Bicephale
join:2005-09-24
 ·TekSavvy Solutions..
| RF Choking (on the WAN side)
Hi DSL customers,
Below you can see a Thomson SpeedTouch 546v6 MoDem/Router
with RF Choking used both on the power lead and its WAN input:
The customer's own wiring, Bicephale, 2008-Jan-14
This was done before i moved to another city. Here are some very
obvious results which i could gather about it, for your convenience:
Tweaks, Bicephale, 2008-Jan-14
The expression "With Choke" (in yellow) means there was a pair of
RF Chokes on the WAN side:
• Two Turns wound on 1" dia. Ferrite Toroïdal core (yellow)
• Ten Turns wound on 1.5" dia. Ferrite Toroïdal core (green)
Those Ferrite Toroïdal cores were refurbished from old PC-XT power
supplies. Such ancient devices had a tendency to use larger cores,
making them better suited for winding phone-line cords. Anyway, it
seems a trade-off was implemented on this "Interleaved" line: a part
of my DS SNR margin was converted into "quietness" (less noise)!...
I moved to Limoilou (Quebec city) a year ago and started to refine
my measurement method. Instead of depending on a single counter
value for my comparisons i adopted long-term (24 h) curves built by
processing five thousand seven hundred sixty samples recorded at a
rate of four per minute (e.g. resolution was fifteen seconds). These
sets of daily curves shown below illustrate how much the Error Rate
(aka "noise") will vary from one day to the next. Such wild variations
make it somewhat difficult to compare sets of curves but it's easy
enough to notice that the rightmost column is significantly "cleaner":
%20.GIF)
From the ground up!, Bicephale, 2009-Jul-17
Thanks to another technique i tagged as "averaging" it's now even
easier to make comparisons for anyone who's not visually impared:
%20.GIF)
From the ground up!, Bicephale, 2009-Jul-17
The conclusion seems evident, DS SNR/Rx CRC variations have been
reduced by using more RF Choking: the DS SNR level improved while
the "noise" level went down... No fancy theories involved, all it takes
is a lot of patience, some work and a bit of common sense!... Oh, by
the way... The Thomson ST5x6 owners should keep in mind that my
daily averaging trick has been included in TerTech's 'DSL_StatScope'
3rd-party trouble-shooting tool - it's unique, powerful and it's free!!!
There's no need to know what a SpreadSheet is if you don't want to
so, what are you waiting for, you people?! Please just don't take my
word for it!!! Simply try to conduct your own experiments and come
back here to share your results! You might find it to be constructive
while having some fun in the process!...
Follow the links provided in this post if you can't find old ferrite cores:
»Re: lousy upstream SN margin
 |
|
TilhasBB
Formally Goden99
Premium
join:2000-08-05
canada | Great post Bicephale! |
|
dbsanfte
join:2005-03-15
Montreal, QC
·Colbanet
| reply to Bicephale
Toroidal cores will cause higher frequencies to attenuate, so the DSL modem loads lower frequencies more highly to compensate. You end up with a lower SNRM because the lower frequencies are loaded that much closer to their max.
If your noise problems are on the high frequency end, then forcing the modem to use lower frequencies, as you've done, would solve them. The modem wouldn't be using that part of the spectrum anymore.
Otherwise I see no point to shitting the higher frequencies of your DSL line, as you have.
Why not just find out what frequencies your intermittent noise is at, and make an RF filter inline with the phone cable to filter it? |
|
Bicephale
join:2005-09-24 | Hi dbsanfte,
Those are great ideas waiting to be explored and validated!!! I'm
looking forward for your future posts, show us what you can do!
 |
|
Bicephale
join:2005-09-24 | reply to TilhasBB
Thank you very much TilHasBB for the "Thumbs Up" rating!
 |
|
TilhasBB
Formally Goden99
Premium
join:2000-08-05
canada
| I see what you mean dbsanfte.
Take a look Bicephale. Your higher frequencies (orange circle) with the choke are gone but the lower frequencies (red box) had a nice gain.
The SNR reading increasing from 7 to 16dB is just how the speedtouch interprets the frequencies your using. |
|
Bicephale
join:2005-09-24
·TekSavvy Solutions..
| Hi,
I wish i had got a record of Bell's portal numbers including
RCOs but i don't. All i have is this pair of 'DMT' captures:
%20.PNG/thumb.jpg)
Tweaks, Bicephale, 2008-Jan-14
%20.PNG/thumb.jpg)
Tweaks, Bicephale, 2008-Jan-14
On top we have the "With Choke" graph and below it's its
counterpart version.
