Hoddem's DIY Linear Actuator

@Hoddem I have experience with control systems and protocols such as RIO, DH+ and DeviceNet. With all of these communication protocols they have an end of line resistor to prevent signals bouncing back up the cable. I know your looking at encoder signals but just wondered if a similar effect could be causing the problem you are experiencing?
Hopefully DE can provide you the answer.

Hiya @Hoddem just wanted to say thanks for the answers regarding the kangaroo and Optical encoders , fingers crossed you have better luck finding out the Limit switch error that seems to come up now and then with the Autotune function Teach, and I tottaly missed the 2 X 60 compatibility with the Kangaroo So never even tried it ... Hoping to have my Platform up and running again in the next week or so will post stuff in my project posting.. and the info about the dip switch setting confusion will save me some grief once im test the optical encoder / Kangeroo setup on spare actuator ..

your quite right about lack of space for the encoder signal pull ups so I actually made a PCB with the 10K pull ups that sat ovet the kangaroo . will try to find them and post a picture .. Cheers and once again thanks for the info ref my question

So do you just add a resistor in line with the signal wire? I can ask DE about that when I get a chance to talk to them. I actually called them twice and left a message, still waiting to hear back from them though.

No problem, I am still waiting for an answer from DE on the limit switch problem. If I get time I might try to run a tune using mechanical limits and just let the actuator crash at each end to establish the limits. I am also tempted to solder the limit switch inputs together and see if the problem still shows up. I know I wont be able to complete the tune, but If it actually got to the point where it hit a limit switch that would be a lot more progress then I have had so far.

I'm determined to get it working though, the sabertooth/kangaroo/encoder combo is awesome when it works.

This website gives a very brief description of its use and why they are needed... http://www.douglaskrantz.com/SCEOLNotTermination.html
I guess I should have called it a terminating resistor instead of an end of line as the two have subtly different purposes.
I've never needed to use on encoder signals but a quick search reveals that I'm not spouting nonsense, it seems that terminating resistors can be used to improve signal quality....
https://granitedevices.com/wiki/Terminating_differential_encoder_lines

I wish you luck in resolving the problem.

Hi @Hoddem found the PCB I made for connecting optical encoders to the Kangaroo . the idea at the time was to allow me to have the Pull up resistors mounted to the pcb and then plug the sensors into that and the whole ting connected to the kangaroo via two 4 pin headers ... the next version will be better laid out but at the time I was pulling my hair out due to the false limit switch error and I even tried the mech limits switch autotune but that played up so kinda lost heart . but the Kangeroo had / Has promise I just did not understand what I was doing wrong ... anyways wont clutter your post , just thought I'd let you see the road I travelled looking into this (bear in mind it was a long while ago but seeing your work has got me thinking again)..

Oh Oh I just remembered , I don't know if your aware that the DE Describe software now has a facility for you to see the limit switch state and check encoder pulses in DESCRIBE .. Charlene sent my an updated version of describe way way back with this function as a pre-release copy but I never checked to see if the latest version of describe has this or if you guys were aware of this

Nice, you must have some experience building your own circuit boards. I have done a few designs, but I always have to send them off to a board house and wait forever to get them back.

Funny, I called DE again today and spoke with a lady, cant remember who but maybe the same person. I asked her about the limit switch issue and she was pretty confident it is a limit switch error an not an encoder error. She said more than likely there is interference coming through the limit switch wires and causing the issue and using shielded cables is a must. just as you said, she also pointed me to the new describe software that allows real time monitoring of both the limit switch status and the encoder and she pointed me to the help desk article on how to do the tune within the describe software.

She recommended that I update the firmware in both the sabertooth and the kangaroo before I do anything and then for testing purposes try to isolate the limit switches from the rest of the wiring.
I will play with the describe software and run the tune with the pc and try to pinpoint the issue. I guess a requirement for this setup will be shielded cables all around.

https://help.dimensionengineering.com/knowledgebase.php?article=18

So I have been looking at performance charts and if I am reading them correctly it looks like the E30-150, E30-400 and A28-400 all have about the same maximum torque for our use.

Here is the torque chart for the E30-400 which is the motor I have had my eye on based on performance and cost. I added the lines in to show the corresponding speed, power, efficiency and torque at 60 and 120A (the peak continuous and peak momentary current of the sabertooth 2x60.



I ran the numbers on these three motors and this is what I came up with.



Somebody take a look at this and let me know if I on the right track here. I gave AMPFLOW a call but they didn't answer, I will try again later to get their recommendations.

