Transistors

Jim Thompson · May 10, 2013

On Fri, 10 May 2013 11:31:33 -0400, Uncle Steve <stevet810@gmail.com>
wrote:

On Fri, May 10, 2013 at 07:42:40AM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

Ok, this is a question that is not so battery-charger specific, but
is obviously related.

[snip]
but that is supposition that the 2n3055 somehow
limits the base current to some proportionate ratio to the collector-
emitter current available to it. Is that in any way a reasonable
explanation for what is occurring here?

Nope, that is your main conceptual error.

Because the TIP147 is PNP and operates completely differently. Is the
issue here that the 3055 is 'fully on' with 5 uA driving the
2n2222?

Perhaps my frustration with arrangements other than the Sziklai Pair
is due to an irrational desire to decouple the 5v circuit from the 18V
section as much as possible. Other than in a darlington arrangement,
I am still having trouble making 2 2n2222 stages do what I want and
the frustration of things not working as I think they should is really
annoying.

But at pennies per transistor at least it's not expensive to fiddle.

Regards,

Uncle Steve

Uncle Steve · May 10, 2013

On Fri, May 10, 2013 at 10:15:00AM -0500, John Fields wrote:

On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

Needless to say I am happy that my $14.00 DVM has a transistor tester,
because with 10k feeding the 2n2222 the 2n2906 doesn't last very long
with a TIP147 in-circuit. Of course, in that configuration the base
of a 2n3055 still shows 38mA. These results suggest that the
base-emitter current of the 2n3055 cannot go any higher without a
larger power supply, but that is supposition that the 2n3055 somehow
limits the base current to some proportionate ratio to the collector-
emitter current available to it. Is that in any way a reasonable
explanation for what is occurring here?

---
No.

In reality it's the base-to-emitter current which controls the
collector-to-emitter current.

The ratio of Ice to Ibe is called 'beta' or 'Hfe' and describes the
current gain of the device.

From the curves you have, you can see that Hfe varies widely with
collector current and temperature, and also from device to device.

So basically I'm approaching this problem backwards. I guess it's
time to step back and re-read some of the tutorial materials I didn't
understand the first time around.

Regards,

Uncle Steve

--
There should be a special word in the English language to identify
people who create problems and then turn around and offer up their own
tailor-made bogus non-solutions designed to completely avoid the root
causes of the situation under consideration. 'Traitor' might be a
good choice, but lacks the requisite specificity. One of the problems
with contemporary English is it lacks many such words that would
otherwise categorically identify certain kinds of person, place, or
thing -- making it difficult or impossible to think analytically about
such objects. These shortcomings of the English lexicon are
representative of Orwellian linguistics at work in the real world.

Jim Thompson · May 10, 2013

On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com>
wrote:

[snip]

Trying to zero in on the object of your endeavor... it sounds like you
want a trickle charger, BUT, when a 20W load is added, you want this
"regulator" to support that load, IF you have 18V available... is that
your target function?

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice

Uncle Steve · May 10, 2013

On Fri, May 10, 2013 at 08:51:42AM -0700, Jim Thompson wrote:

On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

[snip]

Trying to zero in on the object of your endeavor... it sounds like you
want a trickle charger, BUT, when a 20W load is added, you want this
"regulator" to support that load, IF you have 18V available... is that
your target function?

Well, it has to trickle charge the battery if it is fully charged, but
otherwise must also support the load device if it is turned on, which
it will be most of the time.

Stated more generally, the charger has so much capacity, 3A in this
case, with the transformer I am currently using. So, the power
available to charge the battery is whatever isn't being used by the
load device. I realize I may not need PWM to control this thing after
all. The idea I would came before I understood exactly how lead-acid
batteries behave during the charge cycle. This is why I haven't
bothered using the microcontroller in-circuit yet.

Regards,

Uncle Steve

--
There should be a special word in the English language to identify
people who create problems and then turn around and offer up their own
tailor-made bogus non-solutions designed to completely avoid the root
causes of the situation under consideration. 'Traitor' might be a
good choice, but lacks the requisite specificity. One of the problems
with contemporary English is it lacks many such words that would
otherwise categorically identify certain kinds of person, place, or
thing -- making it difficult or impossible to think analytically about
such objects. These shortcomings of the English lexicon are
representative of Orwellian linguistics at work in the real world.

John Fields · May 10, 2013

On Fri, 10 May 2013 11:31:33 -0400, Uncle Steve <stevet810@gmail.com>
wrote:

On Fri, May 10, 2013 at 07:42:40AM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

Ok, this is a question that is not so battery-charger specific, but
is obviously related.

[snip]
but that is supposition that the 2n3055 somehow
limits the base current to some proportionate ratio to the collector-
emitter current available to it. Is that in any way a reasonable
explanation for what is occurring here?

Nope, that is your main conceptual error.

Because the TIP147 is PNP and operates completely differently. Is the
issue here that the 3055 is 'fully on' with 5 uA driving the
2n2222?

Perhaps my frustration with arrangements other than the Sziklai Pair
is due to an irrational desire to decouple the 5v circuit from the 18V
section as much as possible. Other than in a darlington arrangement,
I am still having trouble making 2 2n2222 stages do what I want and
the frustration of things not working as I think they should is really
annoying.

But at pennies per transistor at least it's not expensive to fiddle.

