Thread: Military and aerospace tech's view on electronics in modern vehicles.

You tube video 'What is the most reliable 4WD for Australian touring?'. Usual support for old Toyota, Nissan, Mitsubishi, and one or two lone LRover enthusiasts.

Found one thread reply to the video very interesting, on the question of electronics in modern vehicles. Sorry if this has been highlighted elsewhere on AULRO.

Lifted this from https://www.youtube.com/watch?v=nYLwh6_W6Hw

I am an electronic engineers. 40 years experience now retired mostly military & aerospace tech. Electronic reliability is dependant on design, materials and manufacturing. Reliability comes at cost. It's a sweeping and false statement to say because it has electronics it's unreliable. MOst electronics systems are far more reliable then mechanical system. It all depends on application and how they are used. Buying the cheapest electronic upgrade, I'll guarantee it will fail. The more complexity (functions) to the electronics the more likely it will have issues. In the end KISS. But to blame electronics for reliability is not directly true. It's how it's built, it's how it's wired, it's the connectors (e.g are they water ingress proof). The other issue re electronics reliability is temperature. Sitcking electronics in the enginee compartment on a 50 degree days means it will operate over 70 degrees plus. Most average commercial electronics design in Europe (ECU from Bosch for example) may not be designed for those temperature. The clue is the manufacturer ambient operating temperature specs for the vehicle. Operate outside these and the lifetime of electronics degrades rapidly. It's cummulative too. Land Rovers are 50 degree and -20 degrees ambient and they test them to that extensively. Most others do not. But Amazon electronic upgrades in the bonnet most likely use Commercial rated electronic components that are speced to max 70 deg & 0 deg min.. These will have short live times in the engine bay of your vehicle. Electronics should be made using at least industrial grade components that are -40 to +85 degrees they cost 2 to 10 times more. Then we have the next level of automotive grade components -40 to 110 degrees which if you think about are for the Aussie outback in a engine compartment still not enough. Military grade is next at -55 to 125 degrees and their designed for significant vibration enviroments you'd expect in the outback. cost of these is 10 to 100 times and some have export license restrictions. The next secret to reliabiity is to use the components in a derated design setting. Basically safety factor of 2. So you'd only want components that are rated to 125 degrees, i.e. actually running to a max of 62.5 degrees ambient. This then requires specialised methods of cooling the chips in the engine comparment through fans or conduction to the metal external body not heated by the sun. In the tropics, humidity will induce fungus to grow on electronics components. To prevent that all the electronics boards should be conformally coated with a special type of laquer. No automotive manufacturer does that. That's only done for military. However to use such compnents that give the needed reliabiity in the Aussie outback, automative grade may not good enough. It's the military grade you need. But those come with high cost, paperwork and traceability requirements, impractical for outbackl vehicles. More likely is that you get at best automotive grade. But derate by 2 in an enginee compartment and their still not good enough. The killer of electronics systems: 1/ Temperature 2/ Brine conductive water (clean fresh water may not be a problem) 3/ Vibration (breaks wires unless specifically deisgned for this) 4/ Dust (mixed with moisture) 5/ Humidity re fungal growth 6/ Salt fog (beach atmosphere) creates corrsosion in connections. E.g. Military grade Bushmaster or Hawkei have extensive electronics you smear that includes active electronic suspension control. They are tested to be reliable for our operating environment using military grade components in real battle conditions (ukraine). Electronics is the least of their reliability problems. These vehicle cost are near the millions each. Thats why and that's how you get reliable electronics. Hence mechanical system are cheaper and more reliable in an automotive setting.

Hi,

To answer your questions, first some context. I am (was, long story) building a range rover camper to travel around the world with and as that goes, I got a bit too enthusiastic and decided that I could certainly take up the range rovers electronics! Long story short, my main reasoning was that if I replaced the proprietary hardware with something that I could manage on my own, it would be more dependable, perhaps not more reliable! but I could get myself out of trouble when stranded in the dessert. As it turns out, although not a very easy task perhaps, in the P38 range rover computers were not very developed yet so creating replacement(s) would be at least relatively easy (for instance, everything you do is already a signal with a wire to the BeCM, even honking the horn). Dismantling the BeCM we found the high side switches as mentioned and that gave us inspiration for our own units. The exact items were long since discontinued but newer items were still produced.

The plan was to replace the following computers with arduino mega's: ECU (speeduino), HEVAC and BECM (split into three parts: front and rear PDM and central BeCM). A few development boards were created with two different types of SMD high side switches and they were put through the wringer (I have some pics of those) Turns out that you can even use PWM through the switches as long as the frequency is not too high else they heat up quickly.

The biggest problem I encountered was finding a switch that could reliably handle 100A (or more) so that was done with a 200A relay. Everything else is all electronic.

