TELECOM ACCESS STANDARDS NEWSLETTER NO. 127 March 2001 CONTENTS 1. ADSL LINE FILTER TRIAL 2. ADSL LINE FILTER SPECIFICATION 3. REPLACING 3-WIRING 4. CHECKING POTS CPE COMPATIBILITY WITH ADSL SIGNALS 5. PTC 200 QUALIFICATIONS 6. PTC 253: 2001 TRUNKED MOBILE RADIO 7. TEST LABORATORY CHANGE RETURN TO MAIN INDEX 1. ADSL LINE FILTER TRIAL Newsletter No. 126 mentioned that Telecom has been undertaking a limited trial of line filters connected into each jackpoint connecting POTS CPE. The primary aim of this work is to determine whether customers can be left to purchase and fit these filters, thus avoiding a site visit by Telecom contractors and the associated costs of this work. While we are satisfied that line filters provide isolation between the POTS CPE and the ADSL signals and generally prevent either one interfering with the other, some problems have come to notice (see Item 4). Also, we still have some concerns about radio frequency interference (RFI) from some types of customer premises wiring. There is also, of course, the possibility of customers failing to fit a filter on all POTS CPE, and then having problems with service interruptions. One interesting possibility is interference to the ADSL service caused by telephones with piezo-electric ringers. These can generate quite high level harmonics during incoming ringing on the POTS line and may disrupt the ADSL modem in some cases. A brief in-house trial in Wellington was quite promising and we have since extended the trial to a further 100 customer sites, mostly in Auckland. These also have indicated that RFI may not be as significant a problem as we were led to believe. It appears that in areas of strong AM broadcast signals, any RFI emanating from the wiring is having little or no significant effect on radio performance. The trial results indicate that line filters may be an acceptable solution for many customers, although we still need to determine the range of conditions under which their use can give high assurance of satisfactory operation. In view of this need, we hope to extend the use of line filters progressively so that "JetStream" and "JetStart" customers in general will soon be able to install them subject to some overall guidelines. The exact wiring arrangements in customers premises vary widely as regards the lengths of cabling runs, the number of jackpoints, whether these are 2-wire or 3-wire, whether there are cable stubs on looped runs, the nature of the cable itself (earlier cable was untwisted, a lot of older cable had little twist). Because of these unknowns, there is considerable vagueness over which customers will find the use of line filters to be a satisfactory option. It is more or less a case of "try it and see" at this stage. 2. ADSL LINE FILTER AND SPLITTER SPECIFICATION In view of the promising results to date, we are now releasing the "Draft for public comment" of PTC 280, which covers our interim requirements for customer-connected line filters. Its publication will allow for independent testing of line filters currently on offer and provide an opportunity for potential filter suppliers to assess Telecom's interim requirements as part of their product designs. WARNING: It must be appreciated by prospective filter suppliers that Telecom is not yet in a position to define exactly where and under what circumstances the use of line filters will provide a satisfactory outcome for a customer. This information will be gathered in the light of further experience and it is expected that suitable guidelines will be made available in due course. It may be that the range of customers for which this solution is acceptable is largely limited to the main centres, but this has yet to be seen. We are confident that the installation practice we have used to date is highly effective. This consists of installing a splitter at the network entry point and new wiring to the computer jackpoint. This isolates all the POTS wiring and CPE from the high frequency ADSL signals. However, such installation work can be expensive and time-consuming, as well as requiring a site visit. Filters could offer the much simpler installation alternative of having the customer connect one between each jackpoint and the POTS CPE connected to it. However, filters do not isolate the POTS premises wiring from the ADSL signals and this could result in radiation. Unlike most overseas countries, we are using "full-rate" ADSL, not the "Lite" version. Also, we have a good deal of 3-wiring, which has an inherently poor balance to earth and can thus act as an effective antenna for the ADSL signals. In view of these differences from general overseas practice, we are working with the Radio Spectrum Management division of the Ministry of Economic Development to determine the probability and likely level of radio interference from filter-equipped "JetStream" and "JetStart" installations, especially those with older types of premises wiring. The most likely impact is on radio receivers in close proximity to the wiring, but the Ministry has also been concerned about the long-term effects of widespread xDSL services on background interference levels. This is especially the case with ADSL, as it has potential to be very widely deployed and in large numbers. Each line filter has some impact on the performance of the associated POTS CPE, whether or not the associated CPE is on or off-hook. Any number from one to five filters may be connected on the same line and different types could well be connected in combination. PTC 280 has endeavoured to address these issues, but it is impossible to simulate all possible wiring arrangements, network cable types and lengths, CPE combinations, etc, in a laboratory testing environment. As a result, PTC 280 represents a starting point for the assessment of filter performance. Some conditions and requirements are almost certainly going to be subject to amendment in the light of actual service experience. We will probably also have to cater for filters being integrated into other POTS CPE in the future. Already, some ADSL CPE has a filter integrated into it for connecting a telephone. Unfortunately, those seen so far are designed for 600 Ohm terminations and badly impact on the sidetone of telephones designed to Telecom's PTC 200 specification. Such filters will need to be disabled, as they can also cause problems, even when left unterminated. The "draft for public comment" of this interim Specification is now available free of charge on the www.telepermit.co.nz/PTC280.pdf website. It is recommended that filter suppliers contact Harvest NZ Ltd to arrange testing to this Specification. 