# Bipolar current limit

Started by February 20, 2017
```OK this is perhaps overkill.. see footnote.
So I've got an indicator LED (bi-color) the input is the error
term from a thermal loop, and I turn that into an LED current.

I was thinking it might be nice to turn up the gain on the current,
but then I worried about burning out the LED with too much current from the
opamp.*  So I drew this up... and then added the diodes to stop the
transistors "running backwards".  And the cap to stop the oscillations
when the transistors turn on.

https://www.dropbox.com/s/t14fdf6hng399ou/Current-limit.JPG?dl=0

Don't laugh too loud.
Is there some easier way to do this?

George H.

*Right one easy answer would be to find an opamp with ~20 mA maximum current.
(uA741 ???   :^)
Do you have any suggestions?  All our "in stock opamps are higher currents.
```
```On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:

> OK this is perhaps overkill.. see footnote.
> So I've got an indicator LED (bi-color) the input is the error term from
> a thermal loop, and I turn that into an LED current.
>
> I was thinking it might be nice to turn up the gain on the current,
> but then I worried about burning out the LED with too much current from
> the opamp.*  So I drew this up... and then added the diodes to stop the
> transistors "running backwards".  And the cap to stop the oscillations
> when the transistors turn on.

Assuming that this is for a visual indicator, just use a rail-rail output
op-amp with a series resistor.  In normal operation the op-amp is within
its voltage limits and current is as commanded.  When the input is big,
the op-amp "asks" for too much current, hits the rail, the resistor
limits the current, and Bob's yer Uncle.

There's a lot of reasons not to want to let an op-amp hit the rail, but
as long as you have one that does so in a reasonably well-behaved manner
you should be OK.

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work!  See my website if you're interested
http://www.wescottdesign.com
```
```On Monday, February 20, 2017 at 3:11:11 PM UTC-5, Tim Wescott wrote:
> On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:
>
> > OK this is perhaps overkill.. see footnote.
> > So I've got an indicator LED (bi-color) the input is the error term from
> > a thermal loop, and I turn that into an LED current.
> >
> > I was thinking it might be nice to turn up the gain on the current,
> > but then I worried about burning out the LED with too much current from
> > the opamp.*  So I drew this up... and then added the diodes to stop the
> > transistors "running backwards".  And the cap to stop the oscillations
> > when the transistors turn on.
>
> Assuming that this is for a visual indicator, just use a rail-rail output
> op-amp with a series resistor.  In normal operation the op-amp is within
> its voltage limits and current is as commanded.  When the input is big,
> the op-amp "asks" for too much current, hits the rail, the resistor
> limits the current, and Bob's yer Uncle.
I've done that, but the one issue is the turn on voltage of the LED.
I've got a few volt dead space near zero volts.
(I want to say again that perhaps this is a silly idea, and I'll bag
the whole thing.)
So by feeding back of the current I eliminate that dead space.
Then I dreamed of more gain.. so when the loop is closed but a little
unstable maybe you could see the oscillations in the error term in the
intensity of the LED... or if near zero by having it change color.
But then some student leaves it in high gain with maximum error and the
current from the opamp heats up the led.  (potential failure...)

So I wanted to limit the current to the LED.
(I never can seem to put enough background into my questions...)

>
> There's a lot of reasons not to want to let an op-amp hit the rail, but
> as long as you have one that does so in a reasonably well-behaved manner
> you should be OK.
Hmm well I run opamps into the rails all the time.
For slow stuff, I've never had problems.

In this case I'm talking about running opamps at their current limit.
I've never done that (on purpose) before.

George H.

I
>
> --
> Tim Wescott
> Control systems, embedded software and circuit design
> I'm looking for work!  See my website if you're interested
> http://www.wescottdesign.com

```
```On Mon, 20 Feb 2017 13:35:26 -0800, George Herold wrote:

> On Monday, February 20, 2017 at 3:11:11 PM UTC-5, Tim Wescott wrote:
>> On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:
>>
>> > OK this is perhaps overkill.. see footnote.
>> > So I've got an indicator LED (bi-color) the input is the error term
>> > from a thermal loop, and I turn that into an LED current.
