# Series voltage reference unstable with a series resistor

I am using the MAX6035 as a 5V reference that draws 5mA load current. To protect the input from transient spikes, I thought a simple resistor and zener diode circuit would work well. I noticed that the circuit was drawing more current than expected so I did a little digging. With a 1kOhm series resistor to the input and a 1kOhm load resistor (to simulate 5mA of load current), I was seeing oscillations at the IN pin of the MAX6035. Even in the simulation below, the oscillations occur until I bump up the input capacitance.

The current solution is to add more capacitance to the input pin (1uF vs. 0.1uF) and that seemed to fix the issue at room temperature. However, when running it at -55degC, the oscillations show up again, but to a lesser degree.

I am not necessarily looking for a solution to this specific issue, but an understanding to how a series resistor affects the operation/stability of a series reference. Any references (no pun intended) to reading material is greatly appreciated.

• read the datasheet The MAX6035 typically draws only 73µA of supply current and can source 10mA or sink 2mA of load current Commented Dec 20, 2018 at 1:38

Those regulator are usually a zener with an op-amp.

It is an active circuit with a feedback and thus can oscillate in some circumstances.

What is happening, is that you have a capacitor at the output, and a 1k resistor at the input.

As the system switch on, it needs to fill up C1 and thus drawing more current from the input, as you have the 1k resistor, the voltage at the input will decrease.

As C1, reaches 5V, it draws less current, and U1 input voltage will increase, which in turn changes the operating condition of U1.

Since there is likely some delay for U1 to react to input voltage change, it will decrease the output voltage, C1 will discharge, and U1 will compensate by drawing more current again, reducing U1 input.

If that scenario is correct, you can try to remove C1 and see if you still have oscillation.

Increasing C2 reduces this behavior.

This is generating oscillation at the input which is a sort of Ripple, and voltage reference or regulator are not good at handling ripple.

In a general manner, it is not a good practice to have a resistor at the input of a chip, it will behave on a unexpected matter and you might end-up with a lot of issues. If you need to reduce the voltage before U1, use a voltage regulator like an LDO.

If we disregard the voltage regulator, the two 1k resistors form a resistor ladder that divides the input voltage in half.

The voltage regulator can only add impedance. That means its output voltage can never be more than 50% of input voltage, and will in practice be less. If your voltage supply dips below 10V, the output must dip below 5V.

You are trying to dampen that with capacitors, but all that can do is slow this down a bit. If a <10v state exists for any significant time, the regulator output must brown out.

• The input voltage is 28V Commented Dec 20, 2018 at 3:56
• @Damien did not see that. In that case, pffft, use a diode instead of R1. Supply voltage spikes will only serve to charge C2, which will have the side effect of dampening the spikes too. So the diode should be beefy. I do most of my work in AC voltage where we cannot avail ourselves of options like that. Use em when ya got em! Commented Dec 20, 2018 at 3:59
• @Harper just the cap won't cut it to meet the power input specs. The qual tests being run on this device would easily blow out the reference. I was just looking for some reasoning behind the oscillations due to the resistor. Commented Dec 20, 2018 at 7:03