volt/amp readings using multimeter on ADSL line

@OMN: DMT seems to now be running very well on your router!

I just posted my bits per tone plot above. I'm not exactly sure why there are no bits in the high frequency part of the spectrum. It would be nice to see some plots for a range of both good and bad Telkom lines. In your case your bits per tone and SNR per tone plots seem to vary according to the time of day! Is it possible to get one more plot of the situation when your line is running smoothly - like the one on page 7 of this thread with the bits per tone included? When the interference increases it seems that the baseline for the plots moves upward and only the bins which normally have large SNRs are used. On some of the plots it seems that the pilot tone may even be affected.
 
Last edited:
yip
those bars on the right hand side of the dmt graph...surely its something!!!

clearly visible on the LAST picture - the # of crc errors sampled in 15-30-45min has decreased and the bars/spikes/bits on the 1st graph on the righthand part is away - look at all 3 pics to get a clear picture

http://i210.photobucket.com/albums/bb147/jannie40/dmt20071005_0910.png

http://i210.photobucket.com/albums/bb147/jannie40/dmt20071005_1018.png

http://i210.photobucket.com/albums/bb147/jannie40/dmt20071005_1041.png

so it now appears that the area above 724kHz shows the health changes the best.
 
Well you are getting quite an interesting set of plots. So far your plot on page 7 shows your line performing at its best (RCO ~ 4.6Mbps). Look at that SNR plot quite carefully. You can see that you have two mountains - one around 724kHz. It seems that as the noise increases only those bins that are high up on the mountain can still be used. However, your first plot this morning (RCO ~ 3.9Mbps) does not show any indication of the second mountain around 724kHz which is surprising. In your later plots you get little pieces from the second mountain here and there. Regarding the Tx power - when you look at plots from overseas the Tx powers are usually very high but I think that this is because their lines are running at very high speeds - 8Mbps+. If you are connecting at Telkom type speeds <1Mbps then the Tx power should be lower unless, as you have described, the noise increases and your router is struggling to communicate. This is why I think there is that large increase in Tx power in your second plot. In general your line seems to be worse than it was a short while ago but this may be due to rain? Try to see if things improve this weekend and get some plots if they do. Hopefully somebody else with a 4Mbps line and a DG834GT router may be able to post some plots for you so that you can compare. In your first plot DMT estimates your loop length at between 4-6.5km - maybe you can also keep an eye on this figure to see if this varies when the line deteriorates. If your line always behaved like it did when you did the plot on page 7 then I don't think that you would be complaining. It's not perfect but that would be quite acceptable performance. It would be really interesting to know what is causing the performace to deteriorate so badly.
 
Last edited:
@Y
Sorry only for the late reply but I spent the day at the saltmines with a lousy uncapped 512 IS line. Back home now on ADSL/SAIX 4Mb and its a pleasure.
I think the newer Zyxel modems use Texas Instruments AR7 chipset. Try Whirlpool H/w forums (http://whirlpool.net.au/index.cfm?a=h_models) for the list. There is also a mine of info regarding DMT Tool with Netgear etc (http://forums.whirlpool.net.au/forum-threads.cfm?f=82)
Regarding OMN plots and the change, I think it may be related to weather and physical connection. I have had a problem with the Telkom line and the voice is pretty crappy (filters, phones etc have all been checked out). But my ADSL line is rock solid at 4096/384 with no sync errors However the bits/tone plot from Zymon has changed since the rain as I suspect a bad connection somewhere. My projected speed has dropped from 7300 to 6500 i.e increase in length from exchange I am still trying to find a bits/tone plot I saved before the rain came. I will see if there is a correlation.
Regarding DMT Tool software, I have a sis in the UK and I am thinking about getting here to send me a Speedtouch modem (approx 30 quid). This seems to work pretty well with DMT Tool and syncs well on less than perfect lines.
 
can you guys please try to get an SNR plot for me - say maybe a 2-3 hour sample?

or anyone for that matter - just include your sync rates
 
Last edited:
@OMN
This is my SNR down) margin for a few hours - synching at 4096 up and 384 down. The drop at the end is because I changed the variables.
http://i217.photobucket.com/albums/cc289/tjandies/ADSL/tj-snr.png
This is the Line attenuation
http://i217.photobucket.com/albums/cc289/tjandies/ADSL/Tj_LineAtt.png

This is output from OrbMT before the rain.
http://i217.photobucket.com/albums/cc289/tjandies/ADSL/20071004_014815.png
This is after the rain.
http://i217.photobucket.com/albums/cc289/tjandies/ADSL/20071005_061232.png

@Y
Thanks for links. I saw a few 585s on eBay. Trying to get them for about 20 quid including postage. They are going for abount 20 quid at DSL Depot. Just have to find out if they not netwrok locked for BT or AOL. I am a flash would cure that.
 
i would give my left nut for those kind of flat SNR values....dammit :)

check this:

reply from saix specialist.

