250mA voltage clamp

[Walter Harley]

I need to limit the output of an ordinary unregulated power supply. The
voltage ranges from 18V to 35V, taking load and line variation and ripple
into account, and the load draws up to 250mA. I'd like to limit the output
to 24V (give or take a volt); I don't care if it drops below that, though,
and in fact it will often be right around that. I don't want to introduce
bursts of noise or ringing when the voltage is right around the limit. I
would like to have the input-output voltage differential be no more than 2V,
and would prefer less than 1V. This is a one-off project, so using junkbox
parts is desirable, though not a requirement.

The best idea I've come up with is a power PNP or MOSFET pass transistor,
driven by an opamp that compares the output voltage to a zener reference.
Some potential problems are latchup in the opamp during power-on, and
stability of the opamp (since the feedback loop here includes things like
changes in the upstream voltage). Also, simply connecting the opamp between
the input rail and ground means it sees up to 35V supply, worryingly close
to the operating limit for a jellybean opamp.

Is there an even simpler solution that I'm not seeing?

Should I be worried about the stability of this solution, and if so, what
would be an appropriate way to control it?

Am I right to distrust an LM317 at and below dropout?

Thanks!
-walter

 

[[email protected]]

I need to limit the output of an ordinary unregulated power supply. The
voltage ranges from 18V to 35V, taking load and line variation and ripple
into account, and the load draws up to 250mA.

What's wrong with a 24 volt zener? Pick one that will handle
the dissipation when the supply is current limited by its internal
impedance. What is the current from the supply when it has a load
heavy enough to drop its output voltage to 24 volts?

Jim

 

[Fred Bloggs]

What's wrong with a 24 volt zener? Pick one that will handle
the dissipation when the supply is current limited by its internal
impedance. What is the current from the supply when it has a load
heavy enough to drop its output voltage to 24 volts?

The interesting question is how he plans to limit the I/O differential
to 2V with 35V in and 24V out- looks like 12V to me, and at 1/4 amp
makes for 3W. Sounds like someone is confused, again.

 

[Walter Harley]

The interesting question is how he plans to limit the I/O differential to
2V with 35V in and 24V out- looks like 12V to me, and at 1/4 amp makes for
3W. Sounds like someone is confused, again.

It's a voltage clamp. When the input voltage is above 24V, the differential
will be Vi - 24V. When the input voltage is below (24V + Vdropout), the
differential will be Vdropout. I want Vdropout to be less than 2V,
preferably less than 1V. (Though there is no requirement that Vdropout be
precise or constant.) I'm sorry if that confused you.

I could simply use a LM317 as a preregulator. According to the datasheet
its dropout voltage (at room temperature or higher, at less than 500mA) is
around 1.7V. But I don't know whether to trust it to remain stable around
and below dropout, which is why I asked about that.

And in answer to Jim's question: it takes about 300mA draw for the internal
impedance of the supply to drop the voltage sufficiently. I'd rather avoid
a shunt regulator if I can, so as to avoid unnecessary heating. The
max-load condition is unusual and is expected (though not guaranteed) to
last only a short time; 99% of the time the load is expected to be only 30mA
or so.

 

[Ken Smith]

The best idea I've come up with is a power PNP or MOSFET pass transistor,
driven by an opamp that compares the output voltage to a zener reference.

If the supply is isolated, you can put the pass element in the (-) leg.
This lets you use a low cost MOSFET.

The supply voltage to the op-amp can be regulated by zener. If you really
want, this zener could also serve as the reference.

You didn't rule out putting a large capacitor across the output terminals.
If you do this, you can make the common source MOSFET stage have its high
frequency gain determined mostly by the gm and the output capacitor.
BTW: A source resistor can make the MOSFET's gm easier to predict.

By limiting the supply voltage on the op-amp, you reduce how far it has to
swing as the supply rises past 24V. The op-amp should be a fairly fast
one. The output capacitor should be the important pole in the system.

