Powering up a new airliner?
 

My son is studying electronic engneering and computer science.

He's about to power up a new project (A cube sat) and he's been looking at tips on how to do it without anything going bang.

Which got us thinking...

It's bad enough powering up a new PCB.
Anyone got any info or articles on how an aircraft is powered on first time?

Fire extiguishers at the ready??!:eek:

Follow the "Before Start Checklist"

I think the answer is ‘carefully’ following a checklist special manufacturer designed checklist. Firing up the whole thing would be asking for trouble. I’d also suggest that as many independent systems would be checked as the build progresses to ensure there were no crossed wires.

PM

I don't know, but I would assume that all discrete individual components had some kind of operational test at manufacture. That would leave installed wiring harnesses as the remaining untested components. But it could be I'm making erroneous assumptions about quality control at the manufacturers of components.

I remember going around the A330 line in Toulouse and going aboard one that was still being completed. There was a very old plastic chair where the co-pilot’s seat would evetually be and a worker was sat in it and running tests using the flightdeck screens . I would imagine that the wiring looms are tested before installation, and then power is slowly applied to individual busbars following specific protocols.
Practically a shore supply will be used initially for power up and then probably the apu generator and finally the engine generators

On BAC 1-11 production a very long time ago, power was first applied with all CBs pulled. Then each CB was pushed in one at a time to a schedule and the systems checked for correct operation before progressing to the next one.

In 1976 I did a radio re-fit on a Viscount and checked the supply on the connector to each box before fitting the box and turning on. After about 6 boxes were powered up we got a burning smell, so quickly turned the power off. On inspecting the radio rack by the aircraft door we realised the burning smell was from the first of many forest fires around Bournemouth that long hot summer. Panic over!

backward capacitors
 

I have myself done first power up of circuit boards which I designed a number of times mostly in the middle 1980s. In my case the most troublesome initial bad outcomes involved severe local overheating. I dealt with this possibility by using initially very short connection of the power supplies. I may have started with about one second of time, then doubling. In between applications I would pass a hand over the board seeking to detect local overheating.

After one alarming experience I learned to assure that neither my hands nor anything else sensitive was over the board at the moment of first application of power. In that experience a tantalum capacitor which had reverse power applied to it actually expelled the central pellet at considerable velocity. It hit the wall and not me, but the experience left an impression on me.

While I do recommend short time, I don't think that slowly ramping up voltage is very clever. Lots of things don't like prolonged low voltage on their supplies.

While pellet expulsion is particular to certain types, there are quite a few capacitor types which do bad things if the voltage is reversed--which is a possible consequence of several types of error.

[fixed a typo, and adding note to bow to the subsequent post by MurphyWasRight, who clearly has done a lot more of this than have I, and gave several useful practical tips]

Last point--We all used to call first article turnon the "smoke test" for good reason.

For 'first time' of a new design I would agree with above comments of quick tests followed by inspections with an added caveat of using a current limiting bench supply set to expected full load X 1.5 or so.

This helps prevent thermal events caused by backward caps etc.
Will not save the part in many cases but can prevent melted traces or burned board.
If the supply current limits find out why before proceeding.

Also a careful inspection and reality check measurements with an ohm meter that has high and low sense voltages before applying power.
Do this for every internal voltage rail to ground.
Also check rail to rail! Got literally burned by a short between 2 rails once...

If you have multiple boards measure all of them first and account for any significant differences. If practical allways build at least 3 so you have a tie breaker :)

In any case log the values as reference for the next batch.

The low voltage will not activate semi junctions and will detect shorts and should match a rough calculated expected value, usually fairly high (> 1K) unless bus termination resistor networks connect to both power and ground.

Often you can see the big bypass caps charging so initial reading will be low then quickly climb. Reversing the leads will show the same effect

The 'high', often shown as diode test on the cheap meters, helps detect backward diodes etc, this sometimes takes a bit of digging to understand if not as expected.

If possible bring up the board without the expensive parts first, often not practical unless sockets can be used on first proto.

Connected every meter and scope channel you own to the power rails and watch carefully on power up, especially confusing are sub regulaters that appear about right but are actually cycling between shutdown and overvoltage.

Lastly once the power is OK suspect anything with the word "I2C" in it.

Ah I thought that you meant the pilots

When working on power supply and output electronics, our top bench man uses a mains light bulb, in a socket wired in series with the input power which will allow voltage to build up, but absorb any over-current before something goes "snap". If the device cycles rather than powers on directly, select a bulb of higher wattage.

I've also observed a variac in use. It is a useful tool for AC powered equipment in series with the bulb. One keeps an eye on the current meter and brings up the supply slowly, watching for spikes. Not all high-current devices will abide this method, however, particularly circuits that utilize a power on reset cycle.

For some equipment, I am told there are published resistance checks for many power rail circuits that allow a tech to look for trouble before powering anything on, potentially avoiding the snap, crackle, but perhaps not the odd flying pellet's pop.

A good example of confirmation bias ! you see what you expect to see.

A wheeze that electronics engineers played was to smack two fingers into your other palm when one observed a fellow engineer connecting something to a live circuit or applying test probes etc. The sharp “crack” produced if you get it right sounds exactly like a short circuit or a component exploding, and your hapless colleague jumps a mile back away from the circuit, assuming he has just caused the problem. How we laughed..........

Reminds me of an all time great accidental example of above.

I was project lead for the chassis and backplane for a very (physically and $$$) large IP router joint venture with a Japanese company.
The Japanese built the HW but the initial bring up and integration was done by the USA company.

When the chassis finally arrived I did a very careful pre power check but on request of the Japanese did not apply power until my Japanese counterpart was present.

Word got out so by the time we were ready to flip the switch there were 15 or more people in the lab.

When all was ready I put my left hand in back pocket (an old habit from working on high voltage) and ceremoniously flipped the breaker.

Just then the senior Japanese manager took a flash picture to capture the 'historic' event.

I turned the breaker off while jumping back, almost falling over due to being off balance from the left hand in pocket.

After a couple of seconds I realized what had happened and broke out laughing quickly followed by the rest of the crowd, much to the great relief of the manager.
(For the few who may not know, in Japan embarrassing someone publicly is one of the worst things one can do.)

Ah. That must be why it was such an uncomfortable meeting in the late 90’s when we (BBC engineer chaps) told the NHK (Japanese TV) engineer that the talkback system he had built for the winter Olympics in Japan needed modifying - because he had omitted user volume controls. Oops.

When building something for the vacuum of space, consider heat dissipation. There is no air to convect heat away, leaving only conduction and radiation. Components that get warm on the work bench may become too hot in space.