If I run LTS a few times the metallic noise reading fluctuates as does the AC reading on the A leg. Sometimes it is perfect at 9999 and then drops right down as in below to 2125. As "X" says it is probably a dry joint on the A -leg.

also last email received:

The SI has been logged. Conformance testing to be done on your line. I will keep you posted.

It has been suggested to run a new copper pair from the exchange to your premises, but it seems they want to test the existing pair first.

ok weather - please hold for 1 week :)
 
Last edited:
@tjandies: Your plots look very nice, however, it appears that you may have a small crosstalk/interference problem on your line. Both of your plots are a little ragged, particularly the one that you did after the rain. A good line is supposed to produce a nice smooth plot. Compare your plots with the ones shown in the link below and see what you think. Also see the related links on page 5 of this thread regarding the final solution. This American chap battled for several years to try to resolve his particular crosstalk/interference problem - finally after much frustration he managed to get the phone company to switch his F2 pair to a better quality one.

http://www.dslreports.com/forum/remark,13641311

More information regarding ADSL crosstalk including some spectrum analyzer plots.
http://adslm.dohrenburg.org/

It would be very interesting to see what kinds of plots other Telkom 4Mbit customers are getting so that you have somthing to compare against. If you can actually identify an audible fault on the telephone it may be a good idea to log a fault with Telkom as things could deteriorate in the future to a point where your 4Mbit service degrades. If there is a definite audible fault then Telkom is usually willing to investigate further. If the fault is more elusive then you may struggle to get them to repair it. One advantage that you have now is that you should be able to quickly check your line health after any repair using the bits per tone plot from OrbMT.

As you mentioned, the problem could be caused by a bad joint which is affecting the balance of the line and making it susceptible to noise problems. Alternatively they might have to change the noisy line pair to a better quality one. The link below describes the most common types of ADSL line faults.

http://www.teracomm.com/appnotes/whitepapers/Copper Testing for ADSL & ADSL2+.pdf

You could also try some of the differential and common mode interference fixes on pages 10 and 6 of this thread. It shouldn't cost more than R20 to buy suitable capacitors and inductors. Let me know if you want more details.

@OMN: Any chance you can determine what kind of DSLAM you are connected to with your DG834G?
 
AHEM: check this - http://i210.photobucket.com/albums/bb147/jannie40/dmt20071006_1329.png

@Yotch - Saix specialist said Marconi DSLAM

ok my 834G router is at a customer currently - 4meg line - 1.3km from exchange - same town - here is the telnet result:


BusyBox v0.61.pre (2007.04.16-10:12+0000) Built-in shell (ash)
Enter 'help' for a list of built-in commands.

# cat /proc/avalanche/avsar_modem_stats

AR7 DSL Modem Statistics:
--------------------------------
[DSL Modem Stats]
US Connection Rate: 384 DS Connection Rate: 4096
DS Line Attenuation: 34 DS Margin: 13
US Line Attenuation: 16 US Margin: 13
US Payload : 2316404240 DS Payload: 10769498
40
US Superframe Cnt : 14597850 DS Superframe Cnt: 14597850
US Transmit Power : 0 DS Transmit Power: 0
LOS errors: 0 SEF errors: 0
Frame mode: 3 Max Frame mode: 0
Trained Path: 0 US Peak Cell Rate: 905
Trained Mode: 3 Selected Mode: 3
ATUC Vendor Code: 54535443 ATUC Revision: 1
Hybrid Selected: 4 Trellis: 1
Showtime Count: 1 DS Max Attainable Bit Rate: 11456 kbps
BitSwap: 1 US Max Attainable Bit Rate: n/a
Annex: AnxA psd_mask_qualifier: 0x0000
ATUC ghsVid: b5 00 54 53 54 43 00 00
T1413Vid: 00 00 T1413Rev: 00 VendorRev: 00
ATUR ghsVid: b5 00 54 53 54 43 00 00
T1413Vid: 00 00 T1413Rev: 00 VendorRev: 00

[Upstream (TX) Interleave path]
CRC: 0 FEC: 0 NCD: 1
LCD: 0 HEC: 0

[Downstream (RX) Interleave path]
CRC: 0 FEC: 0 NCD: 0
LCD: 0 HEC: 0

[Upstream (TX) Fast path]
CRC: 11 FEC: 136 NCD: 0
LCD: 0 HEC: 68

[Downstream (RX) Fast path]
CRC: 2159 FEC: 0 NCD: 0
LCD: 0 HEC: 0

[ATM Stats]
[Upstream/TX]
Good Cell Cnt: 137736907
Idle Cell Cnt: 87014896


[Downstream/RX)]
Good Cell Cnt: 290871911
Idle Cell Cnt: 2106478908
Bad Hec Cell Cnt: 1627
Overflow Dropped Cell Cnt: 0

