Self-heated computers for Polar countries & North/South Poles.

Even full blown military spec. (ground and air) equipment can have problems

at very low or high temps. Such equipment may have pre-heaters installed to

warm up the equipment to an acceptable minimum temperature before it is

fully energized, let alone operable. Such heating systems may (by spec.)

allow the equipment to be used in as little as 5 minutes, or in the case of

some precision equipment, several hours.

One of the more temperature sensitive computer devices is, naturally, the

hard drive. Very low temperature causes problems with the mechanicals,

notably the bearings and lubrication.

"Stan Starinski" <China@stealsUSJobsPatentsSoftwareMusicVideo> wrote in

message news:eFwqjEzqKHA.3908@TK2MSFTNGP05.phx.gbl...

>I thought even with a large amount of semiconductors in Controller it would

> not suffer in cold weather.

> I wonder if you actually witnessed such failures, in reality, not in

> theory.

> In fact cold temps improve reliability (heat is the biggest killer of

> electornics).

> I only refer to Controller in the above statement.

>

> However you're right that SCR may NOT trigger properly in very very cold

> weather - that is sometimes specified in datasheets, but is better tobe

> checked by an experiment. Everyone is concerned more with heat, than

> cold,

> so they add "derating" curves to datasheets to show how SCR dissipated

> power

> reduces (DEGRADES) with rising temperatures. Few manufacturers explain

> how

> their devices behave at the opposite extreme - subzero temperatures.

> Generally cold is good for electronics, also leakage current is

> exponentially dependent on temp (leakage is a great evil, MP manufacturers

> always try to minimize it); when I was at Telecom Corp. their

> switches were

> airconditioned to a point where people got sick if they spent a week

> there,

> when you walk out to the street, it's like a hotwave by comparison.

>

> Semiconductors change behavior versus temperature - not necessarily bad,

> but

> certainly a change which designer did not expect.

> So the question is to predict whether SCR Triggering on/off would work as

> expected, not with some extended delay.

> I don't think MP would suffer much, it's the SCR which could suffer.

> Here are possible solutions:

> =====

>

> a) This option will increase parts cost substantially:

> Use extended temperature range parts, usually designated military.

> Military

> does NOT mean it's rare and only sold to Dept. of Defense folks, it's just

> an industry term. There's also a so-called "industrial" rating but I

> prefer

> military. Those devices are sometimes rated for -60 or -40 (?) to +150 F.

> Microchip makes such extended version of your MP (PIC18F4420) - you can

> even

> see it listed at Digikey, Newark, etc.

> It can cost 10-20% more, and it's packaging is NOT plastic. It uses

> ceramic-metal-carbon compound, but it's STILL the SAME CHIP PIC18F4420.

> SCR's are also sometimes available in weather-rugged versions.

> As of resistors - they don't care for cold temperature, they actually are

> more reliable at low temps

> As of capacitors - those which are not "NP0/C0G" type, change their

> capacitance but it's only important for RF circuits, not for power

> controls

> Or you can always use slightly more expensive C0G/NP0 type of capacitors,

> which we use often in product, at Corp, fortunately they're

> rather

> cheap and not necessary anyway.

> The problem however could be with Quartz Resonator (crystal) - it may

> change frequency, if frequency rises that's bad because all delays become

> SHORTER, I hope datasheets specifies how frequency changes at -40F.

> So by using extended temperature devices, the problem of changed behavior

> in

> freezing weather can relieved - that's what Cable & Phone companie suse in

> their overhead cabling - see those CableTV barrels & boxes running along

> roads & highways? They use military style rugged components.

>

> But according to you that may not be enough? So you may still want some

> heating, OK:

>

> b) I am not sure about Mechanical implementation of heating - what heater

> to use, etc.

> But to turn it on when temperature is below certain level is no probelm in

> software. Just remember by running heater all the time in cold weather,

> it

> may kill power saving to a point wher customer may ne be happy. Heater

> draws a LOT of energy and you ened this in every box.

>

> Anyway:

> The problem is in SENSOR - same thermocoule that senses Overheating, may

> not

> be useful to sense the other extreme - "Underheating". because they're

> tuned to a certain narrow range, I guess?

