Searching \ for '[EE] Split Emitter / Split Collector Transistors' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=split+emitter+split
Search entire site for: 'Split Emitter / Split Collector Transistors'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Split Emitter / Split Collector Transistors'
2006\01\05@110257 by Todd Bailey

picon face

 Sages of the PIClist:

 Recently I started designing a discrete BJT op-amp (my first time) for an
audio circuit, and in doing so have been studying the manufacturers'
schematics of different monolithic op-amps.  In several of the circuits (the
741 [Horowitz and Hill, p.189] and the 5532
[http://www.dself.dsl.pipex.com/ampins/webbop/5532.htm] so far) I have seen
components drawn into the schematic which look like a normal BJT with an
extra collector or emitter leg drawn on.

 My questions to you:

 1.)  What are these things called?  I've been calling them "split emitter"
or "split collector" transistors because I've seen that in print before --
but I'm not sure I'm using the term correctly.

 2.)  I assume that these are somehow doped differently than a normal BJT,
but I don't know how.  Does anyone?  Would a "split collector" NPN
transistor be similar to two discrete matched NPN BJTs connected together at
the base and emitter, but connected separately at the collector?

  3.)  Why are these used?  I assume it has to do with maintaining current
ratios in different sections of the op amp by using a common device to set
those currents.

 4.)  Finally, do these devices exist as stand-alone in monolithic
packages, or are they always buried deep in some silicon wafer?  What I mean
is, are these the kind of thing I could (or would want or need to) use on a
breadboard?
 4a.)  Is there some reference book or website that explains these devices
so I don't have to bother you all?

 I know that's a lot of questions, but the only thing I'd seen like this
before had been dual-gate mosfets, which don't look very similar at all.  I
was hoping one of you might know something, because so far the internet and
bookshelf have been pretty dry.

Yours,
and thanks as always-

Todd Bailey


2006\01\05@132712 by David VanHorn

picon face
>
>
> 1.)  What are these things called?  I've been calling them "split emitter"
> or "split collector" transistors because I've seen that in print before --
> but I'm not sure I'm using the term correctly.


That's how I've heard them described, or "multiple collector/emitter".

2.)  I assume that these are somehow doped differently than a normal BJT,
> but I don't know how.  Does anyone?  Would a "split collector" NPN
> transistor be similar to two discrete matched NPN BJTs connected together
> at
> the base and emitter, but connected separately at the collector?


Email Bob Pease at National Semi.  He knows.

 3.)  Why are these used?  I assume it has to do with maintaining current
> ratios in different sections of the op amp by using a common device to set
> those currents.


Yup. The matching is supposed to be excellent.

4.)  Finally, do these devices exist as stand-alone in monolithic
> packages, or are they always buried deep in some silicon wafer?  What I
> mean
> is, are these the kind of thing I could (or would want or need to) use on
> a
> breadboard?



You used to be able to buy them, but they were specialty items.
I haven't seen any in the wild for many years now.


4a.)  Is there some reference book or website that explains these devices
> so I don't have to bother you all?


You've already found AofE, I don't remember anything else that would be all
that helpful, other than studying older National Semi linear databooks.

2006\01\05@142004 by Enrico Schuerrer

picon face

----- Original Message -----
From: "Todd Bailey" <spam_OUTtm_baileyTakeThisOuTspamhotmail.com>
To: <.....piclistKILLspamspam@spam@mit.edu>
Sent: Thursday, January 05, 2006 5:02 PM
Subject: [EE] Split Emitter / Split Collector Transistors


| [Split Emitter/Collector - symbols]

As far as I know this items have the same function as paralled transistors -
so if you see a 2 emitter transistor in real on the substrate these are 2
transistors, where collector and base electrodes are switched in parallel
and 2 emitters are on the substrate.

The symbol comes from the buildup of these 2 transistors on the same
substrate, where automatically (due to the construction) the collectors on
the one hand and base electrodes on the other hand are the same material
respectively.

Regards

Enrico

2006\01\05@144539 by David VanHorn

picon face
>
>
> As far as I know this items have the same function as paralled transistors
> -
> so if you see a 2 emitter transistor in real on the substrate these are 2
> transistors, where collector and base electrodes are switched in parallel
> and 2 emitters are on the substrate.


Yes, but the point is the good matching of current in each split leg, which
you can't get in discrete transistors.

I've done this with discretes, and you can build the circuits, but you have
to match the betas of the transistors by hand, and even then it's not all
that wonderful..

