As mentioned in the previous blog I have created a new variant of my solenoid autothrottle switch that is ‘active on’, i.e. by using a momentary ON toggle switch (1121S ON-ON + Momentary) the electromagnet (12V) must be powered to keep the switch in the ON position.
Without current through the electromagnet the switch flips back to the OFF position.
This is more realistic and means that the switch now only remains in the ON position if autothrottle is allowed in the simulator.
The switch housing has not changed from the previous variant: it still has the two narrow slots for the metal slider that keeps the switch in the ON position.
The slider itself was changed: the end of it has a 90 degree bent now and can be directly attracted by the electromagnet.
The original non-ferro metal slider has been replaced by one milled from a 45 x 65 x 0.6 mm iron plate that normally can be glued to a mobile phone to attach it to a magnet on a car dashboard.
For milling this metal I used depth steps of only 0.1 mm which is quite doable because the plate is only 0.6 mm thick and the outline is also small.
These metal plates have a double sided sticker that for this application can be removed with sticker remover.
In the video there is a metal ring of 3 mm thickness between the end of the frame and the magnet. This is because originally I expected that I could not mill the slider of ferro-magnetic material and had to glue an iron ring to the slider.
To keep the same distance between the slider and the magnet the ring therefore was moved to the back of the magnet.
As space is limited in the MCP in the final version the ring can be left out and the frame shortened with 3 mm.
For testing the SimVim MCP AT_Arm_On output signal was mapped on an Arduino pin that drives a power transistor.
The pins of the toggle switch are connected to the AT_Arm_Switch input.
While testing I found a problem that is not visible in the above video: apparently the AT Arm switch in the simulator has been implemented as a toggle function, i.e. the position of the virtual switch in the simulator toggles each time that the hardware input mapped to it is connected to GND .
Instead I had expected that 0V at the input would always switch it to the ON position and 5V at the input would always switch it to the OFF position.
This causes a problem: as soon as the A/T Disengage button is activated, the AT Arm Switch in the simulator immediately switches off and due to the delay in the solenoid the hardware switch switches off slightly later.
But because this toggles the simulator AT Arm switch again, the simulator switch is now in the ON position again, while the hardware switch is in the OFF position.
This can only be solved in the software and as I expect that all solenoid AT Arm switches must have this problem I want to discuss this with SimVim, ZIBO and/or Laminar Research, the creators of X-Plane.
A disadvantage of an ‘active on’ autothrottle switch is that during long simulator flights the electromagnet will be powered for a long period of time and therefore gets hot.
During my experiments I found that after 20 minutes or so it still can be touched, and according to the manufacturer it can used all day if each time the power is switched off for 1 second after it has been on for 1 second.
A duration of 1 second off is too long for use in the autothrottle switch, but with this pulse width modulation circuit a similar result can be achieved.
I used the second, more simple circuit that is described, but with the mentioned values for U2/C2 the PWM frequency is approximately 1 KHz and my autothrottle switch started to resonate quite loudly at this frequency.
To fix this I have replaced C2 to a value of 220 pF to increase the frequency to approximately 45 KHz.
As the supply for the NE555 circuit cannot be higher than approximately 15 V I have replaced the 24 V electromagnet by a 12 V variant so that both can use the same 12V supply voltage.
There are also complete, single chip solenoid driver circuits available, like the drv103, which is somewhat cheaper and could be used because the current in the electromagnet is only 150 mA, but I’ve chosen for the above NE555 based solution because the components in this circuit are more broadly available.
The good news is that with the PWM circuit the electromagnet temperature now remains at room temperature, so I will integrate this circuit in my new MCP PCB design.
Driving the magnet like this will lower the magnetic force, but I realized that I could increase the force by milling one or more extra iron plates and glueing these behind the current plate so that more magnetic field lines will go through the iron of the slider.
But the experiments show that this is not needed.
Inkscape SVG file for milling the switch slots (for the 18 mm MDF wasteboard as used at the start of the first video)
DXF file for milling the switch slots
Estlcam file for milling the fixation holes for the switch housing during milling of the switch housing slots
Estlcam file for milling the pocket for the switch housing during milling of the switch housing slots
Estlcam file for milling the slots in the switch housing (flip the housing for the slot at the bottom side)
Inkscape SVG design file of the new slider
DXF design file of the new slider
Estlcam file of the new slider
Sketchup 2016 3D model of the complete solenoid AT switch
STL file for 3D printing the frame