Hoddem's DIY Linear Actuator

just start to assemble the actuator to see how everthing fits...

Looking good, Are you 3d printing the plastic parts or are you having them machined?
What are you using for the rods, it looks like polished steel for the main rod and just steel for the smaller ones?
Did you go with 5mm pitch or 10mm pitch ball-screw?

I had some time tonight so I disassembled the actuator and installed the updated version of the main bearing holder.





I also used some white lithium grease on the ballscrew to see if it would quiet down. It didn't help, still very noisy. I used the grease on my bolts that are threaded directly into the 3D printed parts and that works great though.

I am hoping to get the sabertooth and kangaroo all setup tomorrow. I have terminal blocks on the way and then I can install the new connector plate and button it up for good.

Hi,

Yes I'm using polished steel for the main rod.
For the smaller I'm using carbon rods.
The pitch of the ball-screw is 5mm
Also, the plastics are 3D printed.

i solve the issue with moving bearings like this:
Have three 3D-printed parts screwed together to fix the bearing. see picture.
hopefully this is stable enough. What do you think ?

That should work just fine, make sure you measure your actual bearings to get the correct overall length. I noticed that mine don't match the 3d model from the IGUS website exactly and are actually a little bit longer.

@kermitkong is getting close on his actuator, I was getting scared that I wouldn't get mine done first so I got to it.

First test using the Sabertooth and Kangaroo combo (No More Sparks!). I am using the actuator in R/C controlled position mode.
This was at default settings so 75% of full speed running off of a single battery. Everything is cobbled together quick and dirty until my connectors show up.


Second test running with the kangaroo set to maximum speed, I think the battery is struggling to keep up with the draw of the motor on direction changes.


It's still loud as hell, but I can tell it is coming mostly from the motor. It sounds like it is coming from the motor brushes.

Once I get my connectors I will clean up the wiring and get setup to do some load testing.

I'm questioning my choice of the random 800w DC motor (its so loud). I was looking into the ampflow motors and they seem to have a ton of power in a very small package.



I am tempted to pick up an E30-400 for testing, price wise its actually a better deal and I can get one on amazon.

Wondering if anyone has any experience with these motors. I know @Pierre Lalancette has the the E30-150 motors, but I don't think he has gotten to performance testing yet.

In any case I have the stuff to do 24V testing so I will give that a shot with the current motor just to see what happens.

Here is a video of the A28-400 from the ampflow website, I would love to test one out as it may be the ultimate motor.

I can't get these motors out of my head, I noticed that the Ampflow motors have a double shaft. I emailed them and they said they are designed to mount up to an encoder with no modifications.

Here is the encoder they designed them for

https://www.usdigital.com/products/encoders/incremental/rotary/kit/E2

Here is the response I received

Thank you for contacting AmpFlow.

This type of encoder will work:
www.usdigital.com/assets/datasheets/E2_datasheet.pdf?k=634982727183808679

If you order it with the "G" option, it will bolt right to the existing holes in the back of the motor.

Please let me know if there is anything else you need.

Regards,

Carlo Bertocchini
www.AmpFlow.com
650-593-6906

They are between $60 and $70 usd without cables, so not cheap but probably the easiest and best way to do it for guys using the sabertooth/kangaroo package.

for the price's of the motors' sabertooth, power supply's...have you considered to going with AC servo/drive?

Ya, I have been toying with that Idea too. One of the major holdouts for me is the fact that I would need to run a few dedicated circuits just for the simulator. With everything being DC I can throw a couple of car batteries at the problem and bring my total power consumption at the wall to within the 20a maximum for my circuit.

My connectors showed up today so I had a little time to finish out all the wiring and button everything up



The terminations on the wires wouldn't fit so I had to improvise.





Here is my super sophisticated setup, lots of electrical tape.



I'm still just using the batteries, I didn't have time to get the power supplies setup. I am having some issues with the kangaroo now though and I'm not sure if its from the fact that I am using longer wires now or some other problem. I ran the calibration sequence yesterday and all was fine, today I can't get it to go. When I start it it immediately trips and gives me a limit switch warning. Also when I first plug in the sabertooth it tries to do the homing sequence but glitches several times throughout. I figured out that if I move my joystick on my remote around a bunch it will start the homing sequence again. eventually after a few cycles of that it will finally hit the limit switch and then shoot to the middle and all is fine again until I remove power.