 

[[email protected]]

And in answer to Jim's question: it takes about 300mA draw for the internal
impedance of the supply to drop the voltage sufficiently. I'd rather avoid
a shunt regulator if I can, so as to avoid unnecessary heating. The
max-load condition is unusual and is expected (though not guaranteed) to
last only a short time; 99% of the time the load is expected to be only 30mA
or so.

OK, a DC-DC converter, buck/boost type between your source and
load would solve all your problems at 80% efficiency. And be much
easier to implement than some op-amp and pass element scheme.

Jim

 

[Fred Bloggs]

It's a voltage clamp. When the input voltage is above 24V, the differential
will be Vi - 24V. When the input voltage is below (24V + Vdropout), the
differential will be Vdropout. I want Vdropout to be less than 2V,
preferably less than 1V. (Though there is no requirement that Vdropout be
precise or constant.) I'm sorry if that confused you.

I could simply use a LM317 as a preregulator. According to the datasheet
its dropout voltage (at room temperature or higher, at less than 500mA) is
around 1.7V. But I don't know whether to trust it to remain stable around
and below dropout, which is why I asked about that.

A voltage clamp is just a simple emitter follower with a zener from base
to GND. The TIP41 with a 24V zener will work well. The trick is to keep
the power transistor base drive low impedance and somewhat regulate the
zener current against your 18V to 35V input swing. You can use something
minimal like this:
View in a fixed-width font such as
Courier.


..
..
.. V+
.. |
.. .-------------+
.. | |
.. | |
.. [270] |
.. | |
.. | |
.. +------. |
.. | | |
.. | [47] |
.. >| | |/
.. 2N2907A |----+----| TIP41
.. /| | |>
.. | | |
.. | | |
.. | | + '-----> OUT
.. [1K/1W] Z
.. | A 24V [1N4749]
.. | | -
.. | |
.. '------+
.. |
.. |
.. ---
.. ///
..

 

[budgie]

OK, a DC-DC converter, buck/boost type between your source and
load would solve all your problems at 80% efficiency. And be much
easier to implement than some op-amp and pass element scheme.

Surely replacing the highly unregulated power supply would be easier still? I
fail to see the point in spending time building a fix for a simple sub-assembly
that has a problem, when another PSU would solve it.

 

[Walter Harley]

[...]
A voltage clamp is just a simple emitter follower with a zener from base
to GND. The TIP41 with a 24V zener will work well. The trick is to keep
the power transistor base drive low impedance and somewhat regulate the
zener current against your 18V to 35V input swing. You can use something
minimal like this:
View in a fixed-width font such as
Courier.


.
.
. V+
. |
. .-------------+
. | |
. | |
. [270] |
. | |
. | |
. +------. |
. | | |
. | [47] |
. >| | |/
. 2N2907A |----+----| TIP41
. /| | |>
. | | |
. | | |
. | | + '-----> OUT
. [1K/1W] Z
. | A 24V [1N4749]
. | | -
. | |
. '------+
. |
. |
. ---
. ///
.

Thanks!

That makes sense; regulating the current to the Zener makes it possible to
handle the big voltage swing without dissipating too much heat there.

But, this circuit achieves that regulation by shunting the excess current to
ground; it's not a lot of current but I got to wondering whether I could do
it with a series element instead. First idea was to use a JFET current
limiter, but those drop a lot of voltage, which would have meant too-high
dropout voltage.

But doing it with a BJT current source seems to work, although it does have
one more part:

TIP31
-----o---o---- --------
| | \ ^
V .-. ---
- | | |
| | |72 |
V '-' |
- | |
| |< |
o-| |
| |\.-----'
.-. | 2N2907
| | |
22k| | -
'-' ^ 24V Zener
| |
-----o---o---------------

The current source needs to provide the max base current of the pass
transistor, plus enough to keep the Zener in regulation. When the output
load is low, all the current passes through the Zener; but that's still only
(Iout(max)/beta(min) * 24V) or about 1/4W assuming min beta of 25.

I haven't built it yet, just thought about it.

I still wonder whether an LM317 is stable around and below dropout?

 

[Winfield Hill]

Walter Harley wrote...

I still wonder whether an LM317 is stable around and below dropout?

