Maker Pro
Maker Pro

looking for a transistor

C

ChairmanOfTheBored

Jan 1, 1970
0
WTF? I run across them all the time.. *shrug*

Yep. In every package style currently in use for sm signal devices.
 
E

Eeyore

Jan 1, 1970
0
ChairmanOfTheBored said:
There are different data sheets for every maker of the 2N2222, and they
ALL differ in one respect or another from the original.

NONSENSE.

The 2N2222 is a JEDEC registered device. ALL manufacturers must supply 2N2222s
to the JEDEC spec.

THAT is why the Zetex part IS a 2222 as well.

But not a *** 2N *** 2222. The *** 2N *** is a critical part of the device
spec. It indicates a JEDEC registered device with a guaranteed
non-manufacturer dependent spec.

http://en.wikipedia.org/wiki/JEDEC

They (manufacturers) are ALL different,

NOT for JEDEC (1N, 2N etc) parts. They must all comply with the JEDEC spec.

except for the early original makers of the OEM parts.

'OEM' parts are a totally different matter.

After all the other silicon fab houses started making their versions of
it, they all differed on one or more ways. If you can't understand the
concept of how metal conducts heat better than plastic, you need to die
and be reborn in your next life and actually try to learn something in
that one.

How metal conducts is only relevant when a heatsink is fitted. Since shiny
metal radiates heat less effectively than matt black plastic it's not hard to
see how plastic can outperform metal can here. There is also the lead frame to
consider. A plastic device has the die mounted directly onto a lead frame
which then conducts heat away from the die into the pcb. Metal can devices do
not have an equivalent direct conduction path.

You're uneducable in this one.

NO, you're the bloody idiot. You spout shit like this without even the TINIEST
understanding of the underlying science.

Fucking read what I just told you and learn something for a change.

Graham
 
E

Eeyore

Jan 1, 1970
0
ChairmanOfTheBored said:
Yep. In every package style currently in use for sm signal devices.

2N2222s are ONLY available in TO-18. That's the JEDEC spec.

Anything else is NOT a *** 2N2222 ***. It will have a different part number.

Graham
 
L

Lostgallifreyan

Jan 1, 1970
0
How metal conducts is only relevant when a heatsink is fitted. Since
shiny metal radiates heat less effectively than matt black plastic
it's not hard to see how plastic can outperform metal can here. There
is also the lead frame to consider. A plastic device has the die
mounted directly onto a lead frame which then conducts heat away from
the die into the pcb. Metal can devices do not have an equivalent
direct conduction path.

When I pointed out the radiation aspect, John Popelish pointed out that it
is a small part, not significant, and when the Popelfish has translated
things electronical, I tend to listen. You're right about the second part
though, compared to the copper tab formed as a single part with one of the
leads, the thermal conductivity between can and wires is lower. I doubt
it's that much lower though.
 
C

ChairmanOfTheBored

Jan 1, 1970
0
When I pointed out the radiation aspect, John Popelish pointed out that it
is a small part, not significant, and when the Popelfish has translated
things electronical, I tend to listen. You're right about the second part
though, compared to the copper tab formed as a single part with one of the
leads, the thermal conductivity between can and wires is lower. I doubt
it's that much lower though.


The die on the TO-18 is attached to the base of the can... with a
thermally conductive epoxy.

Wanna buy some stock in a stable, not going anywhere any time soon
company? Look at the makers of EpoTEK epoxies. They are what the
industry uses to "glue down" dies to various substrates. I use a very
expensive silver filled variety at work to seal up chassis so they have a
lower EMI radiance, even though it is meant for die attach.

Silver is like $14 an ounce, but the micro-fine silver powders they use
in these media are VERY expensive. Far more than $14 an ounce.

So a one pound epoxy kit costs several hundred dollars.

http://www.epotek.com/SSCDocs/datasheets/H20E.PDF
 
M

MooseFET

Jan 1, 1970
0
When I pointed out the radiation aspect, John Popelish pointed out that it
is a small part, not significant, and when the Popelfish has translated
things electronical, I tend to listen. You're right about the second part
though, compared to the copper tab formed as a single part with one of the
leads, the thermal conductivity between can and wires is lower. I doubt
it's that much lower though.


It is fairly likely that the leadframe makes more difference than the
metal vs plastic housing. The transistor substrate isn't all them
well connected in either thermally. The plastic part ends up having a
lower junction to case thermal resistance than the metal one. I
suspect that this is due to the heat flowing down the lead frame.
 
E

Eeyore

Jan 1, 1970
0
ChairmanOfTheBored said:
The die on the TO-18 is attached to the base of the can... with a
thermally conductive epoxy.

TO-92 still outperforms TO-18 for heat transfer from junction to ambient..

Graham
 
E

Eeyore

Jan 1, 1970
0
Lostgallifreyan said:
When I pointed out the radiation aspect, John Popelish pointed out that it
is a small part, not significant, and when the Popelfish has translated
things electronical, I tend to listen. You're right about the second part
though, compared to the copper tab formed as a single part with one of the
leads, the thermal conductivity between can and wires is lower. I doubt
it's that much lower though.

2N2222 Pd = 500mW (metal)
PN2222 Pd = 625 mW (plastic)

Graham
 
J

Jamie

Jan 1, 1970
0
Eeyore said:
MooseFET wrote:




Not a problem as I'm using them in a DC regulator circuit. Prety OK for any audio apps
too.





Oops ! Yes, the TIPs are a bit slow. Japanese power devices (notably Toshiba IME) are
typically a LOT faster than 'western' designs btw.

Graham
Western designs? Are we on the road of bashing the US again ?

What the hell does Western have to do with it?
 