I will keep looking at it over the weekend, but I may have to give DE a call to see whats up.

So.....Load Testing.

Don't laugh, I do this in the name of progress.



This is running on a single battery, first part is with the kangaroo cranked up to max speed, second part is set at stock (75%).
I weighed myself just after doing this and came in at a hefty 209lbs.

so ya, stop laughing now

That's a guy confident of his work (or just mad)
Impressive! and that is just with one battery (12v?)?

Ride em cowboy . You got to love the Sabertooth 2 X 60 .

It didn't break so lets go with confident.
Yes that was a single battery so 12v. However I have two batteries so I figured why not give it a shot at 24V. lets just say that was not fun, I was about getting air-born and that beam is only 3" wide so not a lot of support for a guy like me.

No video this time, I was just messing around trying to get this thing to auto-tune. still no luck and the profile was way off with 24V, even at center it was jumping up and down like crazy. Also for testing purposes I'm running a single 12g wire and they were getting pretty warm.

Hi @Hoddem quickie question , a few post back you touched on a problem I had a long time ago when I tried to use the kangaroo setup and optical encoders . that was part of the reason I changed to Multiturn pots .. I came across the same issues during the teach process where it would say limit switch tripped but that was not the case ????? were you able to get the teach operation to work repeatedly .

Also did you end up needing 10K Pull up resistors on the encoder outputs to ensure clean signal to the Kangaroo module ?

I also see you have the Kangeroo Connected to the Sabertooth 2x60 ( I did not know this was supported on that card ) or is there a new version/ firmware for the 2x60 that allows the use of the Kangeroo.. ?

I was able to get the tune to work every time without resistors when I was connected directly from the encoder to the kangaroo. Now that I have the ethernet cable as an extension I can no longer get the tune to run. I keep getting the limit switch error and I too have checked the limit switches multiple times and cannot find any issues. I did try adding 1k resistors between the A & B outputs on the Encoder and the 5V source per the kangaroo manual. I suspect the problem relates to the extension cable, but I can try adding 10K resistors when I get a chance.



It's odd though because I had already completed the tune and once I get the system powered up and homed everything works as expected and I can accurately control the position so it seems the encoder is working.


As far a I know the kangaroo has always supported the 2x60 so I'm not sure if that changed before I started working on this. I did not do any firmware updates, in fact I have not even setup my sabertooth in describe. I just plugged the kangaroo in and setup the dip switches per the manual and everything was working.



One thing I did struggle with is it seams the kangaroo manual and the pack-in instructions are contradicting for the sabertooth dip switch settings.
the pack-in says to have the sabertooth set to 1-on, 2-on, 3-off, 4-off, 5-off, 6-off
The manual says to have the sabertooth set to 1-off, 2-off, 3-on, 4-on, 5-on, 6-on





I struggled for quite a while before I figured out It had set per the manual 1-off, 2-off, 3-on, 4-on, 5-on, 6-on

I will continue troubleshooting and give the 10k resistor a shot (not much room on the kangaroo for resistors). If I continue to have issues I will give DE a call and just see if they can point me in the right direction.

I had a little time to play around with the wiring and it seems to be an encoder problem, but I still don't know for sure. I tried changing the 1K resistors to 10K and no change. I did notice though that If I get the actuator to perform the initial home sequence I seemed to be losing encoder pulses. If I run the actuator up and down a bunch it will slowly drift. I changed my 6ft Ethernet cable to a 1ft cable and I am able to consistently perform the homing startup and even got about half way through the tune sequence before it errors out. I also am not losing steps anymore so it appears that I am getting some signal loss from my cable setup.

just to rule out the limit switches I re-wired them to use the 4 connections on the kangaroo pcb and connected my ohm meter to each switch while running the actuator up and down aggressively. I never got a break in the continuity so it doesn't make sense that at the start of the tune where the actuator is barely moving I would be tripping the limit switch.

My next test will be to wire the encoder directly to the kangaroo with the shortest cable run possible and see if the problem goes away. If so I will have to look into shielded cables.

I will be giving dimension engineering a call tomorrow to discuss, because it is odd that I would be getting a limit switch error during the tune if it is in fact the encoder.

If it turns out to be interference I will probably go with a DB9 breakout board and cable instead of the RJ45 setup I have now. It will at least be shielded so it should help or fix the issue.