relays

I see nothing ambiguous about the NEC specs. The only thing to be careful about is the second graph, load current vs ambient temp. If I wanted to push this envelope, I'd be sure to heat sink the output leads to copper pours; that would reduce the fet T_ja by a bunch. These tiny parts can't dissipare much heat from their cans alone.

The application where we're using them, an RTD simulator, the current is low; we're buying them for their size and low "on" resistance.

John

Resorting to personal attacks -- again -- noted. You don't have facts or logic on your side, so you insult.

John Fields just did the math, and your only response was to quote 333 lines of him patiently going through all the numbers of a real-world example with a one line personal attack. You don't have the math on your side, so you insult. Why should I go through the math and risk making an error that you can pounce on when John already did a better job at it than I would have and your only reply was handwaving?

Perhaps they can, but *you* can't. You started this thread by calculating power gain using joules (a unit of energy, not power) and got the infinitely wrong answer of "the power gain of a latching relay in infinite". It isn't and you know it.

So, you concur with JF's calculations? You agree that the power gain of the NEC part is about 280?

John

Ah. I now see where I made a bad assumption. Thank you for correcting my error; I really do appreciate it. While I was correct in stating that a VA rating would by silly in the case of an open contact (A=0, so VA=0) or a closed contact (V~=0, so VA~=0), I failed to account for the case where, at the moment of opening, there can be an arc that has a non-zero voltage and a non-zero current. I also failed to account for the case where, at the moment of closing, some small area of the contact connects first and full contacty happens a very short time later. And, of course, all it takes is some contact bounce to give these effects multiple chances to erode or weld the contacts...

None of this helps John Larkin, of course. He specified a steady-state rating, not a switching rating ("The non- latching relay equivalent for steady-state power gain would seem to be rated contact VA over coil power.") and later claimed that ("relay contacts are usually rated in VA").

Again, thanks for the correction.

I said that "some" relay contacts are rated in VA. I also said that if they're not, a simple mathematical operation will disclose their VA rating. What's all this exotic hand-waving evasion when all we need is a simple power gain calculation?

And all the mumbo-jumbo about arcing and bouncing is silly. The VA rating is that of the LOAD that the relay can control, which is the product of LOAD voltage and LOAD current when the contact is closed. The relay needs to be able to conduct the current and break the voltage. Of course the *contacts* don't see their max rated (closed) current and their max rated (open/switching) voltage at the same time. That's just silly.

So, do you agree with JF's 280 watt result or not?

"Math phobic engineer" is a very strange concept.

John

I concur with his calculations after any errors that are pointed out are corrected. Which, by the way, he seems to have no problem doing. So far he has only made the one small MOSFET voltage drop error that he instantly corrected when he noticed it (I note that you and I both missed it), while you are refusing to admit far more serious errors, many of which are real howlers.

Having very little interest in red herrings such as a solid state relay being inserted into a discussion of latching electromechanical relays, I have not bothered to do the calculations myself, but when I see on one hand an answer from a careful engineer who corrects any errors he or anyone else finds and on the other hand someone who refuses to admit any error even while confusing power with energy, infinite with unbounded, and joules with joules per second, I know who I am most likely to believe.

So the post with your name on it that claimed that "relay contacts are usually rated in VA" was a forgery?

They quit having gain after they limit? Then what are all those parts for?

John

^^^^

Oops, power gain of course.

Well, do you agree?

John

you

never

have

you

Translation anyone?

John

Exactly. You can't do this very simple math, which should take seconds.

That's fine; but why do you and JF *claim* you understand the math, when clearly neither of you does?

The 280 number is preposterous.

I've known a few math-phobic engineers, and a lot of math-phobic techs. The good thing about a university engineering education is that they force you to do the math. I had instructors who gave zero credit for solving a problem and getting the wrong numeric answer, no matter how "correct" the theory was. Rightly so.

The phobia is easily overcome. I think one day I'll write a book that will help.

John

--- Nope, just an error.

What's _really_ preposterous... absurd, even, is someone's refusing to own up to an error when it's as plain as the nose on their face that they're dead wrong. Someone like you, for instance.

The right number is 1418, BTW.

JF

Still disagree, but you're headed in the right direction, in, interestingly, about 4:1 steps. But I'm waiting for Guy's calculations.

John

--- Instead of playing your stupid little troll games, if you see something you think is wrong, why don't you just say what it is?

Or even better: Do it the way you think is right and post your results.

JF

Well, he's been savaging me for my many errors. And he agreed with your past, g=280, number.

Let's see if he can work a calculator.

This doesn't matter, really. The power gain of a relay is just a vaguely amusing number, although the SSRs are impressive. What interests me is how many electronic designers won't or can't do simple math, or won't sanity check their work. THAT matters.

John