Yes, it just saturates to its dropout voltage (which varies with
current, and junction temp, as NSC's datasheet shows in plots.

 

[Fred Bloggs]

[...]
A voltage clamp is just a simple emitter follower with a zener from base
to GND. The TIP41 with a 24V zener will work well. The trick is to keep
the power transistor base drive low impedance and somewhat regulate the
zener current against your 18V to 35V input swing. You can use something
minimal like this:
View in a fixed-width font such as
Courier.


.
.
. V+
. |
. .-------------+
. | |
. | |
. [270] |
. | |
. | |
. +------. |
. | | |
. | [47] |
. >| | |/
. 2N2907A |----+----| TIP41
. /| | |>
. | | |
. | | |
. | | + '-----> OUT
. [1K/1W] Z
. | A 24V [1N4749]
. | | -
. | |
. '------+
. |
. |
. ---
. ///
.

Thanks!

That makes sense; regulating the current to the Zener makes it possible to
handle the big voltage swing without dissipating too much heat there.

But, this circuit achieves that regulation by shunting the excess current to
ground; it's not a lot of current but I got to wondering whether I could do
it with a series element instead. First idea was to use a JFET current
limiter, but those drop a lot of voltage, which would have meant too-high
dropout voltage.

It is not necessary to shunt the current to ground- it can be shunted to
the output. Since you say the load is 30mA most of the time , this works
well. If your loading goes to some minuscule amount, the PNP will just
saturate and the output may go to 24.75V which is not the end of the
world, and the zener receives max current of (35-24.75)/330=30mA for a
750mW dissipation at no load. At nominal loading of 30mA and maximum
Vin, the zener is biased at 0.75/68=11mA (minus NPN base current) and
the PNP collector current is 30mA-11mA=19mA direct to the output leaving
11mA for the NPN.

View in a fixed-width font such as
Courier.

..
.. V+
.. |
.. .-------------+
.. | |
.. | |
.. [330/1W] |
.. | |
.. | |
.. +------. |
.. | | |
.. | [68] |
.. >| | |/
.. 2N2907A |----+----| TIP31A
.. /| | |>
.. | | |
.. | | |
.. '------|------+-----> OUT
.. |
.. |
.. |
.. Z +
.. A 24V [1N4749]
.. | -
.. |
.. |
.. |
.. |
.. ---
.. ///
..
..

But doing it with a BJT current source seems to work, although it does have
one more part:

TIP31
-----o---o---- --------
| | \ ^
V .-. ---
- | | |
| | |72 |
V '-' |
- | |
| |< |
o-| |
| |\.-----'
.-. | 2N2907
| | |
22k| | -
'-' ^ 24V Zener
| |
-----o---o---------------

The current source needs to provide the max base current of the pass
transistor, plus enough to keep the Zener in regulation. When the output
load is low, all the current passes through the Zener; but that's still only
(Iout(max)/beta(min) * 24V) or about 1/4W assuming min beta of 25.

I haven't built it yet, just thought about it.

That circuit is more of a standard, and has lower output impedance at
dropout voltage by about a factor 5. That sounds good initially but in
the case of moderate capacitor filter/load on the output, combined with
a sizable unregulated ripple, the saturated input results in much more
ripple current loading of the NPN. It is better in a DC application but
may be problematic in an AC application.

I still wonder whether an LM317 is stable around and below dropout?

You may be thinking of a conditional stability type of thing - and this
is not an issue with the LM317 design.

 

[Ken Smith]

How about this to improve it. Change the 22K to a 1N5305...1N5313

 

[Walter Harley]

How about this to improve it. Change the 22K to a 1N5305...1N5313

I'm not sure I see why that's an improvement? The current through the 22K
will vary between about 0.7mA and 1.5mA depending on the input voltage, but
all that current is doing is biasing the two diodes in order to program the
current source, and the current source itself is just biasing the Zener, so
the effect on the output voltage should be miniscule (and the clamp voltage
itself is noncritical, it just needs to hold the output low enough to
protect some downstream components). Plus, these current-limiting diodes
expensive - $1.50 or more even in quantity, compared to a penny or two for
the resistor. Is there something I'm missing about this?

