Topics

Closed loop Stepper

forna_daane@...
 

Anyone have tried this motor? 

Khalid Baheyeldin
 

There are lots of discussions in the past week or so on the MKS SERVO closed loop motors.
You have to avoid the SERVO42A, and try the SERVO42B.

They are very promising for OnStep, providing the most precise positioning of all tests that Howard conduction of motor/stepper combo.

One challenge is the need to run 6 logic wires per axis, and the fear of noise.

No one has actually tried them yet.

Serge CLAUS
 

Hello,
I ordered two MKS Servo42B motors.
Instead of waiting for delivery ;-)
My OnStep box is equipped with two DIN 4 and 5 pins for the motors.
I will replace them with 2 DIN 7 pins and network cable.
Wait and see ...

Le 12/01/2020 à 18:13, Khalid Baheyeldin a écrit :
There are lots of discussions in the past week or so on the MKS SERVO closed loop motors.
You have to avoid the SERVO42A, and try the SERVO42B.

They are very promising for OnStep, providing the most precise positioning of all tests that Howard conduction of motor/stepper combo.

One challenge is the need to run 6 logic wires per axis, and the fear of noise.

No one has actually tried them yet.

Khalid Baheyeldin
 

On Sun, Jan 12, 2020 at 12:24 PM, Serge CLAUS wrote:
My OnStep box is equipped with two DIN 4 and 5 pins for the motors.
I will replace them with 2 DIN 7 pins and network cable.
Wait and see ...
We are waiting ...

Please share pictures of how the cables are attached, at both ends.

Serge CLAUS
 

First tests of Servo42b motors.
Ramps 1.4 + Arduino Mega. 12V. Kstars and Indi.
900mA 32 micro steps
Huge couple. Seem very accurate.
Little noisy. The components do not heat up to 900mA.
Slightly warm engines.
More than to test in real conditions on Saturday. If the weather permits.


Drew 🔭📷🚴‍♂️
 

Found this listing on Ebay.

https://www.ebay.com/itm/BIGTREETECH-S42B-V1-0-Closed-Loop-Driver-Control-Board-42-Stepper-Motor-OLED-3D/184123450079?hash=item2ade9ce2df:m:mkyxIYfmnzE-GLtbRxT_gdA

Looks like you can buy the board only and add it to an existing motor? This is very interesting to me. I already (like many others) have a converted mount with the Nema 17 0.9A 0.9degree motors. Adding the board only is cheap and mechanically very doable. Any thoughts or  caveats?

--Oz--
 

I like the ribbon style cable verse lose wires connecting the two, should help with noise. 

Alexander Varakin
 

Drew, thanks for posting, it is indeed interesting, considering that the motors, supplied with the regular kit are very small.
I have larger NEMA 17 motors which I am thinking of using for directly driving GM8. 
BTW, I am hoping to have the first light of my closed-loop drives tonight. 

Drew 🔭📷🚴‍♂️
 
Edited

I went ahead and ordered the board + LCD version. I will be using them on my Vixen GP clone with 0.9A, 400 step Stepper Motors. I am currently running LV8729's  on a ESP32 V2.2 board which occasionally overloads and loses position.

Total cost was $33 for two setups. I have used Bigtreetech for my 3d printer upgrade boards and would recommend as a vendor. It should be about 2-3 weeks from now if they are not observing the New years holiday shutdown (I don't think they are).

I will post my experiences. Please keep posting yours Alexander and Serge.

Alexander Varakin
 

Looking more carefully at the specs, I see that there is a slight issue with these boards - they only go down to 32 micro steps. 
My MKS Servo42Bs go all the way to 256. 
BTW, I already tested my MKS Servo42B, you can see more info in this thread https://onstep.groups.io/g/main/topic/70125677

Alexander Varakin
 

On Thu, Jan 23, 2020 at 12:57 PM, Serge CLAUS wrote:
First tests of Servo42b motors.
Ramps 1.4 + Arduino Mega. 12V. Kstars and Indi.
900mA 32 micro steps
Huge couple. Seem very accurate.
Little noisy. The components do not heat up to 900mA.
Slightly warm engines.
More than to test in real conditions on Saturday. If the weather permits.
Serge, did you have a chance to test them?

bjaffa Jaffa
 

So the external stepper control board from BigTreeTech on ebay that you can add to your existing stepper has a magnetic pickup IC on the
back side of the PCB that mounts on the motor; I assume that you have to add a magnet to the motor back for the magnetic pickup IC to
sense the rotations. How easy is it to add to the existing motor? How accurate or how precise do you need to be in adding the sensor?

thanks  Brent

https://www.ebay.com/itm/BIGTREETECH-S42B-V1-0-Closed-Loop-Driver-Control-Board-42-Stepper-Motor-OLED-3D/184123450079?hash=item2ade9ce2df:m:mkyxIYfmnzE-GLtbRxT_gdA

Drew 🔭📷🚴‍♂️
 

"So the external stepper control board from BigTreeTech on ebay that you can add to your existing stepper has a magnetic pickup IC on the
back side of the PCB that mounts on the motor; I assume that you have to add a magnet to the motor back for the magnetic pickup IC to
sense the rotations. How easy is it to add to the existing motor? How accurate or how precise do you need to be in adding the sensor?"

