In reply to:

---

Well yes, and the notion that an in-spec filter should always work with an in-spec line is naive to say the least.

---

What is the purpose of specifications other than to define what is required of a component in order ot work with other specified components?

In reply to:

---

In fact I'm not at all sure that there is a spec for the ADSL quality of a telephone line, is there?

---

Yes.

In reply to:

---

If there is, it's certainly not guaranteed; the only guarantee is that the line is suitable for low frequency voice communication.

---

Whether it is guarnateed depends on the provider who may guarantee that the product meets the specification. If the line fails to meet the specification then it could be the result of many factors, many of which are under the control of the owner of the line and the premises in which it resides. Nobody else can guarantee that those factors are not involved.

If the filter meets the requirements of the specification (which should be a guarantee by the supplier, if the individual item is faulty they should replace it, it it is designed wrongly the the commercial claim that it does meet the standard is fraudulent), then the failure for the line and the filter to work together is not the fault of the filter. That is all I have said on this subject.

It is possible that another filter may have been designed to its own standards, and may work with the faulty line. I contend that this is not a safe solution to a problem since it relies on the constant performance of an line working outside its specification and the unspecified additional features of a filter.

Please remember that the issue here was not whether the original, Solwise, filter as an individual component worked correctly, but the general statement that these filters were cheap and nasty. In view of the fact that Solwise claim that they meet the defined standard, such an accusation is defamatory. I am very happy that the OP obtained a result by replacing the filter. I am sorry that he had to pay an additional charge for this, since the probability is that a replacement of the original item would have worked just as well unless there was a serious problem with his line (i.e. all the bits of the line up to the point where the filter was attached, and that includes all extensions and other items of equipment on the line). If that is the case, it is of little concern to me but I am would not be happy in his position to rely on the solution being totally effective or long-term.

In reply to:

---

True and there are plenty of examples on here and that I have come across myself were two filters in series is required so the specification is obviously not enough sometimes, SKY boxes being particularly notorious.

---

Is it the specification that is at fault or the equipment placed on the line that causes it to be out of spec? Your mention of a notorious cause of bad lines rather suggests the latter.

Nothing has changed in the low-pass filter specification because of the higher frequencies used in current ADSL. The same low-pass band must be passed by the filter and nothing above is allowed through.

Of course there will be cases where a filter which is in-spec does not work in a particular situation. All that indicates is that the situation is out of spec. There are two ways of dealing with that - correct the problem or patch it up but modifying the associated equipment. I know which I believe to be the safe approach and am not at all happy that others leave the impression that a bodged job is acceptable.

And what is this ADSL line specification? There is a specification for ADSL equipment but there isn't one for the line beyond a line being suitable for voice and a 28Kbps modem. Additionally, you now have lines being provisioned for MAX that would often fail or be refused for fixed speed provisioning. On some of these lines where you might be getting a barely usable signal at the receive end even the 55db rejection of an in spec filter of any noise at relevant frequencies coming from the telephone side of the filter might not be enough to leave an usable signal. The only thing that could be considered out of spec is the line length, it's electrical condition and/or external noise. In such a situation I can't see BT replacing or improving such a length, all else being equal, unless it impinges on voice operation as there is no USO, i.e. minimum spec, for ADLS. I have had a number of such situations and a better specified filter such as an X1 has meant the difference between stable, unstable or even no ADSL, irrespective of what was plugged in on the voice side.

>> >> In fact I'm not at all sure that there is a spec for the ADSL quality of a telephone line, is there?

>> Yes.

Oh no there isn't.

The BT filter spec was defined several years ago, long before IPStream Max was in use in the UK. It was an attempt to define a filter which would work in most (if not all) circumstances, based on the knowledge available at the time. There's no ADSL line spec to match it up against, so there was never any certainty that it would always be adequate. With Max pushing the technology further out, there are now more situations where the minimum spec is not adequate. These situations require a better filter.