2006\01\05@173120 by Gerhard Fiedler

picon face
Todd Bailey wrote:

> Recently I started designing a discrete BJT op-amp (my first time) for an
> audio circuit, and in doing so have been studying the manufacturers'
> schematics of different monolithic op-amps.  In several of the circuits
> (the 741 [Horowitz and Hill, p.189] and the 5532
> [http://www.dself.dsl.pipex.com/ampins/webbop/5532.htm] so far) I have
> seen components drawn into the schematic which look like a normal BJT
> with an extra collector or emitter leg drawn on.

I'm not sure what you are trying to do, but it seems to me that the
techniques to use when building an amplifier from discrete components are
quite different from the ones to use when creating an integrated circuit.
The "components" available are just too different in the two cases. Also
don't forget that the schematic of an integrated circuit is something like
an approximation; this becomes more obvious when you look at the actual
silicone layout.

If you want to build a discrete audio amp, you are probably better off with
some literature about discrete amps, rather than studying schematics of
integrated opamps.

Gerhard

2006\01\06@153309 by Todd Bailey

picon face

Todd Bailey wrote:

>> Recently I started designing a discrete BJT op-amp (my first time) for
an
>> audio circuit, and in doing so have been studying the manufacturers'
>> schematics of different monolithic op-amps.  In several of the circuits
>> (the 741 [Horowitz and Hill, p.189] and the 5532
>> [http://www.dself.dsl.pipex.com/ampins/webbop/5532.htm] so far) I have
>> seen components drawn into the schematic which look like a normal BJT
>> with an extra collector or emitter leg drawn on.

Gerhard wrote:

{Quote hidden}

---------

 I'm afraid I don't know very much about the intricacies of integrated
circuit design, but it seemed to me that, generally speaking, the
manufacturers' datasheets were a lot more exact about the specs and theory
of operation involved with their devices than audio experimenters.  So I
naturally gravitated towards those and H&H rather than audio DIY websites or
Delton T. Horn books or the like.  I'd really like to learn the theory
behind what I'm doing, especially since the components involved in discrete
design can be so pricey and there are so many opinions on the subject
floating around.
 That having been said, what you say seems believable.  Could you recommend
a good book or reference on the technical issues of discrete op-amp design?  
I've found a few web sites, but that's it.

Yours, and thanks again,

Todd


2006\01\06@173406 by Gerhard Fiedler
picon face
Todd Bailey wrote:

>>> Recently I started designing a discrete BJT op-amp (my first time) for
>>> an audio circuit, and in doing so have been studying the
>>> manufacturers' schematics of different monolithic op-amps.  
>
>> I'm not sure what you are trying to do, but it seems to me that the
>> techniques to use when building an amplifier from discrete components
>> are quite different from the ones to use when creating an integrated
>> circuit.
>
> I'm afraid I don't know very much about the intricacies of integrated
> circuit design, but it seemed to me that, generally speaking, the
> manufacturers' datasheets were a lot more exact about the specs and
> theory of operation involved with their devices than audio
> experimenters.  

Yes, probably so :)  But even though they publish their "circuits"
sometimes, they don't publish the components in those circuits :)  So those
data sheets don't help a whole lot.

> I'd really like to learn the theory behind what I'm doing, especially
> since the components involved in discrete design can be so pricey and
> there are so many opinions on the subject floating around. That having
> been said, what you say seems believable.  Could you recommend a good
> book or reference on the technical issues of discrete op-amp design?
> I've found a few web sites, but that's it.

I'm sorry, but I'm not really a discrete analog whiz :) and can't help you
out here. Hopefully some of the others here with more analog design
experience can jump in here and recommend something.

Just out of curiosity: why do you insist on discrete? Just out of
curiosity? :)

Gerhard

2006\01\06@213636 by Todd Bailey

picon face



{Quote hidden}

 Gerhard,

 Honestly, it's mostly for the learning experience, by which I mean for the
fun of it.  In my real job I design the embedded systems for toy prototypes
for a think tank that works for Mattel and Hasbro.  Most of what I do is
embedded C and assembly programming for PIC and Motorola 68k series chips.  
But I originally got that job because the people who ran the company saw
some synthesizers I had built for an art piece that incorporated an analog
RFID system (an LM324 is a lot cheaper than a TI RFID kit for a man on a
budget like I was then) which is why I was really into engineering in the
first place -- because it was fun.  So now that I sit around making Barbie's
eyes blink all day I crave the good old times when I used to sit around
(starving and shivering) with a scope and a big mess of OTAs and mosfets
making weird noises.

