# Which one is better as voltage divider: resistive, capacitive , low pass filter,…?

There are different types of voltage attenuators for AC signals ( a short explanation is here). The most well known is a resistive one. Others like capacitive, inductive or low pass filters are available ( Low passes may include many designs including passive or active .Thanks to Andy Aka who provided a very good link to them in another thread). I know asking which one is better (especially for high frequencies) is not a good question and the answer is :"It depends".

What I want to know is their advantages and disadvantages of them that may lead to a conclusion for selecting the best design.

• Once again, it depends on what you're trying to achieve. If you want effects that are constant across the whole frequency spectrum, resistors are needed. If you need your output to have features of the derivative or integral of the input, capacitors and/or inductors are needed. There are no advantages or disadvantages associated with ideal passive circuit elements. They do what you need them to do, or they don't. – Scott Seidman Oct 9 '13 at 12:21
• Yes that's correct but in the link to Wikipedia that I provided in the question, it says that for higher frequencies a capacitive element should be added to the design. Why we need that? – Aug Oct 9 '13 at 12:27
• To compensate for the effects of the load. The nature of the load is one of those variables that goes in to the "what you're trying to achieve" column. – Scott Seidman Oct 9 '13 at 13:08

There are basically two types of attenuator I would consider and these can be combined in a couple of ways: -

• (A) is used because it offers simplicity with the ability to design it to suit the driving source and what it interfaces to (centre tap).
• (B) is used when you want to "ratio" down an AC voltage whilst not being concerned with the DC levels but for it to work reasonably at low frequencies the capacitances need larger values than for RF signals.
• (C) is a combo of A and B and gives you a broad frequency range of constant attenuation from DC to RF
• (D) I used once for monitoring the output of a high voltage dc power supply - the main top element of the resistive part of the divider was tens of Mohms and due to its size and proximity to high voltage switching circuits picked-up a lot of noise. Adding caps in the same impedance ratio as the resistors was a beginning but the potential for high currents through the capacitors was a worry so resistors were added in series. Because the voltage divider was used as part of a feedback element controlling the high voltage I had to make sure that what was measured was translated accurately else instabilities might occur and at 50kV it didn't need much instability to destroy circuits. The extra resistors in series with each cap also served to limit currents into the op-amp that the "centre-tap" connected to.

This is just a snap-shot of probably many more techniques.

Resistive voltage attenuators are definitively the most used as their attenuation ratio does not change with frequency. Similarly, the current they absorb from the voltage source does not change with frequency.

In some cases, you have to add a capacitor in parallel to one of the resistors to compensate for the presence of an other unwanted capacitor in parallel to the other resistor.

This is what happens with oscilloscope probes. In the schematics below, R2 and C1 represent the oscilloscope input. The probe itself comprises resistor R1 and capacitor C2. C2 is here to compensate for the effects of C1 (and mus be adjusted before use). The compensation is required to have a flat frequency response curve.

simulate this circuit – Schematic created using CircuitLab

In some very special cases, you can use a capacitive attenuator. For example, you want to obtain a small voltage from the main voltage, with a not that small current, and at the same time you do not want to dissipate much power. This can work because frequency is constant here (50 or 60 Hz, depending where you live).

And we can't forget about switched capacitor networks ( a slight variant on a capacitor only network) which are the go to standard in any modern Analog chip design. They have far superior areal density and matching to any of the alternatives.

Of course this is cheating a bit because under Z transform theory, a switched cap IS a resistor.