Hello Rolando, I think the answers you seek are in the message Rurik sent you, on the mailing list, last April, a copy of which I append. You can easily download a comprehensive set of model files and other material for the PhRInGES design at the github links Rurik included in that message. This design uses XAUI links between iBOB and iBOB and iBOB and BEE2.
I would, however, like to emphasize, and indeed expand on, the warning Rurik gave you in that message: > One thing to keep in mind is that since the iBOB is no longer > supported you will need to use an older version of > Matlab/Xilinx/mlib-devel/etc. I believe the last version we could use > was Xilinx 10.x (might have been 11.x but I can't remember for sure at > the moment). Additionally, you will need to use the following version > of mlib-devel to compile the design: > > https://github.com/sma-wideband/mlib_devel/tree/vlbi-devel To paraphrase this: the iBOB is old and basically unsupported hardware, lacking many of the features in the newer CASPER hardware. You will need very old versions of both Xilinx and Mathworks design software, which those companies don’t support, and it may be hard to find. And, perhaps most damning, you need a set of very old and unsupported CASPER libraries. Or to put it another way: All the stuff you need is deprecated. It is strongly not recommended to embark on a new design within this framework. And it is very hard, indeed virtually impossible in fact, for the CASPER community to support your efforts. In my opinion it may be possible for someone who is very independent and experienced with CASPER to get something done using the supplied model files as a starting point. But this would be very hard work compared to using the newer technology, my team and I would not attempt this, and you are apt to feel quite alone in this effort. Best wishes, Jonathan here is Rurik’s message from April with the referenced links to github. > On Apr 21, 2014, at 4:43 AM, Primiani, Rurik <rprimi...@cfa.harvard.edu> > wrote: > > Hi Rolando, > > I suspect Dan may have been referring to the PhRInGES system developed > for VLBI at the SMA and CARMA; it's a phased array with two iBOBs that > sample four inputs each, coherently sum them, and send the sum to a > third iBOB over XAUI links for the final sum (total of 8 inputs). > There is also an XF "calibration correlator" implemented on a BEE2 > board which receives data from the upstream iBOBs to derive the > calibrating delays and phases. > > The design is quite complex and uses multiple "black boxes" to speed > up compile times. The Simulink model files are currently hosted at > Github and I believe the latest version of the code is on Dave > MacMahon's branch located here: > > https://github.com/sma-wideband/phringes/tree/davidm/mssge > > Note: there is some documentation in the "doc" directory. There is > also a poster around somewhere I can dig up for you that describes the > system using a block diagram. > > One thing to keep in mind is that since the iBOB is no longer > supported you will need to use an older version of > Matlab/Xilinx/mlib-devel/etc. I believe the last version we could use > was Xilinx 10.x (might have been 11.x but I can't remember for sure at > the moment). Additionally, you will need to use the following version > of mlib-devel to compile the design: > > https://github.com/sma-wideband/mlib_devel/tree/vlbi-devel > > If you have any questions about the system or need more detail please > email myself, Jonathan Weintroub, and Dave MacMahon. You should be > able to find their emails on the list or I'll be happy to CC them in a > response. > > Best, > Rurik > > On Feb 23, 2015, at 4:37 PM, Rolando Paz <r....@crya.unam.mx> wrote: > > Hello Jonathan > > I am very interested in knowing how to use a XAUI link in two IBOBs. > > Are very interesting the attached documents. > > Do you still have the models files that you used? > > It would be very helpful for me now. > > Thanks in advance. > > Rolando Paz > CRYA-UNAM > > <IBOB to IBOB.png><casper_smbh.pdf> > Hi all. > > Yesterday I was playing a bit with an X-Engine model for the Ibob. > > I attached the model file that I want to use. > > My goal is to split the Pocket Correlator into two IBOBs > > I would appreciate if anyone has any comment > > F_ENGINE > > i_128w_4k_10_r311_cust.pdf > > https://casper.berkeley.edu/svn/trunk/projects/paper/2010_08-PSA32/ > > > X_ENGINE > > x_ibob_ver1.pdf > > Best Regards Rolando Paz