> Therefore in the future a neat solution can be this:

> Get rid of that thermocouple. Don't use a separate input or sensor for

> Overheated or Underheated fault, by the way I don't have an available pin

> on

> MP anyway, even if you wanted it.

>

> Instead do this:

> Use a Thermistor as a sensor. Use a A/D converter channel in MP (many A/D

> channels are still available in our MP) and a piece of software that

> converts it into a digital number.

> The rest is obvious:

> Check if that number is below some level, if yes - turn on heater, if

> above

> some level - turn on bypass and alarm for Thermal shutdown. If that

> number is within normal range, do nothing.

> The Cold/Hot condition should persist for 10-15 minutes before action is

> taken, but I am running out of available Timers in this MP. A hardware

> timer is always necessary for such timed response.

>

> So I will let you know.

> Look, I have NO TECHNICAL PROBLEMS WHATSOEVER, my only problem since a

> year

> ago is MONEY!

>

> If I had enough money to focus on this exclusively, and get parts and

> build

> things, and just pay attention to it every day, we could have great

> results.

> But right now I am away half week and have to listen to from

> Corp because now he is paying some.... it takes attention away

> from and I told him, but you can't say no to him because who knows he

> might blow up

> into a real business.

> Don't worry, I just mention what I am dealing with in life

> There're 2 peoiple behind constantly making it a bit chaotic at home, so

> you

> can't work. You have to go away into some library or Long Island to use

> an

> an "office", at home is hell and I look for money every month....

> Otherwise

> would not leave my own apartment on Long Island., I used to live in a nice

> house in zipcode 11788 - one of the richest in USA. Now I joined American

> bombed-out formerly Middle Class...

>

> Anyway, I hope you can sell this version without a heater for now.

> In the future you just have to remove Thermocouple, replace it with one or

> two THERMISTORS and add a software code which converts Thermistor reading

> into some digital number, and act upon it.

>

> HAVE A GREAT WEEKEND, .

>

>

>

>

>

>

Merci for a very useful information.

As you might've guessed by post had nothing todo with computers, I designed

& keep improving an SCR Controller board, based on Microchip MCU and my boss

is worried about installing it in such cold/outdoor locations as (example)

Chicago street lights - in winter.

I can easily provide a logic signal to heater relay, by design. My problem

is more psychological - I can't stand wastign energy, because the product is

ironically about this - saving energey for HID Lighting. If I got a heater

pumping up some Kilowatts all night, entire winter, this will reduce savings

we're offering to customers.

It kind of defeats the purpose.

But OK I will provide a logic signal to run heater perhaps for 5 minutes

until temp reaches -10C or something.

> One of the more temperature sensitive computer devices is, naturally, the

> hard drive. Very low temperature causes problems with the mechanicals,

> notably the bearings and lubrication.

People report good results with Intel SSD's instead of a mechanical/magnetic

harddrive, and I expect it not only doesn't care, but in fact would work

BETTER at very cold temps. There're pluses & minuses of both types of mass

storage, as usual with anythign in life; but certainly mechanical shock &

weather-ruggedness is where SSD does not suffer.

In fact it would suffer at th eopposite extreme - very hot temperatures,

increasing leakage and decreasing data retention.

Don't need to be an EE Engineer as myself - it's obvious even from basic

Physics, warmer matter means more chaotic/faster "jerkings" of particles

including electonics, etc. which disturb crystal structure.

Yes cold can also render equipment unusuable (likewise one can die from

drinking excess water- everything must be in moderation). Still you

probably with my statement:

Heat is the biggest killer of electronics, and a reliability-oriented

designer is generally more concerned with Heat than with Cold.

Anyways,

My boss still wants a little heater... Jesus Christ, I will give that Logic

pin to him to control a heater, but it'll cut into HID Lighting energy

savings this product is ironically targeting!

I have a specific heating element in mind - currently still in R&D phase &

can't name the company, it's surface-sprayed powder, not a conventional

resistive tube/etc.. extremeley efficient.

Invented by an x-Raytheon Engineer. He was at Raytheon as a materials

scientist (alloys for warheads/missiles), actually, not EE, I met him, great

man, as usual - an immigrant (from Europe this time), now moved up north to

Canada cuz their market & $$$Funding is better for heating stuff....