Home | Articles | Dutch


Preview | Mechanical construction | Firmware and electronic | Housing

How to make a whisky compass yourself:

For a long time the whisky compass or wet compass stood on the list of gifts wanted. Because this instrument couldn't be missed in the DC3 or in any older plane. Certainly when you want to fly like me with limited navigational means. But when I look at the prices of for instance Simkits the need for buying one is quickly over. But thanks to book of Mike Powell about making these instruments yourself it became again a whole different picture. In his book he describes the construction of a gyro compass driven by a stepper motor. The principle is the same, we only need to rebuild it and put in a real wet compass.


Mechanical construction:

First of all some explanation on how we can integrate a real compass and how we are going to let it turn in a natural way. The principle is very simple. A compass reacts on the magnetic north. When we now overrule the magnetic north by a small magnet, the compass will now follow the magnetic north of this magnet. So when we make a disk that is connected to a stepper motor. Put a small magnet on top of it and place all this under need the real compass. The compass will exactly follow the magnet only with some delay, but this is just what we need. Because a real wet compass in an airplane also limps a bit behind.

Tech draw Whisky compass

Like you see on the technical drawing, the construction of the instrument is similar to those I made before. We only made a new horizontal platform to incorporate the stepper motor, the optical interrupter and disk with the magnet and interrupter flag. I bought the compass in a boat shop, removed the original housing. So we ended up with only the compass sphere. On the back of this sphere there was a cylindrical notch witch made it easy to mount. We turned a mounting ring with the lathe and fixed it at front of the first plate. The compass is glued with two component epoxy bond to the aluminium mounting ring. The different decks used for this instrument are the same as in the other two instruments described. The drill holes are made with the same drilling template.


Electronic and Firmware:

The electronic is also based on the same principle as the rest of the instrument. It exist out of a communication part that communicates via a RS422 with the serial port interface witch is connected to the PC. The stepper motor is powered by a L293D H-bridge driver and connected to a PIC16F628. This PIC contains the special firmware who manages all incoming data.

The firmware is again a masterpiece of programming logic. When you turn on the power the compass turns automatically to its zero position. This zero position is sensed by an optical interrupter. The program will half step the motor until the interrupter flag just enters the gap. If the interrupter is already in the gap, the motor is first stepped the other direction until the interrupter flag is moved out of the gap. If this were not the be done, positioning could be off by up to the width of the interrupter flag. After this initialisation we continue with the normal routine. In the original project of Mike Powell, he uses a stepper motor with 400 steps. He uses the half stepping technical to double the amount of steps. During each pass through the interrupt service routine, the difference between the desired position and the current position is calculated. A non zero difference is further examined to determine the best direction of movement. This is simply a case of looking at the sign of the difference. The goal is to move in the direction that minimizes the required number of steps. A movement from 0 to 1 will go in one direction while from 0 to 799 should clearly go the other way even though both are positive differences. In this way the compass reacts like the real thing.
The biggest problem was that in my project I used a different stepper motor. I used a stepper motor with a gearbox of 1/64.
This means that I needed to change two numbers in the firmware. The first value was the total number of steps needed to make a complete turn of 360° and the second value was half of this to determine the direction of movement. All this meant that I needed to type the firmware, 14 pages long, in to the Mplab compiler to be able to make these changes. But by the bad experience with the first instrument where I was unable make a workable HEX file even after tree months trying. I wasn't so happy with it. But there was no alternative once again. After the help of my cousin, a wizard in mathematics and in assembler I discovered later. We succeeded in making a new workable HEX file in two weeks time. After flashing it in to the PIC I made my first test and it worked perfect from the first time. This was a big relief. This meant that the hardware and software worked correct.



After testing the compass thoroughly and adapting the cockpit management program to the new instrument. The only thing missing was the housing to protect the compass and hang it in to cockpit. Indeed, the compass hangs fixed with tree elastics centred above the radio stack. This to eliminate the influence of vibrations of the aircraft to the compass. Otherwise the compass isn't working correct. The housing is made from inox (not magnetic). Of this material we folded a cube and left the back open. At the front we made a rectangular opening to look trough and also made a ring with the some rectangular opening. This ring suggest the ring of the real instrument.


After mounting the instrument in to the housing we determined the point of gravity to see where to put the fixing points. So the instrument would hang straight and not hang backwards of forwards after mounting it.

Now the cockpit looks much more real. Besides we now have a higher degree of realism. We now can reset the gyro compass every 15 minutes with the real compass heading when in level flight and we can also take the magnetic deviation in to account.


<< Previous | Index | Home | Next >>

© Verley Jan 2007-2017