J

John Popelish

Jan 1, 1970
0
MooseFET said:
It is fairly likely that the leadframe makes more difference than the
metal vs plastic housing. The transistor substrate isn't all them
well connected in either thermally. The plastic part ends up having a
lower junction to case thermal resistance than the metal one. I
suspect that this is due to the heat flowing down the lead frame.

I agree. The 2N2222 parts have Kovar or some other iron
alloy leads (gold plated, perhaps) of very low thermal
conductivity, compared to copper, and for the base and
emitter mounted in the can with glass seals. The collector
lead is welded to the can, but the alloy is still the same
low conductivity stuff the other two leads are made of. The
chip wire bonded to the base and emitter leads, so those
don't even have much of a thermal connection to the die
inside the can. Many TO-92 parts have punched copper lead
frames, especially the Zetex E-line, with the die mounted
directly on a flag on the collector lead. Solder that lead
to a large area pad on the board, with a short lead length,
and you carry away more heat than all 3 leads of the metal
can might carry. Plus the thermally conductivity epoxy
carries heat off all the surfaces of the die not bonded to
the collector tab, where the 2N2222 has the die in an
otherwise empty can, so only the collector side has a local
heat sink.
 
E

Eeyore

Jan 1, 1970
0
Jamie said:
Western designs?

US or European.

Are we on the road of bashing the US again ?

What the hell does Western have to do with it?

Slow bipolar power transistors. Don't ask me why but Motorola and RCA never could match
Toshiba or Sanken for device speed combined with good SOA.

Graham
 
L

Lostgallifreyan

Jan 1, 1970
0
2N2222 Pd = 500mW (metal)
PN2222 Pd = 625 mW (plastic)

It's not exactly an order of magnitude, is it? If the thermal resistance
between can and chip were that much greater, I think the Pd difference
would also be greater than that.
 
M

MooseFET

Jan 1, 1970
0
US or European.



Slow bipolar power transistors. Don't ask me why but Motorola and RCA never could match
Toshiba or Sanken for device speed combined with good SOA.

The TIP35 also had a very bad dynamic saturation characteristic. When
the collector voltage got below about 2V, they slowed down to a crawl.
 
E

Eeyore

Jan 1, 1970
0
Lostgallifreyan said:
It's not exactly an order of magnitude, is it? If the thermal resistance
between can and chip were that much greater, I think the Pd difference
would also be greater than that.

No it's not that much but it is the precise reverse of what the resident
FOOL claimed.

Graham
 
M

MooseFET

Jan 1, 1970
0
I agree. The 2N2222 parts have Kovar or some other iron
alloy leads (gold plated, perhaps) of very low thermal
conductivity, compared to copper,

I think it is even worse than Brass. Some TO-92 transistors are
magnetic some aren't. I suspect that some are made with covar lead
frames because the companies are used to working with the stuff.


[..The collector lead ..]
Solder that lead
to a large area pad on the board, with a short lead length,
and you carry away more heat than all 3 leads of the metal
can might carry.

You can solder it to a metal post to do even better. I've seen this
done in an audio amplifier. The TO-92s were soldered onto a 0.025
inch square post that stood up some distance above the point where the
soldering was done.
Plus the thermally conductivity epoxy
carries heat off all the surfaces of the die not bonded to
the collector tab, where the 2N2222 has the die in an
otherwise empty can, so only the collector side has a local
heat sink.

Any epoxy is more conductive than air or just N2 so this may also be
helping.
 
C

ChairmanOfTheBored

Jan 1, 1970
0
No it's not that much but it is the precise reverse of what the resident
FOOL claimed.

Graham
Funny, how the only part that was acceptable in our mil spec designs
was the metal can variety.

Must have something to do with MTBF.

Whom is laughing at whom now?
 
E

Eeyore

Jan 1, 1970
0
ChairmanOfTheBored said:
Funny, how the only part that was acceptable in our mil spec designs
was the metal can variety.

Must have something to do with MTBF.

Very likely so and its ability to withstand adverse environments most likely
because of the hermetic packaging. Irrelevant to anything other than military /
aerospace application and the like of course.

Whom is laughing at whom now?

I'm laughing at your pathetic attempt to wriggle out of your error by changing
the topic.

Graham
 
E

Eeyore

Jan 1, 1970
0
MooseFET said:
The TIP35 also had a very bad dynamic saturation characteristic. When
the collector voltage got below about 2V, they slowed down to a crawl.

Really ? I didn't know that. I've never used a TIP where speed was an issue.

Graham
 
C

ChairmanOfTheBored

Jan 1, 1970
0
Very likely so and its ability to withstand adverse environments most likely
because of the hermetic packaging. Irrelevant to anything other than military /
aerospace application and the like of course.

"and the like"? You're an idiot. The metal housed device is better for
more reasons than the hermetically sealed package, idiot.

It has a much broader range of operation.

I am betting that the epoxy package peters out far sooner in designs
where the transistor's capacities are pushed.
I'm laughing at your pathetic attempt to wriggle out of your error by changing
the topic.

You're an idiot. It is still the same topic. It is still a better
part.
 
F

Fred Bloggs

Jan 1, 1970
0
ChairmanOfTheBored said:
"and the like"? You're an idiot. The metal housed device is better for
more reasons than the hermetically sealed package, idiot.

It has a much broader range of operation.

I am betting that the epoxy package peters out far sooner in designs
where the transistor's capacities are pushed.



You're an idiot. It is still the same topic. It is still a better
part.

IIRC, the collector is internally connected to the metal housing and the
maximum operating Tj is increased from 125-150oC to an easy 200oC with
the TO-39. Then there is the premium hermetic sealing advantage. There's
no question about the superiority of that package to all else....
 
Top