 

[Ken Smith]

I'm not sure I see why that's an improvement? The current through the 22K
will vary between about 0.7mA and 1.5mA depending on the input voltage, but
all that current is doing is biasing the two diodes in order to program the
current source, and the current source itself is just biasing the Zener, so
the effect on the output voltage should be miniscule (and the clamp voltage
itself is noncritical, it just needs to hold the output low enough to
protect some downstream components). Plus, these current-limiting diodes
expensive - $1.50 or more even in quantity, compared to a penny or two for
the resistor. Is there something I'm missing about this?

Perhaps, perhaps not. It depends on how tight you need the line
regulation to be. If the diodes are 1N914s, the voltage drop on each will
vary by, lets say, 40mV each. This means that the 70 Ohm (IIRC) will have
at least 40mV of variation on it.

 

[Walter Harley]

Perhaps, perhaps not. It depends on how tight you need the line
regulation to be. If the diodes are 1N914s, the voltage drop on each will
vary by, lets say, 40mV each. This means that the 70 Ohm (IIRC) will have
at least 40mV of variation on it.

It doesn't need to be at all tight - this is just a preregulator. All it
has to do is make sure that the output voltage doesn't rise above the point
where it would damage downstream circuitry, while not dropping excessive
voltage (or wasting too much current) otherwise. So, simpler is better, in
this case.

But regardless, a 40mV change in voltage on that 70 ohm resistor is less
than 5% change in the current source output; and the only thing that
directly affects is the amount of current through the zener. The current
through the zener will be changed by much more than 5% as the load, and thus
the base current of the pass transistor, changes - in fact I expect at least
10:1 change in zener current. So, precisely regulating the current source
doesn't really buy anything in terms of the regulation of the clamp voltage,
I think.

Cheers,
-walter

 

[Walter Harley]

Walter Harley wrote...

Yes, it just saturates to its dropout voltage (which varies with
current, and junction temp, as NSC's datasheet shows in plots.

To confirm, I hooked up an LM317 to provide 21V, with a load switched
between 75R and 1k by way of a MOSFET controlled by a pulse generator.
Then, with a scope I watched the output as the load switched on and off,
while varying the input voltage from 18V to 35V.

As Win suggests, the LM317 stays quiet and stable when its input voltage is
too low to regulate. There was no hint of ringing or noise around the
transitions. When the input voltage is too low, the LM317 looks like a 2.5V
drop, in series with an ohm or so.

I was a little surprised that the drop was 2.5V, even with only 1k load (in
parallel with the ~4k divider network). From the datasheet I expected it to
be lower. I was using a JRC part rather than National, but the JRC
datasheet still suggests dropout < 2V.

 

[Winfield Hill]

Walter Harley wrote...

Winfield Hill wrote ...

To confirm, I hooked up an LM317 to provide 21V, with a load switched
between 75R and 1k by way of a MOSFET controlled by a pulse generator.
Then, with a scope I watched the output as the load switched on and off,
while varying the input voltage from 18V to 35V.

As Win suggests, the LM317 stays quiet and stable when its input voltage
is too low to regulate. There was no hint of ringing or noise around the
transitions. When the input voltage is too low, the LM317 looks like a
2.5V drop, in series with an ohm or so.

I was a little surprised that the drop was 2.5V, even with only 1k load
(in parallel with the ~4k divider network). From the datasheet I expected
it to be lower. I was using a JRC part rather than National, but the JRC
datasheet still suggests dropout < 2V.

That's interesting.

 

[Fred Bloggs]

I was a little surprised that the drop was 2.5V, even with only 1k
load (in parallel with the ~4k divider network). From the datasheet
I expected it to be lower. I was using a JRC part rather than
National, but the JRC datasheet still suggests dropout < 2V.

With that much dropout it makes no sense to let it drop out. You can
keep the LM317 in regulation by regulating to filtered Vin - 3V up to a
clamped upper limit as required. We are talking about a simple level
shift and diode clamp- and not a lot of excess current. The extra 0.5V
drop is more than paid for by the high performance of the regulator, and
you have all the protection the LM317 offers. Don't forget to add all
the usual diode protection around the IC.
View in a fixed-width font such as Courier.

