This item comes from the field of model railroads. These days many modelers use Digital Command Control (DCC) to power their trains. It operates by putting a voltage between the rails, with a magnitude of about 15V and a waveform that's a non-constant square wave. Timing of the voltage reversals conveys information to receivers in the locomotives. The minimum period between voltage reversals is nominally 58usec, so call it 8KHz. A receiver for DCC contains a rectifier which uses the incoming waveform to produce a d.c. voltage, and then the timing is derived by reading one incoming line's polarity versus the other. I hope that's clear enough.
I just read about a design for a DCC receiver where the rectifier function is performed by a bridge rectifier designed for use with transformers running off the 60Hz power line. My instinct is that this is a bad choice of component, because the diodes won't be specified for rapid changes in polarity. For operation in the kilohertz range, you'd want fast diodes, maybe Schottky type. I couldn't find a spec for the exact component used in this device, but if the diodes are like the ubiquitous 1N4002, they'd have a reverse recovery time of 2usec, and during that time they'll be conductive in reverse. With a sinusoidal waveform, a slow diode has a chance to switch from conductive to non-conductive while the voltage is near zero, and anyway, there are only 120 reversals per second. At 8KHz, and with the voltage reversing rapidly, I'd expect to see the rectifier working inefficiently, as the d.c. filter capacitor would be partly shorted out at each reversal, and it would also put an unnecessary load on the system that generates the DCC waveform. And I'd expect that bridge rectifier to heat up. Does anyone agree that this isn't a good design?
I just read about a design for a DCC receiver where the rectifier function is performed by a bridge rectifier designed for use with transformers running off the 60Hz power line. My instinct is that this is a bad choice of component, because the diodes won't be specified for rapid changes in polarity. For operation in the kilohertz range, you'd want fast diodes, maybe Schottky type. I couldn't find a spec for the exact component used in this device, but if the diodes are like the ubiquitous 1N4002, they'd have a reverse recovery time of 2usec, and during that time they'll be conductive in reverse. With a sinusoidal waveform, a slow diode has a chance to switch from conductive to non-conductive while the voltage is near zero, and anyway, there are only 120 reversals per second. At 8KHz, and with the voltage reversing rapidly, I'd expect to see the rectifier working inefficiently, as the d.c. filter capacitor would be partly shorted out at each reversal, and it would also put an unnecessary load on the system that generates the DCC waveform. And I'd expect that bridge rectifier to heat up. Does anyone agree that this isn't a good design?
