Pot as a divider vs. divider with a rheostat

[CountChocula]

Hi all, I would a little bit of advice about a very basic best practice. When designing a trimmable voltage divider, I usually use the approach below.

My logic is that this is inherently safer, because even if you accidentally turn the pot to either end, a properly-chosen R1 protects the downstream circuit from an excess of current and/or voltage.

The simplest approach, however, would be to simply use a potentiometer, which also reduces the number of components needed. To me, this always feels very risky, because the output of the circuit is unknown until you apply a current to it… at which point it might well be too late :-)

Am I thinking about this problem the right way? I know that it's a small thing, but I'm curious about it. Thanks!

[golden_labels]

The two options are not equivalent. In your scenario the output has a non-linear dependency on wiper’s position and — in most cases — reduced range. See the attachment for charts showing a few outputs for different ratios f between the rheostat and “safety” resistor.  For top resistor being 10Ω and rheostat 10kΩ, you get useless flat line covering upper few percent of the range. For top resistor 1kΩ and rheostat 10kΩ your range covers 70% of of the range, but the curve is logarithmic-like and the shape will slightly change with inaccuracies in both components. Of course that may be what you want, but be aware we are not comparing apples to oranges.

The question is: what are you protecting against?

Voltage surges while the wiper is near the end? Clamp the output or use protection diodes to power rails. The proposed layout, if provides any protection, does that only for shorts to one rail. There is no protection against shorts to ground.

High current when the wiper is close to the end? If that may happen under normal operation, the circuit is likely wrong. Whatever receives a signal from the potentiometer should already have high impedance. Not for safety, but to avoid loading the divider.

High current upon failure? That would imply the downstream got damaged, so there isn’t really much to protect there. But if that is an actual concern and high current could cause dangerous heating, put a resistor at the output of the potentiometer. Mind impedance of whatever accepts the signal. Alternatively you may add resistors at both ends of a normal potentiometer divider. That will decrease its range, but will not affect output curve.

[wasedadoc]

IMHO a trimmer pot should be in the middle of two fixed resistors.  If you need the full range of a trimmer pot without those fixed resistors the overall circuit design is poor, being much too affected by tolerances.

[mcovington]

One thing to think about is what happens if the wiper of the potentiometer goes open-circuit momentarily (noisy pots do that).  The rheostat circuit goes to maximum value.  If the wiper of the potentiometer is the output, then the output disappears.  The latter may be safer, depending on what you're doing.

[TimFox]

IMHO a trimmer pot should be in the middle of two fixed resistors.  If you need the full range of a trimmer pot without those fixed resistors the overall circuit design is poor, being much too affected by tolerances.

For some critical applications, you want to use three high-quality resistors in series, with the pot across the middle resistor.
In that case, the resistance value of the pot should be higher than the resistor directly in parallel with it.
This allows the use of fixed resistors with better temperature co-efficient than the pot itself, but is only useful for a relatively limited range of output voltage (such as adjusting for the tolerance range of a reference voltage).

[MrAl]

Hi all, I would a little bit of advice about a very basic best practice. When designing a trimmable voltage divider, I usually use the approach below.

My logic is that this is inherently safer, because even if you accidentally turn the pot to either end, a properly-chosen R1 protects the downstream circuit from an excess of current and/or voltage.

The simplest approach, however, would be to simply use a potentiometer, which also reduces the number of components needed. To me, this always feels very risky, because the output of the circuit is unknown until you apply a current to it… at which point it might well be too late :-)

Am I thinking about this problem the right way? I know that it's a small thing, but I'm curious about it. Thanks!

The question and any answers are moot until you show what exactly your load is and any specifications.

[CountChocula]

Thanks for all the kind answers! Apologies if the question felt abstract, but I was really trying to figure out if there is a best-practice approach that I could learn.

The two options are not equivalent. In your scenario the output has a non-linear dependency on wiper’s position and — in most cases — reduced range. See the attachment for charts showing a few outputs for different ratios f between the rheostat and “safety” resistor.  For top resistor being 10Ω and rheostat 10kΩ, you get useless flat line covering upper few percent of the range. For top resistor 1kΩ and rheostat 10kΩ your range covers 70% of of the range, but the curve is logarithmic-like and the shape will slightly change with inaccuracies in both components. Of course that may be what you want, but be aware we are not comparing apples to oranges.

Yes, you're absolutely correct, and thanks for pointing that out.

Voltage surges while the wiper is near the end? Clamp the output or use protection diodes to power rails. The proposed layout, if provides any protection, does that only for shorts to one rail. There is no protection against shorts to ground.

Thanks, I hadn't even thought about clamping the output as a possibility! This makes complete sense.

IMHO a trimmer pot should be in the middle of two fixed resistors.  If you need the full range of a trimmer pot without those fixed resistors the overall circuit design is poor, being much too affected by tolerances.

For some critical applications, you want to use three high-quality resistors in series, with the pot across the middle resistor.
In that case, the resistance value of the pot should be higher than the resistor directly in parallel with it.
This allows the use of fixed resistors with better temperature co-efficient than the pot itself, but is only useful for a relatively limited range of output voltage (such as adjusting for the tolerance range of a reference voltage).

Thank you as well—that's another thing I hadn't thought about.

Cheers!

[Terry Bites]

 A string with resistors (with same tempco) above and below the pot will help maintain the output voltage when component values drift. The ratio remains constant(ish).
In the ciruit described the pot and series resistor tempco will not track very well and you will get more drift. So a bad idea.

[CountChocula]

A string with resistors (with same tempco) above and below the pot will help maintain the output voltage when component values drift. The ratio remains constant(ish).
In the ciruit described the pot and series resistor tempco will not track very well and you will get more drift. So a bad idea.

Thank you!