Any joy with your encoders yet?

I originally tried the kangaroo & encoder route on my 2DOF, couldn't get it working properly so switched to pots, couldn't get them working either so ditched the kangaroo altogether and went with an Arduino.

Not yet, I have shielded cables on order and they will be here for the weekend. Hopefully I will get a solution put together one way or another.

I just got off the phone with a gentleman from ampflow and based on my requirements he recommended the E30-400 for testing purposes. He was a little concerned when I told him 60A and 120A current loads, he was recommending about 30A continuous current. He did confirm the more expensive motors do not result in greater torque at the same current, but will have a higher RPM.

I think that was about all the push I needed to get one and try it out, I will wait until after I figure out my limit switch issues though.

Hi,

here an alternative to the end stop switches.
I'm using "HALL-Sensor-Switches" for my actouator design.

Have you done any testing with the sabertooth and kangaroo. I would think the same problem would exist regardless of what type of switch is being used. Note its not a failure of the switch, its an issue with how sensitive the the circuit is to interference. I still haven't had time to test shielded cabling, but the more I think about the problem I keep wanting to just eliminate the limit switches and go with the mechanical stop tuning mode.

also, I figured out why my 25mm linear bearing was pushing out. the 25mm shaft clamp will actually fit inside the 3d printed housing. So when the system crashes it is actually crashing into the 25mm bearing. I'm not sure why I didn't notice that before. I will probably add a feature to create a hard stop so the problem goes away.

Hey @Hoddem ,

Any update on your project ?
How the actuator works ?

I’m in the progress of my actuator build.
But not in the stage to try it out.

Hi @Hoddem Love the idea of the no contact hall sensor for limit switch ( as used in many 3D printers ) , did you not have a problem with the actuator damaging these once the switch was tripped but the inertia overrun would cause hard contact with switch ain the position they are currently mounted ? maybe a spring loaded-4-5 mm diameter pushrod could be fitted to allow switch to be triggered but would allow the rod move against the spring during any gravity /inertia over-run and so as not damage the sensor or end up with mechanical crash which would put unnecessary load on the Leadscrew Nut mount when switch contact was made . not re-read the last few posts so maybe this is no longer an issue in you revised design .. again guys I mention these things not as criticisms but just observation learnt form past experience Damm Fine Work so Keep it up . as yet another great DIY Linear actuator takes the stage ..

Hi,

In my design I use the HALL Sensorsinstead of switches. The detection (switching) is about 10mm before the real physical contact will happen. I hope the 10mm is more then the overrun could be.

@wannabeaflyer2 ,@kermitkong
I haven't had a chance to get back to this project so I have yet to test the new shielded cables or play around with running calibration cycles with the computer. I'm hoping that I can do the calibration without limit switches and just use hard stops or even just use a single limit switch at the bottom of the actuator. I am curious from guys that are already using these how often the actuator hits a limit switch during normal use. From my limited testing I couldn't get the system to overrun far enough to trip one, even with me sitting on the actuator. If the encoder is working properly the system should know when it has traveled beyond the safe zone automatically. The single limit switch would still allow the system to home at startup.

I seem to have more time for designing then building and I have already finished designing R3 of my actuator. R1 was the concept design, R2 was a test case to see if a functional actuator could be built with limited tools and mostly 3d printed parts. I have a little bit of work left, but it looks to be a success. For R3 I am pulling the real tools out and doing a much more robust and overbuilt version.

Hi @Hoddem I can only report on my situation and to be fair to my sim most of the crashes where the actuator went past the limit happened during my setting up the system, to get max travel and speed .. the over travel happened when the limit switch was tripped but the switch actuation levers were the short type so they reengaged when the momentum /gravity of the ballscrew actuator went past he end of the lever and that generally led to hard crash nd panic button being pressed .. hope that explanation came out right so I found it better to use long lever on the switches ,so that in the event of a trip the switch ever actuator would keep the switch in the off state ...downside ment I have to hand wind the actuator back to safe position before re-starting the platform

either way I found that if yhe limit switch circuit killed power to the motors , sometimes the fully load sim and inertia would be enough to backdrive the ballscrew nut at least until it hit a mechanical stop , so based on my limited experience you may need to allow for more than 10 mm .. again just my 2 cents worth based on observation with my build . also note if the feed back pot belt tension gets too loose this may / will cause make the actuator lose its mechanical position . and I found that to be another actor in with limit switch override .. Cheers guys

Awesome actuator design. Very compact and clean.

Yes, that is right, your actuator is perfect. Congratulation for this work