Voltage controlled resistor with minimum threshold

Luke Haberkern

Joined Nov 25, 2020
12
This is an odd problem that I could use help with. I have an input voltage that ranges from 50mV to 10V that I need to use as a voltage controlled resistor to set an output voltage that is pulled up to 10V with a 1k resistor. The tricky part is that I need this output to be driven to 0 for any input voltage less than 100mV.

This cannot be an open drain circuit since the output is pulled to 10V and I cannot change that. I cannot use a diode drop circuit either for the same reason. I considered a JFET circuit but I need it inverted and close to linear. Ie an input less than 100mV needs to set the output to 0V (less than 10mV), 5V input needs to set the output to near 5V, and a 10V input needs to set the output to near 10V.

Any suggestions would be greatly appreciated. I'm sure I'm forgetting some long lost semiconductor... I even considered a PTC thermistor but making it work for this would be challenging.
Thank you

Edit: Thank you all for the various ideas, I think I have a path forward now. I'll be using the P-CH JFET LSJ74A to short the output below ~200mV and create a small offset above that. Should be sufficient for my needs.

Last edited:

crutschow

Joined Mar 14, 2008
33,369
Sorry, but I don't quite understand what the exact relation between the input voltage and the output voltage is that you want.
Tells what you need to do, not how you think it should be done.

WBahn

Joined Mar 31, 2012
29,528
Agree with crutschow. Describe the actual problem you are trying to solve. Treat it like a black-box contract. You need someone to deliver to you a black box that solves your problem. You tell them what behavior you need the box to have and let them worry about what to put inside the box to make it behave that way.

When you say that for an input of 5 V that the output needs to be "near 5 V", how near is near enough? Is 4.9 V near enough?

If the output is 0 V for an input of 100 mV, what should it be for an input of 200 mV? Is an output of 100 mV acceptable?

Sketch a chart having Vin on the x-axis and Vout on the y-axis. Put three lines on it. The first one is what, in a perfect world, you would like the Vout vs Vin to be. On the second one, put the minimum acceptable Vout vs Vin and, on the third one, put the maximum acceptable Vout vs Vin.

Then include a sketch of what this black box is connected to? What is providing the input voltage and what is the output voltage driving. You mentioned that the output is pulled to 10 V. How? By a resistor? What is the value of the resistance? By something else?

ronsimpson

Joined Oct 7, 2019
2,797
What supply voltages do you have? I see +10V. Is there a -10V?
What is the input like? How much load can it drive?
It looks like (input-100mV=output) then limit so the output does not go negative.
I think a simple R-R op-amp circuit will work. It needs to drive the 1k load that is d=connected to +10V.

Luke Haberkern

Joined Nov 25, 2020
12
Sorry, but I don't quite understand what the exact relation between the input voltage and the output voltage is that you want.
Tells what you need to do, not how you think it should be done.
If the input is less tha
What supply voltages do you have? I see +10V. Is there a -10V?
What is the input like? How much load can it drive?
It looks like (input-100mV=output) then limit so the output does not go negative.
I think a simple R-R op-amp circuit will work. It needs to drive the 1k load that is d=connected to +10V.
I have no rails as this will be just a module between two cables. All I have is ground and Vin from 50mV to 10V on the input, and ground and a pulled up 10V output that I need to pull low to the correct voltage. I can draw a representative circuit but it would just be a voltage source with 50 ohm input impedance on the input and a constant 10V voltage source with 1k output impedance on the output. The only way to turn off the device is to pull the output within 10mV to ground.

Luke Haberkern

Joined Nov 25, 2020
12
Agree with crutschow. Describe the actual problem you are trying to solve. Treat it like a black-box contract. You need someone to deliver to you a black box that solves your problem. You tell them what behavior you need the box to have and let them worry about what to put inside the box to make it behave that way.

When you say that for an input of 5 V that the output needs to be "near 5 V", how near is near enough? Is 4.9 V near enough?

If the output is 0 V for an input of 100 mV, what should it be for an input of 200 mV? Is an output of 100 mV acceptable?

Sketch a chart having Vin on the x-axis and Vout on the y-axis. Put three lines on it. The first one is what, in a perfect world, you would like the Vout vs Vin to be. On the second one, put the minimum acceptable Vout vs Vin and, on the third one, put the maximum acceptable Vout vs Vin.

