Automotive Circuit Protection - Part 3

[Garfield Willis]

Yeah, of *course* there's a Part 3. What good's all the blather if you
can't USE it? Ahem, that's what we'll try to cover, or at least get
started; this is where the rubber starts to meet the road.

We need to retrace our steps a bit, tho; way back in Part 1, we said:

>they [transorbs] are there to protect the load
>circuitry just long enough to call in for heavier reinforcements like
>fuses or breakers if needed. Plus, they're there to absorb fast
>transients that are too SHORT to ever actually trip a fuse/breaker.

Also, in Part 2, I mentioned in another vein, that another aspect of
this "distributed" protection you might have to consider is that:

>even tho devices [+5V regulator] like this have been designed with automotive in
>mind, if you also pay close attention to the NUMBERS in the spec, you'll
>see that in some cases the protection may still be barely adequate.
>... you want to USE
>these devices because they are hardened against these kinds of threat,
>but you ALSO probly might want to think about still having a first line
>of defense with transorbs in front of them.

These two issues are the first we'll combine together in dealing with
how to PLAN and SIZE protection for your precious DIY goodies. The first
above is of course, how to choose the proper fusing/breakers to BE that
"second line of defense" (you'll see it's not quite as trivial as you
might first think), as well as choosing the Transorb to both work well
with the fuse/breaker, AND work well all by itself, when the transient
isn't going to be severe enough to trip fuse/breaker.

The second issue that's also involved in the above selection process, is
to take into account the level of protection ALREADY built into the
loads. All Transorbs have a "range" of voltage they work within, a first
voltage, often called the "standoff voltage", slightly above which they
are sposed to begin operating/clamping, and a second (higher) voltage
most often called the max clamp voltage, which is the voltage across the
transorb when it's absorbing the max transient energy, which is referred
to as the "peak impulse current".

OK, let's have an example. Let's say our device is a +5V regulator on
one of our circuit boards, that's connected to +Batt bus, and that it
has written in it's spec, "load dump protected to 40V" [hardly adequate
"load dump protection" all by itself, eh?]. We'll pick a whopper
transorb for overkill's sake, a 15KP device (numbered this way because
it's rated at 15KWatts of Peak Pulse Power; don't worry about all these
terms just yet). A 15KP22A has:
	standoff V 		= 	22V
	max clamp V	=	37.1V

In other words, this device won't start to work until the transient goes
above 22V, but when it's giving its "all" in defense of its clients, at
its peak pulse current, it will STILL keep it's loads protected to below
37.1V.

Now can you see why I picked this device? Because you at least must get
the max clamp V *below* any protection levels that your loads can
supply. In this case, the mythical voltage regulator is self-protected
to 40V, so we just have to make sure we're UNDER that voltage at the
transorb. This of course isn't much of a safety margin, but you can't be
taking the standoff V down much lower than say 17V in a 12V auto system.
So picking a 15KP18A might be a good tradeoff:
	standoff V		=	18V
	max clamp V	=	30.9V
This is a better choice for another reason; the peak impulse current
goes UP as the max clamp V comes down. So essentially, your job in
choosing the "right" one, is if money and space are no object, get the
biggest monster transorb you can find (15KP is about the largest easily
available currently), then choose a standoff V that's a reasonable
margin above your system's normal working voltage (in the case of
automotive, assume 15V as a max normal working voltage). Pick a transorb
with a standoff voltage equal to or greater than this working
voltage.Then just make sure that all the devices you're protecting with
the transorb, are self-protected to at least a couple/several volts
ABOVE the max clamp V.
	
What's all this gonna cost ya? Well, the smallest 15KPs I've seen, in
axial lead package body, is 0.3" long and a whopping 0.375" in diameter.
Also available in a molded rect. package, a huge 1.00" X 0.65" X 0.4".
That's pretty bulbous, either way; are you gonna have room? They're
anywhere from $5 to $10. Worth it? And they're NOT easy to source. Worth
the trouble?

Here's the scenario we're working from: the 15KP18A transorb we've
chosen intercepts and absorbs transients, allowing our delicate load to
see no more than 30.9V at the peak energy level of the transient. As
long as the transient is short enough that the transorb's PPP (peak
power rating) isn't breached, the transorb is fine, the transient has
been absorbed non-destructively, and the fuse/breaker will be blown or
intact, depending on how we sized it (dealt with in Part 4). IF the
transorb's energy rating is overrun before the fuse or breaker has time
to "clear", as long as the transient isn't enough to explode the part,
the transorb will fail SHORT across the bus, and THEN blow the
fuse/breaker. The transorb in this case has died to save its client. In
the unlikely case that the transient is huge and explodes the device,
the transorb will fail open instead. Hopefully the voltage of the
transient isn't large enough to jump any air gaps, and the valuables are
still safe. IF we've been able to size our fuse/breaker properly so that
it trips before the transorb's energy rating is overwhelmed, then both
transorb and client are still intact. The fuse/breaker will be
blown/tripped, but nothing else is harmed.

That's probly enough for now; in Part 4 we'll see if we can get thru the
fuse/breaker selection, and see how that all plays together with the
transorb and it's loads. Pretty much a pain to have to protect against
an event you hope will never likely occur, eh? Trouble is, the downside
risk is so horrendous.

Gar


----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org