Fieldbus: RETA-01¶
The RETA-01 is an Ethernet fieldbus adapter from ABB, installed in option slot 1 of the ACH550 drive. It bridges between the lab PC's Ethernet network and the drive's internal control bus. To the PC it looks like a Modbus/TCP slave on 192.168.50.10:502, unit ID 1.
Identification¶
| Item | Value |
|---|---|
| Firmware revision | 0x0130 (1.30) |
| MAC address | 00:30:11:16:3E:3A |
| IP address | 192.168.50.10 |
| Subnet mask | 255.255.255.0 |
| Gateway | none configured (private link) |
| Modbus profile | ABB Drives (51.16 PROTOCOL = 0) |
| Unit ID | 1 |
These values were recovered during the WT_MS_1 bring-up session and are stored in drive parameter group 51 (see Reference → Parameter index).
DIP switches¶
There is a single block of DIP switches (S1) on the RETA-01 board. All of them are set to OFF. With all switches off, the adapter takes its IP configuration from the drive's group 51 parameters; setting any of them on overrides parameters with on-board defaults and is not what we want.
If the adapter ever powers up with the wrong IP, the first thing to check is that S1 is still all off.
LED indicators¶
| LED | Healthy state | Meaning |
|---|---|---|
| LINK / ACTIVITY | Green, blinking with traffic | Ethernet cable is connected and packets are flowing. |
| MODULE STATUS | Green, steady | Adapter is configured and talking to the drive. |
| NETWORK STATUS | Green, steady | At least one Modbus master has an active connection. |
A red MODULE STATUS is the diagnostic to watch — it means the adapter cannot reach the drive over the internal bus, which is usually a seating or cabling issue inside the cabinet. If it appears, see Troubleshooting → Modbus connection.
Reaching the RETA-01 from the lab PC¶
The lab PC's Ethernet 4 adapter holds the private side: static 192.168.50.100/24. There is no DHCP server on this segment and no need for one. From PowerShell:
should report TcpTestSucceeded : True. If it does not, the problem is one of:
- Ethernet 4 has lost its static IP. Reassert with the connection script in
code/wind_tunnel_control/or via Windows Settings. - The RETA-01 has lost power (drive is off, or the adapter has come unseated).
- The cable is unplugged or damaged.
Why a persistent TCP connection matters¶
The RETA-01 is happy to accept one Modbus/TCP connection at a time. Software that opens a fresh TCP socket for every register read or write — which is how AeroWare was written — ends up retransmitting frequently and stalling on TCP handshake delays of ~9 s. This is the failure mode that ultimately motivated the move from AeroWare to ADCL WinSoft.
The replacement code (code/adcl_winsoft/vfd/modbus_client.py) maintains one persistent TCP socket per VfdClient instance with a threading.Lock around the wire-format writes. The PowerShell tool (code/wind_tunnel_control/WindTunnelControl.ps1) takes the simpler approach of opening a fresh socket per call — that is acceptable at 4 Hz polling but would not scale to a faster loop.
This is captured in the pymodbus quirks note and is the most important architectural decision in the VFD layer.
Parameter mapping¶
To read or write a drive parameter (not a quick-access register) over Modbus, the address is:
$$ \mathrm{address} = \mathrm{group} \times 100 + \mathrm{index} - 1 $$
For example, parameter 03.05 FB REF 1 is at Modbus address 3 \times 100 + 5 - 1 = 304. This is in the register map reference.