View Full Version : Arduino-737NG Project

05-27-2013, 10:39 PM
First and foremost, I would like to apologize to the sponsors of this forum – I enjoy dreaming of the products you create, but I cannot afford them on my small budget. Therefore, I must find an alternative. The arduino and Link2FS appear to be the most promising solution.

The project:
Build a functional 737NG overhead (fwd & aft)

Budget expectations – less than $500 USD (max $1000)

End result:
Build .bmp format for (free) distribution (pending author’s blessing of altered picture)
Display schematics
Build an arduino 737NG library for the arduino

I initially thought I would work out all my bugs before initiating a post. Then, I remembered the purpose of forums - outside input. So, here goes!

The overhead will consists of the three layers:
The main picture of the overhead – with the lamp lights removed via exacto knife
The second layer is the transparent film displaying only the lamp indications
The third layer is the black-out layer; only allowing light to be passed where intended

05-27-2013, 11:13 PM
This is where the project stands thus far:
Left - print w/ LED back lit - not acceptable
Middle - transparent back lit w/ LED and ambient lighting - acceptable.
Right - transparent back lit w/ LED w/ no ambient lighting - acceptable.

The transparent(3M) yielded crisp & clean fonts. This experiment causes me to consider the quality of paper for the over-all picture.

Since the arduino is capable of max 40mA for an individual output, I've used NPNs to drive the numerous LEDs which require > 50mA of current.

05-28-2013, 11:26 PM
I'm just doin' some thinking out loud:

The NG OVRHD has approx 150 indicators; the arduino MEGA is capable of 50 outputs & 10 analog.

Without purchasing (3) or (4) MEGAs dedicated soley to the indicators, how does one efficiently utilize the arduino to accomplish this goal? It's quite the conundrum.

My solution, until a more efficient solution is presented, is to charlie-plex the indicators.
Prior to the final design, I will need to research the maximum number of charlie-plexing LEDs to prevent the brain from recognizing the momentary on-off state of the LEDs.

As stated before, NPNs will be used to drive the LED clusters. However, using the minimum amount of outputs via charlie-plexing created an additional challenge. Fortunately, I had a few Quad - 2 Input 7408s ($ 0.50 ea) (TTL-AND Gates) lying around. After allowing the smoke to escape from one because I accidentally bypassed my 5vdc regulator -LOL, I replaced it with another on hand. The test was successful. According to my calcs, it will require (38 ) 7408s (or equivalent). Knowing this, it's obvious I will need an interface board between the arduino and the OVRHD; it seems I will use Mike Powell's suggestion/link (from a previous thread) to have the PC board etched at a market fair price.

Back to 10 analog outputs, does the NG allow dimming of the warning indicators. I haven't researched it yet; my intuition says "No!"


Jim NZ
05-31-2013, 03:57 AM
Neat mission your on there James ,,, I'm not a 737 fan but I can see that a lot of guys will be watching this.

The lettering actually looks great to me but I'm no expert and you know what you want so just keep on trying things ,, something will come along.

I presume your talking about the 'Ifly' 737NG ,,, I dont have that one so I'm not sure of how much of the 'logic' is available thru simconnect ,, there again you might be building in your own 'logic' in the Arduino. ????

I can easily add extractions or direct inputs in the 'Multi ,,' or there again it might be a better plan just to make a stand alone 'module' just to suit your overhead. ???

Heaps of LED's ,, for an easy extra 64 pins and only using 2 on the Arduino ,, have you looked at ,,,
as Gene talked about in here ,,,

Keep thinking out aloud James ,,, interesting mission for sure.

Good luck with it ,,, Jim

06-01-2013, 12:37 AM
Yes, Jim; the lettering on the transparency was fantastic! The printing on everyday white paper... not so much. The use of high quality paper will correct this since the pics of the lamp indicators were cut from the main picture.

As for the logic - it will be housed in the arduino via a library. I'm thinking this will be the easiest part of the mission; the hardest part for me is hardware construction and securing to the overhead i.e., LEDs. My thought is to purchase a plastic sheet thick enough to hold the LEDs AND the LED retainers. Realizing the LEDs will be slightly above the surface of the back-plane, spacers may be required.

With my breadboard wired up with NPNs, a 7408, and some LEDs - I confirmed my suspicion that 8 common trigger poles, for charlieplexing, was the max for this processor - so as not to detect the blink. Mind you, I lost a bit of luminosity due the multiplexing of the pins. I regained the brightness by using parchment paper (a cookie sheet) instead of using everyday printing paper behind the transparent sheet; this was a suggestion given to me by a wise man - and it worked wonderfully!! Now, I could trim it back the poles to 6 in hopes it will compensate for reading the library and any additional code. With that said, trimming it back to 6 will force me to increase the LED blocks (sets) from (150/8 )19 to (150/6)25; so 25+6=31 - the outputs still remain within one Arduino MEGA.

Cost analysis:
I will have to determine whether (2) Centipede shields, (2) arduino MEGA boards and a custom created PC interface board out-rank the complexities of (1) arduino MEGA board, a custom created interface board, and accompanying components. Thanks to the centipede link and the uncertainties of the timing within arduino, I will prepare the hardware for 2 shields and 2 Megas - yet, work towards a one card solution. This way, if I find additional code interferes with the charlieplexing timing, it will be an easy fix - simply design an alternate interface board.

