Roger,
Thank you for the update on TBR-21.
I find it hard to believe the French stated they need current limiting
because recent Alcatel PABX design requires it. Since the advent of
semiconductors it is a well known fact current limiting protection is
required in semiconductor designs. This will insure the circuit sourcing
the current can not be destroyed by mis-use or something fortuitous.
It sounds to me like the French government has the Alcatel PABX's in
question, installed in a number of locations and need TE's to be
compatible. Or Alcatel has an existing design they do not want to modify
and told their French version of a lobbiest to lobby for current limiting.
As far as I know all SLICS in use today have current limiting built-in for
self preservation. And also have a minimum current detection of 10 mA ,
as Joe stated previously.
To require current limiting in the ancillary terminal equipment is to
insure the SLIC can and will be destroyed sometime during normal use. All
the SLIC circuits I have designed or used have current limiting and
minimum current limit detect.
I have implemented TBR-21 current limiting in a DAA. I didn't have any
problems on EISA cards or external designs. They generally have enough
PCB area for proper heat sinking for 2W.
PCMCIA cards are a another problem. I was able to pass the current
limiting test, but the limit of 49 to 60 mA feeding from 50V through 230
ohms is difficult. A sustained limit will cause hot components to change
characteristics making it difficult to stay within the limiI range.
There is also a problem with holding circuit transistor de-soldering
itself after about 30 minutes of current limiting. To solve this problem
required searching for a SMT transistor large enough to sustain 2W, stay
within the range and not de-solder and/or destroy itself.
After 60 minutes of limiting the PC card metal cover temperature reached
80
degrees C. The inside temperature of the Laptop was 60 degrees C after
40 min.
I had no convenient way to check the temperature of the parts and PCB
with in the covers. For obvious reasons the transistor does not have
contact with PC card covers. As a consequence the heat of the cover is
due to radiation. The 80 degrees C temperature exceeds the modem chip
temperature specs (70 degrees C) . Obviously the French haven't
considered reliable functionality of PCMCIA cards.
The 0.3mA test is a transitional measurement, on-hook to off-hook
reference related to time. It is not a steady measurement. If referenced
to steady state it is taken out of context. The clause actually states as
follows:
4.6.2 Loop Current Characteristics
Justification : 91/2653/EEC. Article 4(f); interworking with the PSTN is
assured by requiring the TE to seize the line.
Requirement: The loop current determined by TE shall:
a) exceed the value of t1 after seizure and
b) remain above tf1 for at least a further (t2 - t01) time and
c) remain above If2 between t2 and t3, for conditions of table 4 and
figure 4.
The limit values t1 (t2 - t01), If1 and If2 are given in the tables 3 and
4 and shown in figures 3 and 4 and:
- (t0) is the reference from the moment of seizure, when the loop current
for the first time 0.1 mA with a feeding voltage of 50 VDC and stays above
this level for longer than 5 ms,
- (t01) is the reference moment, when the loop current exceeds for the
first time the current tf1 with feed voltage of 50 VDC and stays above
this value for more than 5 ms.
- It is permitted for there to be transient periods during which the loop
current drops below the limits stated in this clause, so long as, when
aggregated, they do not exceed 7 ms.
In essence the above clause is describing a template for I (current) Vs t
(time).
It states the following conditions:
Feeding voltage Vf = 50 VDC and feeding resistance Rf = 150K ohms.
Requirements:
a) t1-t0 = 400 ms
b) t2-t01 =400 ms
c) current (mA) = 0.30
Figure 3 is a template that indicates acceptable and non-acceptable areas
from the moment of seizure (0t) to 400 ms vs line current during that
time. Based on parameters above. Sorry I can't show it here.
Your colleagues statements need to be clarified. In this part of TBR 21 I
had no problem meeting the template requirements. My gripe (as I have
stated a number times) is the current limiting in a PCMCIA card. The
cards will simply not dissipate 2W and function or survive. The DAA must
be a dongle that can dissipate the heat. Modems with dongles are
generally not desirable. In other words hard to sell.
Hope this clarifies things
Regards,
Duane Marcroft
Telecom Consultant
On Wed, 23 Apr 1997 [email protected] wrote:
>
> Joe,
>
> The current limiting requirement was introduced in TBR21 in a plain
> oldfashioned horse-trade with the French. This was one of the ways to
> have them vote yes in ETSI.