>> >
>> > I was thinking it might be nice to turn up the gain on the current,
>> > but then I worried about burning out the LED with too much current
>> > from the opamp.*  So I drew this up... and then added the diodes to
>> > stop the transistors "running backwards".  And the cap to stop the
>> > oscillations when the transistors turn on.
>>
>> Assuming that this is for a visual indicator, just use a rail-rail
>> output op-amp with a series resistor.  In normal operation the op-amp
>> is within its voltage limits and current is as commanded.  When the
>> input is big, the op-amp "asks" for too much current, hits the rail,
>> the resistor limits the current, and Bob's yer Uncle.
> I've done that, but the one issue is the turn on voltage of the LED.
> I've got a few volt dead space near zero volts.
> (I want to say again that perhaps this is a silly idea, and I'll bag the
> whole thing.)
> So by feeding back of the current I eliminate that dead space.
> Then I dreamed of more gain.. so when the loop is closed but a little
> unstable maybe you could see the oscillations in the error term in the
> intensity of the LED... or if near zero by having it change color.
> But then some student leaves it in high gain with maximum error and the
> current from the opamp heats up the led.  (potential failure...)
>
> So I wanted to limit the current to the LED.
> (I never can seem to put enough background into my questions...)

I'm not saying don't do the current feedback -- I'm just saying that if
you do that AND put a resistor in series with the LED, then when the op-
amp can it'll regulate current, and when it can't, the resistor will
cleanly limit current.

You could even make the current-sense resistor into your current-limit
resistor, and get plenty of signal thereby.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!
```
```On Mon, 20 Feb 2017 11:14:20 -0800 (PST), George Herold
<gherold@teachspin.com> wrote:

>OK this is perhaps overkill.. see footnote.
>So I've got an indicator LED (bi-color) the input is the error
>term from a thermal loop, and I turn that into an LED current.
>
>I was thinking it might be nice to turn up the gain on the current,
>but then I worried about burning out the LED with too much current from the
>opamp.*  So I drew this up... and then added the diodes to stop the
>transistors "running backwards".  And the cap to stop the oscillations
>when the transistors turn on.
>
>https://www.dropbox.com/s/t14fdf6hng399ou/Current-limit.JPG?dl=0
>
>Don't laugh too loud.
>Is there some easier way to do this?
>
>George H.
>
>*Right one easy answer would be to find an opamp with ~20 mA maximum current.
>(uA741 ???   :^)
>Do you have any suggestions?  All our "in stock opamps are higher currents.

Looks like an OpAmp "snuffer" to me... there is no bound on the
current you are pulling current directly from the output.

...Jim Thompson
--
| James E.Thompson                                 |    mens     |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

Thinking outside the box... producing elegant solutions.
```
```On Monday, February 20, 2017 at 5:07:37 PM UTC-5, Tim Wescott wrote:
> On Mon, 20 Feb 2017 13:35:26 -0800, George Herold wrote:
>
> > On Monday, February 20, 2017 at 3:11:11 PM UTC-5, Tim Wescott wrote:
> >> On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:
> >>
> >> > OK this is perhaps overkill.. see footnote.
> >> > So I've got an indicator LED (bi-color) the input is the error term
> >> > from a thermal loop, and I turn that into an LED current.
> >> >
> >> > I was thinking it might be nice to turn up the gain on the current,
> >> > but then I worried about burning out the LED with too much current
> >> > from the opamp.*  So I drew this up... and then added the diodes to
> >> > stop the transistors "running backwards".  And the cap to stop the
> >> > oscillations when the transistors turn on.
> >>
> >> Assuming that this is for a visual indicator, just use a rail-rail
> >> output op-amp with a series resistor.  In normal operation the op-amp
> >> is within its voltage limits and current is as commanded.  When the
> >> input is big, the op-amp "asks" for too much current, hits the rail,
> >> the resistor limits the current, and Bob's yer Uncle.
> > I've done that, but the one issue is the turn on voltage of the LED.
> > I've got a few volt dead space near zero volts.
> > (I want to say again that perhaps this is a silly idea, and I'll bag the
> > whole thing.)
> > So by feeding back of the current I eliminate that dead space.
> > Then I dreamed of more gain.. so when the loop is closed but a little
> > unstable maybe you could see the oscillations in the error term in the
> > intensity of the LED... or if near zero by having it change color.
