Transistors

[Henry]

Probably a dumb question, but found a site about transistors. They showed a
PNP as being between +V and a load, and a NPN as being between the load and
ground. Is this a conceptual convention or is this really neccessary? Can a
NPN be place between +V and ground. In dabbling with my old TTL chips (7400
series) I have been using NPN for everything. (Partly cause I bought a 15
pack at RS.) But in some cases I have had to invert an output to use an
NPN. Could I use a PNP in the same situation without inverting? I have been
using transistors for all my I/O. (mostly to a camera or a photo flash. Like
a flash delay circuit.) I understand the basic concepts of transistors being
amplifiers and switches (what I am doing) but dont know the practicle dos
and donts. Although my circuits work, I am worried that I might be using the
wrong transistors, or right ones in the wrong way.

Henry

 

[Dan Hollands]

Probably a dumb question, but found a site about transistors. They showed
a
PNP as being between +V and a load, and a NPN as being between the load
and
ground. Is this a conceptual convention or is this really neccessary? Can
a
NPN be place between +V and ground. In dabbling with my old TTL chips
(7400
series) I have been using NPN for everything. (Partly cause I bought a 15
pack at RS.) But in some cases I have had to invert an output to use an
NPN. Could I use a PNP in the same situation without inverting? I have
been
using transistors for all my I/O. (mostly to a camera or a photo flash.
Like
a flash delay circuit.) I understand the basic concepts of transistors
being
amplifiers and switches (what I am doing) but dont know the practicle dos
and donts. Although my circuits work, I am worried that I might be using
the
wrong transistors, or right ones in the wrong way.

Henry

The transistors used as drivers and invertors with TTL (+5V / Gnd) logic are
almost always NPN transistors
The input and output voltages of TTL are most compatible with NPN

As a historical - point in the early days before ICs, PNP transistors were
much less expensive and more available than NPN and most logic system were
Gnd and -12V so that PNP transistors could be used.

--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
[email protected]
www.QuickScoreRace.com

 

[Henry]

The transistors used as drivers and invertors with TTL (+5V / Gnd) logic are
almost always NPN transistors
The input and output voltages of TTL are most compatible with NPN

As a historical - point in the early days before ICs, PNP transistors were
much less expensive and more available than NPN and most logic system were
Gnd and -12V so that PNP transistors could be used.

--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
[email protected]
www.QuickScoreRace.com

Okay, thanks. When I started I picked up packages of each (NPN and PNP).
Only used the NPN so far but am running low. Was wondering if I could, or
should, be using the PNP ones in cases where I was having to invert the
output for the NPNs. Guess I will have to go and buy another package of NPNs
if I want to do more.

Now, what about computer parallel port output? There is a VB DDL Win95IO
that allows a win95/98 (which is what I still use) user made program to
directly read and change the printer port pins. I was thinking of dabbling
with this to interface with a TTL/transistor circuit. Should I still use
only NPN for this?

Henry

 

[Derek Potter]

Probably a dumb question, but found a site about transistors. They showed a
PNP as being between +V and a load, and a NPN as being between the load and
ground. Is this a conceptual convention or is this really neccessary? Can a
NPN be place between +V and ground. In dabbling with my old TTL chips (7400
series) I have been using NPN for everything. (Partly cause I bought a 15
pack at RS.) But in some cases I have had to invert an output to use an
NPN. Could I use a PNP in the same situation without inverting? I have been
using transistors for all my I/O. (mostly to a camera or a photo flash. Like
a flash delay circuit.) I understand the basic concepts of transistors being
amplifiers and switches (what I am doing) but dont know the practicle dos
and donts. Although my circuits work, I am worried that I might be using the
wrong transistors, or right ones in the wrong way.

You really need to look at just two basic configurations, the common
emitter amplifier and the emitter follower. Either NPN or PNP can be
used. The two configurations you describe are both common emitter: if
you turn the circuit upside down they are precisely the same except
that all the polarities are reversed. So a common emitter amplifier
will always invert.

The emitter follower does not invert. It also does not introduce any
gain, it just passes the signal on.

The other configuration you may see is common base. This is primarily
used at RF. However it can also be useful for switching. Imagine
holding the base at, say, 2V. If you pull the emitter lower than 1.4V,
the transistor will turn on (use a resistor in either lead to limit
the current). If the voltage is higher, thje transistor will turn off.
With a load in the collectoer, the output is therefore the same
polarity as the input, though the level will shift.

