The VFS5 servo module was applied to the spindle unit of Cincinnati Sabre 1000 machining center.
Ex. 1 After power on, VFS5 servo module ACTIVE and FAULT lights were on, while SYS OK and CPU were not. Measured the power supply ±18V and +10V on the plug C4, all were normal. And the DC bus 325V also normal. The power supply +5V of microcontroller 136 (N80C196KB16) was normal. The logic analyzer showed that the internal clock signal CLKOUT (pin 65) was normal, but there was no data in the bus. Almost all the port lines were high level. Obviously, the program was not running.
When the power supply is within the tolerance range and the clock signal is stable, the microcontroller can be reset by adding a low level of no less than 4 state times at the reset end (pin 16). Using the logic analyzer tried to capture the falling edge on the reset end, and nothing was found.
The external reset circuit is shown in Fig. 1.
The power on reset circuit is composed mainly of under voltage detection circuit 166A, resistor 193, capacitor 191 and Schmidt trigger 168A. It is integrated by the triodes 156 and 142 with the instruction reset circuit of microcontroller, and directed to pin 16 of microcontroller 136.
Monitoring the logic state of each point with logic analyzer, checked some devices with multimeter, decided that 168 and 158 were poor or damaged, and the reset circuit was normal after replacement (Fig. 2).
Fig. 2 The logic state waveform of each point after the reset circuit was normal
Ex. 2 After power on, the VFS5 servo module FAULT, SYS OK and CPU lights were on, but the ACTIVE light was off. Checked that the voltage of each power supply including DC bus was normal.
According to VFS5 control program (stored in EPROM AM27C256) and logic analyzer tracking, it was known that the cause of fault was the alarm 84 (OUT OF BOUNDS), which is caused by improper setting of VOSPD (overspeed setting) and VMAX (max. speed). Using MotionLink+ to read the variables in the module and compare with the backup file, it was found that the values of most variables had changed (Fig. 3).
Fig. 3 Comparison between real reading and backup of variables
The backup file was transferred to the module. Because the module is spindle servo, the ROTARY variable must be set to zero before transferring to avoid alarm. The fault light was still on after the transfer.
According to the tracking analysis, this time it was the No.18 alarm (BAD TL). Further tracking showed that the value of D672H (word) unit in SRAM DS1230Y had become 9096H (-28522D). The unit corresponds to the variable VOLTS, whose value should be 00E6H (230D).
This variable belongs to Kollmorgen Protected Variable to ensure the correct matching of servo and motor. Users can not modify it through online communication. So 5672H (word) unit was modified by using general programmer. This is because in the VFS5 servo module system, the starting address of SRAM DS1230Y is 8000H.
After the modification, the alarm became No.27 (R/D JUMPERS). The alarm was caused by the change of resolution setting of R/D conversion. The content of D6B9H (byte) unit was changed from ECH (236D) to 0CH (12D) in the same way of tracking and modifying, which is 12-bit resolution.
At this point, the FAULT light went out and the module returned to normal.
Ex. 3 The module FAULT light was on, and the cause of the fault could not be found after routine inspection. According to the logic analyzer, it was judged as No.17 alarm (FEEDBACK LOSS), that is, the feedback signal of revolver is missing.
Checked that the wiring and connectors of the revolver were normal.
The input and detection circuit of VFS5 resolver is shown in Fig. 4.
Fig. 4 VFS5 resolver input and detection circuit
The 8.5kHz sine reference signal REF is generated by the op amp 349, triodes 363, 364 and other components, and goes to the resolver to output SIN and COS signals. SIN, COS and REF are divided and sent respectively to the pins SIN I/P, COS I/P and REFERENCE I/P of the R/D converter 60 (AD2S82AKP). At the same time, the circuit composed of diodes 361, 362, triode 383 and reverser 111D detects the status of SIN and COS signals, which are integrated by pal logic device 171 (TIBPAL16L8) and then sent to the microcontroller.
The sine wave of 8.26kHz and 1.66Vpp at REF end was observed by oscilloscope, but there was no signal at SIN or COS end, and the same was true when the motor was rotated. Thus decided that the resolver was damaged.
Considering the convenience of disassembly, the fault was further judged. Loosened the set screws on the stator of the resolver and pulled it out. Placed it on a wooden block, and put the rotor of another resolver (of the same model) into it. After power on, the FAULT light went out. The sine waveforms of 6.6Vpp could be observed at the ends SIN and COS with oscilloscope, while that of REF was 14.4Vpp.
After replacing the resolver and adjusting the zero point (using MotionLink+), the servo and motor run normally.