[R-390] Old Brown Beauty Statistics Lesson

Wed Jan 12 22:42:03 EST 2005

Hi

A couple of thoughts about why the capacitors crack.

There are a significant number of electronic parts out there that get a 
plastic coating over them when they are manufactured. Some of the 
coatings are relatively soft and will stretch a little. Others are 
fairly brittle and don't stretch much at all. Most plastics are made by 
cross linking long molecules together. The more linking you do the more 
brittle they get and in some cases the more they shrink.

Other plastics are sensitive to humidity. A great example is Nylon. If 
you have ever seen nylon hardware that has been out in a hot dry 
location for a couple of years it's amazing how much the stuff shrinks.

I have personally been involved in a couple of *very* painful recalls 
on epoxy coated parts over the last 30 years. In each case the epoxy 
slowly shrank over time. The net result in one case was thermistor that 
popped open like popping corn. The net result in the other cases where 
tuning capacitors that went out of alignment and shorted out. I suppose 
that if necessary I can dig up the GIDEP's on all of them. In all cases 
these were parts used in military equipment. Rockwell Collins, TI, and 
Magnavox (Ft. Wayne) were the end customers.

Bottom line in each case:

1) The ratio of hardener and resin is pretty important in an epoxy. The 
normal methods mixing the stuff are surprisingly open to error. Is it 
equal parts by volume or by weight. Do you mix each can before 
measuring.

2) Regardless of the chemistry plastics continue to cross link after 
they are judged to be "fully cured". In some cases this is accelerated 
by humidity. In most cases it is accelerated by radiation. In all cases 
that I have been made aware of it is accelerated by heat.

3) Thermal cycling of the shrinking case wrapped around a hard inner 
core will also accelerate the cracking, crushing, or moving process.

So much for the intro, but this is a thread for the long winded.

Now for the variables:

Radio A:

Runs 24 hours a day in a tight rack for 20 years on board a ship. Nice 
salt content in the air even with the high temperature inside the rack.

Radio B:

Sits in a open air hut in far off land for 10 of those years and then 
is in storage in who knows where for 10 years. Used 6 hours a day 4 
days a week when in service. Radio is rack mounted, but the rack has 
wide open ventilation.

Radio C:

Run in the continental US, inside a climate controlled building. The 
guard guys run it two days a month for six hours. Radio is the only 
thing in the rack.

All of these radios have a very plausible carrier in the military. 
Certainly we would all like to get radio C if we could. I would suggest 
that a lot more radios fall in the A and B categories. Somehow I doubt 
we are arguing about radio C so we'll simply drop the "baby doll" 
radios at this point.

If you just look at time and temperature on radio A and radio B there 
is an enormous difference in what happened to each radio. This is not 
to say that is the only variable. It certainly isn't. Radio A saw 10X 
the time on power over the 20 years as compared to radio B. Common 
temperature measurements of the inside of the Navy racks put them up at 
about 60C or so. Best guess on the average hut would be 20C. If the 
acceleration factor is 2X per 10C rise (activation energy below 20C 
...) then radio A gets 2^4 more stress. Net result is that radio A sees 
about 160 times more stress than radio B.

I would humbly suggest that radio A is going to have a *different* set 
of problems than radio B. Most of us would be hard pressed to say that 
radio A has been more abused than radio B by visual inspection.

Regardless of weather the caps crack or something else goes radio A is 
a lot more likely to have had problems of a certain type over it's life 
than radio B. So far I *hope* none of this is to controversial. It's 
pretty much straight out of MIL-HBK-217.

Here's the part that makes for the problem.

Depending on how good the maintenance on the radios was it's a total 
toss up as to weather radio A, B, or C is in better condition today.

It is my contention (and I suspect that you *might* agree with this) 
that we routinely go well past the previous "standard" when it comes to 
working on these radios. I won't argue weather we are going above or 
below the standard, only that a lot of people these days do a lot more 
work on these radios than was done in the past.

Is a leaky capacitor for instance a problem? That depends on who is 
doing the maintenance. In one case a radio that meets minimum 
specifications on the bands of interest never gets pulled or worked on. 
In another case the radio is worked on until it meets "bragging rights" 
specification levels. The first case probably has a very different 
opinion of what is a bad cap than the guy in the second case.

Unless we agree on what is and isn't a bad capacitor I suspect we'll be 
at this for the next hundred years. Somehow I doubt we will come to 
anything other than an agreement to disagree ....

	Take Care!

		Bob Camp
		KB8TQ

On Jan 12, 2005, at 9:20 AM, Dallas Lankford wrote:

> Disclaimer:
>
> "If you all are talking about those rained-on, snowed-on, sun-baked, 
> frozen,
> left in the parking lot for years, or otherwise abused R-390A's, then 
> my
> remarks about capacitors do not apply to them.  I would not touch one 
> of
> those with a 10 foot pole.  There are people who believe they can be
> rebuilt.  I don't.  I wouldn't buy any R-390A nowadays that I couldn't
> personally inspect before I bought it, or that didn't come with a 
> return
> guarantee that it hasn't been abused."
>
>
> The fact that your R-390A was an (old) Motorola [see below] does not 
> change
> the fact that your claims violate the laws of statistics.
>
> Among the R-390A's that I have carefully inspected and rebuilt were 
> two 1956
> Motorolas.  Let's see...  how many brown beauties were in those two.  
> I am
> not sure.  There were about 12 each in the IF decks, and at least 1 
> each in
> the RF deck.  That is a total of 26.  If on the average we would 
> expect  out
> of 10 to be cracked (your claim), what is the probability that none 
> were
> cracked (my observation)?  The answer is simple statistics.  Multiply 
> 3/10
> by itself 26 times.  That is 2.5419 time 10 to the -14 power.  So the 
> odds
> that I would observe none when you observed 70% cracked are 1 in
> 254,190,000,000,000.  This violates the laws of statistics.
>
> The above does not include a 3rd Motorola IF deck that I still own, in 
> which
> none of the brown beauties were cracked or bad.  If I included it, the 
> odds
> would be even more outrageous, namely 1 in 1.3509 times 10 to the 
> minus 20.
>
> BTW, both of those Motorolas are alive and well (about 20 years after I
> rebuilt them), with none of the brown beauties replaced (and none have
> cracked or gone bad in the meantime).  The only problem which has 
> developed
> in either is a switch which will not turn off in one of them.  The 
> owner,
> who lives nerby, is too lazy to bring it by for me to fix.
>
> Dr. Dallas Lankford
> retired Professor of Mathematics and Statistics
> Louisiana Tech University
>
>
>
>
> ----- Original Message -----
> From: <N4BUQ at aol.com>
> To: <r-390 at mailman.qth.net>
> Sent: Wednesday, January 12, 2005 7:29 AM
> Subject: Re: [R-390] Re: Caps and more caps
>
>
> <cut>
>> A large percentage of the "brown beauties" in my '56 Motorola had 
>> cracks
> that could easily be seen.  I don't think mine was abused, but just 
> old.
>>
>> Barry(III) - N4BUQ
>
> _____________________________________________________________
> Win a new Icom IC-756PROIII and help QSL/QTH.net
> Details at: http://mailman.qth.net/index.html
> _____________________________________________________________
> R-390 mailing list
> Home: http://mailman.qth.net/mailman/listinfo/r-390
> Help: http://mailman.qth.net/faq.htm
> Post: mailto:R-390 at mailman.qth.net
> Unsubscribe: http://mailman.qth.net/mailman/options/r-390
>