I've requested comments from Old Computer as he did use
RF Choking. If i've understood him correctly, it's possible RF
Choking prevents local noise sources from "desensitizing" a
DSL device. Why local? My guess is it's because for a same
level of noise being measured at the unit's location a noise
source would need to be abnormally strong if it's not close:
OrbMT on D-Link DSL 2320B, Bicephale, 2009-Feb-24
It would be most interesting if someone with a spectrometer
could use a loosely coupled probe like this to monitor noise:
ST516 - Strange discovery, Bicephale, 2008-Nov-16
Now imagine doing comparison experiments with RF Choking...
Unfortunately, i can't be that someone. |
|
TilhasBB
Formally Goden99
Premium
join:2000-08-05
canada
1 edit | Again the same thing. I think the benefits of doing this is more because your using the lower part of the frequencies (under 500kHz) which is less susceptible to noise.
Hummm... I'd be interesting in seeing one thing thou.
Can you try with and without RF Choking on the phone.
Then Only RF Choking on the power and off the power |
|
Bicephale
join:2005-09-24
·TekSavvy Solutions..
| Hummm...
If i had got access to a controlled environment then it would
have been possible to consider shorter samples but this is the
summer with it's lightning storms, ever lasting humidity, high
winds, etc. What you suggested amounts to an array of four
tests so if each of them must last for six days because of the
variations which i've observed then i would't be done with that
before the 3rd week of August or so, including any unforeseen
mishaps that never fail to occur... As far as i know, the lower
error rate is incentive enough to use RF Choking and i'm quite
satisfied to know it even if everything isn't under control yet.
I'd be tempted to examine the effects of RF Choking in terms
of the numbers of turns but there are other DSL devices i wish
to compare to my Siemens now that i got a comparison ground.
 |
|
Bicephale
join:2005-09-24
1 edit | reply to TilhasBB
About using the lower spectrum, i bet it would be nice to read
TakeTheFifth about it. As you know, the TP-Link TD-8841 (A)
allows its owner to deallocate bins and i read recently that this
might have a similar effect than RF Choking... |
|
planiwa
join:2009-02-19
Toronto M5S
1 edit | And perhaps without the negative (side-)effects of RF choking on the high-end RF-signal being choked. 
(RF-choking of RF-signals to combat RF-noise seems like radiation therapy trying to kill the disease faster than the host. )
--
"Cool Locally -- Warm Globally." |
|
TilhasBB
Formally Goden99
Premium
join:2000-08-05
canada
| reply to Bicephale
Can you resync get your line stats.
Then put a normal phone cable (no RF chokes) just on the power. and resync.
I have a feeling the ST516 Power is whats causing a few of the gaps. If RF choking cleans them up, it would be a no brainer to do one very modem |
|
Bicephale
join:2005-09-24
·TekSavvy Solutions..
1 edit | 
GNet BB0060B vs 'DMT v2.20' capture (2009-Jul-20) |
I guess you may not have noticed this but the ST546v6 records
made using 'DMT v7.35' are over one year and a half old as it
was done during the 2007-2008 winter. My phone-line has some
different characteristics now because i've moved to another city
one year ago (in July 2008). My measuring method has changed
as well since i no longer trust single 'DMT' readings, even if there
was only one "reset" event as displayed on additional captures.
This month's records show that the Error Rate counter doesn't
represent the situation accurately at all times. For example, the
evening of July 10 was real "quiet" while the early morning of
July 9 proved to be extremely "noisy". The same is true of my DS
SNR Margin level which varied somewhere between 4 and 7.5 dB.
Consequently, it seemed to me that one reasonable solution was
to use six-days "averaged" curves for my late comparison. It's a
time-consuming process but i no longer have to worry because
of marginal events which might corrupt my captures otherwise.
If i follow that same logic, the 'DMT' spectral response shown
in a capture will vary as a function of the date and time when
it was taken. This could give rise to a stimulating project but
the SpreadSheet averaging work involved can be overwhelming.
I'm inclosing a 'DMT' capture of my un-tweaked GNet BB0060B,
i put it on-line around 20:30 last night and looking at it you'll
see what i mean by "variations": it's like i've moved elsewhere!
I'm now trying to imagine a way to shield my damned drop-line,
actually: it seems it keeps degrading while i keep cleaning it up!
In any case, to answer your questions, there's an RF Choke on
every transformer used by my MoDems already. There's a strong
AM station around here (CHRC 800 KHz), i'll try to find out if RF
Choking as a significant effect on the corresponding bins...
...stay tuned...
|
|
TilhasBB
Formally Goden99
Premium
join:2000-08-05
canada
1 edit | If your really into noise.
Make sure you have None of these fluorescent light bulbs.
»steadfastfinances.com/blog/wp-co···bulb.jpg
LOL, they are the worst enemy.
»www.gelighting.com/na/home_light···pact.htm
Can compact fluorescent bulbs create interference with electronic equipment, such as radios?