Interfacing to FSX:
Not much thought has gone into the software side yet. I haven't decided whether to use the 'Multi' via keys and special requests or build a total separate module. Personally, I would like it to see it confined to the 'Multi'; I'm certain similar attributes will reach different aircraft. I figure I'll focus on the hardware first and cross that bridge when I get there... hopefully in the year 2013! :)


06-02-2013, 11:36 PM
<dl class="details"><dd class="hdr-info"></dd></dl>Calling all electrical engineers!! I'm not an engineer, nor have I ever claimed to be. But, I need some constructive input. I am contemplating the use of a 14x14 (14 banks of 14 'AND' gates) to drive the 737NG warning lamps. Below is a schematic of the current breadboard using only one output from the 7408 (as a test) and the two inputs of the respective AND gates.

Initially, I was somewhat disappointed in the luminosity loss due to charlie-plexing- yet it was still very usable. I knew if I could hold the 'AND' input higher for period of time more than determined by the ardunio, the luminosity would increase. After consulting a close friend in the field of electronics, he pointed me in the right direction. Due to his input, I can multiplex over 30 pins without detection from the eye. This is twice the amount I require!! My concern is the overall longevity of my modifications - will I have components fizzle out due to excessive current, i.e.-the npn (after a few seconds, the low-power transistor feels warm (not hot) to the touch)?


Jim NZ
06-03-2013, 03:58 AM
Now that's a different circuit to what I thought you were going to build James ,,, must be "plan C' ,, but it obviously works so that's great.

The heat after a few seconds is a bit of a worry though ,, check the type of transistor your using as per its "use" and it's currant handling ability. A bit of heat's OK but only after a few seconds is probably not a good sign for long life.

I'm afraid I'm the 'old' "think of something ,,, try it ,, if it don't work ,, try something else" brigade :) :) :)
Hopefully some-one else will pop in and be of more help James.

Keep at it ,, your close ,,, Jim.

06-03-2013, 10:13 PM
PC bd etchers, please read.

Ha! Plan 'C' ??? I'm already at Plan 'J'!! :)

Anyway, I had a eureka moment during my day job. It happened when I recalled a conversation involving the charlieplex pin to be held high for a bit more after the arduino turns it off. Shift register was casually brought up as a description of my efforts - then quickly disqualified. After combing over the details of that conversation, it hit me - the use of a decade counter (74HC4017).

I believe I have one to test. Now that I can see it start to come together, I wanted to put it on paper first. Using 13pins from the arduino (1pin for the 4017 & 12pins for the LED output logic to the AND gates) will drive 120 lamps; using an additional pin for a 2nd 4017 (14pins) could yield 240 outputs. Sheesh!! Technically, the lamps could be controlled by the tiny Arduino Fio - LOL!!!

Knowing the fwd ovhd alone will require 120 inputs alone, this project has the capability of becoming a hairy mess! This easy method, provided it works, should assist with wire management. I can then purchase the centipede for input use. I read somewhere, probably provided by you, I can cut the needed inputs from 120 to 60 if I set it up correctly. I'll research it when I get to that point.

PC bd etchers,
If this design works, I would rather pay a member of this site to create affordable boards as opposed to using my usual method. If there are inquiries, either post on this thread or PM me.

06-05-2013, 11:22 PM
After allowing my circuit to burn solid successfully for a few days, contemplate the logic design, and perform the cost analysis - I am reconsidering. Not including custom designed circuit boards, the components alone will cost roughly $150 USD do drive the indicating lamps.

Here is the break down for each module:
(4) IDC 14 pin Header receptacle. @ $0.73 each
(4) IDC 14 pin Headers................@ $0.68 each
(3) Quad AND chips ...................@ $0.62 each
(1) 14pin screw terminals for LEDS@$1.84 each
(2) 3pin screw terminals for Power@$0.50 each
(transistors, resistors, capacitors, & LEDS are not accounted for)

That arrives at roughly $10 a module (not counting the custom PC board). Multiplied by the needed 13 LED modules, the figure arrives at $135.

My day job often consists with custom pre-built annunciators with layers of ribbon cable for numerous LEDs and a handful of switches. I know the hassles... is $115 (minus custom PC bd) worth it?? I'm thinking not - the $25 centipede to drive the LEDs is becoming more appealing.


06-30-2013, 08:04 PM
I've been distracted for a few weeks; that's ok. It has allowed my circuit have 'burn time'. Since the last post, the LEDs and
logic remain true. As of today, I exchanged the NPN transistor for the more commonly accessible 2N222 type. Since the hfe was
higher, a few modifcations were needed- the schematic indicates the changes.

Satisfied with original experiment, I began the process of procurement. During the process, I suddenly had an epiphany: why
not apply similar concepts to the input switches? With quad AND gates @ $0.62 ea, I can save $50 as opposed to the centipede
option... and have a cleaner project.

After making the change with the 2N222, I will have to allow 'burn time'. This will also allow me to codify my plan.

LED module layout:

Switch module layout:

Revised schematic:

07-11-2013, 03:01 PM
May i suggest using the free prosim 737 system module for the overhead logic.You are able to send the output information to Arduino with the COM connection thingy and back in threw the com thingy for switches.I really dont know how to program a massive system into Arduino yet but i do know when lets say the fuel pump light comes on it sends a character you put in like "fuel pump" followed by a 1 or 0 for on and off.I do not know about inputs though.

07-12-2013, 08:01 AM
Thanks for the feedback, Conner. I'll have a good study of the features.

07-18-2016, 06:05 PM
hi how did this project end up as i am finally on to building mine...