Eldec transformer rectifier for Boeing 747 (2-418-05)
This article is about one of the transformer rectifiers of an Boeing 747-400 air plane. To be precise, this is one of the transformer rectifiers of PH-BFB, the 747-400 with the most flying hours. The air plane was taken out of service on the 26th of november 2018 and started it's 'next life' as a display piece at the Corendon Hotel in The Netherlands. After 134.000 hours of airtime, the avionics was removed from the plane. After a while of storage, the avionics modules were sold and I bought 33 avionics plug-in modules. And there are five of these transformer rectifiers in the batch. Maybe some transformer rectifier modules were sold already so I don't know of there were more than five modules in one 747-400.
This device is built by Eldec from Lynnwood, Washington. Nowadays the manufacturer is named Crane A&E (Aerospace and Electronics). This device has part number 2-418-05. The transformer rectifier is a rather simple device. The device converts 115 VAC 400 Hz electrical power to 28 VDC. Each transformer rectifier can output 75 Amps of 28 VDC voltage! That's more than 2.100 Watts of electrical power for each unit. That makes more than 10.000 Watts in total. This device was built in 1988. So it's likely this is the original device that lasted 134.000 hours of service. That's an equivalent of more than 15 years in service. Compare that time of service with your smartphone or desktop pc...
Note: Due to contractual obligations is the serial number removed to prevent air usage of this device. The device may never be used again in a flying air plane. The device is bought for simulator/display/reverse engineering/learning purposes and therefore there's a matching price paid. A functional device with all the documentation and history is much more valuable. One of the conditions for selling/purchasing this instrument is that the devices may never ever be used again for it's original functional purpose in the air. The serial numbers are therefore removed so the device is 'unknown' and can't be used in an air plane any more. This is a good thing that retired and 'unknown' devices are not used any more for safety for all of us.
The first step I took was looking inside the device. After removing a bunch of 'ordinary' Phillips machine screws, the alumina housing was removed. (For avionics are regularly torq set screw heads used instead of Phillips type screws.) After opening a lot of dust was found. There's an opening in the bottom of the housing for air ventilation for cooling and dust can enter the device this way. 'Clouds' of dust are gathered here and there. On the images can be seen what it looked like. The good thing is that the dust is 'dry' dust and was easilly removed using a dry and clean paintbrush and a vacuum cleaner. Since there's no static sensitive part inside, a vacuum cleaner can be used safely. The panels and exterior are cleaned using some isopropyl alcohol and a cloth.
|The dust in the air duct under the transformer.
|The result after cleaning with a brush and vacuum cleaner.
This device is rather simple. At a first glance it's just 'a box' with a three phase transformer, a set of rectifier diodes and some filtering. Well, there's also a current sensing shunt, but that's almost all there is I'd think. But the twist is in the details...
The nice thing is that the simplified schematic is shown on the outside of the device including the connections. (Usually it's very hard to find connection information of avionics instruments.) As mentioned the device can handle 2.100 Watts of power. The good thing is that the input frequency is 400 Hz. The higher the frequency, the smaller the transformer gets (at a constant power level). So if the avionics electrical frequency was 50 or 60 Hz like domestic appliances, the device would me much larger and heavier so convert the same amount of energy. It's no surprise that the frequency of the avionics electrical system is 400 Hz to reduce weight...
Well the three phase 115/200 VAC 400Hz input power is fed into pins 2, 9 and 11. The neutral is connected to pin 18.
|115 VAC is the voltage between one of the lines and neutral (Vln).
200 VAC is the voltage between each line. (Vll)
Therefore 115/200 is mentioned sometimes both.
Vll = Vln * square root of 3 which makes 200 = 115 * 1,73
The input energy is fed into the power transformer. On the secondary side of the transformer are two sets of windings. One set of windings in 'star' configuration and one set in 'triangle' configuration. The output of both secondary winding are fed into a set of diodes that act as a full wave rectifier. Until now everything is rather common. The interesting thing is that both rectified voltages are not directly 'tied' to each other but that a interphase transformer is used to 'combine' both voltage portions to one common 28V VDC signal.
|An interphase transformer (IPT) is used 'to absorb the difference between the direct voltages of the individual systems and must be designed for the times integral of this voltage' (Schaefer, 1965).
Interphase transformers are only used in systems that have two rectifier systems being used in parallel like this case. The interphase transformer phenomena was new for me since this is only used in this very specific design. Altough the system can deliver 2.100 Watts, the IPT is surprisingly small. Why there are two sets of windings used is not known to me. I assume that there's some phase difference between the winding sets. If the phase shift is 90 degrees, the resulting ripple on the DC line is probably much less since the 'gaps' between the sinewave tops will overlap. The result is a much smoother direct current voltage that reduces noise in the 28 VDC line. There's no current buffer capacitor found for the 28 VDC output. Maybe this current buffering is done externally or maybe the 400 Hz frequency with the phase shift trick results in a rather smooth 28 VDC signal that can be used directly without a current buffer capacitor.
In de 28VDC output line is a shunt resistor placed. Due the current trough the resistor (with a very low resistance), there's a voltage drop created proportional to the current flow. Both sides of the shunt resistor are connected to the rear connector. Therefore it's likely that there's some monitoring in the avionics system that 'sees' the current flow. This makes is possible to see possible overload of this device or maybe failure where the output current drops to zero for example. Assumed is that the transformer rectifier modules are linked in some way for load balancing/redundancy.
There's also a filter section that consists of a literally 'black box' with some terminals and capacitors on top. Assumed is that there are some inductors in the black box. Inductors and capacitors are commonly used for interference suppression. Later on I'll disassemble one black box to find out what's inside.
The reverse engineering is luckily not the most complex one. The component identification codes are mentioned on a drawing inside the housing and the basic simplified schematic is also given. The hard thing is to find out which wire of the secondary winding is connected to the star and triangle winding. Also the filtering is a literally 'black box'. So it needs some further investigation to complete the schematic. To be continued...
|It's nice that the components are identified on this drawing.
|The reverse engineering in progress.
The simpified schematic is shown here. The related connections are rather self explanatory. Pins 2, 9 and 11 of J2 are used for the 115/200VAC 400Hz power input. The neutral is connected to pin 18 of J2. For safety is an chassis ground wire linked to pin 25 of J2. Top connector J1 is the 28VDC output connector. The contacts are rather big so the current is 75 Amps can be handled. Pin 1 of J1 is for +28VDC and pin 2 of J1 is the 28VDC return. Pins 31 (+) and 32 (-) of J2 are connected to the shunt resistor to measure the voltage drop to determine the current in the 28 VDC line. Remind that these measurement connections are not floating and will be around +28 VDC each. So the output voltage and current of this device can be determined using these two pins.
|Both connectors at the rear of the device; J1 and J2.
|Connector J2 for power input and monitoring signals.
Since I have five of these devices and I only use one for reverse engineering purposes, I have four surplus. If you want to obtain one or more, please contact me via the contact form.
The devices are removed from a functional air plan so the devices should be functional. Remind that the serial numbers are removed and that the devices may never be used in the air again.