Linear actuator using ClearPath Integrated Servo System @ 72V DC

I too am confused even after reading previous posts. I have a 3 phase connection available (Europe). Can I use this by connecting the L1, L2 and L3 and not using the Zero. In my opinion, the star / delta circuit cannot be set on the servo drive.
Who has the redeeming answer in which I don't blow up the drive.

Thanks,
Builder.

Ahhh,... cool! I didn't even try that Never occurred to me. (how to remove the key)

I thought about using 3 phase AC as well, but I think (not sure) my servo controllers are the 240V only version

I did some research and it is now clear to me.
The AASD-15A needs a single phase 230V / 50Hz power supply or a three phase 220V power supply.
The 3-phase 220V is therefore not the 3-phase 400V common in Europe.

How about this now. The 3 phase, which is common in Europe, is structured as follows:

There is 230V between each phase and the ZERO (Nul). But 400V is present between the phases. So if we now connect Fase1, 2 and 3 to L1, L2 and L3 from the AASD-15A, it will probably go up in smoke.

If you still want to use 3-phase, you will have to use a 3-phase transformer in the power supply line that transforms the 3x 400V to 3x 220V.




However, this is a very expensive solution (€ 600 and more)

It is better to use the single-phase 230V, however, not many more users must be connected to the same group to prevent your fuse from tripping.

You can consider connecting 3 separate sockets from a power group. You then connect 2 AASD-15A to each socket to spread the load.



Good luck.

Builder.

Thanks! Great explanation. That's how I understood it as well.

Initially my argument for wanting to use 3 phase 400V was this:
I have 6 motors with 1050W of power. So theoretically the motors could draw 6300W from the mains. Since european mains is 240V, it would have to draw 26A. That would have tripped the 16A breaker.

Then I considered doing just what you suggested. Connect two motors to each 240V phase. That meant I could theoretically supply 11520W.

But eventually I wired all six motors up to a single 240V line, "just to do some testing" and also the computer, 3D printer, lights and monitor all draw from that same line and it still didn't trip the breaker. Ever.

I think that goes hand in hand with what I said earlier, over 99% of the time those motors operate at somwhere around 10-20% load.

Cheers,... Dirty

Hi Dirty...
First of all, congratulate you on the awesome job for your project, not only the actuator itself also for your developments around the software/filters. I feel envious!.
I'm starting with the design phase of 6DOF platform, just first small steps. Your actuator design is the way to go for me, but I would like to make some changes and I would like to get some feedback from you about it. All your design is well thought-out, first reason I choose it, but not sure of what could the impact of those changes that I'm not able to forsee now.
I'm thinking to use a 2020 ballscrew, mostly oriented to reduce the noise coming from the actuators and trying to improve the "feel" (in line with InsectaTech feedback about the "analog feeling" from 20 lead pitch...).
For such purpose, and trying to get a max linear speed about 500mm/s, I thought would be needed to change the reduction belt system to 2:1, to get 4,8Nm for 1500rpm from these actuators (80ST-M02430)
The cost is not much increased from the original design... and the benefits could be worthwhile.
In the same way I heard about of issues related to these motor drivers and 10khz noise issue, so maybe would be worthwhile to investigate driving them with an alternative solution like Odrive or Vesc drivers, to overcome those and at the same time reduce the costs.
Regards,

Hey @brainstormi ,

you very much speak my mind! Back then when I started, the build threads I read used SFU 1605 ball screws and so the most logical thing to do was to follow along with those builds until I eventually diverted by choosing SFU1610.

The reasoning behind that decision was exactly as you said:

• Double the lead --> Half the RPM --> Half the noise
• Double the lead --> Half the RPM --> Half the vibrations
• Double the lead --> Twice the torque required
• Double the lead --> No self-inhibition anymore
• Double the lead --> Twice the speed

...I leave it up to you wether you want to count those points as pros or cons, but in my case the only con to check for was the higher torque requirement. But since my motors (80ST-M04025) were plenty strong, I knew that that was not going to be a factor at all. So, the SFU1610 it were!

One note on the self-inhibition when back-driven: It is not directly the lead that determines wether or not the ball screw self-inhibits. It is actually the gradient of the helix. Which itself depends on the lead and the diameter. Simplified: If you want to compare ball screws in that regard, then calculate the ratio of diameter/lead. Lead values near or greater than the diameter (like SFS 1620 or SFU 1615) can usually be back driven easily without locking, values lower than the diameter (SFU 1610 or SFU 1605) can still be back driven, but generate A LOT of resistance.
Also, the force required to START a backwards motion is much higher than the force to KEEP a backwards motion going. I believe that this manifested itself in a noticeable stuttering on slow retract moves where the ball nut is constantly switching between an inhibited and an uninhibited state.

That was the past. Here's the future:
For exactly the arguments above, on my next build I will be going one step further and use SFS1620. I already have them here on the desk and they feel very promising. Absolutely no self inhibition anymore, super fast, still strong enough and super quiet. I guess I will have the motors running between 20 and 40% torque for most of the time, but I see no reason why they shouldn't be able to do that.

It's hard to describe with purely technical terms, but yeah, they have a very "silky" feeling to them. I didn't know that InsectaTech described it with those words (analgogue), but I could immediately relate to it when I read it in your post. It will be a few more months before I can test them out, but from my perspective I have my hopes up high when it comes to smoothness of those 20mm lead ball screws.

Longevity is certainly a bit of a concern there. SFS1620 will be under significantly higher stresses, so I wouldn't be surprised if wear and tear became an issue. Worst thing that could happen is I will swap them for SF_-2020 or even SF_-2520 at some point.

I think you are on the right track with the 2020 ball screws. Let me know how they work out.

