On 6/15/2012 10:04 PM, Charles Harpole wrote:
> The components which "age faster" (citing you) at full load must age only due
> to increased heat, as per your logical answers.
The information is on the net and is fairly easy to find. I worked in
the industry for over 26 years starting when it was in its infancy. I
mentioned those Silicon wafers. When we stated they were between a half
and 3/4 of an inch across. Now they are 14" and over. The purity of
poly crystalin Silicon is now higher right out of the reactors then we
used to see on multi pass float zone material. Impurities are in tiny
fractions of a part per billion. BTW single crystal Silicon that was
close to intrinsically pure ran about $165 USD per gram. I don't have
direct access to prices any more but it's at least 3 orders of magnitude
less now days while the efficiency for making the material has gone way up.
Actually heat just accelerates the process. Molecular motion, which I
think Carl touched on, is the basis for the aging with impurities
tending to migrate across junctions. Even with heat this is a very slow
process ... unless there is way too much heat and then it's very quick,
but by a different process. Contamination of the junctions affects many
of the transistor's characteristics, but we rarely notice them as actual
aging because it's such a slow process. Their high frequency
characteristics may degrade a bit, but again it's such a slow process
we'd never notice it as such usually before the device fails.
Many "older!" devices which contained far more impurities than most
modern ones will age faster and even at ambient temps. Remember though
that we are still *usually* talking about thousands of hours.
Intrinsically pure Silicon has a resistivity of about 6,000 ohms/Cm^3.
Typically we think of contamination as lowering the resistivity, but
there is such a thing as "compensation" where both the N and P type
dopants work together to give a much higher resistivity.
>
> However, a well designed unit should have adequate cooling provided. I guess
> that the physical size of the transistor, for example, limits how much heat
> it can dissipate regardless of the cooling (not refrigerated, of course_).
Here the problem is the Delta T across and within the transistor. It
takes some exotic heat transfer compounds to get the heat from the
transistor to the heat sink, but only the internal construction can get
the heat from the actual semiconductor through the device to the surface.
The very small size of the Silicon die means we need a very high heat
conductive material that is in intimate contact with that die.
Again this information is on the net and easily available
Refrigeration helps, but only to a point. We need to have the cooler
very cold, but not to the point of creating condensation or frost. That
means the heat sink needs to operate in at least two modes. An idle
that keeps it as cool as possible before power is applied and then kick
into high gear as soon as power is applied. Unfortunately the
semiconductor device can heat a lot faster than the cooling can decrease
the interface temp.
>
>
> If that is true, and it probably is, then one can baby a RF transistor by
> running it at less than full rating. The likely result, however, is so
> small, as you say, that over a normal ham rig's lifetime (it being retired
> for reasons other than failure), the "saving" is trivial.
Agreed as long as not pushing the devices to near their limits.
When driving "the amp" I run my 5000MP in class A yet the heat sink
barely gets warm. Of course class A is the most inefficient mode so what
kind of life I get out of it is yet to be determined.
"I believe" and I emphasize that "believe" as I do not have access to
the actual data, that this is what was happening with the Icom finals in
the 7700 and 7800s. When pushing the limits the fans, heat sinks and
airways need to be kept very clean. A hot transistor is also more
susceptible to voltage and current spikes, (SWR, over drive,
compression, audio settings on the input, and a myriad of other items.)
As time goes on these devices change/improve (we are no longer using
bipolar transistors in finals) they will become much more rugged, more
than likely by become much more power capable so they are operated no
where near their limits while still giving good IM numbers.
73
Roger (K8RI)
>
> Thanks,
>
> Charles Harpole
> k4vud@hotmail.com
>
>
> ----------------------------------------
>> Date: Fri, 15 Jun 2012 00:25:24 -0400
>> From: k8ri@rogerhalstead.com
>> To: amps@contesting.com
>> Subject: Re: [Amps] Babying radios conclusion
>>
>> On 6/14/2012 11:49 PM, Charles Harpole wrote:
>>> >From all generous replies and my reading Google hits, I conclude that the
>>> >only way to "baby" a radio is to leave it always on or always off.
