In a message dated 6/18/2002, Duane Marcroft writes:
> Can someone bring me up to date on Austel..? I've forgotten what the > present Austel Hi-Pot requirements are for modem transformers. Is > 3750 Vrms still in effect...? My client insists it is. But, if > memory serves me the requirement is now IEC950 can some give me an > update. > Hi Duane: The insulation requirements in clauses 2.3.2 and 2.3.3 for Australia (basic insulation from SELV/ground, reinforced from AC mains) are identical to those in IEC 60950. The main differences for Australia are in clauses 6.2.2, 6.2.2.1, and 6.2.2.2. For clause 6.2.2, compliance is checked with BOTH of the tests in 6.2.2.1 and 6.2.2.2 (as opposed to either one, at the manufacturer's choice, in IEC 60950). In clause 6.2.2.1, the test voltage for case (a) is changed to 7.0 KV for hand-held telephones and for headsets and to 2.5 KV for other equipment (as opposed to 2.5 KV for all of case (a) in IEC 60950). In clause 6.2.2.2, the test voltage is changed to 3KV for case (a) and 1.5 KV for cases (b) and (c), as opposed to 1.5 KV and 1.0 KV, respectively, in IEC 60950. Overall, I have not found these modified requirements to be a problem in most designs. The requirements for the physical construction of the transformer itself are the same as in IEC 60950. As long as the equipment is not handheld, the Australian requirements in clause 6.2.2 are not much more stringent than those in IEC 60950. That being said, there is a significant problem with lightning in Australia, and I tend to be conservative in this regard. If your modem does not have protection from tip/ring to ground, I suggest that you design to withstand common mode lightning surges of at least 4000 volts peak. You can usually accomplish this with a transformer that has supplementary or reinforced insulation. For example, one supplementary insulation transformer that I use has been tested and found to have an actual surge breakdown of 7000 volts. Of course, your layout would also need to maintain adequate creepage/clearance for the desired level of performance. Lastly (and I almost forgot this), there was an issue several years ago where "host-independent" modems, such as a PCMCIA card modem that will be sold in retail stores, had to provide reinforced insulation. The argument was that since the AC mains supply could not be evaluated by the safety lab, the reinforced insulation from the AC mains had to be provided in the modem itself. I am 99% certain that this conservative interpretation (also used by the UK at one time) has now been dropped, but it has been so long since I did a true "host-independent" modem that I can't recall for sure. Most applications these days are not host-independent. Joe Randolph Telecom Design Consultant Randolph Telecom, Inc. 781-721-2848 http://www.randolph-telecom.com
