URGENT: Weight support for custom-made robotic arm

Lynxmotion Robot Arms
1

Hello,

I ordered several servo motors, servo brackets, base rotation table, and SSC-32 in order to build a custom robotic arm.

I have one HS-422 for the base rotation, two HSR 5980 SG for elevation the arm’s 1st segment (Shoulder), 1 HS-422 for the 2nd segment (elbow), 1 HS-422 for the “wrist” rotation, and 4 HS-422 for the wrist, performing each one a different task, for my project’s purpose.

The mechanical part is built and is pretty light: aluminum bars and servo-brackets from Lynxmotion, connectors from Meccano, and different screws.

According to my calculations, the two HSR 5980 SG motors have to support less than 1KG of total weight (6 HS-422 motors, 6 aluminum bars, and 6 servo-brackets) at a distance of 25 cm (2 aluminum bars from shoulder to elbow, with 1 motor and 1 bracket) and then 18 cm (2 aluminum bars from elbow to wrist, with 5 motors, 2 aluminum bars and 5 brackets).

Tthe total torque on the base motors is (theoretically) less than what two HSR 5980 SG motors can handle (but more than 1)!!
But, when I try moving the arm, the motors are struggling to elevate !

When I remove the 4 wrist motors, elevation is easier, but still hard to do …

Connections are correct, as I tried the motors at no load (i.e. just connected with aluminum bars).

Logic voltage is 6 V
Servo voltage is 6 V (I tried 7.5, knowing that the HSR 5980 SG can perform better at 7.4, but no better results were observed…)

Anyone has an idea ?!?!?

Thanks for any help … the matter is quite urgent …

Regards,

Germain

2

What is supplying the 6V to the servos? if batteries then what size/capacity, or if a power supply then how much current can it deliver?

3

Actually, the power is being delivered by 2 power adaptors, 1 for logic, and 1 for servo

The one for logic outputs 6 V at 1800 mA
The one for servo has a multiple output selection (3, 4.5, 6, 7.5, 9 & 12) at 500 mA

I thought it might be from the 500 mA being too small, but even by switching them, the results are not much different.

Do you think that using two adaptors outputting 1800 mA each would do the job ? or is it something else ?

Thanks for the answers !!

Regards,
Germain

4

The more power for the servos, the better. The servos will probably need a couple of amps. Also, balancing two servos connected to a common connection can be tricky. One thing to check is the current being pulled under load by each servo. The position signals to the servos may need to be adjusted such that both are pulling equal current at a given load position. Otherwise one will take most of the load and the other won’t be doing much.

5

the logic uses a LOT less power than the servos so put the 500mA output on that.

1800 mA is you are pushing the servos, especially if the shoulder servos are fighting each at all, is probably not enough. you need to be certain you followed the alignment procedure in the assembly guide for the shoulder servos.

Can you get a bigger power supply or a 6V battery to test with? You could use 4 “C” or “D” alkaline cells in series, or a 6V lantern battery would be good too.

6

A quick warning about these types of power supplys: measure the actual output voltage with a meter prior to connecting it to the logic board. Unless the output is regulated, it may actually be putting out 16v when on the 9v setting.

7

Alignment procedure ? I just made sure both servos are at their extremum location and connected them to the arm…

I’m going to buy another power supply with the biggest mA possible.
Should have an answer by the end of the weekend !!

According to their specs, each motor drains 300mA at no load…
with 6 motors, even 1800 is not enough …

Do the type of batteries you gave me deliver more than 1800mA (I’d suppose 3000 or even 5000 mA would be fine, if that doesn’t burn the lab :slight_smile: ) ?

Thanks for the help !

8

It doesn’t matter, I have 2500 mah AA type, if you have 5 amps, it will just last longer. Go on the LM site to get the procedure to the aligment of the servos, I think it’s the most logical thing that’s hapenning (with a 1800 mah, should be fine, more is ok too)

Sometimes, even if both servos are at the extreme of there capacity it doen’t meen that if you put both at 45 degrees they will be at the same place, so one is pushing to go 2 degrees one way and the other the contrary. :arrow_right: Check that out

Maybe Jim or Beth can give you the link because I can’t seen to find it sorry.