>>>
>>>
>>> As my msg say, I asked to eliminate considerations of "mfg defects" or
>>> "misuse" (and poor design that traps heat, etc.). These matters were what I
>>> specifically eliminated in my questions. I wanted to know if the owner
>>> could do anything to baby a well designed, correctly operated radio. The
>>> owner can control the heating and cooling cycles of the radio, which seems
>>> to cause problems, so I do have AN answer, but..............
>>>
>>>
>>>
>>> 1. If a tube is left on, under continuous proper usage and under correct
>>> load, what causes it to fail ? And, what is "soft"?
>>>
>>>
>>>
>>> 2. Same question for transistors.
>>>
>>>
>>>
>>> Especially for transistors because they are inert devices that do not need
>>> to heat up to work, and are made of sand.
>> Not exactly. Only half sand...IE, sand is quartz SiO2, The solid state
>> devices are generally made of either Silicon or Germanium although there
>> is a layer of oxide in there.
>>
>> Solid state devices do age. Older devices age more and faster than most
>> of the newer ones due to impurities in the Silicon.
>>
>> The transistor is created on a Silicon wafer. The wafer is heated and a
>> layer of oxide is formed on the face. With wafers of 14 inches in
>> diameter you can create a lot of devices with closely matching
>> characteristics which reduces the price considerably. Of course large
>> transistors will have a much lower yield per wafer than small ones.
>>
>> BTW every wafer created creates a 50% loss of Silicon from the sawing
>> and polishing operation.
>>
>> Junctions are created by etching a pattern in the oxide and then heating
>> the chips (many devices are usually created on one wafer) while passing
>> a gas containing the dopant material which difuses into the Silicon. In
>> an NPN transistor the base Silicon is doped with a material adding an
>> excess of electrons. Then the P type gate is added.
>>
>> With age, the dopant atoms slowly migrate across the junction which
>> contaminates the junction. Heat accelerates the aging process.
>>
>> This is not the only contributor to the aging but it's a sufficient
>> example and beyond here it gets quite complicated.
>>
>> So, yes transistors do age. Running them near their limits will cause
>> them to age faster. However with modern devices the aging *Usually*
>> takes many thousands of hours. With all the other components mixed in
>> there with them I doubt with normal use it would make little difference
>> whether you leave them on or switch them on and off. The solid state
>> devices are far more reliable than tubes and likely most of the other
>> components in the rig as well. OTOH they are far more sensitive and
>> less tolerant to over voltage or spikes of even short duration than are
>> tubes.
>>
>> 73
>>
>> Roger (K8RI)
>>
>>>
>>>
>>>
>>>
>>> Post script... Running a transceiver at less than its rated RF output DOES
>>> NOT "SAVE" it from wearing out, IF its heat dissipation is correct. However
>>> heat does seem to accelerate the degradation of many components. Too, most
>>> things made with plasticizers (like DVDs and some circuit boards) will see
>>> evaporation of the plasticizers by the nature of the chemistry of the
>>> materials. Chemical/physical interaction of dissimilar metals bonded
>>> together cause the metal lower on the Chart of Elements to gradually lose
>>> its molecules (why that action is not immediate is a puzzle) which is why
>>> antenna connections fail, for example, where copper and aluminum are
>>> joined. Another matter not under the control of operators is "cold flow"
>>> inside a curve stressed coax cable, for example, but it also takes time.
>>> The idea of slowly raising the voltage on a tube filament upon "turn on"
>>> remains a disputed area (to do or not to do, that is the question).
>>>
>>>
>>>
>>> So my questions 1 and 2 remain in my weak mind. But, I am sure right now
>>> that a ham operator can NOT "baby" his correctly designed and correctly
>>> operated radio after he turns it on and uses it.
>>>
>>>
>>>
>>> 73
>>>
>>>
>>>
>>> Charles Harpole
>>> k4vud@hotmail.com
>>> _______________________________________________
>>> Amps mailing list
>>> Amps@contesting.com
>>> http://lists.contesting.com/mailman/listinfo/amps
>>>
>>>
>>
>>
>>
>> _______________________________________________
>> Amps mailing list
>> Amps@contesting.com
>> http://lists.contesting.com/mailman/listinfo/amps
>>
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|