27
\$\begingroup\$

If my schematic calls for a 1% resistor, can I use a 10% resistor that measures to the correct resistance within 1% or is there some quality to tolerance beyond what it measures Ohm-wise?

For example, my schematic calls for a 1% 1000-Ohm resistor. I have a 1000-Ohm resistor with a silver band (10%). I measure the resistor using an Ohm-meter and it reads 1008 Ohms which is within 1% of 1000. Can I use the resistor and meet the designer's intent?

\$\endgroup\$
  • 1
    \$\begingroup\$ will it be exposed to heat much? Make sure the resistor you choose has good heat response. Also the material can affect noise.. So if that resistor is for a precision analog feedback network, look out! \$\endgroup\$ – KyranF Mar 1 '15 at 18:16
  • 1
    \$\begingroup\$ You also want to take into account the measurement error of your ohm meter, and together with the fact that higher rated tolerances are often not part of the low tolerance batches, this makes this a game. \$\endgroup\$ – PlasmaHH Mar 2 '15 at 16:31
  • \$\begingroup\$ I haven't seen a 10% resistor (the fourth colour band being silver) except in ancient valve radios. Do you really have one is your possesion? \$\endgroup\$ – Reversed Engineer Mar 2 '15 at 17:37
  • \$\begingroup\$ @phyrfox Substitute superior for inferior, you mean? \$\endgroup\$ – user207421 Mar 2 '15 at 21:26
  • \$\begingroup\$ [Edited] Generally, you can substitute inferior parts with superior parts, but not the other way around. Since 1% is better than 10%, you could use a 1% resistor when a 10% resister is called for, but you can't use a 10% resistor when a 1% is called for, since it is more likely to go out of the expected resistance values during normal usage. As a computer example, you wouldn't use a 250W PSU when you need up to 500W spikes, but you could use a 1000W when you need 500W. \$\endgroup\$ – phyrfox Mar 2 '15 at 21:48

10 Answers 10

5
\$\begingroup\$

Resistors based on carbon tend to have quite more noise (there are a number of noise types apart from basic thermal noise). So you usually don't want them in an audio circuit even if the accuracy itself may not be much of an issue.

\$\endgroup\$
45
\$\begingroup\$

You don't provide much info regarding the environment that the circuit will be used in or the specific resistor types. If you expect temperature variations (or even temperature change caused by self heating) then temperature coefficient becomes important and the initial measured resistance value may be soon way off.

By 10% you probably refer to either carbon or carbon film resistor types.

Carbon resistors typically have a have a temperature coefficient of 5000 ppm/°C which means 0.5%. That is 5% value change with just 10°C variation.

Carbon film resistors have temperature coefficients typically around 200 to 500ppm/°C, which can give a value variation of 0.2% to 0.5% per 10°C temperature variation.

On the other hand a metal film resistor (which is probably what you refer to by 1% tolerance) has a temperature coefficients ranging between 10 and 100 ppm/°C, so a variation of 10°C will result in a value change of just 0.01% to 0.1%

\$\endgroup\$
17
\$\begingroup\$

Component tolerance specifications generally combine two things into a single number:

  1. How much may the value of a newly-manufactured component vary from specification.

  2. How much may the value of a component change with temperature, aging, and other such effects.

In some cases, a "1%" part may be nothing more than a "5%" part whose deviation from spec is sufficiently below 1% that the part's behavior can be relied upon to be within 1% of spec throughout its lifetime under all conditions where it may be used. In other cases, however, materials whose characteristics are stable within a fraction of a percent may be more expensive than those which are less stable.

In many applications where any resistance within 5% of nominal value would be as good as any other, it may be cheaper to make a part whose initial value is within 3% of nominal, out of materials which could drift 2% due to aging and other factors, than it would be to make a part that was within 4.9% of nominal using materials that would be stable within 0.1%. In such applications, so-called "1%" parts would likely be very different from 5% parts, and substitution of the latter for the former would be ill-advised, even if the parts were pre-screened for initial values. Only if one had precise specifications regarding how parts might change with time, and exactly what changes would be tolerable in the target application, would such substitution be appropriate (e.g. if the target application specified 1% parts, but its actual requirement was that components be within 2.5%, and if parts could be guaranteed to drift by no more than 2%, then one could be safe if parts were screened to be significantly less than 0.5% away from nominal). In most cases, however, the cost differential for simply using 1% parts would be less than the extra effort required to use the lower-quality ones.

\$\endgroup\$
11
\$\begingroup\$

What you want to do is fine, and is fine for any one-off project where you're measuring every resistor. It could be that the value needs to be within 1%, but just as likely, the designer didn't want to screw around with less than 1% resistors. There really is no reason to design in 10% resistors in 2015, even if cost is a concern.

