[automotivetools]

Hi everyone! XD

-- Very sorry for the long message. If you do not want to read everything, please read only my Hypothesis, and What I need section

I have a degree in Electrical Engineer but have only basic knowledge about Servo Drive, I have much more experience in microcontrollers and programming. The company I work for will be launching a product with a servo drive and I was responsible to make a PCB with a microcontroller to read some information from the Servo using RS485/Modbus. I thought the task would not be hard, but after one month struggling I have exhaust all my ideas.

At least I know what is the problem, a strong electrical noise in the Servo Drive that interferes and destroys my RS485 packages/frames

Let me explain what I have:
Xinje Servo Drive + motor 2.4 N.m 400W single phase 220V DS2-P4-AS
Servo - Wuxi Xinje Electric Co., Ltd.-Products-
Manual - http://xinje.pl/pliki/INSTRUKCJE/Ser...anual(DS2).pdf

What I have done
I turned off and unplugged everything and let only the Xinje Servo Drive and the motor. I notice that when I put the servo in any control mode (position, speed, torque or even idle) and turn it ON, there is a very strong noise of amplitude of +- 4 V every 125 us in any Servo output. Please check the pictures and photos.

Pictures and photos
1) noiseRS485-ground.jpg - Measure RS485P to ground
2) noise-RS485P-RS485N.jpg - Measure RS485P to RS485N
3) noiseSwitchingPhaseV-to-Ground.jpg - Switching motor phase V to ground
4) noiseSwitchingPhaseW-to-V.jpg - Switching motor phase W to V

My hypothesis
Well, my conclusion was that the Servo is switching the phases every 8 KHz (therefore at every 125 us), since this is not a very good quality servo, the noise is propagating to every output. A friend of mine told me that this could be generating by something related to a term called "Hard Switching for the IGBT bridge", but if he is right the only solution is buying a servo from another brand.

Finally, what I need:
1) Do you think my hypothesis is right? Xinje engineers in China and Xinje sales person in my country think not.
2) Is there any way to solve this or should I tell my boss to buy another brand of servo? what brand do you recommend?


Do not need to read, but what I already tried in the last two weeks:
1) Design and RC or RL filter
Did not work

2) Using another Xinje Servo
Tried other two, same problem

2) Call Xinje Support
This was a total joke! a chinese girl from the technical support started laughing a lot because she could not speak any word in English.
I called twice and talked to two different persons. There is only one person in the support/sales that can write English, not speak. Ok, I cannot write nor speak English well, but I believe a company such as Xinje should have at least one guy that speaks well at the support.

Well, one "engineer" from Xinje said that I should used shield cables, I said I am using original Xinje cables. He also said the problem is in my PCB, even though I told him my PCB was totally disconnect from the circuit and I was measuring at the servo output.

Thank you for you time! Best Regards and I really appreciate any help, even if incomplete or if you not sure.

[Al_The_Man]

Did you try isolating the microcontroller supply from earth ground and also only use the two differential signal wires for the RS485, i.e. no common conductor.?
I am assuming the servo and drive behave OK and it is only the RS485 you have the problem with?
Also see Ch6 here http://www.automation.siemens.com/do.../emv_r.pdf?p=1
Al.

[mactec54]

automotivetools



It looks like part of their manual, is a direct copy from a Yaskawa manual, If there drives have been designed/copied right, you should not be having any problems with the drives, I believe that these drives are a bad copy of the Yaskawa Drive, the motors, are just a basic servo motor that is available from many different suppliers

[automotivetools]

Hello AI_The_Man,

Thank you so much for you answer. Yes, good question, my PCB board is using a different and isolated power supply. Yes again, I am only using the two differential pairs, no common conductor. I did this exactly to avoid interconnecting the PCB and Servo Grounds. Besides, I have used optocoupler to read signals /A, /B e /Z.

As for the manual you sent, I am going to fully read it. It seems a very good reference.

Thanks again!

Did you try isolating the microcontroller supply from earth ground and also only use the two differential signal wires for the RS485, i.e. no common conductor.?
I am assuming the servo and drive behave OK and it is only the RS485 you have the problem with?
Also see Ch6 here http://www.automation.siemens.com/do.../emv_r.pdf?p=1
Al.

[automotivetools]

Hello Mactec54

Thank you for your time. I believe you're right, their manual seems to me a bad copy from a manual of a different brand. Besides, in the first manual they sent me there was a serious mistake that cost me 3 days to solve. They have indicated wrong ground and digital power supply signals at the servo pins. I also believe their servo is a low cost copy from a good one.

Just for curiosity I have disassembled a DS2-20P4-AS and there are some PCB tracks that seems to have been connected with wires and solder.

automotivetools

It looks like part of their manual, is a direct copy from a Yaskawa manual, If there drives have been designed/copied right, you should not be having any problems with the drives, I believe that these drives are a bad copy of the Yaskawa Drive, the motors, are just a basic servo motor that is available from many different suppliers

[matici]

automotivetools,

did you manage to solve your problem?
As you mentioned, the company where you work uses Xinje servos. Could you share some general experiences with these servos? Could you recommend them for a basic positioning aplication?