Regards,

Uncle Steve

---
In the data sheets for your parts, notice that the values for Vce(sat)
and Vbe(sat) are given when Hfe = 10.

What that means is that if you need about (say) 100mA of collector
current for your load, then 10% of that (10mA) needs to be forced
through the b-e junction in order to force the transistor into
saturation.

Looking at a typical circuit using an NPN with a beta of (say) 100 to
300 at an Ic of about 100mA:

+5>-+--------+
| |
[Rb] [50R]
| |
O-> | |
| C
O--B NPN
E
|
GND>---------+

Assuming a Vce(sat) of 0.3V,

Vs - Vce(sat) 5V - 0.3V
Ic = --------------- = ----------- = 0.094 amperes
Rl 50R

Then, to force 10 percent of that through ther b-e junction we figure
out the value of Rb like this:

Vs - Vbe(sat) 5V - 0.8V
Rb = --------------- = --------------- = 446.8 ohms
0.1 * Ic 0.1 * 0.094A

420 ohms and 470 ohms are the closest 5% values and either would work
perfectly well since the natural beta of the transistor is much
greater than 10.

--
JF

Jim Thompson · May 10, 2013

On Fri, 10 May 2013 12:04:52 -0400, Uncle Steve <stevet810@gmail.com>
wrote:

On Fri, May 10, 2013 at 08:51:42AM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

[snip]

Trying to zero in on the object of your endeavor... it sounds like you
want a trickle charger, BUT, when a 20W load is added, you want this
"regulator" to support that load, IF you have 18V available... is that
your target function?

Well, it has to trickle charge the battery if it is fully charged, but
otherwise must also support the load device if it is turned on, which
it will be most of the time.

Stated more generally, the charger has so much capacity, 3A in this
case, with the transformer I am currently using. So, the power
available to charge the battery is whatever isn't being used by the
load device. I realize I may not need PWM to control this thing after
all. The idea I would came before I understood exactly how lead-acid
batteries behave during the charge cycle. This is why I haven't
bothered using the microcontroller in-circuit yet.

Regards,

Uncle Steve

Naaaah! Lose the micro. Micros are only applied to analog functions
by people who don't understand analog >:-}

What you need is a buck switcher charging you battery. Biggest
problem (for you) is tracking the battery temperature to avoid
overcharging.

I'm wondering if it might be possible to modify an off-the-shelf
alternator regulator to control a buck switcher... I'll pursue the
concept.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice

Tim Wescott · May 10, 2013

On Fri, 10 May 2013 09:13:44 -0400, Uncle Steve wrote:

On Thu, May 09, 2013 at 10:57:58PM -0500, Tim Wescott wrote:

OK. First, the drawing came through mangled. I hope I reconstituted
it correctly.

No, this is actually it:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W | |
/
| | \
+------__v 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

Second, the way you have it drawn, the 2N2222 isn't going to act like a
transistor. If you _really_ did it as drawn, the BE junction of the
2N2222 will act like a zener diode at around 6V, because that's just
what little transistors do when you challenge their BE junctions with
too much voltage.

Third, you've drawn the 2N2906 as an NPN, but my references says it's a
PNP.

Frankly, I don't see how this circuit can do _anything_, which leads me
to believe that either it is not drawn as built, or there's some more
zener breakdown or other oddball things happening.

Like JT said: you've got to put current _into_ the base of an NPN, and
pull current _out_ of the base of a PNP for them to act properly. You
don't seem to be making that happen.

My brain isn't trained to recognize the symbols and circuits with any
fluency yet, so when I did the original drawing while tired, it didn't
register. Apologies for the confusion.

When I make silly errors like that when writing code it's also vastly
amusing.

The biggest problem that I see in that drawing is that you've got the
collector of a PNP connected to a supply voltage that's higher than it's
emitter voltage. So the CB junction is forward biased and the BE
junction is reverse biased. So the 2N2906 is going to work in reverse --
the collector will work (poorly) as an emitter, and the emitter will work
(poorly) as a collector. The current gain will be around 1 or less.

The other biggest problem that I see is that you don't have nearly enough
current-limiting resistors in there.

If you turn the 2N2906 around then when the 2N2222 pulls current from its
base it'll deliver LOTS of current to the 2N3055 base. In fact, you'll
generally get too much current all around -- the 2N2222 will pull current
out of the 2N2906 that's limited only by the trickle of base current into
the 2N2222 and the 2N2222's beta, and the 2N2906 will put current into
the base of the 2N3055 that's limited only by it's base current and
beta. Everything will be hugely device- and temperature dependent, so
you'll constantly be fighting the thing not working or burning up.

You really want a resistor into the base of each transistor. Do that,
and you can significantly reduce the 1M resistance to the 2N2222, speed
up switching, and not burn out transistors.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

Uncle Steve · May 10, 2013

On Fri, May 10, 2013 at 09:17:08AM -0700, Jim Thompson wrote:

On Fri, 10 May 2013 12:04:52 -0400, Uncle Steve <stevet810@gmail.com
wrote:

On Fri, May 10, 2013 at 08:51:42AM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

[snip]

Trying to zero in on the object of your endeavor... it sounds like you
want a trickle charger, BUT, when a 20W load is added, you want this
"regulator" to support that load, IF you have 18V available... is that
your target function?

