Hello, I am a novice but I’m considering buying several these robotic arms for my lab use. Does anyone know that if programmed or controlled properly, is it possible for these arms to tighten/loosen screwed caps on bottles (do not have to be super tight as done by a wrench)? Do I have to purchase the heavy duty rotary wrist for that purpose?
Thank you very much!!
This could certainly be achieved.
Regarding the heavy duty wrist you mentionthis would all depend on how tight you want the lids.
Ordering the arm and a few selected brackets you can choose a decent servo for the wrist rotation… This would be a continuous rotation servo.
The gripper would need modifying i think 'to enable it to grasp the caps. Or even the bottle, if you intend to have another one holding it.
May even need to upgrade that servo also… But again this would depend on how tight it will need to grip.
1 Do you need to just tighten/loosen the screw caps?
2 Or do you need to remove and later replace the screw caps?
If 1, you might be able to substitute a simple plastic funnel with a rubbery inner surface for the expensive gripper. The arm could lower the funnel cone onto the cap and then spin it on or off. A limitted amount of slippage could be tolerated, and the cone shape could accomodate different lid sizes.
If 2, and the caps can be magnetized, perhaps a magnet could be placed in the funnel to retain the caps as needed.
If 2, and the caps cannot be magnetized, you might be better served adapting the cheaper small grippers. You can build extensions for them if you need a wider grip for instance.
Some other factors to consider:
Hi Innerbreed, thank you very much for your advice=) I’ve just placed an order with a heavy duty wrist and I’ll see how tight the cap can be when the gripper starts to slip, and you are definitely right there that I’ll do some work to modify the gripper somehow. Thanks again and hopefully I’ll come with some progress and new questions…
Hi Ted, thank you so much for your idea and advice. Basically I would be like 2, which you mentioned to have the cap removed and later replace it back. I haven’t started playing with the arm but I just assume that remove the cap–> put it down --> deal with the uncapped bottle —>pick up the cap again later —> place it back on the bottle is an easier process to achieve than tightening/loosening (where a lot of torque problems need to be addressed), do you think I’m right? My guess is based on the several videos of the arms on the website.
Also with the other ideas you mentioned, I worried less but it’s a good reminder to quantify the number of turns and my lids are very small and light so they should be OK.
Thank you very much again for your help!!
Hi. Gladiator,
I probably have more thoughts and questions than answers for you. I think the key procedure you must consider first and which will lead you to your other answers is “–>deal with the uncapped bottle”. How do you need to deal with it?
Your original post mentioned the possibility of several arms which could get expensive fast. Also you must remember that these are hobby arms with hobby servos. I have the stronger AL5D model. I haven’t tried to find its weight and useful range of motion limits yet, but they are very limited. The further you reach, the less you can hold, and for a shorter time.
For your process I envision maybe three core devices:
You will also need some way to place a new bottle onto the turntable and to take the old one off. Perhaps your “deal with bottle” arm can do this if you can place everything in its range. Perhaps your cap grip arm can also do this. Perhaps you will need one or two additional arms.
Again, to me the key is what has to happen to the bottle between cap-off and cap-on.
You really have me interested, and there are many others on this forum much more experienced and capable than I. Seek them out. And please post pictures and videos of your solutions.
Hi Ted,
Thank you sooo much for your time thinking about my stuff, and I didn’t expect first that I can get ideas from a guy as help as you=) Thanks again and it’s a great honor and pleasure to discuss with you my thoughts:
How heavy are the filled bottles?
When the bottle contains liquid for the job it weighs about 200 grams. According to the website info, the AL5D should be able to haddle it, correct?
Does the bottle need to be tipped to pour something out? May want your high torque wrist and arm here.
Yes, pouring is needed in some cases and I also assume the AL5D + heavy duty wrist should work.
Or is it simply held vertical so something is poured or pipetted into it?
That’s also needed in some cases.
Must the bottle be shifted in the horizontal plane (translated) from A to B and back to A or on to C? Again you might want high strength/high torque arm holding the bottle.
Yes, location shiftability is anticipated. Again I plan to use AL5D for that.
Are the bottles round cross section (harder to grip without slippage) or square cross section (much less grip needed).
Very good concern! The bottles are manufactured as round, but I plan to modify the bottle by glueing on a squre and solid surface for the gripper.
There are two ways to twist: 1) Have an arm hold and twist the cap, 2) Have an arm grip the cap and have the bottle held in a turntable that can be turned by a continuous rotation motor or modified servo. (This could help avoid repetitive wire twisting and also keep heavy components off of your arms.)
I plan to do 1) you mentioned by buying two arms, because two arms are needed anyway as loosing/tightening is not the only job I expect the arms to do. I just placed an order for a weeker and smaller AL5A to get familiar with everything, and I’ll get a AL5D later to build up a complete system to do jobs with two arms.
Your original post mentioned the possibility of several arms which could get expensive fast. Also you must remember that these are hobby arms with hobby servos. I have the stronger AL5D model. I haven’t tried to find its weight and useful range of motion limits yet, but they are very limited. The further you reach, the less you can hold, and for a shorter time.
I’m glad you mention that AL5D is a promising one for the job which is what I expect. Essentially I need a weeker AL5A and a stronger AL5D working together, like left and right hands, which have different strength and agility.