Then include a sketch of what this black box is connected to? What is providing the input voltage and what is the output voltage driving. You mentioned that the output is pulled to 10 V. How? By a resistor? What is the value of the resistance? By something else?
I will provide this. Thank you for the response

Luke Haberkern

Joined Nov 25, 2020
12
Sorry, but I don't quite understand what the exact relation between the input voltage and the output voltage is that you want.
Tells what you need to do, not how you think it should be done.
Here is a representative circuit that shows the input voltage (from 50mV to 10V), an input impedance of 50 ohm, and the desired output voltage that is being pulled up to 10V with a 1k resistor.

The goal of this is to be a simple, passive circuit that drives the output to 0 without an external source voltage.

crutschow

Joined Mar 14, 2008
33,369
So the 10V is not available as a power source?

Can a resistor be added in series with the 1k pullup?

Luke Haberkern

Joined Nov 25, 2020
12
So the 10V is not available as a power source?

Can a resistor be added in series with the 1k pullup?
That's correct, I will not get any power source access. The current solution I have does use an external power source, but I have been asked to remove this complexity as it creates other issues.
I cannot modify anything outside of the box as they are built into COTS hardware.
I think what I need is just a depletion mode mosfet with a low gate threshold.

Luke Haberkern

Joined Nov 25, 2020
12
What supply voltages do you have? I see +10V. Is there a -10V?
What is the input like? How much load can it drive?
It looks like (input-100mV=output) then limit so the output does not go negative.
I think a simple R-R op-amp circuit will work. It needs to drive the 1k load that is d=connected to +10V.
Unfortunately I do not have access to a power supply for this. (actually the current solution I am using does use an external power supply to drive a bjt/nmos switch, but i've been asked to simplify this as it causes other complications). I believe a depletion mode NMOS with a low gate threshold voltage would work, but I haven't seen any device with less than 1.4V Gth.

Luke Haberkern

Joined Nov 25, 2020
12
So the 10V is not available as a power source?

Can a resistor be added in series with the 1k pullup?
How about a P-Channel JFET with a low Gth ? It wouldn't be a perfect match, but might be good enough

Luke Haberkern

Joined Nov 25, 2020
12
P-channel JFET example. Not perfect, but may be good enough for this application

crutschow

Joined Mar 14, 2008
33,369
How about a P-Channel JFET with a low Gth ?
Offhand, that's the only way I can see a circuit to do what you want without a power supply.

Could you add a small, non-rechargeable lithium battery, such as a CR2032 coin cell, or a CR123A cylindrical, if it would last at least several years without replacement?

Luke Haberkern

Joined Nov 25, 2020
12
Offhand, that's the only way I can see a circuit to do what you want without a power supply.

Could you add a small, non-rechargeable lithium battery, such as a CR2032 coin cell, or a CR123A cylindrical, if it would last at least several years without replacement?
That's not a bad idea, I do that with a different project. If the JFET isn't good enough I'll probably just keep doing what i'm doing now which is pulling a supply reference from a separate source from the output.

Luke Haberkern

Joined Nov 25, 2020
12
Offhand, that's the only way I can see a circuit to do what you want without a power supply.

Could you add a small, non-rechargeable lithium battery, such as a CR2032 coin cell, or a CR123A cylindrical, if it would last at least several years without replacement?
The LSJ74B looks good, but damn those are pricey.

Luke Haberkern

Joined Nov 25, 2020
12
Unfortunately I do not have access to a power supply for this. (actually the current solution I am using does use an external power supply to drive a bjt/nmos switch, but i've been asked to simplify this as it causes other complications). I believe a depletion mode NMOS with a low gate threshold voltage would work, but I haven't seen any device with less than 1.4V Gth.
I'm thinking a p-ch JFET is the way to go, but finding one with a low enough threshold voltage means they are pricey... like the LSJ74B

crutschow

Joined Mar 14, 2008
33,369
You could connect the battery through a few high value resistors to bias the P-JFET's input turn-on bias point to be what you want.

ronsimpson

Joined Oct 7, 2019
2,797
With no power I think it cannot be done.