Well, it has to trickle charge the battery if it is fully charged, but
otherwise must also support the load device if it is turned on, which
it will be most of the time.

Stated more generally, the charger has so much capacity, 3A in this
case, with the transformer I am currently using. So, the power
available to charge the battery is whatever isn't being used by the
load device. I realize I may not need PWM to control this thing after
all. The idea I would came before I understood exactly how lead-acid
batteries behave during the charge cycle. This is why I haven't
bothered using the microcontroller in-circuit yet.

Regards,

Uncle Steve

Naaaah! Lose the micro. Micros are only applied to analog functions
by people who don't understand analog >:-}

I expect it's rather difficult to build an analog circuit to send
"powerok" at 38400bps to a serial port.

What you need is a buck switcher charging you battery. Biggest
problem (for you) is tracking the battery temperature to avoid
overcharging.

PCM still isn't out of the question. The way battery voltage rises,
it might be feasible to use PCM if the input supply is variable, and
have the thing to a proper job independent of the voltage/current
capacity of the transformer. Who knows, perhaps next week a 36V-10A
switching power supply will fall from the sky and I'll want to use it
to charge the battery.

Mind you, I want to charge this battery fairly slowly to minimize
hydrogen-gas production as it will live indoors. Pulse charging
lead-acid may also be good to reduce sulfation of the plates, but I
don't know yet how much.

I'm wondering if it might be possible to modify an off-the-shelf
alternator regulator to control a buck switcher... I'll pursue the
concept.

I wonder if I can make an inductor out of lacquered motor-winding wire.

Regards,

Uncle Steve

--
There should be a special word in the English language to identify
people who create problems and then turn around and offer up their own
tailor-made bogus non-solutions designed to completely avoid the root
causes of the situation under consideration. 'Traitor' might be a
good choice, but lacks the requisite specificity. One of the problems
with contemporary English is it lacks many such words that would
otherwise categorically identify certain kinds of person, place, or
thing -- making it difficult or impossible to think analytically about
such objects. These shortcomings of the English lexicon are
representative of Orwellian linguistics at work in the real world.

Uncle Steve · May 10, 2013

On Fri, May 10, 2013 at 11:40:38AM -0500, Tim Wescott wrote:

On Fri, 10 May 2013 09:13:44 -0400, Uncle Steve wrote:

On Thu, May 09, 2013 at 10:57:58PM -0500, Tim Wescott wrote:

OK. First, the drawing came through mangled. I hope I reconstituted
it correctly.

No, this is actually it:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W | |
/
| | \
+------__v 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

Second, the way you have it drawn, the 2N2222 isn't going to act like a
transistor. If you _really_ did it as drawn, the BE junction of the
2N2222 will act like a zener diode at around 6V, because that's just
what little transistors do when you challenge their BE junctions with
too much voltage.

Third, you've drawn the 2N2906 as an NPN, but my references says it's a
PNP.

Frankly, I don't see how this circuit can do _anything_, which leads me
to believe that either it is not drawn as built, or there's some more
zener breakdown or other oddball things happening.

Like JT said: you've got to put current _into_ the base of an NPN, and
pull current _out_ of the base of a PNP for them to act properly. You
don't seem to be making that happen.

My brain isn't trained to recognize the symbols and circuits with any
fluency yet, so when I did the original drawing while tired, it didn't
register. Apologies for the confusion.

When I make silly errors like that when writing code it's also vastly
amusing.

The biggest problem that I see in that drawing is that you've got the
collector of a PNP connected to a supply voltage that's higher than it's
emitter voltage. So the CB junction is forward biased and the BE
junction is reverse biased. So the 2N2906 is going to work in reverse --
the collector will work (poorly) as an emitter, and the emitter will work
(poorly) as a collector. The current gain will be around 1 or less.

In reality the PNP is connected the right way.

The other biggest problem that I see is that you don't have nearly enough
current-limiting resistors in there.

Well that's my thought too, but...

If you turn the 2N2906 around then when the 2N2222 pulls current from its
base it'll deliver LOTS of current to the 2N3055 base. In fact, you'll
generally get too much current all around -- the 2N2222 will pull current
out of the 2N2906 that's limited only by the trickle of base current into
the 2N2222 and the 2N2222's beta, and the 2N2906 will put current into
the base of the 2N3055 that's limited only by it's base current and
beta. Everything will be hugely device- and temperature dependent, so
you'll constantly be fighting the thing not working or burning up.

10k or 1M, the current into the base of the 3055 is never more than
38mA. Combined beta of the two small transistors is 180000, so Vb-e
of the 3055 sort-of ought to be 90mA. If I used a TIP122 it probably
would be, but then I'd really only need 3mA to turn it on. I suppose
this mean I could use a 2.2M resistor on the base of the 2n2222.

I have five 2n3055 parts so it would probably be better to actually
use them, not that they're all that expensive in the first place.

You really want a resistor into the base of each transistor. Do that,
and you can significantly reduce the 1M resistance to the 2N2222, speed
up switching, and not burn out transistors.

I dunno. I'd be happier if I could just burn out transistors a little
more slowly so I could see it happening and stop it before they are
completely destroyed.

I have a 2n2906 that works in-circuit, but doesn't register at all on
my cheap transistor tester. The good ones show 120Hfe.

I guess it's just part of that "learn by destroying" thing.