Many electronic devices, such as radios, televisions, wireless telephones, and remote controls, use infrared light to transmit signals. Infrequently, these types of electronic devices accidentally interpret the infrared light coming from a compact fluorescent bulb as a signal, causing the electronic device to temporarily malfunction or stop working. (For example, your television might suddenly change channels.) Fortunately, this only happens when light is produced at the same wavelength as the electronic device signals, which is rare.
To reduce the chance of interference, avoid placing compact fluorescent bulbs near these kinds of electronic devices. If interference occurs, move the bulb away from the electronic device, or plug either the light fixture or the electronic device into a different outlet.
check this out
»brneurosci.org/fluorescent-lamp-···nce.html
"This RFI appears at 30-40 kHz intervals up to about 500 kHz. "
Ooo wait thats where you have the best frequencies for your DSL..lol |
|
Bicephale
join:2005-09-24
·TekSavvy Solutions..
| Hi again,
Error Rates aka "noise" have been the focus point of my efforts
for quite a while now... Here's the 1st piece of evidence which
confirmed my doubts about a dimmer located downstairs: this is
the instant when i've predicted & captured what was comming:
%20.GIF)
From the ground up!, Bicephale, 2008-Sep-21
All my dimmers are from Cooper Wiring Devices (Catalog #6441)
as i get no problem with these. In the begining, i thought they
were auto-transformers but it's not the case: their decoupling
is simply done properly, i guess. Perhaps fluorescent lamps do
include bad apples indeed, i'm happy not to have experienced it
1st hand. In fact, i welcome the next generation of LED lamps...
|
|
Bicephale
join:2005-09-24
·TekSavvy Solutions..
| reply to TilhasBB
 .PNG/thumb.jpg "43664 bytes")
Siemens SS4200 - 15 Turns (2009-Jul-22 2h09) |  .PNG/thumb.jpg "43733 bytes")
Siemens SS4200 - No Choke (2009-Jul-22 1h33) |  .PNG/thumb.jpg "47115 bytes")
Thomson ST516 Lin-PSU - 15 T (2009-Jul-22 1h45) |  .PNG/thumb.jpg "46946 bytes")
Thomson ST516 Lin-PSU - No Choke (2009-Jul-22 1h06) |  .PNG/thumb.jpg "46012 bytes")
Thomson ST516 Sw-PSU - 15 T (2009-Jul-22 1h52) |  .PNG/thumb.jpg "45989 bytes")
Thomson ST516 Sw-PSU - No Choke (2009-Jul-22 1h14) |  .PNG/thumb.jpg "47400 bytes")
Thomson ST546 Lin-PSU - 15 T (2009-Jul-22 0h45) |  .PNG/thumb.jpg "47442 bytes")
Thomson ST546 Lin-PSU - No Choke (2009-Jul-22 0h54) |  .PNG/thumb.jpg "43974 bytes")
TP-Link TD-8841 - 15 Turns (2009-Jul-22 2h02) |  .PNG/thumb.jpg "44091 bytes")
TP-Link TD-8841 - No Choke (2009-Jul-22 1h26) |
It seemed tonight my phone-line was going to remain quiet so
here are 'DMT' captures with and without RF Choking. All of my
MoDems have RF Chokes on their power leads. I'm using some
five inches of wiring between the RJ-11 connector box and my
POTS Splitter, the WAN-side feed cable is about the same (if
you exclude what's wound around the toroïdal core, when such
a core is being used, of course!)...
I couldn't generate captures with zero settle time so i got that
roughly equal to five minutes for all of the samples. In my opinion
that's simply insufficient: personally i'd have prefered at least a
few hours but then the test would have required too much time!
|
|
TilhasBB
Formally Goden99
Premium
join:2000-08-05
canada
| Great work !
I am not seeing the choke do much difference as much as the first pictures. Your right it might only make a difference with errors in the long run which we won't see.
But your line is no longer as bad as it was before. I don't think the chokes are even necessary. Your line is pretty good. |
|
Bicephale
join:2005-09-24
·TekSavvy Solutions..
1 edit | Hi TilHasBB,
The problem is it takes time for noise to knock down a few bins,
fresh captures like these may be useful to compare the devices
but it doesn't tell much about RF Choking... Most unfortunately,
i'm afraid the situation has got worst since i began to work on
my phone-line, last year, as i no longer see "flat-liners" like this:
%20.GIF/thumb.jpg)
From the ground up!, Bicephale, 2008-Jul-31
I wish those days had lasted forever, here's what i get instead:
From the ground up!, Bicephale, 2009-Jul-22
That's why i've reached a stage where shielding is an option
i'm now seriously considering... Other than that, i'd need the
drop-line to be changed for a twisted-pair!!!
In the meantime, i don't believe RF Choking has no effect, as
i've demonstrated previously:
|
|
%20.PNG)
%20.PNG)
%20.GIF)