9

well I can’t find it and I have to go at the moment but there is somewhere documented a procedure where you only have one of the shoulder servos screwed down and you set the servos to mid point and DRILL the holes in the horn of the outside servo and then screw it together. This made sure the two servos were properly aligned and did not fight each other. I don’t remember if it was at an extreme end or the middle. my gut feeling would be to align them at mid point as that gives the smallest error at the two end points.

10

Hello !

I couldn’t find more than 1800 mA, so that’s what I bought.

Base motors are working more easily and freely with half the load (3 other motors), but are still struggling with the full load …

I guess if it’s an alignment problem, then it wouldn’t be able to support half the load since the motors wouldn’t be working synchronously (unless 1 of them is taking everything in charge, and the other just following empty…)

Can any1 tell me more abt that “alignment” procedure ?

Thanks :wink:

11

This is not the way to align your servos. You need to make sure they are all at their center point (1.5mS). before you install connect them to the arm. The basic procedure is to send a 1.5 mS pulse (#X P 1500, where X is the SSC-32 servo channel a servo is connected to) and then connect them into the arm. Analog servos like the HS-475HB or HS-645MG have a +/- 90 degree range from center. If you don’t do this centering procedure for each servo before you put it into the arm, your range of motion will not be correct and the joints will not move the way they are supposed to move.

8-Dale

12

…and more importantly, if any two servos are driving a common axis, as in your shoulder elevation joint, they’ll continually be fighting each other, leading at the very least to greatly increased current draw, or to early servo failure as one eventually burns out from constantly resisting the motion of the other.

The physical endpoints of servos are neither consistent, nor are they guaranteed to correspond with any particular pulse width. Locking them down to the joints while they are being powered and supplied with a center-of-travel pulse is the best way to ensure that you get both the motion that you want and the accuracy of position that you expect from them.

13

Here 's the link. Look at step 15…

lynxmotion.com/images/html/build118.htm

14

You really really really need to be using a battery, preferably a big battery, for the servo power. Go buy a 6 volt lantern battery, or 4 C or D cells and a holder, and hook it up. A wall-wart power supply will just not be capable of providing the power required.

  • Jon
15

Lynxmotion sells a 6v 2a wall adaptor. Perhaps you might want to invest in one of these?

lynxmotion.com/Product.aspx? … egoryID=48

16

Not that I disagree with you in principle about batteries Jon, but a little clarification may be in order. The thing that comes to mind when people think “wall-wart” is actually a 12 watt supply. So the 5v at 2.5amps or the 6vdc at 2.0amps ones may be a little punky for this arm. However there are larger supplies that are a bit larger and generally have a PC style power cord. These supplies can be much more capacity. In the range of 25watts. They are harder to find with 6vdc output but they are out there.

The 4x C or D cell battery holder is a really bad idea. Although the batteries can produce a pretty hefty amount of current, the standard black plastic battery holders have like 26awg wire and rely only on a spring to make the electrical connections from one battery to the next. You need big 16awg wires to handle 2 or more amps.

17

Personally, I would try the batteries first.

The HSR 5980 SG has a stall current of 4.2 amps.

robotshop.ca/PDF/hitec-hsr-5 … csheet.pdf

The no-load current is what it will draw when you have nothing attached to it, and its moving. Throw that out the window if you’re running at half the max capacity of the servo. You’re going to be pulling 1-2 amps per servo at least, and that’s just for those two.

Use a battery.

  • Jon
18

Good point. A 6 volt gel cell battery would be a good choice to test this, or a 5-cell RC car NiMh C cell battery pack. Those are designed to handle huge amounts of current…

  • Jon
19

A battery is more capable of handling the high peaks and unpredictable input of the arm too, so I have to agree with Jon but I would go with a sub-c pack in 4x2 format. You could make a pack like this if you want to save money.

20

Wow !
Thanks for all the info !

The robotic arm I’m building is going to be used for medical applications, i.e. I can’t take the risk of using a battery which will (sooner or later) discharge and endanger the application’s precision (It has to be precise) !!

I shifted weight from the end of the arm, and aligned the base (shoulder) motors as explained in the tutorial link that was sent. The arm is now capable of movement (except when at full extension) with two 6V, 1800mA adapters.

The adapter on Lynxmotion website is 2 A, will it make such a difference (compared to 1800mA) ? Does Lynxmotion sell adapters with 6V output and (much) more than 2 A ?

Thank you for your concern !!