Regards,

Uncle Steve

--
There should be a special word in the English language to identify
people who create problems and then turn around and offer up their own
tailor-made bogus non-solutions designed to completely avoid the root
causes of the situation under consideration. 'Traitor' might be a
good choice, but lacks the requisite specificity. One of the problems
with contemporary English is it lacks many such words that would
otherwise categorically identify certain kinds of person, place, or
thing -- making it difficult or impossible to think analytically about
such objects. These shortcomings of the English lexicon are
representative of Orwellian linguistics at work in the real world.

Tim Wescott · May 10, 2013

On Fri, 10 May 2013 13:24:45 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 11:40:38AM -0500, Tim Wescott wrote:
On Fri, 10 May 2013 09:13:44 -0400, Uncle Steve wrote:

On Thu, May 09, 2013 at 10:57:58PM -0500, Tim Wescott wrote:

OK. First, the drawing came through mangled. I hope I
reconstituted it correctly.

No, this is actually it:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W | |
/
| | \
+------__v 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

Second, the way you have it drawn, the 2N2222 isn't going to act
like a transistor. If you _really_ did it as drawn, the BE junction
of the 2N2222 will act like a zener diode at around 6V, because
that's just what little transistors do when you challenge their BE
junctions with too much voltage.

Third, you've drawn the 2N2906 as an NPN, but my references says
it's a PNP.

Frankly, I don't see how this circuit can do _anything_, which leads
me to believe that either it is not drawn as built, or there's some
more zener breakdown or other oddball things happening.

Like JT said: you've got to put current _into_ the base of an NPN,
and pull current _out_ of the base of a PNP for them to act
properly. You don't seem to be making that happen.

My brain isn't trained to recognize the symbols and circuits with any
fluency yet, so when I did the original drawing while tired, it
didn't register. Apologies for the confusion.

When I make silly errors like that when writing code it's also vastly
amusing.

The biggest problem that I see in that drawing is that you've got the
collector of a PNP connected to a supply voltage that's higher than
it's emitter voltage. So the CB junction is forward biased and the BE
junction is reverse biased. So the 2N2906 is going to work in reverse
-- the collector will work (poorly) as an emitter, and the emitter will
work (poorly) as a collector. The current gain will be around 1 or
less.

In reality the PNP is connected the right way.

Not if you have it the way it's pictured!

The other biggest problem that I see is that you don't have nearly
enough current-limiting resistors in there.

Well that's my thought too, but...

If you turn the 2N2906 around then when the 2N2222 pulls current from
its base it'll deliver LOTS of current to the 2N3055 base. In fact,
you'll generally get too much current all around -- the 2N2222 will
pull current out of the 2N2906 that's limited only by the trickle of
base current into the 2N2222 and the 2N2222's beta, and the 2N2906 will
put current into the base of the 2N3055 that's limited only by it's
base current and beta. Everything will be hugely device- and
temperature dependent, so you'll constantly be fighting the thing not
working or burning up.

10k or 1M, the current into the base of the 3055 is never more than
38mA. Combined beta of the two small transistors is 180000, so Vb-e of
the 3055 sort-of ought to be 90mA. If I used a TIP122 it probably would
be, but then I'd really only need 3mA to turn it on. I suppose this
mean I could use a 2.2M resistor on the base of the 2n2222.

The Vbe of the 3055 can't be 90mA, because Vbe denotes voltage.

Unless the 3055 is getting close to saturation you should be flowing
plenty of base current out of that 2906, and get way more than 38mA into
the base of the 3055.

So something's just not right.

Try connecting a 1K resistor to the base of the 2906 and manually
switching it to ground. You should get tons-o-current into the base of
the 3055 _assuming_ that there's enough voltage drop across the 2906 for
it to work.

If that doesn't lead to joy, back up and try connecting the base of the
3055 to +18V with a 2.2 to 10 ohm resistor -- that should turn the 3055
on good and hard. If it doesn't, don't mess around with anything else
until you get that figured out.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

Tim Wescott · May 10, 2013

On Fri, 10 May 2013 13:01:57 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 09:17:08AM -0700, Jim Thompson wrote:
On Fri, 10 May 2013 12:04:52 -0400, Uncle Steve <stevet810@gmail.com
wrote:

On Fri, May 10, 2013 at 08:51:42AM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve
stevet810@gmail.com> wrote:

[snip]

Trying to zero in on the object of your endeavor... it sounds like
you want a trickle charger, BUT, when a 20W load is added, you want
this "regulator" to support that load, IF you have 18V available...
is that your target function?

Well, it has to trickle charge the battery if it is fully charged, but
otherwise must also support the load device if it is turned on, which
it will be most of the time.

Stated more generally, the charger has so much capacity, 3A in this
case, with the transformer I am currently using. So, the power
available to charge the battery is whatever isn't being used by the
load device. I realize I may not need PWM to control this thing after
all. The idea I would came before I understood exactly how lead-acid
batteries behave during the charge cycle. This is why I haven't
bothered using the microcontroller in-circuit yet.

Regards,

Uncle Steve

Naaaah! Lose the micro. Micros are only applied to analog functions
by people who don't understand analog >:-}

I expect it's rather difficult to build an analog circuit to send
"powerok" at 38400bps to a serial port.

