LM317 applications

[powersupplies]

Hello all,

I need some help with my study of a few applications from the
datasheet of the LM317.

First of all, the sheet can be downloaded here
http://www.national.com/ds.cgi/LM/LM117.pdf

My first question is about the High Gain Amplifier application number
00906318. How do I get the formula for the gain. I can see that
constant current will flow into the collector, but how does this
do anything usefull?

My second question is about the 5A Constant Voltage / Constant
Current Regulator, number 00906315. I have managed to figure
out that the opamp will be the current limiter, but how exactly does
this work. For example, if the opamp "kicks-in", current will flow
through R4 from the output to the opamp, but how is that limiting
the current?

 

[Joe McElvenney]

Hi,

My first question is about the High Gain Amplifier application number
00906318. How do I get the formula for the gain. I can see that
constant current will flow into the collector, but how does this
do anything usefull?

Current sources have a high impedance by definition and the gain of
such a configuration is gm x RL where 'gm' is about 40mA/V per mA of
collector current. However I'm not sure that this relationship holds
exactly for the LM195.

My second question is about the 5A Constant Voltage / Constant
Current Regulator, number 00906315. I have managed to figure
out that the opamp will be the current limiter, but how exactly does
this work. For example, if the opamp "kicks-in", current will flow
through R4 from the output to the opamp, but how is that limiting
the current?

Increasing current through R3 causes the increased voltage at its
top end to force the output of the LM301A low via its inverting
terminal. This robs current from the 'adjust' terminal's path through
R8 causing the LM317's output voltage to fall.

R4's function is to supply current to the LED when the constant
current mode starts up. Under normal operation D1 is reverse-biassed
and so out of the game.


Cheers - Joe

 

[powersupplies]

Hi !!

My first question is about the High Gain Amplifier application number
00906318. How do I get the formula for the gain. I can see that
constant current will flow into the collector, but how does this
do anything usefull?

Current sources have a high impedance by definition and the gain of
such a configuration is gm x RL where 'gm' is about 40mA/V per mA of
collector current. However I'm not sure that this relationship holds
exactly for the LM195.

I am used to working with hfe or beta, can you give the gain in that format
?
What is this gm anyway, I thought we use it only for FET's.

My second question is about the 5A Constant Voltage / Constant
Current Regulator, number 00906315. I have managed to figure
out that the opamp will be the current limiter, but how exactly does
this work. For example, if the opamp "kicks-in", current will flow
through R4 from the output to the opamp, but how is that limiting
the current?

Increasing current through R3 causes the increased voltage at its
top end to force the output of the LM301A low via its inverting
terminal. This robs current from the 'adjust' terminal's path through
R8 causing the LM317's output voltage to fall.

R4's function is to supply current to the LED when the constant
current mode starts up. Under normal operation D1 is reverse-biassed
and so out of the game.

Yes, I am starting to see this now. I think the limiting is a result of

"robbing
current" from R8 rather than from ADJ pin. And D2 makes sure that the
led is always "almost on" so than when current limiting starts, it will
imediately come on, otherwise there would be another 0.6V barrier before
the led could come on. What do you say?

 

[Joe McElvenney]

Hi,

Gm, or transconductance, is a device property common to
both transistors and FETs although not often quoted with
the former. If you look at the data sheet for the LM195
though, you will see a curve of transconductance against
collector current - HFE or hfe not being specified.
Remember it is not a simple transistor but an integrated
circuit.

The LM317 has a 2.4 ohm resistor between the output and
the 'adj' terminals so as to give a constant 0.5A output (a
current for which the LM195 appears to have a wopping gm of
3A/V). You would need to study the curves given to come up
with a gain estimate. I must confess that I couldn't do it
in this case without breadboarding the thing first as the
actual AC impedance of the LM317 in that mode is an unknown
quantity.

Re-reading my comment on the shut-down circuit of the 5A
regulator, I can see that it is ambiguous. When I say that
the LM301A robs current from the adjust terminal through R
8, I meant that it robs the current that would normally
flow through R8 but flows into the output terminal of the
op-amp instead; R4 being of too high a value to influence
the output voltage of the regulator. Have a look at the
output voltage formula and you'll see that both terms are
highly dependent on R2 and that is what the current sink
into the LM301A is in effect reducing.


Cheers - Joe

 

[Winfield Hill]

Joe McElvenney wrote...

Gm, or transconductance, is a device property common to
both transistors and FETs although not often quoted with
the former. ...

Just to emphasize the point, BJT transistor g_m values are
perfectly and completely given by the Ebers-Moll equation,
as a function of collector current, g_m = Ic / kT/q, and
therefore does not appear on any transistor data sheets.

This point was repeatedly made earlier (in substantial and
careful detail) to the O.P. by Kevin Aylward in the thread,
"High gain amplifier?"

Transconductance is the single most predictable parameter
of BJTs, and this property is carefully exploited by master
engineers like Barrie Gilbert as Analog Devices.

Sadly, some other valuable BJT device-physics parameters,
like the Early voltage, which kevin also mentioned, also
do not appear on BJT data sheets. This parameter is fab
process dependent and not predictable by the designer, so
we could use more info from the manufacturer. <sigh>

Thanks,
- Win

whill_at_picovolt-dot-com

 

[powersupplies]

Due to server problems I am sending replies to the two previous
posters right here:

Sorry, I incorrectly assumed it is a more or less normal transistor.
Obviously it would not work for a general purpose transistor
as there is no DC biasing. (perhaps it would work for the
positive cycle)
I still don't understand why current sources are high impedance?

Gm, or transconductance, is a device property common to
both transistors and FETs although not often quoted with
the former. ...

Just to emphasize the point, BJT transistor g_m values are
perfectly and completely given by the Ebers-Moll equation,
as a function of collector current, g_m = Ic / kT/q, and
therefore does not appear on any transistor data sheets.

This point was repeatedly made earlier (in substantial and
careful detail) to the O.P. by Kevin Aylward in the thread,
"High gain amplifier?"

If we have the h parameters there is no need to derive gm, this
is done for us in the datasheets.

Transconductance is the single most predictable parameter
of BJTs, and this property is carefully exploited by master
engineers like Barrie Gilbert as Analog Devices.

 

[Joe McElvenney]

Hi,

Connect a 10V supply to ground through a 1 Meg resistor
in series with a further 100 ohm resistor. What value of
current flows?

Now double the 100 ohm resistor to 200 ohms. What
current flows now?

Why are they very nearly the same?

If a circuit actively resists change to the current
through itself while at the same time passing a substantial
amount, then it has a high dynamic (AC) resistance but a
moderate static (DC) resistance. Place this device in the
collector of a transistor and you have a high gain circuit
with relatively low collector load drop. Op-amps do
something like this all the time in the differential input
stage.

You may achieve the same result in a small way by taking
a junction FET, connecting the gate to the source, and
using that as the load.


Cheers - Joe