For clarity, the entire system is based on a hardware abstraction layer that runs on the arduino mega. ie. we do not use the arduino C libraries for anything. You only code the pins that you actually use (and whatever logic you require). This keeps the size down and the speed extremely high. After all, if I depress the brake pedal I expect the brake lights to go on within 10ms, always. Turning on the lights can have a small delay, rather not, but it is acceptable. So there are 3 things that actually can cause an interrupt: brakes, horn, flashing high beam. That's it, everything else can simple wait for the loop to get around to it. Furthermore, all boards have a CAN shield so they can talk to each other (the P38 does not use CAN so no other communication with the vehicles legacy components) and the BECM itself also speaks "land rover" so it can communicate with the dash cluster. Finally we support ISO9441 (K-line) for W.bus, V.bus, etc. (victron shunt/voltage meter, webasto heater).

Since the HEVAC is largely a standalone system (it does use SOME pins on the BECM but they are simple logic voltages and no communication) we decided to concentrate on that first and it also has by far the largest component of logic (and thus memory pressure). I have devised a small board that holds all the components to control the hevac (ie blend motors, distribution servo's, blowers, sensors, etc. The original gear in the range rover was used since that too is fairly simple stuff. A servo with +/- X volts and a potentiometer (3 leads) for the position. The test setup works on the bench and the next phase will be to place all the needed components on a nice shield that will fit the arduino mega. Some of the logic still needs to be fleshed out since it is not at all easy to reverse engineer hevac logic (if ANYONE has an example we might use, please )

Finally, there is an old ipad air 1 that fits exactly in the space where the radio and hevac controls used to live in the dash. They will be removed and a custom app will run on the ipad. From there I can control the hevac and any other things I want to add to the car without have to cut holes in the dash for -whatever-. Most of what you need already has switches (like foggies) but say camp lights or whatever will run of the ipad. The ipad itself has a lightning to can adapter and communicates with the arduino (hevac) bidirectionally in real time. All this already works, it just needs polishing.

Final example before this gets too long and preventing me from finally answering your questions My P38 has been converted to a thermofan (the same type as in the ineos grenadier) and to an electric 18cc AC compressor (for reasons). When the HEVAC instructs the AC to run, it sends a message over the CAN bus to the PDM in the front (Power Distribution Module) which sends a 490hz PWM signal to the AC compressor to determine the duty cycle ie speed. It does the same (well inverse) to the SPAL fan (yes the grenadier has a custom SPAL) and voila, the entire system starts working. The PDM monitors engine coolant temp and condensor temp and communicates that back to the HEVAC. It has an override for engine temp (so even if the CAN bus fails or the BECM fails, it will still cool the engine) and can also in case of emergency (AC pressure too high for example) cut off the compressor. Other than that, there is an algorithm that simply ramps up fan speed according to temps etc. The ipad can see the can messages on the bus so it reports the various temperatures, compressor/fan speed, current, whatever you want/need

This build has been a group effort indeed of like minded people. An embedded software engineer (with automotive experience), an embedded hardware engineer (with aeronautic experience), a car enthousiast and... me I design and build the "supporting" electronics, project manage AND have many excuses to indeed "field test". I am the only land rover owner, both engineers own a... Jeep... Go figure!

I have photo's of the test setup, the development PCB and what not but not the entire finished system as of yet.

The arduino is rated industrial but not automotive, this is however not a real problem since the design is open and an automotive version of the atmega chips exist (atmega328p) and a board can also be made to spec with ease these days (pcbway and the likes). With careful placement I believe automotive arduino's would not be needed.

Yes, I am using a derated design with passive cooling. For ease of manufacturing (and I am oldskool :P) I have chosen to go with through hole components where ever I can. This means that there was only one choice of high side switch left: the BTS50080 and that is rated for nominal load of 9.5A which as discussed in a previous post is WAY more than even the highest load in the car needs (high beam is only half that and that is the largest load in a car generally speaking). There are only very few parts in a car that need more (that need switching) and those would be front and rear window defoggers. I no longer have a rear window and the front is broken as on all older P38's so -eh- Also, the blowers run on their own motor controllers which are actively cooled since they sit in the airflow of the fan (a very old and very common design)

So far the arduino when running on it's default of 16mhz does not get hot at all. I have yet to test it by placing it in an oven at 70c but if need be, as mentioned, we take the open source design of the arduino and build our own.

This also kinda answers your last question, yes, arduino is basically screwed over if you ask me but the big elephant in the room here is: open source license. Their old designs (uno, mega, etc.) are all in the public domain and the atmega chips themselves can simply be bought from anyone so although officially produced arduino's might go the way of the dodo since qualcomm is already pushing in their rather useless chips, there will be plenty of "clones" around for the foreseeable future. The only thing that will change is the arduino IDE but I was the only one using that to begin with to write "pseudo code" or PoP that the other guys would simply implement in the HAL which is vi on linux and a compiler.

Finally, the reason I went with arduino is exactly that: open source, extremely ubiquitous (you can even buy one in Kenia for example), cheap and easy to work on/with. It is also the reason I went for speeduino (it is based around an arduino mega), I can simply carry a spare arduino mega (or two?) and never be left hanging. I can even take the HEVAC arduino out in case of emergency and still I can keep the car running and me to safety.

Perhaps I should make a better write up one day

Cheers,
-P