> > But then some student leaves it in high gain with maximum error and the
> > current from the opamp heats up the led.  (potential failure...)
> >
> > So I wanted to limit the current to the LED.
> > (I never can seem to put enough background into my questions...)
>
> I'm not saying don't do the current feedback -- I'm just saying that if
> you do that AND put a resistor in series with the LED, then when the op-
> amp can it'll regulate current, and when it can't, the resistor will
> cleanly limit current.
>
> You could even make the current-sense resistor into your current-limit
> resistor, and get plenty of signal thereby.
>
Hmm yeah that's my 1 k ohm series resistor.. well 1k to 0 k.
when the opamp is railed, 15V -> 15ma... then when changes
get smaller so ~1V or smaller signals 1V ~15 mA.

George H.
> --
>
> Tim Wescott
> Wescott Design Services
> http://www.wescottdesign.com
>
> I'm looking for work -- see my website!
```
```On Mon, 20 Feb 2017 13:35:26 -0800, George Herold wrote:

> On Monday, February 20, 2017 at 3:11:11 PM UTC-5, Tim Wescott wrote:
>> On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:
>>
>> > OK this is perhaps overkill.. see footnote.
>> > So I've got an indicator LED (bi-color) the input is the error term
>> > from a thermal loop, and I turn that into an LED current.
>> >
>> > I was thinking it might be nice to turn up the gain on the current,
>> > but then I worried about burning out the LED with too much current
>> > from the opamp.*  So I drew this up... and then added the diodes to
>> > stop the transistors "running backwards".  And the cap to stop the
>> > oscillations when the transistors turn on.
>>
>> Assuming that this is for a visual indicator, just use a rail-rail
>> output op-amp with a series resistor.  In normal operation the op-amp
>> is within its voltage limits and current is as commanded.  When the
>> input is big, the op-amp "asks" for too much current, hits the rail,
>> the resistor limits the current, and Bob's yer Uncle.
> I've done that, but the one issue is the turn on voltage of the LED.
> I've got a few volt dead space near zero volts.
> (I want to say again that perhaps this is a silly idea, and I'll bag the
> whole thing.)
> So by feeding back of the current I eliminate that dead space.
> Then I dreamed of more gain.. so when the loop is closed but a little
> unstable maybe you could see the oscillations in the error term in the
> intensity of the LED... or if near zero by having it change color.
> But then some student leaves it in high gain with maximum error and the
> current from the opamp heats up the led.  (potential failure...)
>
> So I wanted to limit the current to the LED.
> (I never can seem to put enough background into my questions...)
>
>
>> There's a lot of reasons not to want to let an op-amp hit the rail, but
>> as long as you have one that does so in a reasonably well-behaved
>> manner you should be OK.
> Hmm well I run opamps into the rails all the time.
> For slow stuff, I've never had problems.
>
> In this case I'm talking about running opamps at their current limit.
> I've never done that (on purpose) before.

This:

Let R = maxcurrent / (Vcc - Vfwd)

It assumes that Vcc = -Vee, and that the two diodes' Vfwd is reasonably
close, but it'll get the job done.  Can be easily rearranged for a single
supply, if you have enough overhead voltage.  If you need more gain put a
resistor in the leg between R and the V- input, with another resistor to
ground.

Vcc
o
|\|
.------|-\
|      |  >----o----.
o--------)------|+/     |    |
|      |/|     |    |
|        o     -    V ->
|       Vee    ^    -
|              |    |
|              o----'
|              |
'--------------o
|
.-.
| | R
| |
'-'
|
|
===
GND
(created by AACircuit v1.28.7 beta 02/28/13 www.tech-chat.de)

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!
```
```On Monday, February 20, 2017 at 5:42:13 PM UTC-5, Jim Thompson wrote:
> On Mon, 20 Feb 2017 11:14:20 -0800 (PST), George Herold
> <gherold@teachspin.com> wrote:
>
> >OK this is perhaps overkill.. see footnote.
> >So I've got an indicator LED (bi-color) the input is the error
> >term from a thermal loop, and I turn that into an LED current.
> >
> >I was thinking it might be nice to turn up the gain on the current,
> >but then I worried about burning out the LED with too much current from the
> >opamp.*  So I drew this up... and then added the diodes to stop the
> >transistors "running backwards".  And the cap to stop the oscillations
> >when the transistors turn on.