For your process I envision maybe three core devices:
-That’s exactly what I planned so far!=)
Again, to me the key is what has to happen to the bottle between cap-off and cap-on.
I think so too and we’ll see soon=)
Hi Ted,
Thanks again for your help and I need more ideas from you here. The arm I just purchased has not arrived and hopefully tommorrow, but I was reading, watching and planning about using the Lynxmotion arm for the past few days. Now I have several concerns you may help?
Can the servos be changed to other ones with a stronger torque or speed if I found the oringinal configuration is not good enough, if the size matches? (may not be needed, just curious)
How much can these arm work? Say, after programming and connected to a PC, for a single arm, is there any problem that it: keep moving/holding objects for 2 minutes (job A)—>rest and wait for 5 minutes—>keep moving/holding objects for 2 minutes (job B)---->repeat this above cycle for maybe up to 10 times, during which it may take total 1 hour or more. Do you think they can do that? If yes, I’ll be very proud of them and I’ll not look at them as hobby arms. If not, do you have any idea how to change or improve the parts the make it possible?
How accurate can these arm be? For a repetitive programmed process, the gripper will reach a certain point in 3D space with what extent of error (+/- 1mm? +/- 5mm? +/- 10mm? +/- 20mm?). If it can be no more than +/- 10mm it should be OK for me, if it can be +/- 1mm that’s fantanstic. It’s a hard question though without trying it personally and if you don’t have the answer I understand 
Is it possible to program motions of two arms into a single programm? Or they can only be programmed seperately and I have to do manual timing to make sure they work in a cooperative way. I bought RIOS, is it possible to do that?
Thank you very much again and I look forward to your ideas,
Min
Bear in mind that the lynxmotion robotic arms are designed for hobby/educational use and are not ment to replace commercial/industrial grade arms used in high useage situations. If you are expecting to automate your lab processes using the hobby servo based robotic arms, you should purchase and test one to see just what its capabilitys and limitations are.
Hi zoomkat, thank you for this reminder. Actually the work in my lab is pretty light and simple, not comparable to industry use and that’s why I want to give it a shot. According to your experience, is it possible to make a several-hours-long program for the arms? The arm doesn’t have to work all the time and actually more than 90% of the time it is resting, but needs to do its job at the time it is programmed, like every 1 hour etc.?
Thank you very much for your help and ideas from other guys who are reading this post now
Programming time into the control program shouldn’t be too difficult. Programming the arm motion might be more of a challange due to the calculations of circular motions (if needed). In place of the “arm” concept, a pick and place type of setup might be simpler in the long run. Below is an X-Y-Z type setup that might be worth consideration.
Hi zoomkat, that’s really cool, thanks for sharing. I’m working on programming now and pretty sure it’s gonna take a while. Thanks again
The arms use the standard SES servo brackets and will accept all the standard Hitec servos. You will have to look up the specs for the servos on the Lynxmotion store servo page to see if any upgrades will suit you. Since you are only slightly more of a novice than I am, I would not recommend trying to substitute some other brand of servo without a lot more specific knowledge.
The short answer is a qualified NO or MAYBE. These are hobby servos, and they DO NOT provide position feed back. When they are turned off, they essentially lose all torque and the device (arm) basically collapses into whatever position gravity takes it. Then when it is turned on, the controller board has no idea what the initial position of any servo is. So the first thing that happens is that the controller moves everything rapidly and violently to a preset “initial” position before allowing any controlled movement. IF you had moved the arm to a “rest” position quite near the “initial” position, this violent startup might not be too bad. BUT if you turned off the arm part way through your “deal with the bottle” cycle, it would probably be disasterous. SO you must keep power applied to all servos throughout your entire operation. If they are extended and under load, they could burn out, so you might need to program some non-extended “rest” positions to protect them.
I don’t have a good answer for this. My AL5D uses the Lynxmotion BotBoarduino controller board which receives signals from me through Lynxmotion’s PS-2 controller. I certainly do not have anywhere near the repeatable accuracy you are seeking. In fact I’m pretty bad at it. Also I gather from the forum threads by the guys developing their own radio controllers that the PS-2 isn’t very accurate anyway. So you would want program control which is outside my experience.
I say YES confidently on the theory that enough programming can accomplish anything. I don’t have RIOS and know nothing about it. Remember you would need a separate controller for each arm, and you would need a way to ensure that each one maintains power to all servos even when your program is addressing another arm’s controller.
At this point, like any good bureaucrat, I have given you the policy, it is up to you to implement it. I apologize for not answering sooner - at this season I am busier than a one-legged man at a butt kicking contest.
If you need more help with the programming, you will need to ask someone else because I just don’t know. Good luck.
One clarification: “Remember you would need a separate controller for each arm”
Not necessarily - the SSC-32 simply controls servos - it does not really care which “arm” each servo is connected to. FlowStone for example easily allows control of two arms from one serial port via the SSC-32.
Hi Coleman,
Thank you very much for your idea! So I assume that to program two arms into the same program the two arms are sharing the same SSC-32 board, correct?
I haven’t check how many servos can RIOS support, if not enough for 6 servo *2 = 12 servos, then maybe RIOS is not the good choice and I should use something like SSC-32 sequencer?
Thanks a lot for the help,
Min
