2
\$\begingroup\$

Suppose, let's assume an IC pin requires a 10kohm 1% tolerant resistor connected to ground for its proper operation.

In that case, can someone tell me whether I can split the required resistors into two or three? Like, can I use two number of 5kohm 1% tolerant resistor in series , so that the equivalent resistance is 10kohm connected to ground?

If not, can someone tell me what could go wrong with this approach?

\$\endgroup\$
1

2 Answers 2

11
\$\begingroup\$

can I use two number of 5kohm 1% tolerant resistor in series , so that the equivalent resistance is 10kohm connected to ground?

Yes, you can.

Not many people do, as one resistor is usually smaller and cheaper.

However, there are good reasons that you will often see series resistors instead of one.

  • You don't have the requisite value in stock, or in your parts system, you do have the smaller values, and it's less trouble to use several than to get the one, correct, value
  • You need to spread the power dissipation over a wider area
  • You need voltage handling larger than a single resistor will provide (voltage shares nicely!)
  • You can't find the value you want in your standard series, especially useful when you use two resistors of significantly different value to trim the larger one up

There are some drawbacks, other than the obvious higher cost and space required.

  • Lower reliability based on component count - but this is a small change in the much larger number of components on your board, so essentially irrelevant
  • Higher capacitance to ground - not going to bother you in anything but an RF circuit, which you would be designing properly anyway
\$\endgroup\$
3
  • 1
    \$\begingroup\$ I'm not sure if this is the case, but could there also be a statistical advantage to using many smaller resistors in series? Assuming a normal distribution (which may already be a flawed assumption), wouldn't many smaller resistors in series tend to cancel out their tolerances and result in a sum value closer to the true value? \$\endgroup\$
    – Mezgrman
    Sep 13 at 7:43
  • 1
    \$\begingroup\$ @Mezgrman no unless you know the distribution, which you don't. Just as an example, what if the 1% resistors are all between +0.5% and +1%, if there is a bias in the manufacturing and best resistors are already taken away and sold as 0.5% resistors. \$\endgroup\$
    – Justme
    Sep 13 at 10:16
  • \$\begingroup\$ @Justme fair, that has also been my experience with resistors. \$\endgroup\$
    – Mezgrman
    Sep 13 at 13:41
8
\$\begingroup\$

5000 ohms + 1% of tolerance is 5050 ohms.

Two of those in series is 10100 ohms, which is 10000 ohms + 1%. So same thing.

However, whether it is recommended to split it or not depends on what is the purpose of the resistor.

Usually things like Ethernet PHY chips or other fast communication chips require such a resistor for setting the reference bias current or voltage for the communication interface.

And usually application notes for such chips may be filled with warnings that any noise that gets coupled to such a reference bias resistor or pin will cause noise on the communication lines.

So the resistor placement, wiring, material and even return path for the current may be critical for proper operation of the chip.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.