Prev: Relay Question
Next: Baxandall class D oscillator, can it produce a triangle likewaveform?
From: Hammy on 23 Jul 2010 15:20
On Fri, 23 Jul 2010 12:22:00 -0500, "Tim Williams"
<tmoranwms(a)charter.net> wrote:

>I noticed, and was appalled, when my recent order of Fairchild TO-220s were *way* under spec.
>
>http://www.fairchildsemi.com/products/discrete/packaging/to220_1.html
>Evidently, they have "single" and "dual gauge" (note F).
>
>LT gives 1.143-1.397mm, a rather imprecise range.
>
>After a quick browse, Fairchild seems to be the only company cheating this JEDEC spec.
>
>Tim

Nope ST is doing the single/dual gauge as well see page 40-41 of the
7800 series regulator data sheet.

http://www.st.com/stonline/books/pdf/docs/2143.pdf

From: Nunya on 23 Jul 2010 15:41
On Jul 23, 12:20 pm, Hammy <s...(a)spam.com> wrote:
> On Fri, 23 Jul 2010 12:22:00 -0500, "Tim Williams"
>
> <tmoran...(a)charter.net> wrote:
> >I noticed, and was appalled, when my recent order of Fairchild TO-220s were *way* under spec.
>
> >http://www.fairchildsemi.com/products/discrete/packaging/to220_1.html
> >Evidently, they have "single" and "dual gauge" (note F).
>
> >LT gives 1.143-1.397mm, a rather imprecise range.
>
> >After a quick browse, Fairchild seems to be the only company cheating this JEDEC spec.
>
> >Tim
>
> Nope ST is doing the single/dual gauge as well see page 40-41 of the
> 7800 series regulator data sheet.
>
> http://www.st.com/stonline/books/pdf/docs/2143.pdf

I think that is funny... "In spite of some difference in tolerances,
the packages are compatible."

Note how that is all under the header "Mechanical data". Is
not thermal performance a mechanical engineering function?
It seems like a vague claim when we all know for sure that
the sinking mass difference will present a performance shift
at the upper end of the operating window.

It *might* be proper if they had classed it as mil spec and
commercial or something like that and actually went
through the trouble of noting the difference.

Going "COTS" was a mistake, at least at the discreet device level.
We should still have hi-rel US fabs and we do not, and even the
overseas hi-rel stuff is evaporating. All folks are willing to pay
for
performance, but makers are getting lame as they shy away
from actual culpability for anything they make.

At the consumer level, Sony is a fine example of just how
disconnected a company can get with their customers. If you
are on their bandwagon, feeding them cash from various
directions (seems everyone's disease these days),
then you are ignored, and they have a hardware warranty
program that is lame beyond pale. My 'was under warranty'
PS3 would have been serviced for me, under warranty, for
a mere $145, and they would not have serviced MY unit, but
merely send me back some lame refurb unit.
That is truly lame behavior. Many folks spend a lot of hours
working on their personal devices, and they would not want
some lame, who knows how many hours it has on it refurb
back instead of them actually repairing your unit. It should be
illegal. Not to mention the fact that the defect 'occurred' after
they changed the spindle speed on the BluRay drive via
firmware, and forgot that some of the units had different
drive motors in them that could not handle the higher rate.
Their screw up, yet they want money from us to fix it, instead
of releasing a proper firmware update that returned the drive
to an operable state. Admitting the mistake is not on their
list of things they want to do, since so many jumped on the
sucker's pay-us-and-we'll-fix-it-bandwagon. It would cost
them a lot.
From: Jan Panteltje on 23 Jul 2010 15:49
On a sunny day (Fri, 23 Jul 2010 15:20:13 -0400) it happened Hammy
<spam(a)spam.com> wrote in <onqj4692p2ciqp7qn9ou9j1k2b2vadf7nt(a)4ax.com>:

