600

Any gripper out there will act like a hammer with a long handle:-(
Either you move very slowly or you have to deal with interesting forces
all over the place:-(

Maybe a gantry setup can work for you?! Not as sexy but requirements are
much more 'linear';-)

Just my ideas;-)
-- Marco

> * Qty needed: 10 (one initially for testing)
> * 6 axis ideally, also interested in 5 and 3 axis robots as well
>
> Your input would be greatly appreciated.
>
>
> Thank you,
>
> Joseph Heady
> HBRC <http://www.hbrobotics.org/> | Programs Director
> +1-408-618-9863
>
> --
> You received this message because you are subscribed to the Google
> Groups "HomeBrew Robotics Club" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to hbrobotics+...@googlegroups.com
> <mailto:hbrobotics+...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/hbrobotics/CAHZUr96F7dBwKLVQ1MOAeSkJoJwc8TnpU61Buw_Vk-S67p%2Ba1Q%40mail.gmail.com
> <https://groups.google.com/d/msgid/hbrobotics/CAHZUr96F7dBwKLVQ1MOAeSkJoJwc8TnpU61Buw_Vk-S67p%2Ba1Q%40mail.gmail.com?utm_medium=email&utm_source=footer>.

Joseph Heady

unread,
Jul 9, 2021, 7:10:57 PM7/9/21
to hbrob...@googlegroups.com
I found a 4 axis robot from Dobot MG400 for >$2900... I have ordered one of these for testing but am looking for additional alternatives to compare. This robot meets most of my requirements.

Check out the spec sheet

Any other recommendations out there?


Thank you,

Joseph Heady
HBRC | Programs Director


To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/586f5cb8-f4ea-14a6-1fa0-2ba9896a59e8%40gmail.com.

Alan Marconett

unread,
Jul 9, 2021, 7:15:39 PM7/9/21
to hbrob...@googlegroups.com
Have you looked at the Sherline line of mills?

Alan

Cheuksan Wang

unread,
Jul 9, 2021, 7:28:00 PM7/9/21
to hbrob...@googlegroups.com

Joseph Heady

unread,
Jul 9, 2021, 7:57:06 PM7/9/21
to hbrob...@googlegroups.com
This one was pretty close to what I was looking for. It's only repeatable to 1mm, so it doesn't meet my specs. Pretty cool nonetheless though.

Thank you!

Joseph Heady
HBRC | Programs Director

On Fri, Jul 9, 2021 at 4:15 PM Alan Marconett <km...@sbcglobal.net> wrote:

Joseph Heady

unread,
Jul 9, 2021, 8:00:24 PM7/9/21
to hbrob...@googlegroups.com
Oops, sorry about that. I responded to the wrong email. These mills might be able to work. But I am looking for a cheap 6/5/4 DOF robot arm that meets the above specs.. or something with specs close to those I mentioned before.

Thank you,

Joseph Heady
HBRC | Programs Director

Chris Albertson

unread,
Jul 9, 2021, 8:12:20 PM7/9/21
to hbrob...@googlegroups.com
Given that you own a shop full of machine tools, you could build a robot arm for your budgeted price.   But you are not going to be able to buy one as you describe for $5K

The budget does not even allow $1K per axis, and you will need some powerful motors and gear (or belt) reductions optical encoders, and quality bearings.   That 1K has to cover not only the bearings but the drive electronic and power supply.

Python is a good language for this although but if the motion is fixed, why not use g-code?   Use python if the robot needs to be connected to (say) a video camera or 3D stereo camera for picking up loose parts from a bin or if the robot is to work in a real-world environment where "anything" can happen.    But for fixed machine-tool-like motion, g-code works.

Here is one option. The "AR3"  https://www.anninrobotics.com/ It's a kit of parts.  The mechanical engineering is good, but the software is VERY primitive.    



--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.


--

Chris Albertson
Redondo Beach, California

Sergei Grichine

unread,
Jul 9, 2021, 9:38:32 PM7/9/21
to hbrob...@googlegroups.com

Chris Albertson

unread,
Jul 9, 2021, 9:39:26 PM7/9/21
to hbrob...@googlegroups.com


On Fri, Jul 9, 2021 at 5:12 PM Chris Albertson <alberts...@gmail.com> wrote:
.... you will need some powerful motors and gear (or belt) reductions optical encoders, and quality bearings. 

