191 lines
7.3 KiB
Markdown
191 lines
7.3 KiB
Markdown
# HEPHAESTUS Wiring
|
|
|
|
This is how I have my LinuxCNC machine wired. I put a linuxCNC box in place of a Southwestern Industries TRAK CNC II control box. I am reusing the TRAK drives, motors, and encoders. Eventually, I switched to using linear encoders instead of the TRAK encoders.
|
|
|
|
| Mesa card | Role |
|
|
| --------- | ------------------------------- |
|
|
| 6i24-16 | FPGA card (PCIe host interface) |
|
|
| 7i52S | Servo interface daughter card |
|
|
| 7i37TA | Isolated I/O daughter card |
|
|
|
|
|
|
|
|
|
|
|
|
# Encoders
|
|
|
|
## TRAK Encoder Pinout
|
|
|
|
The encoders for both motors and the standard TRAK encoders use the same Amphenol CPC connector.
|
|
|
|
The encoders are differential 5V output. They might have an index pulse but I wasn't using it. The Z axis has the same pinout. Peak-peak voltage is 2.7V.
|
|
|
|
| Pin | Signal | Color (in PC) |
|
|
| --- | -------- | ------------- |
|
|
| 1 | B- | White |
|
|
| 2 | 5V | Red |
|
|
| 3 | B+ | Blue |
|
|
| 4 | N/C | |
|
|
| 5 | N/C | |
|
|
| 6 | N/C | |
|
|
| 7 | N/C | |
|
|
| 8 | N/C | |
|
|
| 9 | GND/SHLD | Black |
|
|
| 10 | GND/SHLD | Black |
|
|
| 11 | GND/SHLD | Black |
|
|
| 12 | A- | Yellow |
|
|
| 13 | GND | Black |
|
|
| 14 | A+ | Green |
|
|
|
|
|
|
|
|
## Linear Encoder Pinout (DB9)
|
|
|
|
I switched to some linear encoders that have a D-sub connector. They are also 5V differential output, and nearly the same resolution, resulting in very easy switchover.
|
|
|
|
| Pin | Signal | Color (in PC) |
|
|
| --- | ------ | ------------- |
|
|
| 1 | A- | Yellow |
|
|
| 2 | GND | Black |
|
|
| 3 | B- | White |
|
|
| 4 | N/C | N/C |
|
|
| 5 | Z- | Tan |
|
|
| 6 | A+ | Green |
|
|
| 7 | 5V | Red |
|
|
| 8 | B+ | Blue |
|
|
| 9 | Z+ | Maroon |
|
|
|
|
|
|
## DB37 Pinout
|
|
|
|
The DB37 cable contains drive signals and motor encoder signals and connects to the TRAK drive box.
|
|
|
|
*This pinout was determined by probing. I'm not sure what some of the pins do; some of the ??? pins might actually be an index pulse or something.*
|
|
|
|
|
|
| Pin | Signal | Notes |
|
|
| ----- | ---------- | --------- |
|
|
| 1-5 | ??? (N/C?) | |
|
|
| 6 | YB- | Y Encoder |
|
|
| 7 | YB+ | Y Encoder |
|
|
| 8 | YA+ | Y Encoder |
|
|
| 9 | YA- | Y Encoder |
|
|
| 10 | XB+ | X Encoder |
|
|
| 11 | XB- | X Encoder |
|
|
| 12 | XA+ | X Encoder |
|
|
| 13 | XA- | X Encoder |
|
|
| 14 | XENW | X Enable |
|
|
| 15 | XENR | X Enable |
|
|
| 16 | YENW | Y Enable |
|
|
| 17 | YENR | Y Enable |
|
|
| 18 | ZENW | Z Enable |
|
|
| 19 | ZENR | Z Enable |
|
|
| 20 | ??? (N/C?) | |
|
|
| 21 | 5V | |
|
|
| 22 | 5V | |
|
|
| 23 | GND | |
|
|
| 24 | GND | |
|
|
| 25 | 5V | |
|
|
| 26 | 5V | |
|
|
| 27 | GND | |
|
|
| 28 | GND | |
|
|
| 29 | 5V | |
|
|
| 30 | 5V | |
|
|
| 31-37 | GND | |
|
|
|
|
# MESA Cards
|
|
|
|
|
|
|
|
## 7i37TA
|
|
|
|
Three enables are used - one for the X and Y axes, one for the z axis, and one for the spindle. This allows the z axis to be disabled independently for manual drilling operations.
|
|
|
|
Each works by connecting a pair of wires. Made = enabled.
|
|
|
|
The spindle only has enable; direction and speed is manually controlled.
|
|
|
|
A custom `M101` and `M102` g code enables/disables the z-axis.
