Dayton G2

[wb4jfi at knology.net]

Apache Labs IS the company that took Hermes over from TAPR.  I believe Abhi, 
the person that did the final layout of Hermes may be involved.  You can 
check out the Hermes page on openHPSDR.org to find out more on it.  They 
also make a dual-ADC version.

While the HPSDR hardware uses a DDC-design with a high-speed A/D and a big 
FGPA, the bandwidth had always been limited to 192kHz.  This goes back to 
the original HPSDR boards (Ozy/Janus) that were essentially a fancy sound 
card, and how the signals on the motherboard (Atlas) were propagated.  When 
the multi-board DDC boardset came out (Mercury/Ozy), they kept the same 
signaling on the backplane.  That kept the inter-board signaling at the same 
slow speed.  Also, the computer interface board (Ozy), used USB, a newer 
computer interface board (Metis) came out that is ethernet based.

While Hermes is a single-board design, it is very similar to a Mercury & 
Metis combined.  The openHPSDR version of PowerSDR - in fact almost ALL 
versions of PowerSDR - only support up to a max of 192kHz.  All based on the 
original Flex PowerSDR, which was designed to support soundcard-based SDR 
hardware.

Recently, work has been done to speed up the signals between the FPGA and 
interface chip on both the Mercury/Metis and Hermes versions of HPSDR 
hardware.  There is also an effort to add wider bandwidth displays to 
PowerSDR in the HPSDR group.  They are doing at least 384k,and working on 
more.  There is also FPGA code to allow multiple receivers (slices) inside 
the FPGA, for multiple ham band display.  And other things.

The dual-A/D design is primarily for diversity reception.  Since many of the 
DDC designs can support multiple 192k bandwidth frequency swaths or slices, 
the dual-A/D design is not necessary for just doing this.  Dual-A/D allows 
multiple antenna inputs.

As an aside, much of the Ozy/Mercury/Janus/Atlas design was originally done 
by Phil Covington.  He separated from the HPSDR group, and sells his 
QuickSilver QS-1R receiver (and matching QS-1E exciter)(models may not be 
exact).  His design does NOT use the serial data path between FPGA and 
interface chip, since his was a one-board design.  The QS-1R can do up to at 
least FIVE slices at once.

I have an openHPSDR boardset and case using Mercury/Ozy/Atlas/Penelope.   I 
am acquiring the parts to build a Hermes.

The HPSDR designs are very nice overall.  There have been some bumps in the 
road, however.  I have been looking for a Metis for my boardset to upgrade 
from USB to ethernet.  OTher parts can be difficult to locate.  Apache Labs 
is a great solution to many of the problems.  Not only is there the PowerSDR 
software, but there's a CuSDR, which uses the main computers GPU for 
processing.  There's also KISS, advertised as a simpler SDR software 
package.  And, there's a version for Macs as well, Heterodyne?
73, Terry, WB4JFI


-----Original Message----- 
From: Mike ODELL
Sent: Monday, May 20, 2013 10:19 AM
To: Mike O'Dell
Cc: Tacos
Subject: Re: Dayton G2

hmm - their web site is confusing.
the info doesn't seem consistent, ESP
about the ADC and size of the FPGA.
it may be their "commercial" unit that has
the two channel ADC. it seems the ham version
is the Hermes design with a nice PA/filter card.
in one spot, the realtime bandwidth looks suspiciously
like there's a sound card in there somewhere, but
that makes absolutely no sense based on everything else.
so..... yet more dust to settle. encouraging dust, but dust
none the less.

sigh

     -mo

Sent from my iPad so please excuse the jammy fingers.

On May 20, 2013, at 9:44 AM, Mike O'Dell <mo at ccr.org> wrote:

> Dunno how many folks went to Dayton,
> but for those who did, perchance did you
> stop by the "Apache Labs Pvt Ltd" stand?
>
> they are selling some interesting SDR hardware.
> one of them is a 100 watt HF+6 SDR transceiver
> with twin-channel A/D (130 MSPS) front-end and a *big*
> FPGA. 14 receiver strips uses about 30% of the FPGA.
> it uses GigE to speak to the host so that interconnect is
> free of silliness. It also has a very thoughtfully designed
> filter section which can be used to exploit the dynamic
> range in the area of interest. And the board can be
> populated with a *second* independent front-end string.
>
> It's not cheap, but it is less expensive than the Flexen
> and a more modern design.
>
> it gets more entertaining every day!
>
>    -mo
>
> _______________________________________________
> Tacos mailing list
> Tacos at amrad.org
> https://amrad.org/mailman/listinfo/tacos
_______________________________________________
Tacos mailing list
Tacos at amrad.org
https://amrad.org/mailman/listinfo/tacos