Re: Sudden K3 AGC pop-off after adding P3

Posted by Joe Subich, W4TV-4 on Oct 08, 2011; 8:15pm
URL: http://elecraft.85.s1.nabble.com/Sudden-K3-AGC-pop-off-after-adding-P3-tp6872073p6873188.html


 > Why should polling the AGC randomly turn it off on the K3?

Because HRD uses the *K22* mode.  If HRD would use the K3 commands
and not enable the K2 extended command set, the GT command would
require an explicit off command.

What appears to happen is a data clash between HRD and the P3 ...
P3 data or an echo gets appended to the the GT; poll to the K3.
The the corrupted data is "just right" AGC is turned off.

Why does HRD use an old command set and why does it poll for AGC
status *THREE TIMES*?  Polling for the same data three times in
31 milliseconds is stupid, sloppy, and increases the chance of
false triggering exponentially.

73,

    ... Joe, W4TV


On 10/8/2011 2:54 PM, Alan Sewell N5NA wrote:

> I'm not a programmer nor do I play one on TV.  I'm just an observer.
>
> Why should polling the AGC randomly turn it off on the K3?  None of the
> other parameters that are polled at the same frequency are adversely
> affected, only the AGC.
>
> AGC should be able to be polled without it turning off.
>
> Alan
>
>
> On 10/8/2011 12:29 PM, Guy Olinger K2AV wrote:
>> The real deal would be why the over-polling is necessary in the first
>> place.  I'd guess it's done to service ALL the different rig types
>> (not just Elecraft and K3) with a single strip of code, instead of
>> pushing those constants at program startup based on the target or
>> configured hardware.
>>
>> If it's not cheep code, then it's certainly not state of the art, nor
>> at all close.  One always should have a constants library which
>> carries the array of values necessary to drive any hardware timing
>> sensitive code.
>>
>> Where I worked at my last TWO places of employment, going back to an
>> AT&T school in 1976 (!), embedding target-variable constants in
>> program code and using code switches would get your code summarily
>> rejected by a reviewer or auditor, subject one to an auditor approved
>> redesign and recode (with prior effort flushed down the turlet), and
>> have an effect on one's pay grade. There was a brutal "standards are
>> just that" attitude.  Putting in the table driven code costs some in
>> the beginning, but saves huge in maintenance, and a maze of switches
>> eventually means recoding in its entirety.
>>
>> That's probably what we're really fighting with the HRD code.
>> Otherwise it's a simple table change that gets done in seconds, and
>> then recompiled.
>>
>> 73, Guy.
>>
>> On Sat, Oct 8, 2011 at 12:31 PM, Alan Sewell N5NA<[hidden email]
>> <mailto:[hidden email]>>  wrote:
>>
>>      This is a problem not only for HRD but any software in which the AGC
>>      status is polled continuously.  I frequently use CQ/X by NO5W and had
>>      the same problem.  Chuck turned off the AGC polling in CQ/X and the
>>      problem no longer occurs.  Of course, now the AGC status can't be
>>      shown
>>      in his remote interface.
>>
>>      I sent Wayne, N6KR, a couple of log files from portmon back in July
>>      showing what was happening but so far I haven't seen any updates
>>      on this.
>>
>>      73,  Alan  N5NA
>>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html