Solid State Fuse

[Jim Thompson]

[snip]

I think you might find that Middlebrook had something to do with it.......
Mind you I have no idea about the historical perspective.

http://ardem.com/

Ooooh

Input Filter Considerations in Design and Application of Switching
Regulators
IEEE Industry Applications Society Annual Meeting, 1976 Record, 366-382
(IEEE Publication 76CH1122-1-IA).

BTW, who the **** is Fred Lee? Perchance.... does he take Middlebrooks
explanation of the problem and generate shitloads of IEEE papers about how
to solve it with an AVR implementing an appropriate PID controller in
suftware?

DNA

DNA

Good man, Middlebrook. Met him once when he spoke to the MIT Club of
Phoenix.

My LoopGain simulation element (free to download, on my website) is an
adaptation of his paper on measuring loop gain in the lab without
perturbing the loop.

...Jim Thompson

 

[Fred Bloggs]

I think you might find that Middlebrook had something to do with it.......
Mind you I have no idea about the historical perspective.

http://ardem.com/

Ooooh

Input Filter Considerations in Design and Application of Switching
Regulators
IEEE Industry Applications Society Annual Meeting, 1976 Record, 366-382
(IEEE Publication 76CH1122-1-IA).

BTW, who the **** is Fred Lee? Perchance.... does he take Middlebrooks
explanation of the problem and generate shitloads of IEEE papers about how
to solve it with an AVR implementing an appropriate PID controller in
suftware?

DNA

DNA

Get with it, limey.
http://www.ece.umr.edu/~power/PEEC/lee.html

 

[Genome]

Get with it, limey.
http://www.ece.umr.edu/~power/PEEC/lee.html

OK, thanks....(not really, in the nicest possible way) I did have a poke
about and as far as I am concerned Lee is one of Middlebrooks students.

Anyway, assuming you are an American and at least five hours behind then
it's stil Jiff Lemon Day so go fry one off for yourself.

DNA

 

[Genome]

The rocket part of it is being handled by people who are experts in
that aspect, and
everything happens with a view to safety first, and then success.

There is a review process to get permission to launch something, and we
will be in the middle
of a desert probably. I think some kind of horizontal excursion is
out of the question, as the rocket will be equipped to recognize
attitude with substantial reserve resources, and made to cancel the
flight in any such event.. and at any time... despite my confidence in
my immediate group, I remain convinced that the more critical control
funcitions are going to work precisely because I'm the fall guy for the
works, no matter who designs which parts.

Serious rocketeers always think first about safety of others.

Also, rocketry is a subject unto itself. A casual observer would have
little idea of physical operation and what goes into development.

The basic process is a simple execution of events based on sensing.
What I'm adding to it is
failsafe operation, and some calling of the tune for the band.

These guys are very clever. Some RF, some software, analog, digital,
math, physics, and then most are very strong in technical electronics.


It stands to be a significant achievment in both global networking and
amateur rocketry, and could influence future directions in satellite
deployment methods.

Ok, 400 miles, will do.

Plus, they also seem to test their rockets upside down.

DNA

 

[Fred Bloggs]

OK, thanks....(not really, in the nicest possible way) I did have a poke
about and as far as I am concerned Lee is one of Middlebrooks students.

Well of course he is- that VPEC thing in Blacksburg has been going for
at least 15 years now that I recall, and just one of several high tech
electronic applications oriented research centers for the University.

Anyway, assuming you are an American and at least five hours behind then
it's stil Jiff Lemon Day so go fry one off for yourself.

DNA

Look up "baked pancake", make one, and don't use less than a whole bar
of butter. Then tell me about Jif Lemon pancakes.

 

[Genome]

Well of course he is- that VPEC thing in Blacksburg has been going for at
least 15 years now that I recall, and just one of several high tech
electronic applications oriented research centers for the University.


Look up "baked pancake", make one, and don't use less than a whole bar of
butter. Then tell me about Jif Lemon pancakes.

