Tutorial: How to match resistors up to 0.01% using a cheap DVM
Feb 13, 2008 at 6:15 AM Thread Starter

#### jERiCOh

I guess everyone who will use that bridge knows how to build it. Just to give you some hints, I used perfboard, 2 2x2 dip sockets to easily swap the resistors, 2 banana posts to plug in the DVM probes and as a source, I used a 9V battery.

Before being able to use the bridge you need to get a pair of 1K resistors matched at least to 0.1% or even better 0.01% to populate R and S. The way I did to get a pair of 0.01% resistors without spending 60\$ is not proven by an engineer but mathematically it should be right.

1- Get 25 1k resistors matched to 1%. (You don't necessarily need 25, the objective is to find the closest match possible. Get enough resistors to find a perfect match)

2- Plug in your DVM probe between X and Y, the other probe between R and S. Set it to its 200mV range.

3- Populate Y, R and S with any of those 1k resistors.

4- Populate X and you should get a value 0.0 and 15.0 on your DVM or maybe a little more.

5- Identify the resistor with Its result.

6- Repeat step 4 and 5 until you get 2 resistor with the same result.

Now that you have 2 matched resistors, put them in position R and S. You are now ready to match your own 0.01% resistors.

Use positions X and Y to test your to-be-matched resistors. The objective is to get the DVM to display 00.0 mV to have a pair of matched resistors.

I hope it is clear enough to be useful to someone...

Any further information about the Wheatstone bridge can be found on Wikipedia

Feb 14, 2008 at 8:19 PM
After reading Tangent's article on matching resistors down to 0.1% Hand-Matching Resistors to Tighter Tolerances , and the need for a pretty high resolution (count) meter, I ended up getting a Fluke 179, 6000 count. This will NOT match to 0.1% below 100 ohms, and will not match to 0.1% at the low end of any of the ranges. It will match to 0.167% down to 60 ohms, and hit the full range in the rest of the ranges, if my little spreadsheet is correct.

Given all the above blather, I wondered about using a Wheatsone Bridge to do the same thing fairly simply. Your approach to finding X and Y seems reasonable to me. The only variables are the resistors you are swapping out, and the resolution of your "cheap DMM." This resolution may be the only "flaw" in your system. You can only match to the accuracy of your DMM to measure current repeatably. My Fluke is 0.09% + 2 counts on the mA scale and should then measure current repeatably to about .2 mA at the bottom of the mA scale to about 0.75 mA at 600 mA. At 15 mA currrent reading, 0.75 mA is about 5% accuracy. This has me wondering if the Wheatstone Bridge is really lets one match to 0.1% with a meter as accurate as my midrange Fluke?

Disclaimer here - I'm an ME, not an EE, so... SOMEBODY CHECK MY RESULTS!!!!

Feb 14, 2008 at 8:57 PM
Assuming R and S are perfectly matched, a difference of 0.1% on X and Y will cause 2.25mV to be measured. Most decent meters should be able to measure to better than the required accuracy.

There is still absolute uncertainty about which resistor is off by how much (since the voltage isn't constant or measurable to such high accuracy), but it should be sufficient to find two very close values.

Increasing the voltage on the bridge increases the voltage drop, and makes it more accurate.

Feb 14, 2008 at 11:30 PM
/me hugs his hp3456A, 4wire 6.5 digit meter. If you got room for the full size 19" rackmount case, they come up on eBay now and then for very good prices

Sep 6, 2013 at 12:26 AM
This is a classic example of how to use classical ideas and a little common sense to do an apparently impossible task!
Thank you jERiCOh, for the article. It is not only useful for something I trying to do, but also triggers my thoughts on several angles.