 I feel like the op amp is sort of the holy grail of analog design -- I use
them all the time but never *really* knew how they worked outside of the old
Diff Pair, VAS, Output Stage.  So I figured the best way to get a good
handle on the beast was to build one, with the added bonus that I could then
use that op-amp in all my artsy-fartsy designs from here on out.

And I'd love to make that animal monolithic, but until I learn VHDL I'm
afraid solder is the way I've got to go.

Thanks again,

Todd


2006\01\06@222620 by David VanHorn

picon face
>
>
> Honestly, it's mostly for the learning experience, by which I mean for the
> fun of it.  In my real job I design the embedded systems for toy
> prototypes
> for a think tank that works for Mattel and Hasbro.


Sounds interesting! :)  Any room for an AVR guy?

2006\01\06@223859 by Dmitriy Kiryashov

picon face
Hi Todd.

Todd Bailey wrote:
>
> I have seen components drawn into the schematic which look like a normal
> BJT with an extra collector or emitter leg drawn on.
>
>   My questions to you:
>
>   1.)  What are these things called?  I've been calling them "split emitter"
> or "split collector" transistors because I've seen that in print before --
> but I'm not sure I'm using the term correctly.

Multi emitter or multi collector BJT transistors.


>   2.)  I assume that these are somehow doped differently than a normal BJT,
> but I don't know how.  Does anyone?  Would a "split collector" NPN
> transistor be similar to two discrete matched NPN BJTs connected together at
> the base and emitter, but connected separately at the collector?

It wont be similar to two discrete BJT connected together. When you run
technological process on the piece of silicon and transistors are located
close to each other it is quite possible to dope the areas with high
precision and thus obtain very close charateristics of transistors.
It is very difficult and impractical to attempt to achieve that with
discrete transistors (unless they were made on the same piece of silicon
in the same cycle and equipment was setup to produce matched transistors )
You will need to find a book about technological processes used in seconductor
industry, properties od materials used as dopes etc. I had such a huge book
once ( it was in russian ) while was studing things at university but lost it
somewhere later when had no practical implementation to that knowledge.


>    3.)  Why are these used?  I assume it has to do with maintaining current
> ratios in different sections of the op amp by using a common device to set
> those currents.

Multi-emitter circuit were quite commonly used in TTL digital logic ICs
awhile ago ( who remember that raise you hand ? :)
Multi-collector are quire commonly used in integrated circuits for various
purposes.


>   4.)  Finally, do these devices exist as stand-alone in monolithic
> packages, or are they always buried deep in some silicon wafer?  What I mean
> is, are these the kind of thing I could (or would want or need to) use on a
> breadboard?

I doubt anybody would pack just a few matched transistors in very expensive
( compared to the cost of silicon piece inside of it ) package and be able
to still sell while others will be selling complete multifunctional more
complex circuits for the same price.


>   4a.)  Is there some reference book or website that explains these devices
> so I don't have to bother you all?
Art of Electronics you mentioned already, "semiconductor circuits design"
Titse Shenk? not sure if I spelled names correctly it was in russian.
Also good to read " op amps and linear integrated circuits" Honeycutt.
Some sort of "sourcebook of modern transistor circuits" might also help.
Try to search in nearest areas of knowledge as well.

Good luck with research. :)


WBR Dmitriy.

2006\01\06@232246 by Spehro Pefhany

picon face
At 09:36 PM 1/6/2006 -0500, you wrote:

>   I feel like the op amp is sort of the holy grail of analog design -- I use
>them all the time but never *really* knew how they worked outside of the old
>Diff Pair, VAS, Output Stage.  So I figured the best way to get a good
>handle on the beast was to build one, with the added bonus that I could then
>use that op-amp in all my artsy-fartsy designs from here on out.

There are some detailed op-amp circuits (and descriptions of the building-
blocks you need) in _Analog Design for CMOS VLSI Systems_ by Maloberti (Klower)
It's pretty much all CMOS, as the title suggests, although there are some
interesting applications of the parasitic well BJTs. The building blocks are
similar in function in bipolar and CMOS, just slightly different
implementation.

My ancient university text _Modern Electronic Circuit Design_ by Comer has a
bunch on bipolar building blocks and op-amps, but it's not as well organized
as Maloberti.


