Neoden TM245P Pick and Place Teardown and Upgrade

• Neoden TM245P Pick and Place Teardown and Upgrade
  Posted by luizrenault on 21 Oct, 2018 01:44
• Hi.
  
  Tired of the painful process to setup the Neoden TM245P Pick and Place, I decided to make a mod to OpenPNP to export a CSV file compatible with the Neoden machine.
  
  This way I can reuse component database and check if required components are present or need to be installed on free feeders.
  
  But that isn't enough. I really want to install a vision system and make it compatible with OpenPNP to control the machine.
  
  Today I opened the bottom plate and collected some information to check if the upgrade is possible using the original electronics.
  
  Here I share my findings...
  
  The guts of the TM245P consists of:
  
  

  • a motherboard (model TY164) with TFT LCD and resistive touch
  • a power supply and control board (model TY131)
  • two microstep driver
  • two vacuum pumps
  • one air blower
  • a transformer

  
  
  
  
  The motherboard have a STM32F407ZGT6 and a RA8875L3N TFT LCD Controller. The SD card slot can be seen on the left side of the board.
  At the center there is a 26 pin IDC connector that goes to the power supply and control board.
  On the top a 3 pin SWD connector and on the right side a 2 pin connector that powers the vibration feeder can be seen.
  
  The board uses as power supply 24V that comes from the power supply and control board. There is a 5V step down regulator (LM2576S-5.0) that supplies the uC through a 3.3V Linear reg.
  The 5V rail are sent back to the power supply and control board.
  
  
  
  This is the microstep driver:
  
  
  
  The power supply and control board have 4 diode bridge rectifiers (GBU808 on bottom layer) with +100V, +30V, +12V and -12V labels.
  
  There are four step down regulators (LM2576T-ADJ): +30V->+24V, +30V->+24V, +12V->+9V e -12V->-9V
  
  And the following connectors:
  
  

  • P1 -> Transformer (label 220V)
  • P2 -> air blower (connected to AC input)
  • P3 -> Transformer (label 110V)
  • P4 -> AC input and on/off switch
  • P5 -> Transformer taps labeled 70V, 23V, 11V, 11V
  • P6 -> To motherboard
  • P7 -> Power Supply to step controller X axis (connected to 100V labeled line)
  • P8 -> Power Supply to step controller Y axis (connected to 100V labeled line)
  • P9 -> X axis step controller signals (EN, DIR, STEP)
  • P10 -> Y axis step controller signals (EN, DIR, STEP)
  • P11 -> To DB9 connector at the back of the equipment (CAN-H CAN-L JTMS e JTCK
  • P12 -> Endstop Y axis switch
  • P13 -> Pick and place head
  • P14 -> Vacuum Pump A
  • P15 -> Vacuum Pump B
  • P16 -> Endstop / origin Y axis switch
  • P17 -> Feeder block 1
  • P18 -> Feeder block 2

  The feeders and the head are controlled through a CAN bus connection. The same bus goes to the back DB9 connector.
  
  
  
  
  
  This is the 26 pin connector pinout from the motherboard:
  
  
  
  I tried to read the firmware but the uC have Level 1 Read Protection.
  
  The next step is to log the CAN communication while sending simple commands using the equipment GUI.
  
  Then I will check if the X and Y axis can be controlled through CAN messages.
  
  Does anybody have the CAN message specification?
  
  Anybody tried to mod this machine?
  
  Best Regards,
  
  Luiz Renault

  ---

  

  

  

  

  

  
• #1 Reply
  Posted by luizrenault on 21 Oct, 2018 19:33
• Hi today I placed a scope on the CAN_H and CAN_L signals on the DB9 back connector.
  
  But they are too noisy to be useful and I had open the Pick and Place head and monitor on the inner (uC) side of the CAN transceiver (SN65HVD230).
  
  I assumed a bit rate 512 kbit/s because the smallest pulse length.
  
  But the oscilloscope wasn't able to decode it. The problem is that the CAN bit stuffing is not present.
  
  Does anyone have any clue?

  ---

  

  scope_11.txt

  

  scope_12.txt

  

  scope_13.txt

  scope_14_csv.txt

  scope_14.txt
• #2 Reply
  Posted by luizrenault on 21 Oct, 2018 20:38
• Although it uses a CAN transceiver, I could see on the Placement Head board that the CAN_TX and CAN_RX signals were connected to the USART signals of the STM32F103C8 (PA9 and PA10).
  
  Then I was able to decode de USART 512kbit/s signal. It uses a binary format.

  ---

  
• #3 Reply
  Posted by szymon019 on 31 Dec, 2018 12:08
• Hello,
  
  I am also interrested with your mod and I am offering help.
  
  Kindly Regards
  
• #4 Reply
  Posted by charliex on 31 Dec, 2018 21:07
• i only have the tm220a and the neoden 4 , i've though about redoing the tm220 since neoden no longer support it and it has firmware bugs, i just dont know if its CAN internally though, packets might be similar.
  
  probably be easier to use a can sniffer device than a scope to listen to that bus though,
  
  a totally wild guess its going to be something like 6-8 byte with a crc16/crc8 , or bottom of a crc16 at the end with a modified crc8 single byte reply