Eric

In reply to:

---

These situations require a better filter.

---

Which frequencies should not be filtered out which were not required to be filtered out in the past and at what level?

The standard for filters lets through what is required for the voice channel. What parts of that should be blocked now that ADSLMax has been introduced?

In reply to:

---

And what is this ADSL line specification?

---

You appear to be asking a specific question about a technical matter. I am prepared to answer that on a professional basis. I provided a complete answer to the original question that was asked.

I would like to add a few things to this thread about filters.

There are huge differences in the designs of filters, some have been known to claim they meet BT spec such as Tekhal that clearly do not. In the case of the Tekhal branded filter there was a wire that passed the ring signal from the line through the filter, an error that is specifically highlighted in the BT spec. So there are some companies out there that are making claims that are not correct. I don't know if this is deliberate or just they have assumed that all filters meet the standard so have put it in thier product description. In the case of Solwise they have sold two types of faceplate, one made by Austin Taylor and one by Pressac. It would seem that they have mostly run down their stock of the Austin Taylor one and have started supplying the Pressac version. The Pressac one was designed in association with BT to meet the spec and has been tested by BT exact labs to prove it meets the spec, so they are not misleading anyone by stating that for the faceplate. I would add however that their microfilters and extension sockets are very basic and I can not imagine they could get them through test, but I'm not in a position to know if they have been tested and passed by some mirical.

It is very common for people to find that filters that seemed to work fine with a fixed speed connection now need replacing with Max. The reason being that often with a fixed speed line your SNR margin would be high, say 20db. Using the phone might make it drop to 10db but that wouldn't cause a problem because you still have enough margin to maintain a connection at the speed required. However with ADSL Max the line syncs at the highest speed possible without letting the SNR margin drop below 6db. Now if you use your phone and you losse 2db of your margin your link is lost as the margin is not suitable for the speed you are connected at and the modem re-syncs. If you replace your filters with another model that does not couse a change in the SNR margin then you will not losse sync. Equally if you install better filters that boost the SNR margin available you will get fater speeds on Max.

Filter technology has evolved since the days when the Pressac faceplate was designed but it is still a very capable filter so it should work ok for most people.

Deleted

Edited by deleted (Mon 09-Oct-06 16:34:44)

As far as I know there are only two manufacturers of these 'High Quality' filters for use on a BT NTE5 master socket, Pressac and Austin Taylor. By 'High Quality' I mean they are based on a three stage coupled inductor design. In fact all filters will have thier own idiosyncracies and personally I prefer a single stage design with a less agressive high cut as this seems to be more Caller-ID friendly.

At Solwise we stock both manufacturers devices and based on test results we consider them to be interchangeable, though, probably as a result of misguided forum postings, our customers seem to have developed a preference for the Pressac one. This is the model with the curved top. The ADSLNation model looks like a re-badged Austin Taylor one, with a sloping top.

We sell about 600 of these units per month and have very few returns of either type, maybe 3 or 4 in the last 12 months, of course your unit could be one of the faulty ones.

Now, ADSL frequencies are inaudible. If you can hear 'static' on the line this can only be because the ADSL frequencies are being translated down into the audio band as a beat frequency caused by an interaction with another high frequency signal. Typical sources of such frequencies include cordless phones and anything with a switched mode power supply. Before condeming your filter it is worth switching off a few electrical devices which may be getting some radio-frequency energy on to the raw ADSL side of your line. Sometimes just moving your cordless phone base-station will get rid of the noise.

There is further coverage of these issues on the Solwise forum. I hope this rather wafly and over-technical posting will help some-one!

Thanks for that reasonable comment. The only thing I would add is that the ADSLNation faceplate uses an active filter with transistors, so I don't think that your guess at its identification is right. I've seen tests which suggest that active filters are not so good in plug-in units because they don't play well together when there's more than one, but that's not a problem with an NTE5 faceplate filter of course.

Eric