[SAR AAL5 Stats]
Tx PDU's: 15580358
Rx PDU's: 15527652
Tx Total Bytes: 1920857611
Rx Total Bytes: 740483414
Tx Total Error Counts: 0
Rx Total Error Counts: 222


[OAM Stats]
Near End F5 Loop Back Count: 0
Near End F4 Loop Back Count: 0
Far End F5 Loop Back Count: 0
Far End F4 Loop Back Count: 0
#
#
 
Last edited:
@Y
Thanks for the advice. I do have an audible problem on voice which (i think) is affecting the SNR. Just some background. Built my house about 4 years ago and ran a sealed conduit with drawwire to the marked Telkom point outside my property. Telkom connected up and telephone service (incl ISDN) was perfect. About 2 years ago, the scratchy line/voice problems started. Telkom have been here approximately 6 times and they claim that the conduit I put in is the problem (even though they did the physical connection and house-wiring)
A year ago they ran a new cable from the connection point to my house (approx 8 metres long) which is not underground. However after heavy rain, the line quality deteriorates i.e the I suspect the connection has suffered again.
The reason why I have not logged fault again is that I am redoing paving on that side of the house. I will thus run a new conduit + cable (properly sealed etc) from my house junction to the Telkom point. I will then log a fault and ensure that the connection is done properly this time.
Just a last question : Since my ADSL installation (2 weeksa ago), I now have a humming on the line that was not there before. All microfilters are installed on all telephone points. Any ideas where this would come from?
see ADSL line stats from this morning. PLot looks a bit better and not as ragged as before.
http://i217.photobucket.com/albums/cc289/tjandies/ADSL/20071007_091817.png
 
my 2cents : with all the techs having been at my place - they said they "scotch-lock" joints where possible (small plastic enclosure that they put the 2 wires in and "crimp it" - to help against water.

a typical humm (like when you connect big speakers to an amp(sounds like a sound leak)) - i would say - water in the connectors outside.

have u tried using another phone just to be 200% sure?
 
@tjandies: Are you 100% sure that the problem is with your current conduit. It should be really easy for Telkom just to disconnect the cable to your house and then to check the line without the influence from the conduit for the buzzing noise with one of their blue telephones. Once the conduit cable has been disconnected you could easily test with a multimeter to check for any resistance between the two wires which would indicate a wet section - just make sure that all telephones, lightning protectors, faxes etc. have been disconnected before checking the resistance. You should try to get to the bottom of the buzzing noise on the telephone line first before looking any further.

Are the ends of the conduit still sealed? Do you think that the conduit is currently wet inside? It is possible that it is now full of water but this shouldn't happen unless there is a reasonable sized hole somewhere.

Telkom technicians often try to blame line faults on damaged cables in the customer's roof etc. to avoid actually investigating the problem!

Here is some information regarding DMT plots and wet sections. An SNR per tone plot might also help if you can get this information from your router using SNMP etc. Your plots would seem to suggest more of a crosstalk problem than a wet section but I could be very wrong. You could also try to measure the on hook and off hook DC voltages as well as the off hook loop current to see if these are below the recommended values.

"If slow service is indicated,technicians can look at the bits per tone measurement. Dips in the bit per tone graph indicate interference. Checking the frequency of the interference often makes it possible to identify its cause.The accompanying chart shows the frequencies associated with common interference sources, such as T1, E1 and HDSL. If the bits per tone are taking a hit and it is determined that noise is not the cause, then the cause could be a DC fault, such as a bridge tap or a wet section. Very low or nonexistent bits per tone in the high frequency band indicate the presence of a long loop. If there is a major dip in the bits-per-tone graph but the noise power-per-tone graph does not show any abnormalities, the most likely troubles are bridged taps or wet sections on the span. Use a loop-troubleshooting tool with a TDR to find and fix these problems. If the bits per tone are low across the whole bandwidth, the cause is most likely DC troubles on the loop, such as shorts or grounds. A major dip in bits per tone indicates AC trouble on the loop.To verify the type of AC trouble, compare the bits per tone to SNR. If the noise influence and bits-per-tone dip occur at the same frequency, the degraded performance is most likely due to a transmission influence. The particular frequency will help identify the source of influence or crosstalk. An example of possible shared frequency zones could be T1 centered at 770 kHz or HDSL at 196 kHz."

http://www.jdsu.com/test_and_measur...ts/whitepaper/ADSL_Technology_White_Paper.pdf

If you are planning to redo your conduit wiring it may be a good idea to use proper twisted pair lan cabling otherwise you may need to redo it again in the future if line speeds ever increase.