Ignore Jim's ranting. He's just a superannuated old geezer. Everyone
knows that analog circuits are only applied to analog functions by people
who don't understand microprocessors.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

Uncle Steve · May 10, 2013

On Fri, May 10, 2013 at 12:50:41PM -0500, Tim Wescott wrote:

On Fri, 10 May 2013 13:24:45 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 11:40:38AM -0500, Tim Wescott wrote:
On Fri, 10 May 2013 09:13:44 -0400, Uncle Steve wrote:

On Thu, May 09, 2013 at 10:57:58PM -0500, Tim Wescott wrote:

OK. First, the drawing came through mangled. I hope I
reconstituted it correctly.

No, this is actually it:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W
| | /
| | \
+-----v__| 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

Second, the way you have it drawn, the 2N2222 isn't going to act
like a transistor. If you _really_ did it as drawn, the BE junction
of the 2N2222 will act like a zener diode at around 6V, because
that's just what little transistors do when you challenge their BE
junctions with too much voltage.

Third, you've drawn the 2N2906 as an NPN, but my references says
it's a PNP.

Frankly, I don't see how this circuit can do _anything_, which leads
me to believe that either it is not drawn as built, or there's some
more zener breakdown or other oddball things happening.

Like JT said: you've got to put current _into_ the base of an NPN,
and pull current _out_ of the base of a PNP for them to act
properly. You don't seem to be making that happen.

My brain isn't trained to recognize the symbols and circuits with any
fluency yet, so when I did the original drawing while tired, it
didn't register. Apologies for the confusion.

When I make silly errors like that when writing code it's also vastly
amusing.

The biggest problem that I see in that drawing is that you've got the
collector of a PNP connected to a supply voltage that's higher than
it's emitter voltage. So the CB junction is forward biased and the BE
junction is reverse biased. So the 2N2906 is going to work in reverse
-- the collector will work (poorly) as an emitter, and the emitter will
work (poorly) as a collector. The current gain will be around 1 or
less.

In reality the PNP is connected the right way.

Not if you have it the way it's pictured!

What picture?

The other biggest problem that I see is that you don't have nearly
enough current-limiting resistors in there.

Well that's my thought too, but...

If you turn the 2N2906 around then when the 2N2222 pulls current from
its base it'll deliver LOTS of current to the 2N3055 base. In fact,
you'll generally get too much current all around -- the 2N2222 will
pull current out of the 2N2906 that's limited only by the trickle of
base current into the 2N2222 and the 2N2222's beta, and the 2N2906 will
put current into the base of the 2N3055 that's limited only by it's
base current and beta. Everything will be hugely device- and
temperature dependent, so you'll constantly be fighting the thing not
working or burning up.

10k or 1M, the current into the base of the 3055 is never more than
38mA. Combined beta of the two small transistors is 180000, so Vb-e of
the 3055 sort-of ought to be 90mA. If I used a TIP122 it probably would
be, but then I'd really only need 3mA to turn it on. I suppose this
mean I could use a 2.2M resistor on the base of the 2n2222.

The Vbe of the 3055 can't be 90mA, because Vbe denotes voltage.

Base-emitter current is what I meant.

Unless the 3055 is getting close to saturation you should be flowing
plenty of base current out of that 2906, and get way more than 38mA into
the base of the 3055.

I'll measure it again. 37mA with one 2n3055 and 39mA with another.
I have a TIP122, and unlike the PNP TIP147, it shows .5mA when used as
a drop-in replacement for the 2n3055. So it seems that the base
current is being limited to what is required to saturate the power
transistor

So something's just not right.

Try connecting a 1K resistor to the base of the 2906 and manually
switching it to ground. You should get tons-o-current into the base of
the 3055 _assuming_ that there's enough voltage drop across the 2906 for
it to work.

44mA.

If that doesn't lead to joy, back up and try connecting the base of the
3055 to +18V with a 2.2 to 10 ohm resistor -- that should turn the 3055
on good and hard. If it doesn't, don't mess around with anything else
until you get that figured out.

10 ohms: 38mA

At this point the circuit is pretty minimal so it's difficult to say
whether there could be a component defect. Ignoring the LED/2.2k pair
on the output, that leaves a 10 ohm resister, the 2n3055, and a 5 ohm,
10 watt resistor for a load -- plus some wire.

Regards,

Uncle Steve

--
There should be a special word in the English language to identify
people who create problems and then turn around and offer up their own
tailor-made bogus non-solutions designed to completely avoid the root
causes of the situation under consideration. 'Traitor' might be a
good choice, but lacks the requisite specificity. One of the problems
with contemporary English is it lacks many such words that would
otherwise categorically identify certain kinds of person, place, or
thing -- making it difficult or impossible to think analytically about
such objects. These shortcomings of the English lexicon are
representative of Orwellian linguistics at work in the real world.

Tim Wescott · May 10, 2013

On Fri, 10 May 2013 15:07:05 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 12:50:41PM -0500, Tim Wescott wrote:
On Fri, 10 May 2013 13:24:45 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 11:40:38AM -0500, Tim Wescott wrote:
On Fri, 10 May 2013 09:13:44 -0400, Uncle Steve wrote:

On Thu, May 09, 2013 at 10:57:58PM -0500, Tim Wescott wrote:

OK. First, the drawing came through mangled. I hope I
reconstituted it correctly.