|
|
|
|
| Enable | 7i37TA Pin | 6i24-16 Pin | Wire Color |
|
|
| ------- | ---------- | ----------- | ---------------------- |
|
|
| XY | OUT6+ | 46 | White & Green to black |
|
|
| Z | OUT4+ | 44 | Red to black |
|
|
| Spindle | OUT3+ | 41 | White to Green |
|
|
|
|
## 7i52S
|
|
|
|
#### Encoder connector
|
|
(This is for connector 0; carries over to channels 0-4 as well)
|
|
|
|
*Five encoder inputs are used - x, y, and z each use one without index pulse, and x and y have a secondary linear encoder that does use the index pulse*
|
|
|
|
*Sometimes the A-channels and B-channels are flipped, at least color-wise.*
|
|
|
|
| Pin | Signal | Function | Wire Color |
|
|
| --- | ------ | -------- | ---------- |
|
|
| 1 | QA0 | A+ | Green |
|
|
| 2 | /QA0 | A- | Yellow |
|
|
| 3 | GND | GND | Black |
|
|
| 4 | QB0 | B+ | Blue |
|
|
| 5 | /QB0 | B- | White |
|
|
| 6 | +5V | 5V | Red |
|
|
| 7 | IDX0 | I+ | Maroon |
|
|
| 8 | /IDX0 | I- | Tan |
|
|
|
|
#### Differential Output Connector
|
|
|
|
*Three differential outputs are used - one for each x, y, and z*
|
|
|
|
| Pin | Signal | Function | Color |
|
|
| --- | ------ | -------- | ----- |
|
|
| 1 | GND | | N/C |
|
|
| 2 | GND | | N/C |
|
|
| 3 | TX0A | TXA+ | White |
|
|
| 4 | /TX0A | TXA- | Green |
|
|
| 5 | TX0B | | N/C |
|
|
| 6 | /TX0B | | N/C |
|
|
| 7 | +5V | | N/C |
|
|
| 8 | +5V | | N/C |
|
|
|
|
#### HAL Channel assignments
|
|
|
|
| Axis | Encoder ch | PWM ch | Notes |
|
|
| -------- | ------------ | ----------- | ------------------------- |
|
|
| X linear | `encoder.00` | — | Joint position feedback |
|
|
| Y linear | `encoder.01` | — | Joint position feedback |
|
|
| Z linear | `encoder.02` | — | Joint position feedback |
|
|
| X motor | `encoder.03` | — | Velocity/damping feedback |
|
|
| Y motor | `encoder.04` | — | Velocity/damping feedback |
|
|
| Z motor | (none) | — | No dual loop on Z |
|
|
| X drive | — | `pwmgen.00` | Differential PWM out |
|
|
| Y drive | — | `pwmgen.02` | Differential PWM out |
|
|
| Z drive | — | `pwmgen.04` | Differential PWM out |
|
|
|
|
## Servo Drive Config
|
|
|
|
### PWM
|
|
|
|
- Carrier: **15.5 kHz** (`hm2_5i24.0.pwmgen.pwm_frequency 15500`)
|
|
- `output-type 1` (PWM + Direction pins)
|
|
- `offset-mode 1` → locked anti-phase / bipolar single-signal: duty cycle directly encodes signed command
|
|
- `scale = OUTPUT_SCALE`: X = +1, Y = -1, Z = -1 (sign flips drive polarity to match wiring)
|
|
- peak-peak voltage of 3.3V from the TRAK CNC II control
|
|
|
|
|
|
When the enable line is deasserted the pwmgen output is forced off (not 50%), so the amp sees a distinct disabled state vs. "commanded zero". You can hear the drives hum/whine at the 50% state.
|
|
|
|
My notes say that setting duty cycle to 0 or 100% cause the drive to disable. I'm not sure if this is the case or not. I don't know exactly what the functional band of the PWM is.
|
|
|
|
### Enable
|
|
|
|
Enable lines are switched by 7i37TA isolated MOSFET outputs (`OBITn+`/`OBITn-`). Per the 7i37 manual, a **low** FPGA pin turns the MOSFET **on**. HAL sets `invert_output true` on the enable GPIOs, so HAL `enable=TRUE` → FPGA-low → MOSFET on → contact **closed**.
|
|
|
|
| HAL enable | FPGA pin | MOSFET | Contact | Drive |
|
|
| ---------- | -------- | ------ | ------- | -------- |
|
|
| TRUE | LOW | ON | made | enabled |
|
|
| FALSE | HIGH | OFF | broken | disabled |
|
|
|
|
Fail-safe: cable cut, 5 V loss, or hm2 unconfigured all leave the MOSFET off → contact open → drive disabled.
|
|
|
|
 |