That sounds like some excuse for toad in the hole where you fill the middle
bit with shit to cover up the fact that it deflated and went soggy.

4. Bake for 15 minutes at 450º; lower oven temperature to 350º and bake for
another 10 minutes. It will have puffed dramatically but will soon settle
down a bit, so show it to your guests quickly so they can exclaim with
delight. Also, the edges end up higher than the middle, making a crater
perfect for any toppings of fresh or sauteed fruit or this simple fix:

First hit on Google, are you telling me Google is wrong?

DNA

 

[John Larkin]

That sounds like some excuse for toad in the hole where you fill the middle
bit with shit to cover up the fact that it deflated and went soggy.

4. Bake for 15 minutes at 450º; lower oven temperature to 350º and bake for
another 10 minutes. It will have puffed dramatically but will soon settle
down a bit, so show it to your guests quickly so they can exclaim with
delight. Also, the edges end up higher than the middle, making a crater
perfect for any toppings of fresh or sauteed fruit or this simple fix:

First hit on Google, are you telling me Google is wrong?

DNA

A thin puff pastry, with lots of butter and curled edges? Mo makes
those and calls them Big Fat Dutch Babies, don't know why. They're
good with powdered sugar and fruit, especially sauteed bananas and
whipped cream.

Somebody just discovered that butterfat is good for you.

John

 

[ehsjr]

Going to space for 1 minute, effectively.. but will be low atmosphere
for 8 minutes or so.. I'm not going, but our project is.
It's not a theoretical problem.

The problem with a polyswitch is it resets after the fault is cleared
and power is removed. I need something that will reset automatically,
the idea is a nut will rattle loose and short something, or similar
event, and the circuit needs to get power restored
so I'd imagine a device that would cycle it in an out often (but not
RF often).

Imagine a can with some circuits in it, a battery, and some pocket
change, a few washers, nuts, and. Between say three circuits, a
minimum of two of them have to function to execute an event upon
certain condition or time greater than Tx. As the can is rattled,
wasker, nuts, screws, change, and a dash of fire and brimstone interact
with the circuits occasionally only at the connectors to the boards.
The electronics are put inside, the switch turned on, and the can is
shaken vigorously for a minute, during that minute, the 'mains' have to
be protected from total failure, and certain events have to take place
irrespective of individual or temporary failures.
We haven't worked short-proof into the scheme yet, and can't risk
melting the mains while assuring that anything that can use power gets
it. It's probably the most obsessive project I'll ever be involved
in, it has to work on absolute terms, even though we aren't sending
people up. If we don't have a clean power protection solution we
have to go with hard wire and crossing fingers, and I don't want to go
that way, though it merits consideration.

The specific heat of the chips is often adequate, but at any
appreciable power (>100mW) potting or heat sinks will be used.
There is another cooling idea.. we might bring along a gas cannister.
The trick is running the transmitter in the vacuum, when it is needed
most.. or highly desired.. at some 20 watts.. but that will be dealt
with.


Thanks,
Geoff

I have no idea if this will work for you - this just
allows the poly to reset if it opens by insuring
no current through it when the relay drops out.
You need a delayed on relay for a 2 part solution.

Vcc-----+----POLY-----+----- N/O
| | ^------------ To Circuit
+----- N/C |
^--------+----[DelayRly]---Gnd

I don't know if either the poly or your circuit
can take the possible "abuse" this might cause.
(A "bolted" short comes to mind ...)

Ed

 

[Winfield Hill]

Rene Tschaggelar wrote...

Have a look at the LT1153, a PMOS control circuit that
gives it the PMOS fuse like characteristics.

You meant LTC1153, right? Actually, that's a CMOS chip
to control N-channel MOSFETs, with a built-in gate-bias
generator, and a "Programmbale" Auto-Reset Time.

I note a complete LTC1153 circuit takes at least 6 parts,
more if a big lamp or large capacitive load is driven.