> >
> >https://www.dropbox.com/s/t14fdf6hng399ou/Current-limit.JPG?dl=0
> >
> >Don't laugh too loud.
> >Is there some easier way to do this?
> >
> >George H.
> >
> >*Right one easy answer would be to find an opamp with ~20 mA maximum current.
> >(uA741 ???   :^)
> >Do you have any suggestions?  All our "in stock opamps are higher currents.
>
> Looks like an OpAmp "snuffer" to me... there is no bound on the
> current you are pulling current directly from the output.

Right.  I'm not sure where the voltage drop is.. (opamp or transistor)
but the current is limited by whatever the opamp delivers.

All my opamps (in stock) are ~30-40 mA max output current.
They are all rated to have the outputs shorted continuously.
(It's not something, I've looked at carefully.)
I thought that current might kill the leds.. even the
spec of 25 mA max probably limits the lifetime.
So I'm snuffing the opamps above ~14 mA (0.7 V is about
where it turns on hard.)

George H.
>
>                                         ...Jim Thompson
> --
> | James E.Thompson                                 |    mens     |
> | Analog Innovations                               |     et      |
> | Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
> | STV, Queen Creek, AZ 85142    Skype: skypeanalog |             |
> | Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
> | E-mail Icon at http://www.analog-innovations.com |    1962     |
>
>      Thinking outside the box... producing elegant solutions.

```
```On Monday, February 20, 2017 at 7:15:03 PM UTC-5, Tim Wescott wrote:
> On Mon, 20 Feb 2017 13:35:26 -0800, George Herold wrote:
>
> > On Monday, February 20, 2017 at 3:11:11 PM UTC-5, Tim Wescott wrote:
> >> On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:
> >>
> >> > OK this is perhaps overkill.. see footnote.
> >> > So I've got an indicator LED (bi-color) the input is the error term
> >> > from a thermal loop, and I turn that into an LED current.
> >> >
> >> > I was thinking it might be nice to turn up the gain on the current,
> >> > but then I worried about burning out the LED with too much current
> >> > from the opamp.*  So I drew this up... and then added the diodes to
> >> > stop the transistors "running backwards".  And the cap to stop the
> >> > oscillations when the transistors turn on.
> >>
> >> Assuming that this is for a visual indicator, just use a rail-rail
> >> output op-amp with a series resistor.  In normal operation the op-amp
> >> is within its voltage limits and current is as commanded.  When the
> >> input is big, the op-amp "asks" for too much current, hits the rail,
> >> the resistor limits the current, and Bob's yer Uncle.
> > I've done that, but the one issue is the turn on voltage of the LED.
> > I've got a few volt dead space near zero volts.
> > (I want to say again that perhaps this is a silly idea, and I'll bag the
> > whole thing.)
> > So by feeding back of the current I eliminate that dead space.
> > Then I dreamed of more gain.. so when the loop is closed but a little
> > unstable maybe you could see the oscillations in the error term in the
> > intensity of the LED... or if near zero by having it change color.
> > But then some student leaves it in high gain with maximum error and the
> > current from the opamp heats up the led.  (potential failure...)
> >
> > So I wanted to limit the current to the LED.
> > (I never can seem to put enough background into my questions...)
> >
> >
> >> There's a lot of reasons not to want to let an op-amp hit the rail, but
> >> as long as you have one that does so in a reasonably well-behaved
> >> manner you should be OK.
> > Hmm well I run opamps into the rails all the time.
> > For slow stuff, I've never had problems.
> >
> > In this case I'm talking about running opamps at their current limit.
> > I've never done that (on purpose) before.
>
> This:
>
> Let R = maxcurrent / (Vcc - Vfwd)
>
> It assumes that Vcc = -Vee, and that the two diodes' Vfwd is reasonably
> close, but it'll get the job done.  Can be easily rearranged for a single
> supply, if you have enough overhead voltage.  If you need more gain put a
> resistor in the leg between R and the V- input, with another resistor to
> ground.
>
>
>                     Vcc
>                      o
>                    |\|
>             .------|-\
>             |      |  >----o----.
>    o--------)------|+/     |    |
>             |      |/|     |    |
>             |        o     -    V ->
>             |       Vee    ^    -
>             |              |    |
>             |              o----'
>             |              |
>             '--------------o
>                            |
>                           .-.