3. REPLACING 3-WIRING The introduction of ADSL ("JetStream") and other developments into home computer networks over telephone wiring bring increased focus on the advantages of our latest residential premises 2-wire wiring practises, as outlined in PTC 103:1998. This Code of Practice covers our 2-wire version of the British Telecom 3-wiring system, which is based around the UK-designed 6-way modular plug and socket, originally introduced here in 1983. ADSL places a lot more emphasis on the need for balanced 2-wiring throughout the customer's premises, as distinct from earlier unbalanced 3-wiring. 2-wiring greatly reduces noise induction, as well as reducing potential RF radiation for the wiring. When we introduced 3-wiring, telephones were the predominant form of CPE, with few fax machines, modems, etc. Decadic signalling was virtually the only type in service, although there were a few DTMF telephones in use. Plug and socket connection of CPE was relatively rare, with most telephones fixed-wired, using a wide range of complex connection options within each telephone. All telephone installation work was thus done by Post Office staff. Hardly seems credible these days, does it? The BT plug and socket allowed customers to connect their own CPE and prepared the way for an explosion in the number of types of telephone and CPE used. Since that time, technology changes, deregulation and open competition have had a vast impact on the types of CPE available. As at the end of 2000, we had Telepermitted over 6200 different products! This is even more impressive when the small overall size of the New Zealand market is considered. 3-wiring catered for much simplified wiring arrangements and provided the solution to an annoying problem in 1983 - "bell tinkle" during dialling. This was caused by a combination of inductive power feeds in exchanges and the line breaks of decadic signalling. Both have now gone as the result of electronic telephone exchanges and DTMF signalling, although there is a very small percentage of decadic signalling phones still in service. 3-wiring brought considerable benefit at the time, but it also resulted in the premises wiring having poor balance to earth. This was not usually a serious issue in 1983, when the great majority of CPE were telephones. However, problems started to show up as the use of fax and data increased and such items as 56 kbit/s data modems put much more emphasis on the need for noise-free lines. This need is even greater with ADSL now coming into wider use, as it operates at radio frequencies. We introduced our "2-wiring" in 1996, as a means of increasing reliability and line balance. This had the benefit of improving data performance in noisy line conditions. Our 2-wiring is more in line with international practice, but it had to cater for the many 3-wire connected telephones in service. It does this via the "floating" 1 (F capacitor in each 2-wire jackpoint. We "guesstimate" there are probably as many as 2 million 3-wire connected telephones still in service in New Zealand, so we cannot just ignore them. Needless to say, we strongly recommend that customers convert to 2-wiring, especially if they wish to avail themselves of high speed data services. The continued use of 3-wiring is expected to be a limiting factor in the use of line filters for ADSL. It is also known to reduce analogue modem performance in the presence of electrical noise. 4. CHECKING POTS CPE COMPATIBILITY WITH ADSL SIGNALS We have had only a few reports so far of POTS CPE is being disrupted in some way by the presence of ADSL signals on the same line. Nevertheless, with well over 6 000 different products granted Telepermits, it seems likely that more such reports will be received as "JetStream" installations increase. An ADSL splitter typically introduces around 50 - 60 dB attenuation on the "POTS cabling" of any signals from the local modem. This has so far been found to avoid interference problems between ADSL and POTS services in almost all cases. However, this may not always be the case, as most items of POTS CPE were never tested to see if they are susceptible to ADSL signals. As a result, we simply don't know whether a product will be affected or not, nor just how serious any such effect may be. The most serious case so far was with a handsfree telephone, which lost its handsfree capability as soon as an ADSL modem recently connected on the same line synchronised with the network. This telephone continued to operate correctly in handset mode, but no longer provided the main function it was purchased for. The proposed use of line filters, dealt with in item 1 above, puts more emphasis on the need to check out POTS CPE compatibility. A filter typically inserts about 25 dB loss between the local ADSL modem signals and the POTS CPE connected behind it, so there is increased risk of interference relative to the situation with a splitter. There is also the possibility that a customer will plug in POTS CPE without the necessary filter and thereby connect the full modem output onto that CPE. In the case of the