The bearings really matter if the goal is 0.1 mm repeatability.  Most of the cheap robot arms use cheap or undersized bearings. Many use the bearing built into the motor as load-bearing joint bearings.  You are going to have to look for a design that uses large oversized precision bearings.  Very tiny amounts of "play" gets multiplied over the length of the arm to a lot more than 0.1 mm   

Many low-cost robot arms in your budget don't bother with expensive-to-make 0.1mm precision because they will use video feedback or even human-in-the-loop teleoperation to guide the arm.     It seems like you are looking for open-loop positioning

 

Chris Albertson

unread,
Jul 9, 2021, 9:52:51 PM7/9/21
to hbrob...@googlegroups.com
Aber B1 is a Kickstarter project now and when it actually is finished the MSRP will be $15K, three times out of budget.

Also, the stated accuracy is 0.5mm, five times out of spec.

But look about what I wrote about bearings.   Amber B1 uses the bearings built into the harmonic drives.  They look good, but you will need some MUCH larger ones to get the rigidity required for 0.1 mm

You might want to re-think the big-picture goal so you can use a lower precision and lower cost robot.  Perhaps some like of sensor could measure the end effector on the arm and provide feedback?  Depends on the application.

Joseph Heady

unread,
Jul 9, 2021, 10:12:03 PM7/9/21
to hbrob...@googlegroups.com
Sergei Grichine: This robot you recommended is super cool. I definitely would like to check it out. I do need something that is immediately available, is well supported, and I need to be able to buy 10 or so in the immediate/near future. The price after the kickstarter is well above my price range. The repeatability is advertised at 0.1mm and on their kickstarter page I believe it is listed as 0.05mm.. which is great. This overall is a great find. Thank you for your help!

Thank you,

Joseph Heady
HBRC | Programs Director

Stephen Huber

unread,
Jul 10, 2021, 1:48:43 AM7/10/21
to hbrob...@googlegroups.com
I’m not sure what your specific application is but another way to reduce the need for precision movements is to use some creative fixturing or limit witches near to the target to re-reference the robot.

One robot I have built is the AR3 by Chris Annin.  Really good instructions and intuitive SW make it pretty easy to assemble and get going.  Their are also kits available for purchase.  The full build will cost you well under $5k but I’m not sure how close it gets to meeting your specs. 


Stephen Huber

Sent from my iPhone


On Jul 9, 2021, at 7:12 PM, Joseph Heady <gothic...@gmail.com> wrote:



Mark Johnston

unread,
Jul 10, 2021, 3:00:06 AM7/10/21
to HomeBrew Robotics Club
The 0.1mm made me step back and say YOWZA!   That is like a really big ask for 6 axis and $5k.   
Chris is spot on in his discussions, I have nothing to add. (You Rock Chris!)

If a kickstarter says 0.1mm they may also offer a really nice bridge painted a golden orange just off of San Francisco, wonderful view I am told.
It's easy to say something like that in a kickstarter cause you are going for funding, then after they make 20 of them they find that there is no way they retain that initial accuracy or only 3 out of 10 can get near that 0.1mm or it costs them 5x to actually produce them than they estimated.    

Anyway, this thread has a lot of nice things being discussed just the same so is a great thread here.   Lovin it.
Mark



Bob Smith

unread,
Jul 10, 2021, 1:51:41 PM7/10/21
to hbrob...@googlegroups.com

>>>>> On Fri, Jul 9, 2021 at 2:09 PM Joseph Heady <gothic...@gmail.com>
>>>>>> I am working on a project that requires that I move an object linearly
>>>>>> along x/y/z, axially about x/y/z, in a figure eight motion, and similar. I
>>>>>> am reaching out to see if anyone has any recommendations for a desktop
>>>>>> 6-axis robot (could also possibly get by with a 5 or 4 axis robot!)
>>>>>>
>>>>>> The robot should match the following specifications:
>>>>>>
>>>>>> - Reach of >=500mm
>>>>>> - Payload capacity of >500g
>>>>>> - Repeatability <0.1mm
>>>>>> - Able to program using python
>>>>>> - Desktop size
>>>>>> - Price <$5000
>>>>>> - Qty needed: 10 (one initially for testing)
>>>>>> - 6 axis ideally, also interested in 5 and 3 axis robots as well