 

[Walter Harley]

With that much dropout it makes no sense to let it drop out. You can keep
the LM317 in regulation by regulating to filtered Vin - 3V up to a clamped
upper limit as required. We are talking about a simple level shift and
diode clamp- and not a lot of excess current. The extra 0.5V drop is more
than paid for by the high performance of the regulator, and you have all
the protection the LM317 offers. Don't forget to add all the usual diode
protection around the IC.
View in a fixed-width font such as Courier.

.
. +--------+
. Vin>--+-----+--------------|IN OUT|-----+----+-->
. | | | | | |
. | | | ADJ | | |
. | | +--------+ [220] |
. [4.7K] [10K] | | |
. | | | | |+
. | | .--|>|--+----------+ ===
. | | | | | |
. | | | | 1N4748 | |
. | | | |< ---/ |
. +----------------+-----| Q2 // \ |
. |+ | | |\ --- |
. === | |/ | | |
. 100U +--------| Q1 | | |
. | | |> | | |
. | | | | | |
. | | [620] | | |
. | | | | | |
. | '-|>|-|>|--+ | | |
. | | | | |
. +----------------+-------+----------+----+-->
. |
. ---

That's a nice circuit. Am I right that the function of the diode is to
protect Q2's base-emitter junction from reverse breakdown? Or does it do
something else that I'm missing?

I like the idea of having the LM317's protection. But I'm still leaning
toward the discrete approaches, because I don't like having this much
input-output differential in the non-clamped case; it means that the piece
of equipment this is powering won't handle brownouts as well. Probably a
non-issue, but I hate for my stuff to be the weak link. I'll probably go
with one of the discrete circuits instead.

Thanks!
-walter

 

[Fred Bloggs]

With that much dropout it makes no sense to let it drop out. You can keep
the LM317 in regulation by regulating to filtered Vin - 3V up to a clamped
upper limit as required. We are talking about a simple level shift and
diode clamp- and not a lot of excess current. The extra 0.5V drop is more
than paid for by the high performance of the regulator, and you have all
the protection the LM317 offers. Don't forget to add all the usual diode
protection around the IC.
View in a fixed-width font such as Courier.

.
. +--------+
. Vin>--+-----+--------------|IN OUT|-----+----+-->
. | | | | | |
. | | | ADJ | | |
. | | +--------+ [220] |
. [4.7K] [10K] | | |
. | | | | |+
. | | .--|>|--+----------+ ===
. | | | | | |
. | | | | 1N4748 | |
. | | | |< ---/ |
. +----------------+-----| Q2 // \ |
. |+ | | |\ --- |
. === | |/ | | |
. 100U +--------| Q1 | | |
. | | |> | | |
. | | | | | |
. | | [620] | | |
. | | | | | |
. | '-|>|-|>|--+ | | |
. | | | | |
. +----------------+-------+----------+----+-->
. |
. ---

That's a nice circuit. Am I right that the function of the diode is to
protect Q2's base-emitter junction from reverse breakdown?

Yes- the NPN current source keeps the base of the PNP at ~Vin-5 while
the zener clamps at 22V, so the PNP BE junction starts to reverse bias
at Vin=27V.

Or does it do
something else that I'm missing?
Nope.


I like the idea of having the LM317's protection. But I'm still leaning
toward the discrete approaches, because I don't like having this much
input-output differential in the non-clamped case; it means that the piece
of equipment this is powering won't handle brownouts as well. Probably a
non-issue, but I hate for my stuff to be the weak link. I'll probably go
with one of the discrete circuits instead.

Thanks!
-walter

I suppose you could adapt the same approach to an LDO, pre-regulating
say 0.2V-0.5V down up until clamp level,- there are a few that handle
40V IIRC, or use a more common and cheaper one and install a zener
over-voltage bypass. If you clamp at 24V and Vin,max=35 then the LDO
only needs to hold off 11V worst case. The trick is to find one in
something easy to heatsink like a TO-220 , and the 250mA current
requirement is not an issue.