No, this is actually it:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W | |
/
| | \
+-----v__| 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

Second, the way you have it drawn, the 2N2222 isn't going to act
like a transistor. If you _really_ did it as drawn, the BE
junction of the 2N2222 will act like a zener diode at around 6V,
because that's just what little transistors do when you challenge
their BE junctions with too much voltage.

Third, you've drawn the 2N2906 as an NPN, but my references says
it's a PNP.

Frankly, I don't see how this circuit can do _anything_, which
leads me to believe that either it is not drawn as built, or
there's some more zener breakdown or other oddball things
happening.

Like JT said: you've got to put current _into_ the base of an
NPN, and pull current _out_ of the base of a PNP for them to act
properly. You don't seem to be making that happen.

My brain isn't trained to recognize the symbols and circuits with
any fluency yet, so when I did the original drawing while tired,
it didn't register. Apologies for the confusion.

When I make silly errors like that when writing code it's also
vastly amusing.

The biggest problem that I see in that drawing is that you've got
the collector of a PNP connected to a supply voltage that's higher
than it's emitter voltage. So the CB junction is forward biased and
the BE junction is reverse biased. So the 2N2906 is going to work
in reverse -- the collector will work (poorly) as an emitter, and
the emitter will work (poorly) as a collector. The current gain
will be around 1 or less.

In reality the PNP is connected the right way.

Not if you have it the way it's pictured!

What picture?

_your_ picture. The emitter of the 2N2906 should be connected to +18V,
the collector should connect to the emitter of the 2N3055. You have the
emitter and the collector swapped.

The other biggest problem that I see is that you don't have nearly
enough current-limiting resistors in there.

Well that's my thought too, but...

If you turn the 2N2906 around then when the 2N2222 pulls current
from its base it'll deliver LOTS of current to the 2N3055 base. In
fact, you'll generally get too much current all around -- the 2N2222
will pull current out of the 2N2906 that's limited only by the
trickle of base current into the 2N2222 and the 2N2222's beta, and
the 2N2906 will put current into the base of the 2N3055 that's
limited only by it's base current and beta. Everything will be
hugely device- and temperature dependent, so you'll constantly be
fighting the thing not working or burning up.

10k or 1M, the current into the base of the 3055 is never more than
38mA. Combined beta of the two small transistors is 180000, so Vb-e
of the 3055 sort-of ought to be 90mA. If I used a TIP122 it probably
would be, but then I'd really only need 3mA to turn it on. I suppose
this mean I could use a 2.2M resistor on the base of the 2n2222.

The Vbe of the 3055 can't be 90mA, because Vbe denotes voltage.

Base-emitter current is what I meant.

Unless the 3055 is getting close to saturation you should be flowing
plenty of base current out of that 2906, and get way more than 38mA
into the base of the 3055.

I'll measure it again. 37mA with one 2n3055 and 39mA with another. I
have a TIP122, and unlike the PNP TIP147, it shows .5mA when used as a
drop-in replacement for the 2n3055. So it seems that the base current
is being limited to what is required to saturate the power transistor

So something's just not right.

Try connecting a 1K resistor to the base of the 2906 and manually
switching it to ground. You should get tons-o-current into the base of
the 3055 _assuming_ that there's enough voltage drop across the 2906
for it to work.

44mA.

If that doesn't lead to joy, back up and try connecting the base of the
3055 to +18V with a 2.2 to 10 ohm resistor -- that should turn the 3055
on good and hard. If it doesn't, don't mess around with anything else
until you get that figured out.

10 ohms: 38mA

At this point the circuit is pretty minimal so it's difficult to say
whether there could be a component defect. Ignoring the LED/2.2k pair
on the output, that leaves a 10 ohm resister, the 2n3055, and a 5 ohm,
10 watt resistor for a load -- plus some wire.

Hah. OK, 10 ohms implies 380mV. That, plus the VBE of the 2N3055 is a
bit above 1V. You probably have something similar with the 2N2906, where
the VCB of the 2N3055 drops, dropping the VCE of the 2N2906, which makes
it go into saturation and reduces base current to the 2N3055 --
particularly if the 2N2906 is backwards.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

Tim Wescott · May 10, 2013

On Fri, 10 May 2013 10:22:39 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 11:00:13PM +1000, Phil Allison wrote:

"Uncle Steve"

Nevertheless,
there is nothing like making measurements to show what is in fact
going on with these devices.

** Give this man a Kewpie doll.

I've always wanted a Kewpie doll.

It's interesting that all of the literature and web-pages I've read so
far have failed to impart an accurate sense of what these things do. I
don't know whether that is because the learning curve is so steep, or
whether the terminology is truly confusing to the uninitiated. I think
it might be easier to understand electronics if circuit diagrams
represented electron flow and electron charge potential more obviously.
After all it's the electrons moving around that gets work done, correct?

Yes, it's electrons moving around that gets the work done. But in the
end it's usually easier to just think of current as a magic fluid that
goes in the direction indicated.

I'm not sure what you mean by "accurate sense of what these things do".
If you mean transistors, yes, the subject is both deep and wide. You can
do most anything with one of two models, though:

Model 1: there's a fixed drop from base to emitter of 0.65V, the
collector current equals beta * base current, and odd things happen in
saturation.