The newer LTC4213 eliminates the current-sense resistor,
by sensing the voltage drop across the FET's Rds(on),
which is a cute idea to save parts. The LTC4213 also
has a faster shutoff than the LTC1153, however it's only
rated to work up to 9 volts, compared to 18 volts.

 

[Fred Bloggs]

Since there are so many technologically informed people here, I'll try
an interesting question.

What kind of device could be contrived to make a self-resetting (after
say, 3 seconds) circuit breaker? It has to be 'solid state' - no
moving parts, handle a few amps of current at 10 volts or so, and work
1000% of the time.

A simple, unstoppable power switch that functions only to interrupt
current for a few seconds upon overload.

What comes to mind is a mosfet with some kind of optoisolator /
comparator setup.. involving a button cell to power the opto.. but this
is the naive view.

I'm thinking of a device which can be spliced into a DC carrying
wire... a self-resetting circuit breaker. It should cost less than $10
to implement in parts,the big requirement is that it could never have
any failure modes except when hit by lightning, so to speak. It has to
stay cool at 1 to 10 amps, in a vacuum.

No moving parts.

Geoff

Your description of the application implies a simple fixed current limit
with duty cycle monitor for pass element over-heat protection. And this
duty monitor will consist of, and get this, a three component analog
computer, aka R-C network driving a comparator with hysteresis, that
simulates the constant rate of heat dissipation into the pass element
thermal mass of fixed capacity. Said monitor then trips the pass element
OFF to ON or LIMIT and back to OFF, as dictated by operating conditions,
attempting to maintain the pass element junction temperature at Tj,max.
This approach will be then be optimal in the sense that it is minimally
adaptive to an otherwise unpredictable intermittent fault set. If
battery charge consumption is a consideration then there are any number
of ways to work this into an analog computation also, say by short term
averaging of the fault events and charge gauging combined with a simple
time-into-flight prioritization of the instantaneous current limit,
which need not remain constant as either a threshold or limiting cycle
parameter. But if all you can think of is "fuse" and "no moving parts" (
damned joke that one)- go back to reading your popular mechanics magazines.

 

[Mark]

For an incandescent, more voltage always makes more current, at least
until it burns out. For the polyfuse, more voltage makes less current
in the active region, approaching a
constant-power/constant-temperature mode, so there's incremental
negative resistance, like a tunnel diode but a lot slower. As you
point out, it must have a negative resistance region to be bistable.

So, is it possible to make an oscillator from a polyfuse?

John

John...

yes, I think you are correct, the polyswitch in the active region
tries to maintain a constant temperature therefore constant power, so
as the voltage across it goes up, the current through it must go down
which indicates it is a negative resistance (over a long time constant)
.. So I agree, even if the load is a straight constant resistance,
after an overload is removed, the polyfuse can latch and hang and fail
to reset until the power is removed i.e. the system IS bi-stable.
It's the same as foldback current limiting. Thank you for pointing
this out about polyfuses..., I never realized it before.

Mark

 

[engineer]

hehe.. well, no moving parts meaning I can't have relays in a rocket
because the vibration possibilities are wide, also I was thinking
someone would tell me to use an automotive flasher and I didnt want to
see that one.

For the record, I can take about 5 seconds of popular mechanics before
I put it down.. but I'm not sure why you would offer some good
suggestions and then in following suggest I read a low-grade magazine,
to each their own I suppose there is some general rule in force to
take part in online hazing or be subject to escalation by the more
brainwashed marginals addicted to it.

My conclusion is I'll have suggestions for the guy most likely to
implement the DC-DC converters and approach to distribution, and he
will do what he thinks is best.

 

[Mark]

John
Plot the VI curve for the load and for the poly, if they cross at one
point, then the system is monostable will RESET without removing the
power. If they cross at 2 or more points then the system is bi-stabe
and can hang up as you describe. But for the load lines to cross at 2
points, one has to exhibit a negative resistance region. Does the
poly exhibit a negative resistance region? I don't think that it
does... like an incandescent light bulb, the resistance changes but
it never goes negative.