It sounds as though you need to go over basic transistor theory, the
idea of current gain, the meaning of the characteristic curves and how
to use them with load lines. It's not difficult but it will save
worrying whether you're using the transistors in the "wrong" way.

On the subject of that, have you noticed that the symbol for a
transistor is almost symmetrical? The emitter has an arrow head on it
but is otherwise the same as the collector. This is not due toi
whimsy, the transistor is basically a symmetrical device. You can
often swap the emitter and collector. However, transistors are
actually made asymmetrically so if you do this you need to keep the
circuit voltages below about 6V for most devices. the performance is
usually a lot worse the "wrong way round" in all but one respect. When
a transistor is switched "on" in the normal way, the collector voltage
rises to a few millivolts even with no current. With C and E swapped,
the "on" voltage drops right down. This used to be useful in some
applications, though MOSFETs go to nanovolts and have displaced
bipolars for this sort of thing.

 

[Dan Hollands]

Okay, thanks. When I started I picked up packages of each (NPN and PNP).
Only used the NPN so far but am running low. Was wondering if I could, or
should, be using the PNP ones in cases where I was having to invert the
output for the NPNs. Guess I will have to go and buy another package of
NPNs
if I want to do more.

Now, what about computer parallel port output? There is a VB DDL Win95IO
that allows a win95/98 (which is what I still use) user made program to
directly read and change the printer port pins. I was thinking of dabbling
with this to interface with a TTL/transistor circuit. Should I still use
only NPN for this?

Henry

I'm not an expert on parallel ports but I believe they output 0-5v and may
be able to interface to TTL. I wonder why you are using TTL. Most designs
today use CMOS which are much easier to interface to as they have much
higher input inpedance.


--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
[email protected]
www.QuickScoreRace.com

 

[Henry]

I'm not an expert on parallel ports but I believe they output 0-5v and may
be able to interface to TTL. I wonder why you are using TTL. Most designs
today use CMOS which are much easier to interface to as they have much
higher input inpedance.


--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
[email protected]
www.QuickScoreRace.com

I am simply trying to make use of what I already have. Mostly due to budget,
partly because TTL is what I dabbled with 25 years ago and is what I still
know a little bit. I am not interested in re-learning electronics in general
as much as just working on a specific project. I am out of work, so I have
some time to tinker and came across my old box of electronics that gave me
some ideas. I cant really spend money except for a few resistors,
transistors, ect from Radio Shack.

Thanks,
Henry

 

[Derek Potter]

To turn a transistor fully on (into saturation), the base voltage has to be
slightly past collector voltage. for a NPN with collector connected to the
positive supply that'd require a drive voltage greater than the supply
voltage, which is usually inconvenient to implement.

You can saturate a transistor with the collector voltage below the
base. You typically need just a few tens of millivolts VCE though you
still need the .6V or so on the base (VBE). In the configuration
described, ie an emitter follower, the difficulty is getting the base
voltage .6 above the emitter. Once that has been done, the BE junction
conducts like any other diode.

The way it works is this. The emitter doping level is much higher than
the base, so most of the carriers that diffuse across the junction
come up from the emitter into the base not down from the base into the
emitter. Once in the base region they diffuse around and a few come
out of the base connection. The majority fall into the collector.
Hence the current gain. But unlike getting the B-E junction to
conduct, the carriers do not need any helpto fall into the collector.
Hence it will remain conducting down to very low voltages.

The voltage objection to the emitter follower does not apply if the
circuit is AC coupled. You can then get an output swing which is
practically rail-to-rail.

probably not a good idea...

Bang!

I think Henry probably means "between +V and the load, with the other
side of the load to ground". That would be the emitter follower again.

15 pack? so few?

Indeed. At a few pence each it's worth having enough to burn

maybe.

Depends on the logic swing. TTL doesn't usually pull to 5V so it may
need a pull-up resistor. Also it depends what Henry needs the output
to do. If he wants to CURRENT switch something, going from NPN to PNP
will mean that what was "on" becomes "off" and vice versa. If he
wants a ogic VOLTAGE drive out of the system then either type of
transistor is going to invert.

It's also important to make sure the logic swings are translated into
a swing past the .6-.7V threshold of the transistor. The same resistor
chain can also be used to limit the current into the base. Switching
speed will be rotten compared to TTL.

possibly. I mainly learned transistors by copying others.

