Adjustable voltage clamp circuit with minimal voltage drop below clamp threshold

[Rack201]

Hi - could anyone suggest a simple voltage clamp circuit where the clamping threshold is adjustable via a pot, and where there is no/minimal voltage drop when input voltage is below the threshold?  I started down the path of the following with a TL431, but it seems this will only pass a minimum of 2.5V.  My input voltages will be in the order of 1-4.5V and I need to clamp in the range for 3-4.5V, and pass any voltages below the threshold.

Cheers

[SiliconWizard]

What is the max current for the output of your circuit?

[Rack201]

This will sit between a vehicle ECU and MAP sensor.  I wouldn’t expect more than 50mA

[SiliconWizard]

You can have a look at this: https://www.electronicdesign.com/technologies/analog/article/21801600/op-amps-make-precision-clipper-protect-adc

[Rack201]

Thanks for taking the time to respond.  I simulated what I think is the relevant portion of that circuit for my use case, and appears to work.  I have a couple of questions though:

Is the benefit of this op-amp design clipper vs. a simple zener approach just the accuracy at which you can control the clipping point?
To set the clipping voltage on the input to the op amp, is a simple voltage divider as shown ok?
Does D1 need to be a zener as shown in the link you provided, when I am only dealing with positive DC voltages?  In the sim I replaced it with a regular diode and it seems to have no impact



Many thanks - learning here and this probably belongs in the beginner section!

[Zero999]

How fast does it need to be?

[magic]

Hi - could anyone suggest a simple voltage clamp circuit where the clamping threshold is adjustable via a pot, and where there is no/minimal voltage drop when input voltage is below the threshold?  I started down the path of the following with a TL431, but it seems this will only pass a minimum of 2.5V.  My input voltages will be in the order of 1-4.5V and I need to clamp in the range for 3-4.5V, and pass any voltages below the threshold.

TL431 draws less than 0.4mA when Vref < 2.5V. That's 0.4V lost across your 1kΩ resistor. Maybe lower resistance would work?
There low power substitutes which draw less current below the threshold.
Note that the resistor divider also draws some current and decreases output voltage.

Whenever using pots, think about what happens when the wiper gets disconnected.

[Rack201]

I built the circuit I posted above using an OP07C op amp https://www.jaycar.com.au/medias/sys_master/images/images/9825340817438/ZL3974-dataSheetMain.pdf.

It works (i.e. it clamps the output), however:

My typical (e.g. when idling) input signal from the sensor is around 2.26V.  I've noted that as I increase the clipping voltage threshold, the output signal increases somewhat (with no change to the input).  E.g. for an input of 2.26V, with a clipping threshold set at 3.8V, I get an output of ~2.8V whereas it should be 2.26V.  If I raise the clipping threshold further, the output also rises.  I wonder if this op-amp is not suited for a low input voltages, but the datasheet seems to indicate its ok as far as I understand it.  My supply voltage is a well regulated 5V.

Grateful for any suggestions.

[Rack201]

As I learn more, wondering if my problem is that the OP07C is not a 'rail to rail' op-amp...

[magic]

Your problem is components Q21~Q24 (schematic on the front page of Analog Devices datasheet).

If you don't have very strict requirements for offset voltage, even old LM358 could work if a higher supply rail is available.

But with only 5V and a requirement to work with 4.5V signals, you are looking at chips which work up to the positive rail (quite rare, maybe TI's new TL071H?) or with R2R input, indeed. Maybe some MCP6001 or whatever, they seem to be quite popular.

[Rack201]

Thank you magic. 

Would you think an LMC6482  would likely be a suitable alternative to the MCP6001 as I can get it locally?

https://www.ti.com/lit/ds/symlink/lmc6482.pdf?HQS=dis-mous-null-mousermode-dsf-pf-null-wwe&ts=1675461434030&ref_url=https%253A%252F%252Fwww.mouser.mx%252F

[magic]

Yes.
I think if you are satisfied with OP07 except for the problem of limited input range and one input pulling up the other, this one should be good enough.
Its precision specifications are weaker, but I suppose you can live with a few mV offset voltage and thermal drift doesn't look too scary at 0.001mV/°C typical.

edit
You may want to add a resistor (like 1MΩ) which pulls the input to ground or VCC or some other fixed fail-safe level in the event the pot goes open circuit.

[T3sl4co1l]

Yes, that is a fine way to do it.  The response will be slowed by the amp's slew rate as it saturates to the far rail when inactive (this takes ~10us for a typical general-purpose (couple MHz, couple V/us) amp), but those signals sound quite slow (even ~ms would be fast) so that's fine.

Be sure to check that the op-amp tolerates voltage between its inputs.  Some amps use back-to-back diodes, and you have to read pretty far into the datasheet to find them -- they might be current-limited with internal resistors so that the pins themselves still achieve rail-to-rail ratings (i.e. one tied to +V, other -V) but they're drawing huge current (~mA) in the process.

Even an LM358 is likely fine here.

I would however be much more concerned about the EMC and thermal environment, in an automotive context.  Where will this be placed?  Will it be within a larger module, behind filtering and shielding?  Will it be placed in the engine bay, near (or on!) hot materials?

Tim

[Rack201]

I rebuilt the circuit with the LMC6482 and it appears to be working well.  Thanks for the help.

[Rack201]

Hi Tim - I looked at the LM358 and saw in the datasheet that that the output swing is 0V to V+ - 1.5V.  As my supply is 5V and I need to clamp at ~3.8V I figured this wouldn't suit.

I was planning to case the board in a metal enclosure as I run a couple of HAM radios in the car (which is an older diesel).  I haven't yet determined where I'll mount it - would be simpler to mount it on the cold side of the engine bay.  The datasheet seems to suggest a 150 degree max, and it wouldn't get neat that on the cold side.

[magic]

Because of the diode, the output only needs to swing 3.2V to clamp at 3.8V.

But there is a similar limit on the input side, which would be more problematic. Particularly at very low temperature.