[hpsdr] Hermes - initial powerup

[Joe Martin K5SO]

Erik, 

The way the communication protocol works is that the PC program specifies to Metis/Hermes or Ozy/Magister how many receiver IQ streams should be interleaved for inclusion into the ethernet (Metis or Hermes) or USB (Ozy or Magister) packets that are sent to the PC.  The PC program specifies the number of receivers to implement, via three bits in the Command and Control bytes sent by the PC to Metis/Hermes or Ozy/Magister; namely, the three bits C4(5), C4(4), and C4(3) when C0 is zero.  That's all there is;  no firmware lists are used for this nor are such lists needed. 

 However, for this protocol to function (with multiple receivers) it is necessary that the firmware for Metis/Hermes and Ozy/Magister be coded to recognize those three C&C bits and then interleave the IQ streams according to the three-bit value (000 = 1 receiver, 001 = two receivers, etc).  Metis v1.6 and Ozy/Magister v2.1 firmware do that already but Hermes v1.3 firmware does not.  The next version of Hermes firmware will have that functionality implemented; then, along with several other changes to the Hermes code, Hermes will be able to implement multiple receivers in its FPGA, just as a single Mercury board is able to do that now.  

73,  Joe K5SO

On Mar 29, 2012, at 7:01 PM, Erik Anderson wrote:

> It does actually, thank you.  It sounds like the only way for receiver software to know that something like this is happening is to have a list of firmwares and its capabilities somewhere (I'm guessing the single-receiver version branch is running its own version numbers or something, think I remember reading that on the list).  Part of me is thinking that it might be nice if there was (yet another) CC0 code listing the number of receivers in the datastream, although this would likely be mooted if the only receivers that supported it were multi-receiver.
> 
> 73, Erik km2g
> 
> On Thu, Mar 29, 2012 at 5:08 PM, Joe Martin K5SO <k5so at valornet.com> wrote:
> ***** High Performance Software Defined Radio Discussion List *****
> 
> Hi Erik,
> 
> I didn't notice a response to your questions so I'll give them a go.
> 
> As I recall, if multiple receivers are requested from Hermes v1.3 by a PC program Hermes will continue to output a data stream that IQ values in the stream that is appropriate for single receiver operation but the PC will be expecting "interleaved" IQ values from multiple receivers.  This will usually cause "distorted" noise to be heard in the audio output channels and/or a large central spike in the panadapter display with no discernable signals present in the display, unless of course the PC program does some kind of "firmware version checking" in which case the PC program will usually halt execution of the PC program as the expected firmware version number will not be received by the PC.
> 
> No interleaved data streams of entirely zeroes will be produced by Hermes v1.3 when a PC program asks for multiple receivers.
> 
> A multiple-receiver-compatible version of firmware for Hermes will be produced and released to replace Hermes v1.3 as soon as possible, I'm sure.
> 
> I hope that answers your questions satisfactorily.
> 
> 73,  Joe K5SO
> 
> 
> On Mar 29, 2012, at 2:32 PM, Erik Anderson wrote:
> 
> > ***** High Performance Software Defined Radio Discussion List *****
> >
> > Any particular behavior to watch out for if multiple receivers are requested?  do they end up in the datastream as zeroes or does it just assume you requested only one receiver?
> >

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20120329/47e3eb9e/attachment-0004.htm>