Mark

Correction..

I should have said negative INCREMENTAL resistance...
Mark

 

[Rene Tschaggelar]

Rene Tschaggelar wrote...



You meant LTC1153, right? Actually, that's a CMOS chip
to control N-channel MOSFETs, with a built-in gate-bias
generator, and a "Programmbale" Auto-Reset Time.

I note a complete LTC1153 circuit takes at least 6 parts,
more if a big lamp or large capacitive load is driven.

The newer LTC4213 eliminates the current-sense resistor,
by sensing the voltage drop across the FET's Rds(on),
which is a cute idea to save parts. The LTC4213 also
has a faster shutoff than the LTC1153, however it's only
rated to work up to 9 volts, compared to 18 volts.

Yes, the LTC1153. Some chips have the "C" others don't,
at random. The LTC4213 is specified to 6V only.
I wonder if it could be lifted with a Zehner underneath
or such.

Rene

 

[Terry Given]

I think you might find that Middlebrook had something to do with it.......
Mind you I have no idea about the historical perspective.

http://ardem.com/

Ooooh

Input Filter Considerations in Design and Application of Switching
Regulators
IEEE Industry Applications Society Annual Meeting, 1976 Record, 366-382
(IEEE Publication 76CH1122-1-IA).

BTW, who the **** is Fred Lee? Perchance.... does he take Middlebrooks
explanation of the problem and generate shitloads of IEEE papers about how
to solve it with an AVR implementing an appropriate PID controller in
suftware?

DNA

VPEC are busy puking out hundreds of papers on some nice ways to prove
the stability of smps loaded with many other smps - "distributed power"
or words to that effect. basically just measuring output impedance.

Yep, Middlebrook was there at the beginning, his seminal (now now, dont
be grubby) paper ought to be required reading.

Cheers
Terry

 

[Winfield Hill]

Rene Tschaggelar wrote...

Yes, the LTC1153. Some chips have the "C" others don't,
at random.

Not random. LTC's scheme is this: The LTCxxx parts are all
CMOS, whereas the LTxxx parts are bipolar. But LTC doesn't
reuse numbers between their LT and LTC parts, so in a sense
both prefixes are extra characters.

 

[Fred Bartoli]

VPEC are busy puking out hundreds of papers on some nice ways to prove
the stability of smps loaded with many other smps - "distributed power"
or words to that effect. basically just measuring output impedance.

Yep, Middlebrook was there at the beginning, his seminal (now now, dont
be grubby) paper ought to be required reading.

Did you read his book? Worth reading?

 

[Terry Given]

Did you read his book? Worth reading?

which book? I've got the stuff he wrote with Cuk, its pretty good
(although I personally find the equivalent-circuit approach annoying,
whats wrong with a little calculus). I've also got his book on
differential amplifiers

Cheers
Terry

 

[Rene Tschaggelar]

Rene Tschaggelar wrote...




Not random. LTC's scheme is this: The LTCxxx parts are all
CMOS, whereas the LTxxx parts are bipolar. But LTC doesn't
reuse numbers between their LT and LTC parts, so in a sense
both prefixes are extra characters.

Thanks that was informing.

Rene
*** Free account sponsored by SecureIX.com ***
*** Encrypt your Internet usage with a free VPN account from http://www.SecureIX.com ***

 

[Fred Bartoli]

which book? I've got the stuff he wrote with Cuk, its pretty good
(although I personally find the equivalent-circuit approach annoying,
whats wrong with a little calculus). I've also got his book on
differential amplifiers

I was once intrigued by what he calls DOA and kind of remember he was
planning to write a book about this.
But I just checked its website and just found some DVDs, so I must be wrong.
This DOA-low entropy... thing seems to me a lot of buzz words for things we
naturally do but I got curious about he had to say.