>And I'd love to make that animal monolithic, but until I learn VHDL I'm
>afraid solder is the way I've got to go.

I don't think VHDL will help you much, but Spice might. There is a certain
amount of math behind such things as compensation, and once you've
synthesized the design, Spice will help you wring it out (such as changes
with ambient temperature etc.). Of course it's only as good as the device
models you use, and being mathematical in nature it will give silly results
such as 0V output for a gain of 4,000,000 amplifier with the inputs shorted,
when in reality we know that the output will most likely be sitting at one
PS rail or the other.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffspamspam_OUTinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\01\07@064105 by Gerhard Fiedler

picon face
Spehro Pefhany wrote:

> I don't think VHDL will help you much, but Spice might. There is a certain
> amount of math behind such things as compensation, and once you've
> synthesized the design, Spice will help you wring it out (such as changes
> with ambient temperature etc.). Of course it's only as good as the device
> models you use, and being mathematical in nature it will give silly results
> such as 0V output for a gain of 4,000,000 amplifier with the inputs shorted,
> when in reality we know that the output will most likely be sitting at one
> PS rail or the other.

Wouldn't that be because you forgot to add a few noise sources in your
models (and the circuit)? Which of course is not really common to do, but
probably necessary once you get into that range.

Gerhard

2006\01\07@065037 by Gerhard Fiedler

picon face
Dmitriy Kiryashov wrote:

>> 4a.)  Is there some reference book or website that explains these
>> devices so I don't have to bother you all?

> Art of Electronics you mentioned already, "Rsemiconductor circuits design"
> Titse Shenk? not sure if I spelled names correctly it was in russian.

That would probably be U. Tietze/Ch. Schenk "Halbleiter-Schaltungstechnik"
(the German original :)  

Gerhard

2006\01\07@092523 by Spehro Pefhany

picon face
At 09:40 AM 1/7/2006 -0200, you wrote:
>Spehro Pefhany wrote:
>
> > I don't think VHDL will help you much, but Spice might. There is a certain
> > amount of math behind such things as compensation, and once you've
> > synthesized the design, Spice will help you wring it out (such as changes
> > with ambient temperature etc.). Of course it's only as good as the device
> > models you use, and being mathematical in nature it will give silly results
> > such as 0V output for a gain of 4,000,000 amplifier with the inputs
> shorted,
> > when in reality we know that the output will most likely be sitting at one
> > PS rail or the other.
>
>Wouldn't that be because you forgot to add a few noise sources in your
>models (and the circuit)? Which of course is not really common to do, but
>probably necessary once you get into that range.
>
>Gerhard

Typically your model doesn't have an offset voltage (the modelled transistors
are mathematically *identical* to each other by default, which doesn't
happen in
real life, even if they are beside each other on the same die). Of course
you can
fiddle the device-level models a bit or add an offset to the op-amp
macromodel, but
since the average Vos is zero, mostly you'd just leave it at zero for practical
purposes (the behavior of the op-amp open loop with exactly zero volts
input is not
of much interest anyhow- even if you're using it as a comparator or
something you'd
typically have a test input that moves through the 0V region.

As you mention, noise (and startup transients) are important factors in
getting some types of oscillators to start, and that can be an issue with
Spice.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
@spam@speffKILLspamspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2006\01\07@093030 by Peter Todd

picon face
On Fri, Jan 06, 2006 at 09:36:31PM -0500, Todd Bailey wrote:
>   Gerhard,
>
>   Honestly, it's mostly for the learning experience, by which I mean for the
> fun of it.  In my real job I design the embedded systems for toy prototypes
> for a think tank that works for Mattel and Hasbro.  Most of what I do is
> embedded C and assembly programming for PIC and Motorola 68k series chips.  
> But I originally got that job because the people who ran the company saw
> some synthesizers I had built for an art piece that incorporated an analog
> RFID system (an LM324 is a lot cheaper than a TI RFID kit for a man on a
> budget like I was then) which is why I was really into engineering in the
> first place -- because it was fun.  So now that I sit around making Barbie's
> eyes blink all day I crave the good old times when I used to sit around
> (starving and shivering) with a scope and a big mess of OTAs and mosfets
> making weird noises.

Do you have a webpage or something with your work? I'm an electronics
artist myself, still in the going to arts school and in the starving and
shivering mode, and am always interested in seeing other peoples work,
especially when they actually know the technology!

--
KILLspampeteKILLspamspampetertodd.ca http://www.petertodd.ca

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...