>
> The official "technical" reason is that there are some quite new Alcatel
> PBX:es out there that need this feature i.e. not old stuff.
>
> However, the requirement will eventually be removed (e.g. the limiter
> function could be implemented in a separate unit instead) but as the
> current political approach is to persuade the no voters using the
> existing text, it will be in there when TBR21 becomes a CTR.
>
> It would be interesting to hear other comments from manufacturers who had
> the opportunity to implement the current TBR21 standard in actual
> designs.
>
> Our experience so far (after two TBR21 approvals) is that when you
> actually have a design on the table and you start pre-testing it, some
> very strange requirements pop up.
>
> One example is the Loop current characteristics (ยง4.6.2 in prTBR 21: June
> 1996), which is measured down to extremely low currents (0.3 mA!!!).
> These currents simply don't exist in the real world and this requirement
> shows that the standard was not prepared with semi-conductor based
> designs in mind. I find it quite amazing..........
>
> Can someone explain why this requirement is needed (not to mention why it
> should be justifed by Art 4(f) in the Directive)????
>
>
> Roger Magnuson
> TAD GRUBER COMMUNICATION AB
>
>
>
> ----------
> From: [email protected]
> Sent: Wednesday, April 23, 1997 06:58
> To: [email protected]; GECAB/EST/ROGER
> Subject: Re: PABXs & TBR21 current limiting
>
>
>
> In a message dated 97-04-22 11:24:43 EDT, you write:
>
> << I heard one colleague changed his French DAA current limiter from
> 45mA (i.e right in the middle of the range of values (26-60mA) that
> France requires you to perform your current limiting for NET4) to
> 55mA (i.e dangerously close to the high end of the scale using low
> tolerance components), because some PABXs tried to force >45mA into
> this DAA and therefore wouldn't work (No Off-hook detect).
> >>
>
> John:
>
> This story is an interesting one, but I am surprised at the reported
> problem
> with off-hook detect.
>
> Most solid-state SLIC circuits these days (PBX and central office)
> incorporate current limiters. The reasons are twofold:
>
> 1) Conserve battery power
> 2) Prevent excessive heat dissipation in the SLIC (self protection)
>
> This function is a current limiter, not strictly a current source. That
> is,
> the current is not allowed to exceed a specified maximum limit. However,
> there is no problem if the loop current is below the limiter threshold.
>
> The SLICs that I have worked with tend to have off-hook detection
> thresholds
> around 10 mA, although this parameter is programmable in many devices.
> Maybe
> there was something strange about the specific PBX in the story you
> heard. I
> would like to hear more if this is a real problem.
>
> And while I'm on this subject.........
>
> The whole issue regarding current limiting in TBR 21 really annoys me.
> Clearly, there is no need for current limiting in the TE in any country
> except France. Even in France, which supposedly has a modern
> infrastructure
> that is 100% ISDN-ready, I would expect that the vast majority (perhaps
> all?)
> of the PSTN lines are fed with modern, solid state SLIC circuits that
> incorporate current limiting at the central office.
>
> Maybe, somewhere in the south of France, there is a rural central office
> that
> still relies on the TE to limit the loop current. This would probably
> have
> to be an old electro-mechanical switch that has not yet been changed out.
>
> I question whether this is sufficient justification to force all of
> Europe to
> adopt a TE design that contains a current limiter which must be capable
> of
> dissipating 2 watts.
>
> Surely, if there really are some old central offices remaining in France
> that
> might actually be damaged by a non-current-limited TE, the French might
> be
> able to tell us when these last few offices will be replaced. Then, we
> could
> contemplate a CTR 21 that includes a phase-out period for the current
> limiter
> requirement.
>
> My experience suggests that in regulatory matters, reason usually
> prevails in
> the end. Thus, I continue to hope (wish?) that the next version of TBR
> 21
> will do a better job of dealing with the current limiter issue.
>
> I noticed in an earlier posting that the second attempt at TBR 21 has
> gone
> out for voting. Does anyone on treg have a copy of the version that went
> out? If so, can you tell us whether there is any relief on the current
> limiter requirement?
>
>
> Joe Randolph
> Telecom Design Consultant
> Randolph Telecom, Inc.
> <<File: UUCP_ENV.TXT>>
>
>
>