At $2K the Liteplacer has four axis and meets most of your
specs. I do not know if it has a Python API or not.
https://liteplacer.com/

Chris Albertson

unread,
Jul 10, 2021, 2:29:54 PM7/10/21
to hbrob...@googlegroups.com
On Sat, Jul 10, 2021 at 12:00 AM Mark Johnston <mjst...@gmail.com> wrote:
The 0.1mm made me step back and say YOWZA!   That is like a really big ask for 6 axis and $5k.   
Chris is spot on in his discussions, I have nothing to add. (You Rock Chris!)

If a kickstarter says 0.1mm they may also offer a really nice bridge painted a golden orange just off of San Francisco, wonderful view I am told.
It's easy to say something like that in a kickstarter cause you are going for funding, then after they make 20 of them they find that there is no way they retain that initial accuracy or only 3 out of 10 can get near that 0.1mm or it costs them 5x to actually produce them than they estimated.    

Anyway, this thread has a lot of nice things being discussed just the same so is a great thread here.   Lovin it.
Mark



On Friday, July 9, 2021 at 10:48:43 PM UTC-7 ste...@uppump.com wrote:
I’m not sure what your specific application is but another way to reduce the need for precision movements is to use some creative fixturing or limit witches near to the target to re-reference the robot.

One robot I have built is the AR3 by Chris Annin.  Really good instructions and intuitive SW make it pretty easy to assemble and get going.  Their are also kits available for purchase.  The full build will cost you well under $5k but I’m not sure how close it gets to meeting your specs. 

So you actualy have an AR3?   I don't butwhat impressed me is the video here the robot touches a dial indicator, does a movethn comes back tothe dail indicator and we see the position repeatabilty is VERY good.

Looking at how he did it we see he is using tapered roller bearings that were designs for truck axels and in the AR3 used at something like 1% of their design limit.   Basically 100X overkill,    Being tapered we can adjust the preload and I assume they are screwed down tight.

So I ask, Are they?  is that how Chris Annin got the "slop" out of the joints?  I think tapered bearings was the key innovation here.

Using this assumption I designed an arm and shoulder system using modular parts that could be used to make a robot for way-cheap.   It could be printed or machined from aluminum.   It's cheap because it reused the same parts in each joint.  It is untested, but borrows from AR2/AR4
Here is the whole assembly https://a360.co/3wr5ozu
Here is just one block (that is used 6 times) https://a360.co/2KuCmLF

My point is that you can build sorting that uses the ideas from a commercial product for 25% of the cost. and if you own a CNC milling machine or printer, you can make one yourself.   

The green flange bolts directly to the next red frame.  Do it 6 times and you have 6DOF.  The bearing are standard parts and cost something like $10 each and the "slop" is controlled by torque on the screws.  Motoers are the same as AR2/3 and come from Stepper Online.

This was part of "Jake Bot" a robot designed to do "anything a person in a wheelchair could do". Which is quite a lot if the environment were designed to ADA requirements.


Sergey Pankov

unread,
Jul 10, 2021, 2:36:59 PM7/10/21
to hbrob...@googlegroups.com
Check ufactory xarms, this one https://www.ufactory.cc/products/xarm-5-lite-2020?variant=31539074891894 is 5DoF, but otherwise meets or exceeds all your requirements.  Note 3kg payload and 1m/s linear speed. Gripper is sold separately,  or you use your own. They also have xarm 6, with 6DoF and 5kg payload but for $8k or so.

--

Joseph Heady

unread,
Jul 10, 2021, 3:30:40 PM7/10/21
to hbrob...@googlegroups.com
Nice find Sergey!

It meets my specs. Super cool. I'll see if I can order one and try it out.

Thank you,
Joseph

Sergey Pankov

unread,
Jul 10, 2021, 5:19:50 PM7/10/21
to hbrob...@googlegroups.com
Glad if it helps. The numbers are probably the best case scenario, like max speed for no payload, and max payload includes the end effector weight. Still impressive specs for the price. They say they use harmonic drives, so perhaps repeatability and durability are also good.

Sergey

Reply all
Reply to author
Forward
0 new messages