Delta Tau GEO BRICK LV User Manual
Turbo PMAC User Manual
68
Setting Up Feedback and Master Position Sensors
Conversion Table Processing Setup – Turbo PMAC Interface
16-bit data from an Acc-28x converter connected to a Turbo PMAC is processed for servo feedback using
an Acc-28 A/D entry (method digit $1) in the Turbo PMAC’s conversion table. Fundamentally, this just
copies the data from a Y-register to an X-register (where the servo loop can access it) and shifts the data
so that the LSB out of the converter will be treated by the servo as a count. The conversion table sets all
of the fractional-count bits to zero. Unlike the parallel data conversion, it will not roll over and software-
extend the data. The integrated Acc-28 A/D entry (method digit $5) does the transfer and shift too, but
also adds the previous result to the present cycle’s data. In either case, if the A/D data is unsigned, as
from an Acc-28B or 28E, the bit-19 mode switch must be set to 1 to handle the data properly.
an Acc-28 A/D entry (method digit $1) in the Turbo PMAC’s conversion table. Fundamentally, this just
copies the data from a Y-register to an X-register (where the servo loop can access it) and shifts the data
so that the LSB out of the converter will be treated by the servo as a count. The conversion table sets all
of the fractional-count bits to zero. Unlike the parallel data conversion, it will not roll over and software-
extend the data. The integrated Acc-28 A/D entry (method digit $5) does the transfer and shift too, but
also adds the previous result to the present cycle’s data. In either case, if the A/D data is unsigned, as
from an Acc-28B or 28E, the bit-19 mode switch must be set to 1 to handle the data properly.
Data from 12-bit A/D converters is processed using the parallel data conversion (method digit $2 without
filtering or $3 with filtering). The source address is the de-multiplexed memory register, not the address
of the ADC itself (unless you are only using a single channel of the ADC). The width/offset setup word
in the entry is $00C00C (width 12, offset 12) to use data in the high 12 bits of the register. This
conversion transfers the data to an X-register and shifts the data so that the LSB out of the converter will
be treated by the servo as a “count”. The conversion table sets all of the “fractional-count” bits to zero.
filtering or $3 with filtering). The source address is the de-multiplexed memory register, not the address
of the ADC itself (unless you are only using a single channel of the ADC). The width/offset setup word
in the entry is $00C00C (width 12, offset 12) to use data in the high 12 bits of the register. This
conversion transfers the data to an X-register and shifts the data so that the LSB out of the converter will
be treated by the servo as a “count”. The conversion table sets all of the “fractional-count” bits to zero.
Note that this conversion does permit rollover and software extension, so make sure there cannot be an
11-bit change (half of the input range) in a single servo cycle, or Turbo PMAC will roll over the result.
11-bit change (half of the input range) in a single servo cycle, or Turbo PMAC will roll over the result.
For details of setting up the encoder conversion table to process analog-to-digital converters, consult the
section Setting up the Encoder Conversion Table in this manual and the specification for variables I8000
– I8191 in the Software Reference Manual.
section Setting up the Encoder Conversion Table in this manual and the specification for variables I8000
– I8191 in the Software Reference Manual.
Conversion Table Processing Setup – MACRO Station Interface
16-bit data from an Acc-28x converter connected to a MACRO Station is processed for servo feedback
using an “Acc-28 A/D” entry (method digit $1) in the MACRO Station’s conversion table.
Fundamentally, this just copies the data from a Y-register to an X-register (where the servo loop can
access it) and shifts the data so that the LSB out of the converter will be treated by the servo as a count.
The conversion table sets all of the fractional-count bits to zero. Unlike the parallel data conversion, it
will not roll over and software-extend the data. The integrated Acc-28 A/D entry (method digit $5) does
the transfer and shift too, but also adds the previous result to the present cycle’s data. Because the only
16-bit ADCs that can be connected to a MACRO Station are unsigned, the bit-19 mode switch must be set
to 1 (making the second hex digit 8) to handle the data properly. The resulting data is copied to the
MACRO ring using MI10x, and then treated in the Turbo PMAC as any MACRO feedback data would
be, using the unshifted parallel conversion, with the MACRO node address as the source.
using an “Acc-28 A/D” entry (method digit $1) in the MACRO Station’s conversion table.
Fundamentally, this just copies the data from a Y-register to an X-register (where the servo loop can
access it) and shifts the data so that the LSB out of the converter will be treated by the servo as a count.
The conversion table sets all of the fractional-count bits to zero. Unlike the parallel data conversion, it
will not roll over and software-extend the data. The integrated Acc-28 A/D entry (method digit $5) does
the transfer and shift too, but also adds the previous result to the present cycle’s data. Because the only
16-bit ADCs that can be connected to a MACRO Station are unsigned, the bit-19 mode switch must be set
to 1 (making the second hex digit 8) to handle the data properly. The resulting data is copied to the
MACRO ring using MI10x, and then treated in the Turbo PMAC as any MACRO feedback data would
be, using the unshifted parallel conversion, with the MACRO node address as the source.
Data from 12-bit A/D converters is processed in the MACRO Station’s using the parallel data conversion
(method digit $2 without filtering or $3 with filtering). The source address is the de-multiplexed memory
register, not the address of the ADC itself (unless you are only using a single channel of the ADC). The
bits-used mask setup word in the entry is either $000FFF to specify the low 12 bits, or $FFF000 to use the
high 12 bits. This conversion transfers the data to an X-register and shifts the data so that the LSB out of
the converter will be treated by the servo as a count. The conversion table sets all of the fractional-count
bits to zero.
(method digit $2 without filtering or $3 with filtering). The source address is the de-multiplexed memory
register, not the address of the ADC itself (unless you are only using a single channel of the ADC). The
bits-used mask setup word in the entry is either $000FFF to specify the low 12 bits, or $FFF000 to use the
high 12 bits. This conversion transfers the data to an X-register and shifts the data so that the LSB out of
the converter will be treated by the servo as a count. The conversion table sets all of the fractional-count
bits to zero.
This conversion does permit rollover and software extension, so you must make sure there cannot be an
11-bit change (half of the input range) in a single servo cycle or Turbo PMAC will roll over the result.
11-bit change (half of the input range) in a single servo cycle or Turbo PMAC will roll over the result.
For details of setting up the encoder conversion table to process analog-to-digital converters, consult the
MACRO Station manuals, the section Setting up the Encoder Conversion Table in this manual, and the
specification for variables I8000 – I8191 in the Software Reference Manual.
MACRO Station manuals, the section Setting up the Encoder Conversion Table in this manual, and the
specification for variables I8000 – I8191 in the Software Reference Manual.