Re: Tidbinbilla digital VLBI backend with Mark5C
- Contemporary messages sorted: [ by date ] [ by thread ] [ by subject ] [ by author ] [ by messages with attachments ]
From: <Chris.Phillips_at_email.protected>
Date: Fri, 18 Oct 2013 06:16:14 +0000
Hi Shinji
This is good and bad news. Its great you have a new wider bandwidth option but this is now another VLBI backend to support. If we use this for the LBA we will have to decide how we support this in sched. If it looks like an RDBE (for example) if may be OK. We need to discuss this with Cormac before we use it for production use (tests etc are not a big issue as we can always hand edit vex files). We will want to run a fringe test sometime.
We should still be able to send modules to Parkes to copy off from the Mark5B+ there.
We definitely need to consider running all LBA observations on the new system. There are a number of possible issues though
- What happens is there is JPL and LBA VLBI at the same time?
- This has to be supported by Sched in some way without major hacking (see some comments below)
> Then the questions for you are:
> 1) Should the new DVP/Mark5C system be sufficient/compatible to observe some of LBA observations at Tid using Mark5A, such as the v271 series for Leonid and some others, e.g. for Yuri Kovalev, Oleg Titov, and Michael Bietenholz as we did in the past? With the DVP we can record up to 2Gbps, so we can increase sensitivity if we establish the use of the new system for LBA. We need a fringe test if we decide it’s feasible. In that case, can anyone provide us a Mark5C diskpack to be used for LBA?
As long as the disk packs are the same as we use with the Mark5B+ at Parkes, sure. We need to do a disk audit at Parkes and sort this out.
> 2) If we should keep the existing Mark4-DAT & Mark5A backend for LBA experiments, is anyone able to provide us an unused Mark5A module? Do you think it’s worthwhile?
I doubt it, certainly not if we can use the Mark5C.
> 3) If none of Mark5C and Mark5A options as mentioned above is possible, how bad is living only with the LBA-DR at Tid for the next several years? We can record fine up to 512Mbps, which I think the maximum data rate for any LBA stations to support. Perhaps only issue for Tid with only the LBA-DR is limited number of IFs for experiments that require wide frequency ranges.
The LBADR may not last too much longer - the Ceduna one has failed (motherboard). This can probably be repairs or a new server purchased but these are not issue free. The rest of the LBA has various other upgrade plans (we should discuss this in February).
>
> Tips for DVP Data Format: We don’t support Mark5B-emulated data. We write VDIF format to the Mark5C disk packs. We can output 16 Complex I/Q Channels, or 32 USB/LSB channels. Each data packet contains data from all channels. We can record up to 2 Gb/sec on one disk pack.
Are these two different modes? DIFX supports complex data but the setup tool chains do not completely. We want to avoid too much hand editing of setup files.
Thanks for the update - this is an exciting development.
Cheers
Chris
Received on 2013-10-18 17:16:27
Date: Fri, 18 Oct 2013 06:16:14 +0000
Hi Shinji
This is good and bad news. Its great you have a new wider bandwidth option but this is now another VLBI backend to support. If we use this for the LBA we will have to decide how we support this in sched. If it looks like an RDBE (for example) if may be OK. We need to discuss this with Cormac before we use it for production use (tests etc are not a big issue as we can always hand edit vex files). We will want to run a fringe test sometime.
We should still be able to send modules to Parkes to copy off from the Mark5B+ there.
We definitely need to consider running all LBA observations on the new system. There are a number of possible issues though
- What happens is there is JPL and LBA VLBI at the same time?
- This has to be supported by Sched in some way without major hacking (see some comments below)
> Then the questions for you are:
> 1) Should the new DVP/Mark5C system be sufficient/compatible to observe some of LBA observations at Tid using Mark5A, such as the v271 series for Leonid and some others, e.g. for Yuri Kovalev, Oleg Titov, and Michael Bietenholz as we did in the past? With the DVP we can record up to 2Gbps, so we can increase sensitivity if we establish the use of the new system for LBA. We need a fringe test if we decide it’s feasible. In that case, can anyone provide us a Mark5C diskpack to be used for LBA?
As long as the disk packs are the same as we use with the Mark5B+ at Parkes, sure. We need to do a disk audit at Parkes and sort this out.
> 2) If we should keep the existing Mark4-DAT & Mark5A backend for LBA experiments, is anyone able to provide us an unused Mark5A module? Do you think it’s worthwhile?
I doubt it, certainly not if we can use the Mark5C.
> 3) If none of Mark5C and Mark5A options as mentioned above is possible, how bad is living only with the LBA-DR at Tid for the next several years? We can record fine up to 512Mbps, which I think the maximum data rate for any LBA stations to support. Perhaps only issue for Tid with only the LBA-DR is limited number of IFs for experiments that require wide frequency ranges.
The LBADR may not last too much longer - the Ceduna one has failed (motherboard). This can probably be repairs or a new server purchased but these are not issue free. The rest of the LBA has various other upgrade plans (we should discuss this in February).
>
> Tips for DVP Data Format: We don’t support Mark5B-emulated data. We write VDIF format to the Mark5C disk packs. We can output 16 Complex I/Q Channels, or 32 USB/LSB channels. Each data packet contains data from all channels. We can record up to 2 Gb/sec on one disk pack.
Are these two different modes? DIFX supports complex data but the setup tool chains do not completely. We want to avoid too much hand editing of setup files.
Thanks for the update - this is an exciting development.
Cheers
Chris
Received on 2013-10-18 17:16:27