# How accurate are kitchen scales?

I read about weighing coffee to the 0.5 gram, but what's the margin of error for a typical 5kg/11lb kitchen scale?

• What recipe calls for 0.5 gram of coffee ?
– Max
Mar 16 '16 at 10:37
• Maybe they aren't weighing coffee but caffeine? But then, being 0.5 gram off would make the difference between defeating mild fatigue and defeating the borg.... Apr 6 '18 at 23:21

Every scale should indicate its accuracy in its manual.

A good guess if you have no manual is that the accuracy is approximately half of the size of the smallest unit it reports (for example, my digital scale reports down to grams, so its accuracy is probably within about 1/2 gram). If the scale is analog, you can get an idea by the smallest interval on the dial or indicator.

Still, only the manual will tell you for sure.

Scales that go up to 5 kg are unlikely to be good for measuring 0.5 gram quantities like coffee. You will want a scale specialized for that level of quantity.

The exception is laboratory quality scales, which have higher accuracies and greater accurate ranges, but they are not pretty and are expensive.

• If you need accuracy without spending a lot of money, search for "American Weigh Scales" on Amazon. You'll find a 0.1 g scale that goes to 1kg and a 0.01g scale that goes to 100g. Both for under \$10 at the moment. This is good for measuring things like xanthan gum. For coffee, I'd guess an extra half gram wouldn't hurt anything. Jul 26 '13 at 5:52

You'll need a sub-gram scale for that such as this one by Jennings:

The bigger scales' are at best accurate to a gram or two (the expensive german one in our kitchen is good to 2 grams). That defeats the point of weighing coffee given that amounts effect on the result.

Mine gets it right to the nearest gram.

It's easy enough to test the accuracy of your scale using water. For tiny increments, you can use coins. A US quarter weighs 5.67 grams, a nickel weighs 5.0 grams, a dime weighs 2.268 grams. I'm pleased with mine. It doesn't do fractions of a gram, but when I weigh 2 quarters it comes up 11 grams occasionally bouncing up to 12 grams. I can't imagine any kitchen application (outside of molecular gastronomy) that needs more accuracy than that. My scale is "typical" as it has a capacity of 5 kilos.

EDIT: I've been meaning to do this anyway, the following will demonstrate just how accurate kitchen scales can (and should) be. This will also soon be a rave review on Amazon, especially since I only paid \$16 for the scale.

Big pan of cans

OK, 4.585 Kilograms, pushing max capacity

Tare (weight - 0 grams)

OK, I said before that a US nickel weighs exactly 5 grams, right?

One nickel: (weight - 5 grams)

Two nickels: (weight - 10 grams)

Three nickels, four nickels, five nickels, six nickels: (15g, 20g, 25g, 30g)

My scale can beat up your measuring cup!

• So, I appreciate the detail here, but do remember that the question was "how accurate are they typically", so there's some value in giving a straight answer. (Also, keep in mind that your tests demonstrate only that it can weigh differences to the nearest gram; as unlikely as it may be, the 4.585kg could be off by much more.)
– Cascabel
Oct 17 '13 at 19:11
• It correctly weighs to the nearest gram a tiny weight (two quarters). It correctly weighs differences to the nearest gram of 6 consecutive 5 gram nickles when weighed on top of 'ballast' pushing the maximum capacity of the scale. I'd call that pretty compelling evidence of remarkable accuracy to the nearest gram. Oct 17 '13 at 19:29
• Like I said, it's unlikely, but you could get the results you're seeing with a systematic error of say 1%. It'd weigh 30g as 30.3g and display 30. It'd weigh the difference between 4000g and 4030g as 30.3g and display 30 (since you zeroed it). But it'd weigh 4000g as 4040g.
– Cascabel
Oct 17 '13 at 19:38
• Well, we agree on one thing - it's unlikely. For what it's worth I have double checked it with different amounts of water dozens of times. I just looked on Amazon, I bought it in July 2011. I've used it almost daily since. Oct 17 '13 at 19:43