Hi Treg Yes Joe Randolph is correct regarding max power dissipation, I did not word my example very well as I was thinking about total power dissipated (mainly in the feed circuit) which could peak at 6 W + (3W in each 200 Ohm resistor) for a 50v CV supply. Of course, most modern PABX feeds now have CC supplies typically operating at around 32 mA. Use this data in designing a test jig if you want to measure the dc characteristics on the test bench!
Joe is also correct with regard to the current sink in the TE. I would add the note that, both the resistor used to determine the slope value for the current sink and the regulating transistor should be assessed for worst case current (60 mA). A design capable of handling 2 to 2.5 W total dissipation should be adequate but don't ignore the I squared R calculation for any resistors in the sink. I Understand that the French PTT say that the Feed inductors / relay coils used in their network saturate at 60mA +, when saturated no ac is passed so transmission becomes highly attenuated. This was too powerful an argument for allowing current to exceed 60mA in the TBR / CTR. Cheers: Bill Ellingford . ---------- From: [email protected][SMTP:[email protected]] Sent: 12 March 1999 18:28 To: [email protected] Cc: [email protected] Subject: Re: TBR 21- 60 mA max. loop current In a message dated 3/12/99, [email protected] writes: > Has anybody out there also "taken the heat" 8-) and if so, > been able to defy Mr. Ohm and/or Mr. Kirchoff and somehow > dissipate less heat (and still meet TBR 21, of course)? Bob: Your analysis of the situation is correct. Unfortunately, there are few ways to avoid having to dissipate about 2 watts for the worst-case feeding condition (50v and 230 ohms) in CTR 21. I don't quite agree with Bill Ellingford's statement that this situation has been common in Europe for some time. In the example he cites from the UK, the worst case feeding condition is 50 volts and 400 ohms, which is very similar to the worst case feed in the USA. In the old UK specifications, testing was performed at "whatever current the TE draws" when connected to a 50 volt and 400 ohm feed. In practice, this usually came out to roughly 90 mA, for a dissipation of about 1.3 watts. You could only get the 125 mA that Bill refers to if the TE presented a short circuit, in which case the power dissipation in the TE would be zero. In the past, the only country where 2 watt power dissipation came up was France. France has long had a 60 mA current limit requirement. The reason that CTR 21 has a similar requirement is that France insisted on it as a condition of their support for CTR 21. Reportedly, some of the central offices in France rely on the TE to limit the loop current. In the French national specifications, the worst case feed is 54 volts and 300 ohms, which can result in dissipation of up to 2 watts if the current limiter is set to 60 mA. However, with the French national requirements, there is some opportunity to reduce the power dissipation by setting the current limiter in the TE to a low current (say, 25 mA). This results in about 1.2 watts dissipated in the TE. Unfortunately, when the French requirement was transcribed into CTR 21, clause 4.6.2 of CTR 21 wound up requiring that the TE draw a minimum of 49.6 mA with a 50V, 230 ohm feed. This comes out to 1.9 watts. My view is that this was probably an oversight, since a lower limit of 20 mA or so should be fine to ensure that the C.O. recognizes the seizure. Maybe this requirement will be reconsidered in the future (any comments on this possibility from others on treg?). Strictly speaking, the 49.6 mA requirement in CTR 21 only applies for the first 1.2 seconds after going off hook. This means that you could have a separate "seizure shunt" that you activate for the first 1.2 seconds after going off-hook, after which you fall back to something like 25 mA. Thus, with some additional hardware and extra software control, you could reduce your steady-state power dissipation to about 1.1 watts. In practice, I have not implemented this extra complexity in any of my CTR 21 designs. I simply use a current limiter circuit that can handle 2 watts continuously, while keeping the transistor temperature under 100 degrees C. To accomplish this, I use a higher power transistor than I would otherwise use, and I provide a copper heat sink pad in the board layout. There are other circuit complications introduced by this approach that relate to the transient response requirements in CTR 21, but these can be overcome without adding too much to the cost or board area. As an engineer, the extra power dissipation of this approach bothers me, but I tell myself that the 2 watt feeding condition (50 volts and 230 ohms) is probably very rare in the field. For instance, I don't think that a 230 ohm resistance will be encountered anywhere outside of France (400 ohms minimum is more typical). Even in France, you would have to be on a very short loop get the high heat condition. Lastly, I think that almost all of the new solid state SLICs being deployed in central offices limit the feeding current to less than 40 mA or so, just to protect the SLIC chips in the central office. I have no concrete information about the newer central offices in France, but I suspect that this is true in France as well. In summary, your concerns about the 2 watt power dissipation are well-founded. There is a way to reduce the power dissipation to about 1.1 watts, but this requires additional circuitry and software complexity. I have found it easier to simply design for 2 watt dissipation, and hope that in the majority of installations, the actual power dissipation will be somewhat less. Joe Randolph Telecom Design Consultant Randolph Telecom, Inc. 781-721-2848 (USA)
<<application/ms-tnef>>