I had the exact same questions. I found this on the Github repository for the board.

https://github.com/bigtreetech/BIGTREETECH-S42B-V1.0/blob/master/3d%20printer%20-ender3%20closed%20loop%20driver%20module%20installation%20instructions.pdf

The magnet comes with the kit, it is shown on Ebay once you know what it looks like. You install it with superglue according to the instructions. Doesn't seem to bad.

Being open source (along with the MKS version) I would hope the 32 microsteps limitation is to be increased as it gets updated and revised. Also,  being open source I would hope that the motor/board interface issues (if any) are resolvable. I will get into this when I get the board and attempt to run OnStep with it.

Alexander Varakin
 

On Sun, Jan 26, 2020 at 02:15 PM, bjaffa Jaffa wrote:
How accurate or how precise do you need to be in adding the sensor?
I don't think it needs to be very accurate - the motor has a calibration procedure, which I think is implemented in order to compensate for the inaccuracy of the magnet installation.

Drew 🔭📷🚴‍♂️
 

On Sun, Jan 26, 2020 at 07:54 PM, Alexander Varakin wrote:
I don't think it needs to be very accurate - the motor has a calibration procedure, which I think is implemented in order to compensate for the inaccuracy of the magnet installation.
Right, I found this is the encoder data sheet.

The TLE5012B is a 360° angle sensor that detects the orientation of a magnetic field. This is achieved by
measuring sine and cosine angle components with monolithic integrated Giant Magneto Resistance (iGMR)
elements. These raw signals (sine and cosine) are digitally processed internally to calculate the angle
orientation of the magnetic field (magnet).

In other words, once it is calibrated to the field generated by a particular magnet it is good to go.

Drew 🔭📷🚴‍♂️
 

The other thing is the polarization of the magnet. These are not like the neodymium magnets we are used to. They are polarized perpendicular to those. The North and South poles are on the diameter of the magnet, not the flats.

--Oz--
 

I don't think it needs to be very accurate - the motor has a calibration procedure, which I think is implemented in order to compensate for the inaccuracy of the magnet installation.Right, I found this is the encoder data sheet.The TLE5012B is a 360° angle sensor that detects the orientation of a magnetic field. This is achieved bymeasuring sine and cosine angle components with monolithic integrated Giant Magneto Resistance (iGMR)elements. These raw signals (sine and cosine) are digitally processed internally to calculate the angleorientation of the magnetic field (magnet).In other words, once it is calibrated to the field generated by a particular magnet it is good to go
I have used analog and digital hall sensors and radial flux sensors (instead of pots) in my home built PC sim racing wheels/pedals, also in my RC transmitters for my quad-copters/planes.
The magnet used is called diametric magnetized.


Centering the magnet on the shafts axis is important and then positioning the sensor, but less so on this type of sensor (flux lines verse flux density). Calibration can take care of some errors, I think the key is "how much" can calibration take care of the mechanical errors. I am curious exactly what the calibration is doing, the datasheet quote does not really say anything about calibration. Howard's testing shows very good results, but those magnets were mounted at the factory and most likely with a small jig setup to align the magnet. I used a micro scope, goop glue (gives time to align) and a tooth pick to use as a reference to center the magnet. Also the quality of the magnet linearity .Howards setup should be the truth detector :) 

Drew 🔭📷🚴‍♂️
 

If you look at the Ebay pictures it appears that the magnet is the same diameter as the washers that go underneath it. This implies that there is a shaft centering recess in the magnet. This means centering is not an issue with the supplied magnet. I will post if indeed this feature is present when I receive the kit.

--Oz--
 

This implies that there is a shaft centering recess in the magnet.
The picture I found I did not see this, got a picture or link? tia
In this picture the magnet is under the right bottom corner of the LCD, looks like a standard shape disk magnet.


Maybe it does not matter at all. I am still curious what the calibration function does, anyone?

Drew 🔭📷🚴‍♂️
 
Edited

Here you go. This is the "Type 1". It is not hidden in this picture, the magnet is between the washers and screws. While it is physically the same as a disk magnet it is polarized as in the previous postings.  The calibration routine "trains" the encoder to the magnetic flux lines of the particular magnet.