How Lionel's whistles work, how QSI expanded it, and what can go wrong.
Original: Lionel made a relay that had a shield blocking the AC magnetic flux from reaching the armature. When DC was applied, the 1 direction flux would pass thru the shield and the armature pulled down. Things could also overheat, both the relay and the transformer's selenium rectifier.
To counter the overheating effect, Lionel created a seamless 2 step whistle-on switch. The first position put just the rectifier in series with the center rail's line.
The second position connected a resistor around the rectifier. This changed the 1/2 wave rectified voltage back to almost full wave, with one 1/2 of the wave being smaller that the other.
The 1st step kicked up the relay's armature and the second step was sufficient to hold it up.
The combination of resistive rectification and the additional load of the relay and blower motor resulted in the loco going slower, so the second step also inserted an additional in-phase winding of 4 Volts which kept the hyper Voltage sensitive AC steamer motors happy.
The Lionel relay operated for either polarity of DC.
Note: Whenever an electronic whistle operated by an
older linear transformer doesn't respond, try pressing the whistle
button only part way.
Electronic Expansion: QSI's whistle detection circuit looks for any DC offset in the track power. However, the track power is seen thru the rollers, which have the highest rectifier resistance of any part of the power circuit supplying the loco.
The micro Processor will respond quickly to any offset voltage of over 2 Volts. It will wait several seconds to respond to a low value steady state offset of a little as 0.2 Volts, 200 mV. This process avoids over 99.9% of all bogus signals.
It is also capable of distinguishing between a positive offset and a negative offset, something that Lionel could not achieve with relay technology.
It was very fortunate that the Lionel transformers were wired in such a way that the horn signal put a positive bias on the center rail. That became the defacto standard for electronic operation. That consistency allowed for a negative bias to become the Bell signal.
Now back to the rollers. By their very nature they have both a resistance with a small rectifying quality. This is the result of the contact system being made with dissimilar metals, i.e. rail, roller, rivet, arm and pivot, and all being slightly lose (allows for flexibility and unfortunately, bouncing). The more the current the greater the Voltage drop and rectification. Having 2 rollers divides the current, cutting the rectified voltage in 1/2. Having clean rail, & clean rollers with graphite lubrication, reduces the resistance & rectification even further. Most systems keep the rectified voltage well below the QSI sensing threshold of 200 mVolts for the normal range of currents.
Hence the micro Processor does not respond to the all too abundant stream of minor DC offsets.
So how can the horn blow in a "clean" environment?
All it takes is for some part on the loco's electronics to pull more current on 1/2 of the AC cycle than the other.
The more common problems turn out to be partially damaged bridge rectifiers that are used to rectify the AC to drive a DC powered circuit, such as a smoke unit or a headlight and or maker lights, and very occasionally even the QSI electronics or DC motor bridge rectifier.
With more current being pulled on 1/2 of the cycle than the other, the rollers' rectifier resistance comes into play as the higher current half cycle will create a higher offset that the other half cycle, the difference of which can easily go over the 200 mV threshold. Depending upon which half is greater, either the horn will go on or the bell will go nuts while the motor is powered.
In the case of grounding a part of the motor's power circuit, 1/2 of the motor's current is shorted, cutting motor power and creating a large DC offset current thru the rollers yielding a high DC offset voltage for the micro Processor to sense and respond to. It is due only to the rollers' resistance and the remaining diodes' resistance that the transformer's circuit breaker does not open.
For the case where the 1/2 cycle remaining is the Horn signal, one will get no change in the bad behavior. However if the remaining 1/2 cycle is for the bell, the loco will stop due to lack of power.
The loco may not return to normal behavior due to the very erratic action of the supplied power, as it "sees" it.
|
|