>                           | | R
>                           | |
>                           '-'
>                            |
>                            |
>                           ===
>                           GND
> (created by AACircuit v1.28.7 beta 02/28/13 www.tech-chat.de)
>
> --
>
> Tim Wescott
> Wescott Design Services
> http://www.wescottdesign.com
>
> I'm looking for work -- see my website!

Right.. did my dropbox link not work?

George H.
```
```On Mon, 20 Feb 2017 16:29:53 -0800, George Herold wrote:

> On Monday, February 20, 2017 at 7:15:03 PM UTC-5, Tim Wescott wrote:
>> On Mon, 20 Feb 2017 13:35:26 -0800, George Herold wrote:
>>
>> > On Monday, February 20, 2017 at 3:11:11 PM UTC-5, Tim Wescott wrote:
>> >> On Mon, 20 Feb 2017 11:14:20 -0800, George Herold wrote:
>> >>
>> >> > OK this is perhaps overkill.. see footnote.
>> >> > So I've got an indicator LED (bi-color) the input is the error
>> >> > term from a thermal loop, and I turn that into an LED current.
>> >> >
>> >> > I was thinking it might be nice to turn up the gain on the
>> >> > current,
>> >> > but then I worried about burning out the LED with too much current
>> >> > from the opamp.*  So I drew this up... and then added the diodes
>> >> > to stop the transistors "running backwards".  And the cap to stop
>> >> > the oscillations when the transistors turn on.
>> >>
>> >> Assuming that this is for a visual indicator, just use a rail-rail
>> >> output op-amp with a series resistor.  In normal operation the
>> >> op-amp is within its voltage limits and current is as commanded.
>> >> When the input is big, the op-amp "asks" for too much current, hits
>> >> the rail, the resistor limits the current, and Bob's yer Uncle.
>> > I've done that, but the one issue is the turn on voltage of the LED.
>> > I've got a few volt dead space near zero volts.
>> > (I want to say again that perhaps this is a silly idea, and I'll bag
>> > the whole thing.)
>> > So by feeding back of the current I eliminate that dead space.
>> > Then I dreamed of more gain.. so when the loop is closed but a little
>> > unstable maybe you could see the oscillations in the error term in
>> > the intensity of the LED... or if near zero by having it change
>> > color.
>> > But then some student leaves it in high gain with maximum error and
>> > the current from the opamp heats up the led.  (potential failure...)
>> >
>> > So I wanted to limit the current to the LED.
>> > (I never can seem to put enough background into my questions...)
>> >
>> >
>> >> There's a lot of reasons not to want to let an op-amp hit the rail,
>> >> but as long as you have one that does so in a reasonably
>> >> well-behaved manner you should be OK.
>> > Hmm well I run opamps into the rails all the time.
>> > For slow stuff, I've never had problems.
>> >
>> > In this case I'm talking about running opamps at their current limit.
>> > I've never done that (on purpose) before.
>>
>> This:
>>
>> Let R = maxcurrent / (Vcc - Vfwd)
>>
>> It assumes that Vcc = -Vee, and that the two diodes' Vfwd is reasonably
>> close, but it'll get the job done.  Can be easily rearranged for a
>> single supply, if you have enough overhead voltage.  If you need more
>> gain put a resistor in the leg between R and the V- input, with another
>> resistor to ground.
>>
>>
>>                     Vcc
>>                      o
>>                    |\|
>>             .------|-\
>>             |      |  >----o----.
>>    o--------)------|+/     |    |
>>             |      |/|     |    |
>>             |        o     -    V ->
>>             |       Vee    ^    -
>>             |              |    |
>>             |              o----'
>>             |              | '--------------o
>>                            |
>>                           .-. | | R | | '-'
>>                            |
>>                            |
>>                           === GND
>> (created by AACircuit v1.28.7 beta 02/28/13 www.tech-chat.de)
>>
>> --
>>
>> Tim Wescott Wescott Design Services http://www.wescottdesign.com
>>
>> I'm looking for work -- see my website!
>
> Right.. did my dropbox link not work?
>
> George H.

Yes it did -- but you have a bunch of extra stuff in that schematic
that's not necessary if you size R correctly.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!
```