Fullscreen HTML5Med Res (??MB)HTML4

250VAC capacitors for LED circuits We had some correspondence last year on the danger of using DC-rated capacitors as the voltage-dropping elements in 240VAC circuits (Ask SILICON CHIP, November 1994) with particular refer­ence on my part to driving LEDs from the mains. However, I notice that since that time, mains-rated capacitors of suitable size have become readily available; eg, from DSE and Jaycar. Using these, would you now regard such applications as “safe” (I suppose nothing connected to the mains can be complete­ly safe) two requirements which make the design somewhat more complicated. If you have the timer on and then you decide to overrule it by manually switching to LPG or gas, then that setting will be maintained until you remember to switch over to the timer – which rather defeats the purpose of the exercise. We are also concerned about the possibility of the engine faltering if the timed changeover takes place at a critical time, such as when pulling out to pass another car at high speed. With these thoughts in mind, we are reluctant to publish a circuit without more information on the subject. Uninterruptible power supplies for PCs I would like to comment on the matter of making up an unin­terruptible power supply for computers as raised in the “Ask SILICON CHIP” pages of the June 1995 issue. I have had a UPS running for over one year. It was interrupted once when Feather­foots the cat jumped on a charging lead! It consists of a home made battery charger, a 10A Variac to adjust the charging voltage via the battery charger primary, a 1kW 24V modified square wave inverter and a 24V 80A/hr deep draw battery set. The battery charger consists of a 300 watt toroi­dal power transformer driving a 40A bridge rectifier with suitable heatsink and transformer cooling is by small computer fan. A steel case was used as a plastic one 92  Silicon Chip or would there still be some reservations about it? (J. K., Kenmore, Qld). • Provided these capacitors do have the correct 250VAC rating and also a suitably rated limiting resistor is placed in the circuit, such circuits should be safe. The resistor should be designed to fuse in the event that the capacitor becomes a short circuit and thereby fail-safe. However, we are still not really happy with LEDs running directly from the mains and prefer the old-fashioned and reliable neon indicator lamp with inbuilt current-limiting resistor. may melt if the components overheat. A small value high wattage resistor is used in series with the charging lead to limit the possible current draw to that rated for the transformer; in my case, 8A. Don’t be an idiot like me and use a number of power resistors in paral­ lel; the whole lot will fail in cascade! The batteries were specified with bolt type connections and were strapped together with aluminium bar, 3mm x 25mm. Plenty of petroleum jelly was used over the terminals and lead ends at the batteries. Charging and inverter wiring was heavy duty electrical earth cable. In practice, the battery is charged to its rated maximum voltage and the Variac is adjusted, over a day or so, until with the computer load, the batteries stay fully charged. My rather extensive 386 system, less monitor, draws just on 3.4A at around 27V. I can plug the monitor in when I need to; the supply stops my Unix system crashing and on test kept it running for 19 hours continuously. After that, it does take a day or so for the setup to drop right back to 3.4A. I aimed at a charging capacity about double the load draw. The 1kW inverter means that I can run everything except the laser printer for reasonable periods, about six hours if the 20-inch monitor uses about 200 watts. The batteries have needed no maintenance whatsoever since installation thirteen months ago. They have an expected life of five years, limited by internal terminal corrosion. The cost to me of this setup was about $1000 plus an afternoon or so of work and a bit of experimentation. Nothing has had to be done to the unit for many months. It is deliberately made simple for high reliability, hence the Variac rather than automatic voltage control. I have had a big regulated power supply fail to regulate. The device it was driving caught alight when the power Mosfets blew! (R. H., Tranmere, SA). • While we understand your desire for simplicity, using a Variac to adjust the charging current and hence, the final charg­ing voltage, is a little crude. There is a risk that the mains voltage could run at a high level for several days and you could be seriously overcharging without knowing it. Variable rejection filter wanted I was interested to see your simple 2-transistor CW filter in the March 1995 issue of SILICON CHIP and wondered if this could be easily modified to act as a variable frequency rejection filter. I want one which could be used to reject the 9kHz whis­tles on the broadcast AM band and also the various whistles on the international short­ wave bands. (A. S., Denmark, WA). • The CW filter in the March 1995 issue is not suitable as a variable rejection filter. To effectively reject 9kHz whistles without removing too much of the wanted audio signal, you need a filter with a very deep and sharp null, exactly at 9kHz. If the filter is just 10Hz or more off the exact frequency, the rejec­tion is greatly reduced. For that reason, a variable rejection filter is unlikely to be effective as it is too difficult to set it to the exact frequency. We did publish a 9kHz whistle filter as part of the circuit for the wideband stereo AM tuner described in the February, March & April 1991 issues. Notes & Errata Walkaround Throttle, Ask Silicon Chip, page 93, May 1995: the suggested wiring diagram for a centre-tap transformer shows the 10µF capacitor near the 7812 reverse biased. The capacitor’s negative connection should go to SC the centre pin of the 7812.