>On Fri, 23 Jul 2010 12:22:00 -0500, "Tim Williams"
><tmoranwms(a)charter.net> wrote:
>
>>I noticed, and was appalled, when my recent order of Fairchild TO-220s were *way* under spec.
>>
>>http://www.fairchildsemi.com/products/discrete/packaging/to220_1.html
>>Evidently, they have "single" and "dual gauge" (note F).
>>
>>LT gives 1.143-1.397mm, a rather imprecise range.
>>
>>After a quick browse, Fairchild seems to be the only company cheating this JEDEC spec.
>>
>>Tim
>
>Nope ST is doing the single/dual gauge as well see page 40-41 of the
>7800 series regulator data sheet.
>
>http://www.st.com/stonline/books/pdf/docs/2143.pdf
>
I did not know that.
In my view .51 is too thin, as then when bolted against a heatsink,
say on a PCB, the slightest pressure on the leads will push it away from the heatsink

heatsink
|
||| package
|--| screw
|||
| \\ -> force away from heatsink
| \
===0============
solder VIA
Bit exagerated view.
From: Nunya on 23 Jul 2010 16:20
On Jul 23, 12:49 pm, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote:

> I did not know that.
> In my view .51 is too thin, as then when bolted against a heatsink,
> say on a PCB, the slightest pressure on the leads will push it away from the heatsink
>
> heatsink
> |
> ||| package
> |--| screw
> |||
> | \\ -> force away from heatsink
> |  \
> ===0============
>    solder VIA
> Bit exagerated view.

You are supposed to attach planar surfaces (heat sink mating face)
first,
then perform the solder operations on the leads, and there should be
no stress on those leads before (or after) said solder process.

Then, the leads and solder joints actually serve to bolster the
planar
heat sink interface placement.

Most folks these days use a pressure foot placed upon the package
itself, directly above the device die location, to apply pressure
where it
achieves the greatest work product for the application., i.e. heat
sinking of the heat source.

Never solder leads, then force the part over to the sink, and never
attach to the sink and then apply force opposed to that by way of
failing to remove any back pressure the leads may be placing on
the part by less than optimal lead forming practices.

This problem amplifies itself when these devices get ganged
onto a common sink. It is also one of the causes of a specific
form of solder creep circuit failure.
From: Martin Riddle on 24 Jul 2010 15:18


"Jan Panteltje" <pNaonStpealmtje(a)yahoo.com> wrote in message
news:i2cna2$ck1$1(a)news.albasani.net...
> On a sunny day (Fri, 23 Jul 2010 10:41:59 -0700 (PDT)) it happened
> Nunya
> <jack_shephard(a)cox.net> wrote in
> <1c4de387-f1c0-48cf-a9c4-bdf154b3bc2a(a)w31g2000yqb.googlegroups.com>:
>
>>On Jul 23, 9:19 am, Jan Panteltje <pNaonStpealm...(a)yahoo.com> wrote:
>>> My old 7805 stabilisrs have 1.3 mm thick copper.
>>> The new ones I bought today are .5 mm.
>>> The new ones you can just bent with plyers, even break the tab off.
>>>
>>> Wonder when we go to .25 :-)
>>
>> It needs to conform to the thermal spec of that part to be able to be
>>legally called that part. i.e. the original data sheet that the oem
>>chi maker claims to be compliant with in order to call it that part.
>>Otherwise, it is a fake or a counterfeit, no ifs ands or buts. OR
>>they
>>have to put a DIFFERENT moniker on it and tout it as a
>>"suitable replacement" or such.
>>
>> A claim of being fully compliant with the spec that part
>>originally had requires true full compliance with said specs.
>>
>> Cutting away sinking mass leads to obvious alterations in
>>those specs. There is no way around it. They are cheap fakes.
>
> Yes, I suppose the heat conductivity changes, also the mounting
> becomes less
> reliable in my view.
> Here is a picture that compares this thin 7805 with a normal IRF TO220
> MOSFET,
> the 7805 is on the right:
> ftp://panteltje.com/pub/thin_7805_compare_img_2090.jpg
>
> This is the markings it shows:
> ftp://panteltje.com/pub/thin_7805_img_2092.jpg

We have 7912's that are thinner than usual. But havent seen 7805's that
thin yet.

Cheers


First  |  Prev  |  Next  |  Last
Pages: 1 2 3
Prev: Relay Question
Next: Baxandall class D oscillator, can it produce a triangle likewaveform?