The 10KHz whining noise is definitely there and audible, but it is not so bad that you couldn't use those servos. They are still very good bang-for-buck-wise. Maybe @Thanos knows other servo manufacturers that have control loop frequencies that are higher and thereby inaudible, but I reckon they be more expensive.
If I'm in VR, even with Index' and Reverb G2's near-ear speakers I can basically not hear the noise anymore.
To me a compromise I was willing to make.

Cheers,... Dirty

Hi @Dirty ,
You are a well of wisdom, I'll bring with me your tip about the ratio diameter/lead and its impact to back drive the actuator. I'm desiring to see what your new build get over the table. Unfortunately from my side I'll progress in a slower pace, this would be my first build and I don't like to put the cart before the horse, so I'll start playing with motors/electronics/programming to grasp the inners, maybe with a ffb joystick/shifter design meanwhile I find a bigger place to allocate the moving platform.
Regarding the motor driver discussion, as I see it, it seems the noise problem is related exclusively to the driver itself and its control frequency, because this issue is not present when you use an IONI driver board with these servos. It's not only a matter of costs/noise, also could bring other benefits to use a different driver like reducing the footprint installing PSU + drivers on the legs of the platform, reducing EMI due to shorter cables ... I already see and interesting open source driver/controller project, OpenFFboard, that it's already driving these Mige motors successfully. It's an interesting step to dig a bit more in my step learning curve...
Cheers,

Getting very close to start my thread and show the assembly of the entire rig.

Today I got finally the package with the rest of the motors and drivers.

BUT, to my surprise... DRIVERS ARE DIFFERENT to the AASD15A

When requesting the purchase, I sent pictures and model numbers. They supposed to understand and ship the right stuff.

The manual recieved is in Chinese and I found some kind of similar English manual here:

http://www.topcncworkshop.com/data/download/201912/20/53099b6d4c63185bf17b3b9925ec6e93.pdf

Seems that the specs are similar (or better... 30A?)... But worried about the parameters.

Already left a message to @Thanos in RaceDepartment.com, but wanted to post here in case someone has seen this type of driver.

Thanks to all as always for any idea if this would work or I shall return!

Charlie

Definitely return them. Not only there is no DB25 plug and play connector (has a 50pin connector instead), it doesn't support automatic calibration against hard stop either...

Make sure when you order to get from seller that their drive looks like this externally:




Link to this for example

Thanks
Thanos

Thanks so much @Thanos

Will return them!

CI

Hey there ,

I have done some tests to confirm my FE analysis:


Since I designed those actuators for a nominal load of 1000N (~100Kgs / 220lbs) in pull direction and I wanted the usual safety factor of 3, I had to put 300kgs of pulling load on those U-joints.

I used a (pretty beefy) tension belt from my slackline kit to create the tension and a crane scale to measure it, but even though I am certainly no pencil-neck, I was unable to create the necessary 3000N with the wrench.

So, I printed one isolated half of the forkhead as a test sample to measure on it's own.


That way, I only needed to generate 1500N of load to confirm the safety factor of 3


First step:

1600N, held up over 1h. It appears the belts have given a little, so I had to re-tighten it twice to keep the load up.

Second step:

2000N, also held up over 1h.

That's the equivalent of 400Kgs/880lbs pulling on those U-joints!

Given that over 99.9% of the time those joints will only be under ~30Kgs of compression load, I'd say that is a reassuring result

What I will do now is keep up that load for a few hours, just to be sure and then comes the fun part:

Pulling them apart to determine the ultimate failure load

Oh,. btw,

I printed those on a Prusa i3 Mk3 in Prusament PETG, 260°C, 0.3mm layers, 5 perimeters with 0.4mm each and 40% infill.
I also made some test samples in Polycarbonate (Polymax PC) which could almost double the strength, but I think that won't be necessary. How would I even test those?!?

Cheers,...

Dirty

Oh, and I think one major contributing factor to those positive results is the fact that I am not introducing the load directly into the PETG, but via a bearing. That distributes the loads nice and evenly over the surface and prevents load peaks:

Hi Dirty, have you got more experience about 2020 ballscrews? How much them need torque to keep actuators static?

Hey @S2.

Sorry for replying so late. In my latest actuators I am using SFE1620. Going from 10mm to 20mm lead means that your motors will need twice the torque to keep the static load. In my case that was not an issue:

...but you should definitely check for your motors. If you are using 2020 ballscrews (and no belt reduction) then you will be able to push 314N or ~30Kgs for every Nm of motor torque.

I can definitely reccommend 20mm lead on the ballscrews, because it significantly reduces vibrations since they spin at half the speed compared to the 10mm versions. It's really hard to describe. They just feel smooth as silk

Whether or not you really need the 20mm in diameter is another question. So far my 16mm diameter screws work fine, but for commercial applications 20mm or even 25mm diameters might be the way to go. I know one supplier of custom industrial actuators that uses 2525 ballscrews. But those actuators were almost 6000€/pc.

In short:
1. Go for the 2020!
2. Use a strong enough motor, or....
3. use a belt drive to get enough torque.

...you'll be fine.

Thank you for reply.

I have ordered SFS2020 700mm screws. Do you still use M04025 motors?
I have 80st-M04025. Think I try them first with 1:1.

Yes, I am using an M04025 (4NM/2500rpm) and the 1:1 transmission onto a 1620 spindle works fine. Another big advantage of a foldback configuration, besides the shorter overall length, is that you have the option to easily change the transmission ratio to make maximum use of the motor power.

Let me know if you start a build thread!

hello I want to ask what is now the maximum lift weight of your platform with ballscrew 1620.thx