Comp science student and Embedded

Good point. In fact you might expand it to say familiarity with the objective business is also valuable. In fact, very little is useless.

OK, so I should have capitalized Post Office, and they did the phones as well. Was that intended as an answer to my question?

Thank you! _That_ answered my question.

... snip ...

I am having a problem envisioning the purpose of the moths. :-)

They provide the discipline obviously.

Obviously, that's how they teach them to use the debugger.

Please remember that "embedded systems" can be anything from a greetings card playing a favorite melody to dozens of 19" racks doing a single task.

Anyway, I would expect that anyone applying for embedded engineering jobs should be familiar with at least some degree with circuit diagrams. For industrial applications, you should understand how and why 4-20 mA circuits work. i.e understanding some degree of electronic theory.

Paul

What is a 4-20mA circuit?

The Lizard

In fact, get an EE degree in addition to the CS degree. Or perhaps instead of the EE (which at some schools may go into physics of silicon and assume you're going to run a fab or something), it might be better to get a less-well-known but growing-in-popularity "Mechatronics" degree.

For some projects, add 50% DSP to that. Do they cover DSP in CS thesedays? When I was in college I think only EE's learned about DSP.

It's a long-time analog signaling and control convention. 4mA represents the low end of your sensor or other analog value, and 20mA represents the high end. Since the value is represented by current rather than voltage, signal wiring losses aren't an issue under normal circumstances, and a broken wire is easy to detect (0 mA not being a valid value).

Mike

Thanks. Got it. A Google search turned up several useful pages.

Final question: the simplified diagrams don't seem to show the case when the sensor itself has some power requirements. How is that handled? Separate power supply at the sensor? Clearly, 4mA isn't enough to power anything practical unless you mean 4mA at 10kV or something. : )

After this comment, I would not expect anyone hiring You for an embedded position :-).

A _huge_ power for a sensor.

A few volts across a typical sensor system at 4 mA is quite sufficient to power the sensor.

In fact beaming up a few hundred mW using an optical fiber is enough to power a current transformer+converter hanging at 400 kV line in order to get the current transformer measurement back on an other fiber.

Paul

That's right, dude. I'm unemployed, unemployable, and I hang out on comp.arch.embedded.

OK, I believe you. It is just that an electronics genius like me -- on a good day -- is able to bias a blue LED without assistance. That is somewhere around 3 volts at 10mA. So, when I think 4mA, I think "not enough for a typical LED".

I know nothing about sensors ... I just imagined that there were some sensors that needed a bit of power (to heat a filament, or light a light, or whatever). Not 40W, but I could believe 10W in some applications.

That is interesting. I would have thought you could power it off the 400kV directly.

There was a farmer some decades back who powered a well pump or something like this by winding a bunch of copper and placing it below high tension lines that ran across his property. It resulted in court action by the power company. His argument was that the fields and waves and so on radiated onto his property and thus belonged to him. The power company brought in some physicists and engineers who explained that it wasn't so simple -- he actually was exerting a reciprocal influence and stealing the power company's power. In the end the judge ruled for the power company.

But I thought 400kV would be enough that you could do away with one of the two fibers.

The Lizard

At those voltages, you definitively do not want to power any low power systems between phases.

I have seen airplane warning lights close to airports hanging on 400 kV lines with takeout points only a few meters from each others on the same phase conductor.

Paul

Yes.. The Post Office did al the telecoms and much of the computing and their courses were among the best in those days. IT was not a correspondence course. AFAIK they only did them for their own staff. However they had a lot of technical people. The were one of the few centres in the UK and indeed globally, that did some of this work.

No guys - it's software so we should expect them to be trained to deal with bugs !

Michael Kellett

en

I've seen plants run 24 VAC which would be rectified and filtered and run through a DC-DC converter to provide power to devices. But as has been mentioned, lots of sensors can be found that just need the 4-20mA to run. The market demands them, so manufacturers make them.

Mike

Weren't there some Americans involved in the invention of computers?

I'm clearly confusing the science in CS with something that can be done without mathematics. Well, without sufficient for embedded systems, which is generally not that much.

Peter