On 3/25/2018 1:25 PM, M. George wrote:
Joe, your 2010 data (which you haven't looked up and provided, but I
assume exists? from 8 years ago) isn't a good comparison to changes
at the OS, CPU / hardware level between 2010 and 2018. You are very
good at telling everyone the way things are, but not very good at
showing folks the way things are / work with actual experiments and
data / video.
1) I don't need to "look up" that data ... it exists in the RTTY
archives for those who haven't already seen it *and* participated in
discussions of the issues with jitter when using software generated
FSK.
2) The issue of jitter is just as important today as it was eight years
ago as it causes the same loss of signal to noise ratio as Chen takes
very great care to explain to you on the RTTY list. Even today the
number/percentage of amateurs using liquid cooled, hex core 3 GHz i7
processors like you used for your first "demonstration" is exceedingly
small. As I have told you multiple times, based on my support work
the average amateur system is something like an 2.4 GHz Core2Duo with
1 - 4 GB of RAM and typically a single USB Root Hub to serve CAT, CW,
FSK, digital sound card, *and* software panadapter.
*NONE* of your demonstrations showed that level of system under *FULL*
load. Your first demonstration may have been running rig control and
software panadapter but it wasn't processing a cluster feed at contest
rates (your panadapter was clearly visible with only one or two signals
on the band) and your CPU did not exceed roughly 40% utilization. Your
second demonstration did not include rig control, panadapter or cluster
yet by, your own measurements, had more than 10% *per bit* jitter which
is enough according to Chen to reduce SNR by several dB.
In the 2017 CQWWRTTY contest (using a 5 year old Ivy Bridge i7 Intel
cpu) with my K3S / internal FSK keyed by the same FTDI serial
interface in my video presented this last week, I'm shocked I was
able to work 1285 QSO's and 184 band countries.
You're far enough out on the advanced end of the CPU curve that you can
get away with software FSK and work plenty of strong stations. JA7UDE
made jitter measurements <http://www.qsl.net/ja7ude/extfsk/jitter.html>
with EXTFSK64 using an Ivy Bridge i7-3770 which showed 75 baud jitter
at +/-40 uSec in a lightly loaded system, increasing to -67/+340 uSec
when the USB system was loaded by data transfer. Oba also reported
EXTFSK on that same lightly loaded system at +/- 50 uSec (similar to
your lightly loaded liquid cooled system).
Again, jitter is not a significant issue isn't with software FSK on high
spec CPUs with light CPU loading and dedicated USB Root Hubs ... it is
the +/- 4 msec jitter that is typical on slower, dual core systems, a
single USB Root hub and heavy processor/USB loading.
You do a disservice by insisting that jitter is not a problem because
*it doesn't reach a critical level* on your high end computing
platforms.
73,
... Joe, W4TV
On 3/25/2018 1:25 PM, M. George wrote:
Joe, your 2010 data (which you haven't looked up and provided, but I assume
exists? from 8 years ago) isn't a good comparison to changes at the OS, CPU
/ hardware level between 2010 and 2018. You are very good at telling
everyone the way things are, but not very good at showing folks the way
things are / work with actual experiments and data / video.
This exercise isn't productive until you actually watch all of my content
and start reproducing my results on a scope etc... If you are not willing
to do that in 2018 with your own equipment and continue to rely on
microHam's numbers from 2010... (or someone else's data) There is no point
to the discussion. A lot has changed since 2010. Plus I made it crystal
clear in my video that my main computer setup is much faster than most
other Ham shack computers at the moment (2018). You act like I expected
everyone to have the same setup... which clearly shows that you cherry
picked sections out of the video to confirm your preconceived conclusions
before every clicking on the video links. You insist on leaving out all
the detail provided in the Video, because it can't be conveyed via a few
paragraphs of text in an email.
As usual, (which is well known), you will get the last word in... Until you
produce something new for 2018, a response from myself isn't productive at
all. And will not take place. For now I'll let my video speak to the
experiment and 2018 current state of what I recorded. As I do other tests
and present other findings, I will be willing to change my
understanding/conjecture/opinion based on data collected and presented.
In the 2017 CQWWRTTY contest (using a 5 year old Ivy Bridge i7 Intel cpu)
with my K3S / internal FSK keyed by the same FTDI serial interface in my
video presented this last week, I'm shocked I was able to work 1285 QSO's
and 184 band countries. My FSK generated signal on the other end of the
QSO's must have been horrible with this setup (jitter all over the place
I'm sure).
Band QSOs Pts ZN Cty SP Pt/Q
3.5 62 74 8 6 30 1.2
7 338 599 27 58 50 1.8
14 689 1519 31 84 50 2.2
21 194 370 23 34 37 1.9
28 2 4 2 2 1 2.0
Total 1285 2566 91 184 168 2.0
Score : 1,136,738
I look forward to your updated data and visuals / video using data gathered
from 2018. That will be very interesting indeed.
73 de Max NG7M
W4TV:
​Unfortunately, your first video is completely unrealistic as the vast
majority of amateurs uses computers significantly less powerful than
your lightly loaded (less than 40% CPU utilization by your own video)
six core, 3+ GHz CPU with the EXTFSK port on a dedicated motherboard
USB port with no loading from multiple (high priority) sound cards
(they are on a different USB Root Hub) and no contest level cluster
spots.
From my customer support support experience, the typical amateur station
is a 2-2.4 GHz Core2Duo (two cores, 4 execution units) with 1 GB RAM and
all USB ports (typically 4) served by a single USB Root Hub. The system
typically runs a logging program that polls one or two transceivers for
eight parameters every 50 to 100 msec along with one or two 96 or 192
KHz sample rate USB sound cards for a software panadapter (or
equivalent USB I/Q SDR receivers) and another 16 bit, 48 KHz sample
rate USB sound card for digital operation. In addition, those systems
are connected to a DXCluster node with CW/RTTY Skimmer providing a
net spot flow of > 100 spots per minute.
I've provided jitter data as measured by microHAM in 2010 multiple times
on the RTTY list ... and that data has been verified by JA7UDE (author
of EXTFSK an ETFSK64) who also confirms the added jitter when the USB
system (single USB Root Hub) is loaded with heavy data transfer. Oba's
results have also been reported on the RTTY list and are available on
his EXTFSK64 page.
73,
​
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[email protected]
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [email protected]