handsfree telephone mentioned above, the full signal was enough to completely upset its processor chip. The product became totally non-operational until it was disconnected from both the line and its plug-pack power supply. After this simple operation, it reverted back to normal. As "advance warning" to suppliers, we are considering a requirement for all new types of POTS CPE submitted for Telepermit to be checked for their susceptibility to ADSL signals. It appears from our tests of the product mentioned above, that there are three sets of conditions to consider; b. checking with a typical filter, and c. checking with a direct connection across the modem. Each such test may reveal a different level of disruption from "unaffected" to "disabled". In the meantime, it is recommended that CPE suppliers carry out their own checks on their current and superseded POTS products to see if any are affected by ADSL signals. This is a precautionary move, as it is better to be aware of such problems and to know how to react if products are returned for servicing after the customer concerned has subscribed to one of our ADSL services. 5. PTC 200 QUALIFICATIONS A number of issues have come up in the past few weeks regarding aspects of PTC 200, which were not previously defined. a. Stutter dial tone A recent service problem revealed that our specification for stutter dial tone duration is incorrect for the older NEAX 61 Kai exchanges. PTC 200 and TNA 102 quote a stutter tone duration of 2.5 seconds, after which the tone reverts to standard continuous dial tone. Measurements made during investigations into the service problem revealed that these earlier exchanges provide the stutter signal for only 1.6 seconds. This has not previously been a problem, as most CPE intended to detect stutter dial tone tests for it during the first few bursts of tone. In one case though, the product did not verify the signal until close on 2 seconds, thereby getting a false indication. b. Message waiting Although it is not defined in TNA 102 or PTC 200, it should be stated for the avoidance of any doubt, that the FSK Message Waiting signal is sent in Multiple Data Message Format. This is not particularly significant in that PTC 200 requires CPE to support both SDMF and MDMF. c. Caller Number delivery With our Caller Display signalling, we send both "time of day" and "calling number" following the end of the first ringing cadence. Although there are essentially two lots of information, these are sent in Single Data Message Format (SDMF). As mentioned above, PTC 200 requires CPE to support both SDMF and MDMF, so the use of SDMF has no impact on Telepermitted CPE. However, there is another subtle change. According to TNA 102, clause 10.2, there is a silent period of not less than 500 ms before the CSS is sent to the called party. Measurements have since revealed that this quiet period may be as low as 350 ms in some circumstances. The actual duration of the CSS signal is still 250 ms, as specified and this is consistent for all ringing cadences. d. Integral ADSL filters While we have not seen any POTS CPE with integral ADSL filters so far, these may well come up for Telepermit in the future. We will require that these will be tested to PTC 200 in the usual way and shall comply with the specification, with no allowance for the existence of the filter. This approach is being adopted because customers are likely to purchase and connect such CPE, whether or not they have ADSL service. To avoid any customers who do subscribe to our ADSL services also fitting external filters to such CPE, it is proposed that a suitable label be fitted along the lines "This equipment incorporates a filter for use with Telecom "JetStream" service. An external line filter is not required". Since an integral filter would probably be regarded as a marketing feature, such a label would help to emphasise this benefit. Should a customer not understand this notice, any performance degradation caused by a second series-connected filter will be the same as for any other non-filter equipped products. ADSL modems with integral line filters for connecting POTS CPE are expected to be tested in accordance with PTC 280, but with one to four reference line filters connected in parallel, thus simulating the test conditions of PTC 280. 6. PTC 253: 2001 TRUNKED MOBILE RADIO A completely revised Issue 3 of this specification is now available free of charge on our website. This document has relatively narrow readership, but its publication is being recorded in the newsletter for any suppliers of trunked mobile radio equipment who have not had recent involvement with Telecom's Mobile Radio Group. Access Standards is not involved in any part of the testing process, but will issue Telepermits to suppliers following confirmation by Telecom Mobile Radio that their products comply with the new PTC 253 specification. As with cellphones, the first stage of the Telepermit process is to submit two samples of each product to Telecom Mobile Radio for testing and assessment, along with the supporting technical information. These two samples are to be pre-programmed for operation on the Telecom network in accordance with PTC 253. As usual with such radio equipment, one of the pre-requisites for a Telepermit is evidence that the product has been covered by a Declaration of Conformance to the Ministry of Economic Development, stating that the product complies with RFS 26. Full details are given in www.telepermit.co.nz/PTC253.pdf. 7. TEST LABORATORY CHANGE We have just received notification that one of our accredited testing laboratories, KTL Ottawa Inc, has been bought by the Norwegian company Nemko AS. The test lab will now trade as Nemko Canada Inc. The contact details are otherwise the same as before this purchase. These are shown on our website. DOUG BURRUS Manager Access Standards