\$\endgroup\$
  • \$\begingroup\$ Depends on your volume if 10% makes sense. Agree with your answer part though. \$\endgroup\$ – Some Hardware Guy Mar 1 '15 at 15:46
  • 4
    \$\begingroup\$ Just did a quick comparison of a 0805, 1k resistor in 1% and 5%(no 10% available). The piece price difference on a reel is $0.00057. Management at my company would laugh about trying to save 57 hundred thousandths of a penny. I would imagine that is the case for anyone not selling millions of something. \$\endgroup\$ – Matt Young Mar 1 '15 at 15:58
  • 6
    \$\begingroup\$ You mean 5.7 hundredths of a penny \$\endgroup\$ – Scott Seidman Mar 1 '15 at 16:02
  • \$\begingroup\$ Right that's my point millions of units times multiple resistors in a low margin business. More if you are not in the US. \$\endgroup\$ – Some Hardware Guy Mar 1 '15 at 16:03
  • 1
    \$\begingroup\$ @ScottSeidman Yeah, that, orders of magnitude. \$\endgroup\$ – Matt Young Mar 1 '15 at 16:08
9
\$\begingroup\$

Depends on if you need the stability of a 1% part. A 10% will almost certainly be something unstable like carbon film while a 1% will be metal film.

\$\endgroup\$
8
\$\begingroup\$

Simple answer, For a one off project use the 10% resistor that measures in spec

Two reasons to use a 1% resistor 1. Accuracy, resistance value is within +/- 1% of what you want 2. More values available, With 1% resistors you have more nominal values to work with.

With 5% and 10% resistors the catalog of values is a smaller number. The 1% resistor values go in smaller steps.

For high accuracy the ultimate resistors are matched resistors. Most times in a circuit it is the ratio of two resistors that matter, not the absolute values. So, the ratio of 2 1% resistors have a 2% accuracy. For meters and special amplifiers you can get thin film matched resistors down to .1% matching ratios. Funny thing, on matched resistors the absolute value tends not to be that great but you don't care.

\$\endgroup\$
3
\$\begingroup\$

A 10% error means that the value will fluctuate between +-10% of the specified value. So it can go anywhere from 900 to 1100. Just because you read 1008 once does not mean you can count on it. If conditions change and is within the working specifications, you might get a different value.

\$\endgroup\$
1
\$\begingroup\$

You have to be careful - sometimes a "1%" spec implies other things, even if it's a poor way of specifying things. For example, the designer might have had metal film resistors in mind - they have lower noise and better tempco than carbon composition or even thick film resistors.

It'd help to see the circuit around that resistor. It'd also help to know what kind of a 10% resistor are you trying to use - what's the material of the resistive element?

\$\endgroup\$
1
\$\begingroup\$

It completely depends on what the circuit is and what the purpose of the resister is. Designers don't always give much thought to tolerances when they identify a device, and in many cases it really doesn't matter.

On the other hand, there are cases where it can matter. It's often the case that a designer picks a high tolerance device. This minimizes the need to be concerned about doing any sensitivity analysis to parametric changes in device characteristics. It also might be driven by the designers assumption that the circuit will be built using a particular process (e.g., pick and place with thin film chip resistors), so the tolerance spec comes with the type of devices that are expected to be used. It becomes a question of how the designer wants to spend his time.

It would be better if you showed the circuit you are referring to before anyone gives you a specific answer to your question.

\$\endgroup\$
1
\$\begingroup\$

Yes, you can use your measured resistor.

Your schematic does not specify a resistor stamped with a 1% specification or with a gold band, but a resistor with an actual value within 1% of the value indicated on the schematic.

If you purchased a 1% resistor, you can rely on its indicated value and should not have to verify the actual resistance value yourself. If you purchased a 10% resistor, or wanted to reuse an old resistor with all color bands worn off, you should measure it first.

Regarding stability: simply writing 1% on the schematic does not imply that your device will have to operate for X nr of years within spec. Nor that it will operate over a wide range of ambient temperatures . . .

That said, I would like to add that in lots of cases the values of resistors are not critical. If someone calculated that a resistor of 1000 ohm would work fine, (s)he might specify a 1% because this does not cost more than a 10% one, if you could get one. But there is no promise that the device would not work with a 1011 ohm resistor. One of 1500 ohm might work just as well, perhaps even better, that is completely dependent on the design. A designer might specify all resistors of the same value of 1000 ohm, even if other values might work just a little better, because it is easier to have lots of equal parts. Again, it depends on the design. Use your own skills and judgement and don't let you be intimidated by the "1%" label.

\$\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.