a United Technologies Company

Tasks varied from being a team member of the Real Time Embedded Propeller Control Systems group to a independently creating electronic test equipment that interfaced with a new propeller program which was used for balance calibration (pictured below). My time was split between the software group as well as working with a project engineer to assist in managing various projects as well as extensive amounts of paper work. The bulk of my work consisted of debugging, testing, and code reading software for the propeller control systms of military aircrafts such as the C130, E3, C3, and NP2000 to name a few. A brief synopsis of the balance calibration box that I designed while there can be found below. Due to confidentiality and non-disclosure agreements the schematic is not available. Every other propeller on the NP2000 and A400M has two magnetic pickups on them. One blade has three magnetic pickups and thus designating it as Blade 1. On the hub of the propeller assembly is another, fixed, magnetic pickup. As the propeller spins each of the nine magnets pass the fixed magnet on the hub. As each one passes this fixed magnet, it creates a pulse. This results in a pulse train of nine pulses in groups of two with one group having three. In between these groups is a single pulse that is created from a different “tooth”. This tooth’s position between the groups varies depending on blade angle. The final pulse train consists of the groups of two, and one group of three (referred to as the “triple tooth” from here on), as well as a single pulse in between each group.

The purpose of this circuit is to recognize the “triple tooth” and to output a pulse at the beginning of the third tooth. To do this, the conditioning circuit from the EPC-100 (ES804827-7, Sheet 7) was used to step-down the voltage to the five-volt range and also to “square-up” the pulses. The output of this conditioning circuit is a stream of five-volt pulses that can be used in the rest of the CMOS logic. This output is fed into a Synchronous 4 Bit Binary Counter with an Asynchronous Reset Pin (54LS161). The first two bits of this counter are used since only a count of three is needed. The first output bit is fed into a 3-Input NAND Gate (54LS10) and also fed into a Monostable Multivibrator (54LS221). The inputs and outputs are configured in a way that one HIGH-level pulse is outputted on the rising edge of the input pulse, which in this case is Bit 1 of the binary counter. This HIGH-level output pulse has been calculated to be 3.78mS long. This 3.78mS can be looked at as the window for the triple tooth to happen. If three pulses do not happen in this 3.78mS then the counter will be reset and the circuit will restart. This 3.78mS output pulse is inputted into the 3-Input NAND Gate and into a second Monostable Multivibrator. The purpose of this second Monostable Multivibrator is to detect the falling edge (the end) of the 3.78mS pulse and to output a 7.2mS pulse which connects to the Reset pin of the counter.  The three inputs of the NAND gate are Bit 1 and Bit 2 of the Counter as well as the 3.87mS output pulse coming from the first Monostable Multivibrator. When all three of these inputs are high this says that the counter is outputting a decimal 3 (or a binary 0011) and the 3.78mS pulse is still occurring. The output of the NAND gate is an active low. This active low pulse is fed into a third Monostable Multivibrator that will detect the falling edge and output a HIGH-level output pulse with a duration of 38.2mS.

These pulse-widths (tw) mentioned above are based on the values of the resistors and capacitors connected to the 54LS221s. (tw = R x C x 0.69) The only value that should vary from the NP2000 to the A400M is the resistor value of the first Monostable Multivibrator. The resistor value will determine the window for the triple tooth. In the NP2000 the window was found to work best at 3.78mS. This finding was done as a worst-case scenario (that being 250RPM and a blade angle of –16.0°). The A400M will likely be different. The other pulse times were calculated to the minimum pulse width while still being long enough for an accurate read. The pulse timing is drawn on the following page, though it is not drawn to scale it serves as a graphical representation of the timing.


Pictures of the final product (two were made)