Henry should hang onto his TTL chips. In another 25 yeares they'll be
collectable.

 

[Henry]

You can saturate a transistor with the collector voltage below the
base. You typically need just a few tens of millivolts VCE though you
still need the .6V or so on the base (VBE). In the configuration
described, ie an emitter follower, the difficulty is getting the base
voltage .6 above the emitter. Once that has been done, the BE junction
conducts like any other diode.

The way it works is this. The emitter doping level is much higher than
the base, so most of the carriers that diffuse across the junction
come up from the emitter into the base not down from the base into the
emitter. Once in the base region they diffuse around and a few come
out of the base connection. The majority fall into the collector.
Hence the current gain. But unlike getting the B-E junction to
conduct, the carriers do not need any helpto fall into the collector.
Hence it will remain conducting down to very low voltages.

The voltage objection to the emitter follower does not apply if the
circuit is AC coupled. You can then get an output swing which is
practically rail-to-rail.


Bang!

I think Henry probably means "between +V and the load, with the other
side of the load to ground". That would be the emitter follower again.

Yes, I meant between +V and load. Sorry.

Actually I am now trying to interface my camera hot shoe and flash sync to
an RS232 parallel computer port connector. It turns out that under Win95/98
(which is what I am still using) you can use a DLL (Win95IO) to peek and
poke at the parallel port pins. I have an educational version of VB6. I
think I can do this with a simple resistor on the input side (status pins)
and a combination of resistors and a PNP transistor on the output side (data
pins). The Pins, both in and out, seem to default to high.

Indeed. At a few pence each it's worth having enough to burn


Depends on the logic swing. TTL doesn't usually pull to 5V so it may
need a pull-up resistor. Also it depends what Henry needs the output
to do. If he wants to CURRENT switch something, going from NPN to PNP
will mean that what was "on" becomes "off" and vice versa. If he
wants a ogic VOLTAGE drive out of the system then either type of
transistor is going to invert.

It's also important to make sure the logic swings are translated into
a swing past the .6-.7V threshold of the transistor. The same resistor
chain can also be used to limit the current into the base. Switching
speed will be rotten compared to TTL.


Henry should hang onto his TTL chips. In another 25 yeares they'll be
collectable.

I actually found one little old electronics store near me (about 25 miles
away) that still has some TTL chips on stock. Sales person said that no one
has bought any for 10 years or more.

Henry

 

[Derek Potter]

Actually I am now trying to interface my camera hot shoe and flash sync to
an RS232 parallel computer port connector. It turns out that under Win95/98
(which is what I am still using) you can use a DLL (Win95IO) to peek and
poke at the parallel port pins. I have an educational version of VB6. I
think I can do this with a simple resistor on the input side (status pins)
and a combination of resistors and a PNP transistor on the output side (data
pins). The Pins, both in and out, seem to default to high.

Very likely. I used Win95IO.dll ages ago. It was a bit iffy so I
upgraded to WinRT but even that didn't work properly with NT/2000/XP.
under 9x you should be able to peek/poke the parallel printer port. I
doubt whether you can directly access the pins of the serial ports,
though maybe the handshaking lines will be accessible as bits in the
USART registers.

Watch out what you get out of a camera - they often handle a couple of
hundred volts on the flash contacts if a flash gun is connected. This
is bad news for you and worse for electronic bits and pieces.

 

[Henry]

Very likely. I used Win95IO.dll ages ago. It was a bit iffy so I
upgraded to WinRT but even that didn't work properly with NT/2000/XP.
under 9x you should be able to peek/poke the parallel printer port. I
doubt whether you can directly access the pins of the serial ports,
though maybe the handshaking lines will be accessible as bits in the
USART registers.

Watch out what you get out of a camera - they often handle a couple of
hundred volts on the flash contacts if a flash gun is connected. This
is bad news for you and worse for electronic bits and pieces.

Yes, the high voltage of some flash units is one of the reasons that I want
to eventually go to opto-isolators. But my current flash units are 4v for
one, and 15v for the other. They are different types of flash units.

Henry

 

[Derek Potter]

Yes, the high voltage of some flash units is one of the reasons that I want
to eventually go to opto-isolators. But my current flash units are 4v for
one, and 15v for the other. They are different types of flash units.

I learned the hard way - when the gear got wet underground. Ouch!