PC 104 - Low temp specs
Qsine Corporation Limited
I've been trying to find PC104 controllers and modules for outdoor use in
the extremes of Canada. Most product I've found is only rated from 0-55°C
(some down to -20°C and some up to 70°). Can anyone point me to a vendor
with good low temperature h/w?
TIA
Kev
James Meyer
<qs...@cadvision.com> wrote:
>I've been trying to find PC104 controllers and modules for outdoor use in
>the extremes of Canada. Most product I've found is only rated from 0-55°C
>(some down to -20°C and some up to 70°).
There is a big difference between a system *rated* for a wide
temperature range and a system that will *work* over a wide
temperature range.
I suggest that you find a system that meets your needs at room
temperature in regards to performance, features, and price. Then test
the system over the range of temperatures you're concerned with. You
will probably find that many systems that are not rated for the
temperature range will continue to work just fine.
Jim
Bill Sturm
>
> I've been trying to find PC104 controllers and modules for outdoor use in
> the extremes of Canada. Most product I've found is only rated from 0-55°C
> with good low temperature h/w?
Have you tried Winsystems or Octagon systems, I believe that they both
make boards with extended temperature ranges.
mk
James Meyer wrote in message <3746c87...@netnews.worldnet.att.net>...
> There is a big difference between a system *rated* for a wide
>temperature range and a system that will *work* over a wide
>temperature range.
>
> I suggest that you find a system that meets your needs at room
>temperature in regards to performance, features, and price. Then test
>the system over the range of temperatures you're concerned with. You
>will probably find that many systems that are not rated for the
>temperature range will continue to work just fine.
Without wishing to be contraversial I would suggest that this is a very
dangerous practice.
We have found that some processors (MC68HC11 variants) rated 0 to 70 fail to
start up correctly at -5.
Electrolytic capacitors often perform very poorly at low (less than -20)
temperatures and this can cause the decoupling to fail. The solution is to
use more expensive parts rated for the full temperature range.
It is common for component manufacturers to select out the better performing
parts from time to time and this can mean that a batch of 0-70 rated parts
may contain many parts that work to -40 and yet annother batch may contain
none at all.
If the application is in any way importatant it is much better to use parts
actually specified for the conditions of use.
Michael Kellett
martin lytz
and you try that you are asking for trouble. Many design "features" crop up
when exercising hardware over voltage extremes. Start out with rated hardware
first then uncover those bugs first.
good luck
James Meyer wrote:
> On Fri, 21 May 1999 22:29:50 -0600, "Qsine Corporation Limited"
> <qs...@cadvision.com> wrote:
>
> >I've been trying to find PC104 controllers and modules for outdoor use in
> >the extremes of Canada. Most product I've found is only rated from 0-55°C
> >(some down to -20°C and some up to 70°).
>
> There is a big difference between a system *rated* for a wide
> temperature range and a system that will *work* over a wide
> temperature range.
>
> I suggest that you find a system that meets your needs at room
> temperature in regards to performance, features, and price. Then test
> the system over the range of temperatures you're concerned with. You
> will probably find that many systems that are not rated for the
> temperature range will continue to work just fine.
>
> Jim
martin lytz
James Meyer
wrote:
>James Meyer wrote in message <3746c87...@netnews.worldnet.att.net>...
>>
>> I suggest that you find a system that meets your needs at room
>>temperature in regards to performance, features, and price. Then test
>>the system over the range of temperatures you're concerned with.
>
>Without wishing to be contraversial I would suggest that this is a very
>dangerous practice.
>
>If the application is in any way importatant it is much better to use parts
>actually specified for the conditions of use.
>
>Michael Kellett
Parts are specified only after testing. Whether this is done
by the manufacturer or by the user is mostly a matter of economics. A
particular part may be perfectly able to function over a range of
conditions that were not tested for or guaranteed by the manufacturer.
Putting too many constraints on manufacturer guaranteed
specifications will result in fewer parts being available and a
greater cost for those that are.
Engineering of any sort is always a game of "trade-offs".
Using *parts* guaranteed to meet some specific conditions is no
guarantee that the whole *system* will perform satisfactorily. Only
good engineering will accomplish that and testing has always been a
part of good engineering.
