Hi all,

I am tinkering with some LM358 opamps and voltage clamp diodes. I am not trying to accomplish anything, just tinkering/learning here. So here goes:

I have a LM358 fed by the output of a LM35. The opamp gain is roughly 10x, so at room temperature, 21 degrees Celsius is 0.21V coming out of the LM35 and 2.1V at the opamp output. Increasing the temperature of the LM35 brings the output up in a linear fashion. I add a diode to the output with the cathode side on a voltage rail thinking that as the opamp output increases past the forward voltage of the diode, it will eventually conduct to the voltage rail and "clamp" the opamp output at that voltage (increasing the current drawn from the opamp output which I know is not a good idea...Just tinkering here). So say the diode's cathode is on a 3.3V LDO regulator output, the Vf for a 1N914 is somewhere in the 400mV to 600mV range depending on temperature and current so as the output passes (3.3V + Vf, which is say 400mV = 3.7V) I figured the voltage wouldn't surpass that level.

What I find is that the voltage on the output of the LDO regulator increases along with it and I'm not sure I understand why. Touching a soldering iron tip to the LM35 to heat it up briefly (I have tons of these things, so if it dies it's a sacrifice I'm willing to make) will send the 3V3 rail up past 4.5 to 5V or more.

Why no clamp? I know sometimes there are resistors with Overvoltage protection diodes to limit excess current from the opamp but I would imagine that decreasing the current would make it harder for the diode to conduct and probably increase it's Vf threshold. Does it have something to do with the sink/source of the LDO regulator maybe?

 

Don't "clamp" the Output of the Op-Amp,
if anything needs clamping, do it on the Input,
or, reduce the Power-Supply-Voltage and use a modern "Rail-to-Rail" Input and Output Op-Amp.

You may smoke the Output of the Op-Amp by forcing it to drive a short-Circuit for extended periods.

.
.
.

.

 

as for LDO drifting....
your LDO is a series regulator. it can supply more current if load needs it but ... it has no way to control the output voltage (and bring it down) if some other (external) source is responsible bringing that voltage up.

same is with gravity (just opposite direction...):
you can place something on the table (a fork or knife). table will support the knife and maintain the position with great accuracy even if mass of the knife was to change a lot.... but... this support only acts one way - resists objects from going down. if there is something to pick the knife up... table cannot do anything to stop that.

 

Does it have something to do with the sink/source of the LDO regulator maybe?

Per panic's response above, it is exactly that - the LDO can source current to supply a circuit, but cannot SINK current to hold the rail down to 3.3V

 

years ago had large board with CPU, tons of io etc. board was supplied by 24V (also used for I/O). and onboard was 5V regulator for CPU etc. it was working great until one day someone was trying to connect another device to one of the spare outputs while board was powered up. the output was tested and left on but before the other side of the wire could be secured - it was accidentally dropped and it touched 5V rail on the board. it smoked CPU and couple other ICs... for same reason the series regulator operation was just explained. CPU alone was over $50 but did not have spare and had to wait for a week to get replacement. lets just say client was not very happy. from that day i always include protection not only for I/Os but also add clamping to each power rail..

 

If you want to clamp the output of an op amp then put a 5kΩ resistor in series with the output before the clamp to limit the current to a couple mA or so.

Adding a 1kΩ resistor load at the output of the LDO will allow it to stay at its regulated voltage while absorbing the current from the clamp diode.

 

This makes sense, thanks everyone. So I know the clipper diode is a thing, so in what scenarios would it work since LDO regulators are so widely used? Is there a way to have a regulator work with it? Now that I think about it, in the circuits I've seen that use it I didn't get a good look at the power supplies involved.

EDIT: Actually crutschow answered this. Nevermind

 

The effectiveness of a "clamp" voltage limiter is limited by the current sinking ability of both the clamping device, ( the diode) and also the clamp voltage reference voltage supply. In addition, the forward voltage drop of a 1N914 diode increases with temperature. AND it is not s Schottky device, so the symbol is wrong in the drawing. In addition, often a shunt capacitor is located just above the clamp diode to provide a lower impedance for the rising edge of the clamped voltage.

 

the forward voltage drop of a 1N914 diode increases with temperature

A nit, but a diode has a negative temperature coefficient.