The Altitude Indicator Conundrum

The price of Spitfire Altitude Indicators (AI's) are currently a little on the high side for my liking so I have been looking at alternative cheaper Instruments. I have managed to buy this WW2 Halifax bomber Altitude Indicator 6A/1319.

The AI is going to be very tricky to alter, it is a precision instrument which will need watchmakers skill to modify. Somewhere deep within the mechanism the small linear barometric bellows movement is transferred first to a shaft with a gear arc that rotates through say roughly 36 degrees. That 36 degrees of movement is then translated through a watch like gear mechanism to move the 3 needles (100's feet, 1000's feet, and 10000's feet).

This gear arc is the logical and simplest point to attach to using a single stepper motor.

But to get from 0 to 20,000 feet the 100's feet needle has to move through 20 full rotations or 7,200 degrees. So the AI's internal gear mechanism will have a multiplier of roughly 200 to 1 on the original gear arc, ie 1 degree of gear arc movement is just over half a rotation of the 100's feet needle!

Most cheap stepper motors have an 18 degree step ie 20 steps for 360 degrees. So in this case a single step without gearing would be equivalent to 10,000 feet ! If we say a visibly acceptable movement of the needle is 10 feet per step then we will need a 1000 to 1 gearbox on our motor and with little or no lag in it.

The highest gear ratio stepper motor that I have seen on eBay is 275:1 or roughly 37 feet per step. So with an additional 4:1 reduction gear set this could be doable...

The second question is can we move it quickly enough for real time performance. The early Spitfires had a maximum climb rate of just over 2,000 feet per minute or 33 feet per second. So in this case we would need to make 3 steps per second which is easily achievable.

The maximum dive speed is 450mph which is equivalent to 660 feet per second (but I doubt that was vertical). In that case we would need to be able to move the stepper motor at a rate of 66 steps a second and that may also be marginal for an Arduino Nano's processing loop. Time for some more experimentation I think......

The ground level pressure adjustment is also a problem - as in a typical mechanical altimeter this is achieved by simply rotating the entire innards. But there is a calibration stop on the adjuster knob so this could be used to rotate the numbers independently.

Aside: One alternative starting point might be a more modern Slaved Altitude Indicator which has the look of an AI but is already stepper motored and geared but these are also beyond my budget.

26 Jan 19 - Still need to solve this one soon - I have just bought a later Smiths AI to play with so when that arrives I will have a good look and see what can be done.

28 Jan 19 - One alternative that has cropped up is to use a servo motor with a 12 bit controller. 12 bits gives 4096 positions so with a 2.5 turn sail winch servo that would work for the 2 of the hands, 1000 and 10000 ft if I leave them coupled, as this would give around 4 positions per degree. Then I would need a direct drive on the 100's feet hand. Mind you if I can get to the 100’s hand shaft it should be possible to drive the others in reverse – but maybe just not fast enough…..

02 Mar 19 - I have joined a Facebook Cockpit builders Group, it is mainly for the big metal guys building 737's, but I put this conundrum to them. I got a pretty good answer within 2 hours pointing me at a 3D printed Altimeter on Thingiverse that I can plagiarise for my own design. They also pointed me at HeritageFlightSim.com who are way ahead of me in building their own Spitfire Cockpit.