Overengineering

[Shocker]

How much do you overspecify the components when building electronics???

[japlytic]

Re: Overengineering

It depends on what it is, but in most cases, I overengineer, but not excessively.

[Shocker]

Re: Overengineering

I understand the concept of "wasted durability". But if it only adds a few % to the overall cost, why not???

Also, you never know how the equipment is going to get abused. Like dropping hard drives to the floor. No, seriously, I had a Barracuda ATA III fall about 1m and it still worked fine.

Personally, I would like equipment to be designed with an operational lifespan of at least 20 years at 40°C ambient. I'm only interested in technologies that make a real and significant improvement to usability, reliability, efficiency, robustness, or compactness, without compromising. Or health and safety, but note that lead-free solder does not make the list. In fact, chances are it's causing more harm than good, due to tin whiskers and BGA problems. Not to mention that solder only makes a small percentage of the total lead in electronics anyway. Two things: CRTs, and car batteries.

[PeteS in CA]

Re: Overengineering

My dream is for the @#$%ing Envirocrats who dreamed up RoHS to find their lives dependent on medical electronics built with lead-free components and solder!

[dmill89]

Re: Overengineering

My dream is for the @#$%ing Envirocrats who dreamed up RoHS to find their lives dependent on medical electronics built with lead-free components and solder!

I'll second that, how eco-friendly is something that ends up in a landfill due to early failure anyway?

[Shocker]

Re: Overengineering

I'll third that...

For the record, I once saw a circuit board labelled "RoHS" that was soldered with ordinary leaded solder. It was in a cheapo PSU with a seized fan though so fake RoHS was the only good thing about it.

Oh yeah, health and safety. Despite being in Australia, I stumbled across a North American type IEC cable. They really need to get rid of that too since on the plug I'm looking at, there's all of 3.5mm from the outside edge to the pins. Which aren't insulated even today. Nice job, NEMA .

For the record, my favourite plug design is BS 1363 with massive pins (a prime example of overengineering), mandatory shutters, and built-in fuses. In case you're wondering, I do have (multiple) BS 1363 cords too. Which plug is your favourite???

[dumpystig]

Re: Overengineering

You have these (British) plugs in Australia?

[Shocker]

Re: Overengineering

What I'm talking about that is, but I don't quite get the question.

[dumpystig]

Re: Overengineering

Hmmm now I'm puzzled - how can you not get the question?

BS1363 is a British Standard. Are you saying that the same plugs/sockets are installed in Australia?

[Shocker]

Re: Overengineering

No, I just happened to get my hands on an entire bag of the cords (as well as CEE 7/7).

[dumpystig]

Re: Overengineering

Ah, right.

Yeah, it is a very good standard, the plugs/sockets are polarized and are very safe especially compared to the basic plugs used in most of Europe.
The only drawback is the size/shape of the plug - it's very awkward and makes it hard work when packing in a small case for a portable device.

[budm]

Re: Overengineering

Since the UK plug has built-in fuse, do they install the Time-delay, or fast-blow type fuse, current rating?
Is the Neutral also bonded to Ground at the panel?

"North American type IEC cable"
By the way, NEMA does not set the standard for the IEC.
http://www.iec.ch/

[Shocker]

Re: Overengineering

If you want to be formal, it's a "NEMA 5-15P to IEC C13" cable.

But you get the idea.

EDIT: Anyone willing to bring this thread back on topic???

[tom66]

Re: Overengineering

Some of my mentality:
- LDOs/Linear regs not allowed to exceed 60C case temp under normal conditions
- If an LDO would dissipate more than 1W, consider a switcher.
- Axial power diodes: 2A max, derate 50% from datasheet. Otherwise you NEED a heatsink. Vbrr 90% of maximum.
- Power MOSFET not to exceed 80C case (110C for 150-175C rated devices), Vdss 90% of maximum (to allow for ringing at drain for inductive loads.)
- If a resistor needs to dissipate more than 5W in your design, you are probably doing something wrong.
- If you make something idiot proof, the universe will invent a better idiot.
- Fuse stuff above 50V or where available power exceeds 50VA. (Except in certain circumstances where short circuits are rare or can be guaranteed to not happen within design parameters.)

[Shocker]

Re: Overengineering

I think 90% of reverse breakdown voltage is too close for comfort. Remember that reaching the breakdown voltage will lead to catastrophic failure. I'd aim for ⅔.

If you make something idiot proof, the universe will invent a better idiot.

++

[tom66]

Re: Overengineering

I think 90% of reverse breakdown voltage is too close for comfort. Remember that reaching the breakdown voltage will lead to catastrophic failure.

Not catastrophic, no. It will lead to breakdown of the PN junction which is reversible if pulse energy limits are maintained. Usually it will cause excessive heat in the diode and lousy efficiency in certain types of converters (a flyback, for example, would discharge through the transformer, and back into the primary snubber, I think.)

[Shocker]

Re: Overengineering

And if the dissipation is high enough, and you keep it up for long enough...



Always best to play it safe.

[tom66]

Re: Overengineering

Since the UK plug has built-in fuse, do they install the Time-delay, or fast-blow type fuse, current rating?

There are five current ratings for the fuse: 1A, 3A, 5A, 10A and 13A; but the most common are 3A, 5A and 13A. The fuse is intended to stop a ring circuit overload or a plug overload which would not be stopped by a breaker.

Is the Neutral also bonded to Ground at the panel?

Nope, but it is bonded at the substation or sometimes the electrical pole. In our house earth measures 6VAC relative to neutral, so the bonding is probably very close.