For VSI recording, 1-bit sampling and 8 MHz bandwidths (e.g. R1 and R4 experiments):
mk5=mode=ext:0x55555555:2
For VSI recording, 1-bit sampling and 4 MHz bandwidths (e.g. APSG, CRF, OHIG and T2 experiments):
mk5=mode=ext:0x55555555:4
To check which data recording mode is being used by the Mark5 recorder we run mk5=mode?
in eRemoteControl. The output from which will be:
/mk5/!mode? <return code> : <data source> : <bit-stream mask> : <decimation ratio> : <FPDP mode> ;
Eg. The expected output from mk5=mode?
is:
/mk5/!mode? 0 : ext : 0xffffffff : 2 : 2 ;
for the 2nd and 4th R1/R4 experiments each month;/mk5/!mode? 0 : ext : 0x55555555 : 2 : 2 ;
for all other R1 and R4 experiments;/mk5/!mode? 0 : ext : 0x55555555 : 4 : 2 ;
for APSG, CRF, OHIG and T2 experiments; and/mk5/!mode? 0 : ext : 0xffffffff : 1 : 1 ;
for AUST experiments.
When mk5=mode?
returns successfully, <return code>
will be 0
, anything else ought to be investigated. Descriptions for what the other possible return codes indicate may be found on Page 2 of the Mark5B command set pdf (Rev 1.12).
Possible values for <data source>
are:
ext
which indicates an external data source (data on VSI 80-pin connector);tvg
which indicates an internal text vector generator; andramp
which indicates an internal ramp generator.
Currently ext
is used for all experiments, it should be investigated if either tvg
or ramp
are returned when mk5=mode?
is run.
There are 16 channels (10 X-band and 6 S-band) that may be used, <bit-stream mask>
contains 8 bits (each bit encodes 2 channels) that indicate which channels, and how many bits per channel, the Mark5 will record.
Recall that the conversion to binary from the hexadeximal number:
5
is 0101
(which indicates one-bit sampling); andf
is 1111
(which indicates two-bit sampling).
One-bit sampling uses the sign of recorded voltages when recording, two-bit sampling partitions the interval of possible voltages up in to 4
pieces to obtain sign and magnitude components.
Possible values for <bit-stream mask>
are:
0x55555555
; and0xffffffff
.
Each channel is being sampled at a rate of up to 32 MHz. The maximum bandwidth per channel is 16 MHz. The decimation ratio sets the proportion of samples thrown away (or alternatively, the proportion of samples recorded) by the Mark5 recorder.
Possible values for <decimation ratio>
are:
1
which indicates we are keeping all of the data (16 MHz bandwidth);2
which indicates we are keeping one half of the data (8 MHz bandwidth);4
which indicates we are keeping one quarter of the data (4 MHz bandwidth); and8
which indicates we are keeping one eighth of the data (2 MHz bandwidth).The procedure file for each experiment should contain what bandwidth is being used (an example here would be good).
Possible values for <FPDP mode>
are:
1
which indicates a maximum data recording rate of 1 Gbps; and2
which indicates a maximum data recording rate of 2 Gbps.
Recall that the nyquist rate is the minimum rate at which a signal can be sampled without introducing errors, which is twice the highest frequency present in a signal. Consequently, the sampling rate is double the maximum bandwidth per channel (which the decimation ratio determines).
The data recording rate for an experiment can be calculated as <no. bits per channel>*<no. channels used>*<sampling rate>
.
Eg. The data recording rate is:
2*16*16 = 512 Mbps
for the 2nd and 4th R1/R4 experiments each month;1*16*16 = 256 Mbps
for all other R1 and R4 experiments; and2*16*32 = 1 Gbps
for AUST experiments.