Model 2: There's a diode from base to emitter, and a diode from base to
collector. Both either point to or away from the base (that's the
meaning of the arrow on the emitter lead). There's a current source in
parallel with the collector. When current flows through the base-emitter
diode, the current source in parallel with the collector flows almost the
same amount of current (beta / (beta - 1), in fact, also known as alpha
in the old literature).

With that model you not only predict beta, but you can predict a lot of
behavior around saturation because that base-collector diode gets forward-
biased in saturation and starts to flow its own current.

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

John Fields · May 10, 2013

On Fri, 10 May 2013 15:07:05 -0400, Uncle Steve <stevet810@gmail.com>
wrote:

I'll measure it again. 37mA with one 2n3055 and 39mA with another.
I have a TIP122, and unlike the PNP TIP147, it shows .5mA when used as
a drop-in replacement for the 2n3055. So it seems that the base
current is being limited to what is required to saturate the power
transistor.

---
Since the power transistor is wired as an emitter follower, you'll
never be able to run it into saturation, but it'll probably be good
enough for your application.

The best you'll be able to do is to get the emitter voltage equal to
the supply voltage minus the sum of Vbe of the power transistor and
Vce(sat) of the PNP.

If you like, try the circuit I posted for you; it just works.

--
JF

Jamie · May 10, 2013

Uncle Steve wrote:

On Thu, May 09, 2013 at 07:11:43PM -0700, Jim Thompson wrote:

On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

Ok, this is a question that is not so battery-charger specific, but
is obviously related.

So here's a circuit fragment:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W
| | /
| | \
+------__v 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

So this circuit shows ~38mA going in to the base of the 2n3055, and
~13.9V across the 5 ohm resistor. It gets warm quickly.

If I change the 2n3055 for a TIP147 (PNP, beta 1k) and make the
obvious changes to the 2n2906 to make it work, there is ~80mA current
at the base of the TIP147, and ~14.1V across the 10W resistor.

WTF, over?

Needless to say I am happy that my $14.00 DVM has a transistor tester,
because with 10k feeding the 2n2222 the 2n2906 doesn't last very long
with a TIP147 in-circuit. Of course, in that configuration the base
of a 2n3055 still shows 38mA. These results suggest that the
base-emitter current of the 2n3055 cannot go any higher without a
larger power supply, but that is supposition that the 2n3055 somehow
limits the base current to some proportionate ratio to the collector-
emitter current available to it. Is that in any way a reasonable
explanation for what is occurring here?

Draw it up real purty, so you can tell PNP's from NPN's and can see
that current can only flow OUT of the NPN emitter and INTO a PNP
emitter... and DO THE MATH ;-)

Sorry, I was tired and really mangled the schematic. Fixed above.

Regards,

Uncle Steve

The 2N2906 is orientated incorrectly. The emitter should be on the

18Volts and 2N3055 collector and of course the collector to the base
of the 2N3055.

As for the base drive to the 2N2906 from the 2N2222 collector, you
should be using a resistor, otherwise you'll let the blue smoke out.

Jamie

Jim Thompson · May 10, 2013

On Fri, 10 May 2013 12:53:49 -0500, Tim Wescott <tim@seemywebsite.com>
wrote:

On Fri, 10 May 2013 13:01:57 -0400, Uncle Steve wrote:

On Fri, May 10, 2013 at 09:17:08AM -0700, Jim Thompson wrote:
On Fri, 10 May 2013 12:04:52 -0400, Uncle Steve <stevet810@gmail.com
wrote:

On Fri, May 10, 2013 at 08:51:42AM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve
stevet810@gmail.com> wrote:

[snip]

Trying to zero in on the object of your endeavor... it sounds like
you want a trickle charger, BUT, when a 20W load is added, you want
this "regulator" to support that load, IF you have 18V available...
is that your target function?

Well, it has to trickle charge the battery if it is fully charged, but
otherwise must also support the load device if it is turned on, which
it will be most of the time.

Stated more generally, the charger has so much capacity, 3A in this
case, with the transformer I am currently using. So, the power
available to charge the battery is whatever isn't being used by the
load device. I realize I may not need PWM to control this thing after
all. The idea I would came before I understood exactly how lead-acid
batteries behave during the charge cycle. This is why I haven't
bothered using the microcontroller in-circuit yet.

Regards,

Uncle Steve

Naaaah! Lose the micro. Micros are only applied to analog functions
by people who don't understand analog >:-}

I expect it's rather difficult to build an analog circuit to send
"powerok" at 38400bps to a serial port.

Ignore Jim's ranting. He's just a superannuated old geezer. Everyone
knows that analog circuits are only applied to analog functions by people
who don't understand microprocessors.

Jasen Betts · May 10, 2013

On 2013-05-10, Uncle Steve <stevet810@gmail.com> wrote:

On Fri, May 10, 2013 at 07:38:44AM -0700, Jim Thompson wrote:
On Fri, 10 May 2013 08:58:49 -0400, Uncle Steve <stevet810@gmail.com
wrote:

On Thu, May 09, 2013 at 07:11:43PM -0700, Jim Thompson wrote:
On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

Ok, this is a question that is not so battery-charger specific, but
is obviously related.

So here's a circuit fragment:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W
| | /
| | \
+------v_| 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

PNP collector must be more negative than its emitter... swap emitter
and collector on the 2N2906.

