[hpsdr] Hermes protocol 2 Firmware.

[Kjell Karlsen]

Hi Scott.

There is nothing wrong with the link you sent me. I answered Dave before I  
received your mail. And I have always imagined that I get ALL mails on the  
Apache labs forum but when I read through the list I see when I followed  
your link, I have not got any of the mails concerning the Hermes Protocol  
2. And if I try to answer or send a mail, all bounce back to me.

So I was not aware that there was a separate firmware developing team that  
was working with this. Anyway, all of the updates should be sent to the  
same place and not spread all over the net.

I have tested all versions of Hermes 16 bit firmware for protocol 2 and  
all works in RX mode. I can run 1,5 Mhz sampling with good results but in  
TX there are no sign of signals from the FPGA to the DAC. When I mentioned  
this for Phil, he was sure it is due to the FPGA speed in my Hermes boards.
I do not think I am the only one that has this problem but I have not seen  
anyone mention this.

Can you help me to be able to get the mails from Apache labs forum? I can  
not log into it using what I believe is my user name and password.

73, Kjell



På Tue, 18 Feb 2020 17:09:41 +0100, skrev Scott Traurig  
<scott.traurig at gmail.com>:

> Kjell,
>
> What was wrong with the link I sent you?
>
> https://apache-labs.com/community/viewforum.php?f=32   
> In this case no one can spoon feed you the "one true link". There are  
> many variants of the firmware that have been produced for each hardware  
> >platform. In your case you need to read the Hermes related thread at  
> that link, understand which versions have been successful and for whom,  
> and >then start experimenting with various versions (all posted in the  
> platform specific discussions) yourself.
>
> The problem is not that the FPGA is not fast enough, or big enough (size  
> is a problem on the 14-bit Hermes, though). The problem is that a  
> universal >timing closure that works for all serial numbers has not been  
> discovered. This problem is essentially limited to the MAC logic in the  
> FPGA, which is >the only part of the build that runs at speeds higher  
> than 125MHz. There is sufficient variability in the silicon, power  
> supplies, and perhaps even >board trace lengths, such that a build that  
> works for Person A with serial number X may not work for Person B with  
> serial number Y. Many people can >even observe temperature related  
> effects whereby the build will not be stable until the hardware warms  
> up, or vice versa.
>
> There just hasn't been sufficient hardware engineering on the design at  
> GigE speeds, and using the free version of Quartus is also a significant  
> >handicap. All honor to Rick, N1GP, who took us out of a nasty Alpha  
> test stage and has gotten us into a pretty good Beta state, But it's  
> still not easy >for everyone to run Protocol 2.
>
> 73,
>
> Scott/w-u-2-o
>
> On Tue, Feb 18, 2020 at 4:43 AM Kjell Karlsen <la2ni at online.no> wrote:
>>
>> Hi Dave.
>>
>> That is were I have been looking but the version 10.8 is not there. It  
>> was mentioned on the Apache labs forum but I am not able to find it.
>>
>> I have problems with at least two of the Hermes boards that we bought  
>> from TAPR (or built ourself from the PCB´s we got). They works OK in RX  
>> >>but nothing out in TX. I asked Phil about this and he say that it is  
>> the FPGA that is not fast enough.
>> I had hoped that someone comes up with firmware for Hermes that solves  
>> this problem (if it is possible at all).
>>
>> Thanks,73, Kjell



-- 
Sendt med Operas e-postklient: http://www.opera.com/mail/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20200218/6ebcb76a/attachment.html>