I assume that the buzzing noise on the line is still there with all routers, faxes, lightning protectors, filters etc. disconnected with only one working telephone connected?
 
@OMN: OK so you broke a new speed record today (RCO ~5.6Mbps). On page 7 of this thread your SNR downstream values are much more stable than mine - nearly a straight line - would indicate that if the problem can be fixed your line should perform quite well.

Is the DG634G router connected to an ADSL2+ DSLAM at the moment? The TI chips seem to be very popular in ADSL2+ DSLAMS - could be a Marconi. Take a look at this link about using the hybrid data to locate line faults. You may also be able to find out a few things by monitoring how the hybrid changes after a resync. The hybrid data is listed in the information that you posted although you say it is not from your line. The choice of hybrid 4 does not sound good.

http://www.dslreports.com/faq/sbc/3.1_Speedstream_5100b___4100
 
Last edited:
ok...

was a "big" day today - had a specialist FINALLY getting better testing equipment here - a big yellow device and a portable black router/tester..

shame...they struggled from 9am up until 4pm.

-they diverted my e-side 4 times - fault persists
-tested cable from home to dist.box - no fault
-tested cable from dist.box to exch - fault

now - having the cable diverted 4 times in 1 day with no visible effect - we came to the concl. that it is a TOTAL SECTION OF CABLE and not just A joint..

so he phoned up a manager in Pretoria - who told me - to monitor until tomorrow ( time out for the day) - and if not a BIT of improvement - its a given - total replacement..

and in his words: " new copper cable from exch to dist.box or optics "..

i'd rather settle for optics - because that means they cannot connect new copper to older possible faulty copper.

specialist was DUMBSTRUCT when he saw the # of errors this line generated - and the manager freaked when i told him this had been going for 3 years...just getting worse...

aaaaaaaah

i can see a light at the end of the tunnel
 
@OMN: Let's hope they actually repair the fault now. At least they have now accepted that there is a problem. Does this mean that Telkom will either be installing 3.2kms+ of copper cable or optical fibre plus a mini DSLAM cabinet? Any idea of how long this may take?

Q: If the problem is between the exchange and the distribution box how does your friend B (see map on page 6) manage to get a reliable service though?
 
Yotch said:
@OMN: Let's hope they actually repair the fault now. At least they have now accepted that there is a problem. Does this mean that Telkom will either be installing 3.2kms+ of copper cable or optical fibre plus a mini DSLAM cabinet? Any idea of how long this may take?

Q: If the problem is between the exchange and the distribution box how does your friend B (see map on page 6) manage to get a reliable service though?
Click to expand...

i don't know the answer to your question (i wish i could)........the tech did use 2 testing equipment - sunrise sunset tester and a 3m dynatel 965 loop tester...and with BOTH he saw errors even with e-side- (3-4 seperate) diversions..

so he recons - that my "friend" in B has a better but not complete healthy cable - hence 384 having higher slack in SNR

fibre optic has been ruled out - due to demand in local dist.box .

I'm afraid this might take a while as i had a SERIOUS and rather unpleasant phonecall from a person from capetown (senior person) who recons my line should be OK now (after admitting he didn't know the detail in my scenario)- a day after tests showed everything faulty (after i HEATEDLY told him that he'd better read his own investigation's comment section properly before making me angry)

he wants to go screw around in the manholes (10 of them) - i told him - listen - how about for ONCE we try and IGNORE the problem and just put a complete new copper pair 2400meter in.... he was silent.

however - i've put him to "Rest" for now .

tomorrow i'm gonna try organize something with the special investigation...a requirement for closure of case - a signature from my side..because i see they CAN't stand a problem being open for more than 24 hours..

shame really.

i've got 3-4 senior contact persons - i'll give 200% effort for resolution - or they must supply PROPER alternative.
 
Last edited:
@OMN: Latest doesn't sound too good but maybe you'll be lucky. Yes, 384K is a lot more tolerant than 4Meg but I think that in your case you may even have problems with that. Any idea of the cost of a new 2.4km pair? If you ever see a technician working at the distribution box I would ask him privately to try to find you a better pair to the exchange in the meantime now that you know where the problem is as Telkom may keep stalling you for months. If you ask Telkom to do this now they will probably just assume that they don't need to replace the cable anymore. Unfortunately any viable alternative to Telkom still seems a long way off.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X