I wish I could run that fancy ASCII schematic CAD tool, but I don't
use Windows.

hmm, your newsreader claims to be "nn" some sort of *nix prolly.

Assuming you're using X and something with an X86 processor
install "wine" and "xchm" then having downloaded the installer do

wine swcadiii.exe

after the installer ends do

wine 'C:\Program\ Files/LTC/SwCADIII/scad3.exe'

to launch it.

I wouldn't characterize 3A as a trickle, but in essence I hope to end
up with a battery charger that will charge the battery while it is
under a moderate 20W+ load.

Many off-the-shelf regulated chargers will do that reasonably well.
Jim's the expert on lead-acid battery charging.

I have bit-banged serial out on the
microcontroller, so it will log status to the load device that way,
and inform it when mains power fails.

that bit is probably going to need to be custom, or overpriced.

--
ââ 100% natural

--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

Jasen Betts · May 10, 2013

On 2013-05-10, Uncle Steve <stevet810@gmail.com> wrote:

On Fri, May 10, 2013 at 11:13:48AM +0000, Jasen Betts wrote:
On 2013-05-10, Uncle Steve <stevet810@gmail.com> wrote:
Ok, this is a question that is not so battery-charger specific, but
is obviously related.

If you want to use a microcontroller you have to design
the low side switch part so that current flows out of
the 5V node, not into it.

so dont do this:

| |
| |
\ SW |
\ |
| |
5V +-----------+ |
|
GND +-----------+-------------------------+

When I actually hook this up to a microcontroller the 5V will be the
output from the pin designated for PWM, assuming I stick with that
strategy. As I said earlier, the switch is there to simulate a
microcontroller.

but it doesn't even come close to simulating a microcontroller
output.

|
|
|/
5V--+--[10K]--|
| |>|
| |
`----|<-----+
|
.---->|-----+
| |
| \ sw
| \
| |
+-----------+----------

Not sure what the purpose of the second diode is there. Doesn't look
like it can do much of anything.

it's part of the static protection circuitry typical of
microcontrollers in your circuit it does nothing. It's
there for completeness. if I left it out there would
be four of five people telling me. These diodes are
pretty wimpy, it's usualy best that they never conduct
any current.

--
ââ 100% natural

--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

John Fields · May 10, 2013

On Fri, 10 May 2013 17:47:58 -0400, Jamie
<jamie_ka1lpa_not_valid_after_ka1lpa_@charter.net> wrote:

Uncle Steve wrote:
On Thu, May 09, 2013 at 07:11:43PM -0700, Jim Thompson wrote:

On Thu, 09 May 2013 20:59:01 -0400, Uncle Steve <stevet810@gmail.com
wrote:

Ok, this is a question that is not so battery-charger specific, but
is obviously related.

So here's a circuit fragment:

18V +----+--------__^------/\/\/--------+
| 2n3055| 0R5 |
| | 5W /
| | \ 5R10W
| | /
| | \
+------__v 2n2906 |
| |
| |
2n2222 +---| |
| |>----+ |
| | |
/ | |
1M \ | |
/ | |
\ | |
| | |
| | |
\ SW | |
\ | |
| | |
5V +-----+ | |
| |
GND +---------------+------------------+

So this circuit shows ~38mA going in to the base of the 2n3055, and
~13.9V across the 5 ohm resistor. It gets warm quickly.

If I change the 2n3055 for a TIP147 (PNP, beta 1k) and make the
obvious changes to the 2n2906 to make it work, there is ~80mA current
at the base of the TIP147, and ~14.1V across the 10W resistor.

WTF, over?

Needless to say I am happy that my $14.00 DVM has a transistor tester,
because with 10k feeding the 2n2222 the 2n2906 doesn't last very long
with a TIP147 in-circuit. Of course, in that configuration the base
of a 2n3055 still shows 38mA. These results suggest that the
base-emitter current of the 2n3055 cannot go any higher without a
larger power supply, but that is supposition that the 2n3055 somehow
limits the base current to some proportionate ratio to the collector-
emitter current available to it. Is that in any way a reasonable
explanation for what is occurring here?

Draw it up real purty, so you can tell PNP's from NPN's and can see
that current can only flow OUT of the NPN emitter and INTO a PNP
emitter... and DO THE MATH ;-)

Sorry, I was tired and really mangled the schematic. Fixed above.

Regards,

Uncle Steve

The 2N2906 is orientated incorrectly. The emitter should be on the
18Volts and 2N3055 collector and of course the collector to the base
of the 2N3055.

As for the base drive to the 2N2906 from the 2N2222 collector, you
should be using a resistor, otherwise you'll let the blue smoke out.

Jamie

---
Wow, an echo claiming to be source?

--
JF

Transistors

Jim Thompson

Guest

Uncle Steve

Guest

Jim Thompson

Guest

Uncle Steve

Guest

John Fields

Guest

Jim Thompson

Guest

Tim Wescott

Guest

Uncle Steve

Guest

Uncle Steve

Guest

Tim Wescott

Guest

Tim Wescott

Guest

Uncle Steve

Guest

Tim Wescott

Guest

Tim Wescott

Guest

John Fields

Guest

Jamie

Guest

Jim Thompson

Guest

Jasen Betts

Guest

Jasen Betts

Guest

John Fields

Guest

Log in

Welcome to EDABoard.com

Sponsor