It was once required that systems manufactured for the
military use only "MIL-spec" parts. Those requirements have been
relaxed lately so that "ordinary" parts can be used as long as the
system still functions properly.
Jim
Ivan Baggett
<qs...@cadvision.com> wrote:
>Hello!
>
>I've been trying to find PC104 controllers and modules for outdoor use in
>the extremes of Canada. Most product I've found is only rated from 0-55°C
>(some down to -20°C and some up to 70°). Can anyone point me to a vendor
>with good low temperature h/w?
>
>TIA
>
>Kev
>
>
We have a PC/104 SBC called the Chickadee XL which was tested by one
of our customers to -40 C with no problems. Of course, it is rated
for 0 C. What you will find with most SBCs is that the parts (chips)
are not available to go down to low temperatures. That is why most
products you see don't span a wide range. With Pentium systems, the
top end is not very high since the chip itself runs very hot even at
room temperature.
When a company does rate its systems for low temperatures, it usually
operates the chips outside of their rated temperatures. This is not
as bad as it sounds. As long as the oscillator can start up when
cold, things will be OK. CMOS actually works better at low
temperatures.
Another thing to watch out for is the performance of any on-chip PLL
circuitry. This can be adversely affected by low temperatures.
You can check out our products at www.bagotronix.com.
Best regards,
Ivan Baggett
Bagotronix Inc.
Christian G. Taranti
James Meyer wrote:
You are somewhat right. At least in the Brazilian Air Force, all the parts used
are usually rated for the MIL spec (at least for embedded systems that go into
the planes). If you have one of two parts, that usually means that every batch
of your components must be tested for climatic cycle. That you need to take a
sample of your product for every batch of components and put it for two or three
days on a climatic chamber changing from freezing to boiling. That sometimes is
considerably expensive and the approach normally used is "use MIL parts".
It is also very hard to certify embedded avionics and weapons if you are not
using MIL parts. That's because the tests that were done by the fabricant of the
part will now need to be done in-house.
Even so, I (myself) can see that the automobile specs are possibly going to
takeover some of the MIL market. The automobile range is even more strict that
the industrial. But not as wide as the MIL.
Christian
mk
James Meyer wrote in message <37470db3...@netnews.worldnet.att.net>...
> Parts are specified only after testing. Whether this is done
>by the manufacturer or by the user is mostly a matter of economics. A
>particular part may be perfectly able to function over a range of
>conditions that were not tested for or guaranteed by the manufacturer.
I feel I must come back on that one. Some parts are specified after testing
but others are not - some examples:
Intel's multi level Flash parts are rated -20 to +70 , the technology just
is not reliable outside that range
Many electrolytic suffer a dramatic increase in ESR alt low temperatures
If you decide to test a part yourself you should check with the manufacturer
that there is no fundamental physical reason that it will not work - some
manufacturers are more helpful than others on this.
Remember that the parts you need may have been selected out. We have
observed distributions of filter characterisitics in production with two
peaks - this is because we used 10% capacitors. All our production was
within spec but because we working for a big automotive customer we had to
explain the strange statisitical distribution of our production. It turned
out that all the 5% tolerance capacitors were selected out by our supplier
which we confirmned by measuring the values of the parts we recieved.
MIchael Kellett
Daniel Norton
of a heater or insulation that retains the heat generated by the components
themselves? Maybe the heater could be focused just on the components or
sub-components that have problems operating at extremely low temperatures
(e.g. the oscillator), or placing "hot" components near those that need the
heat.
Wayne Johnston
The problem with heaters is they consume energy. In a remote application
powered by solar panels, heaters raise the cost of the system (larger panels
needed) or make it impossible to power. The same applies to a battery
powered device.
The problem with insulation is that it is always there. It's not a problem
when it is cold, but when it's hot any additional insulation might make the
device overheat.
In much of Canada, the specification -40 to 60 C (-40 to 140 F) is not that
different from the actual working environment. You can commonly expect an
unheated, uncooled and unvented enclosure sitting outside to experience -35C
to 55C (or more) over the course of a year. Adding a sun shade helps with
the high end a bit.
Although testing a device with a smaller temperature range is feasible,
generally this is a job you want to leave to the manufacturer.