9
\$\begingroup\$

enter image description here

Normally we neglect the resistance added by the leads of a resistor because it's too low compared to the resistor's resistance value. But can't this be an issue when it comes to a very low resistance like 0.1 ohm (like the one in the photo)or even less?

\$\endgroup\$
6
  • 5
    \$\begingroup\$ Yep. On current sense resistors you often need a Kelvin connection, so it makes difference where you measure. \$\endgroup\$ – Wesley Lee Apr 16 '17 at 19:18
  • \$\begingroup\$ A decent resistor will list in the datasheet where the connection should be made for highest accuracy. \$\endgroup\$ – Ignacio Vazquez-Abrams Apr 16 '17 at 19:22
  • 3
    \$\begingroup\$ depending on what your requirements are, you need to account for it. If you are measuring current "accurately": yes. If you just need an idea when you reach a certain current (like an over-current protection circuit), probably not. \$\endgroup\$ – pgvoorhees Apr 16 '17 at 19:25
  • 2
    \$\begingroup\$ Following what others, including @Janka below, have said, I have used Ohmite TGHG precision current sense resistors. These have four terminals: two for the current path and two for use in current sensing. Both pairs connect directly across the resistance, 500 mohm in our case, so that the second pair are outputting the voltage drop across just the resistance element, not the current-carrying terminals. \$\endgroup\$ – TonyM Apr 16 '17 at 19:52
  • 1
    \$\begingroup\$ look up resistance of 1mm wire! , more important , what is inductance? \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 16 '17 at 22:18
16
\$\begingroup\$

Your photo is of a 5% or 10% resistor, so if it is 0.1 ohm then there is +/-0.005 ohm tolerance.

This (cement resistor) is not a good kind of resistor to measure current with- it's got poor tolerance, has only two connections (not four) and is inductive (the latter characteristic can cause serious problems in some kinds of fast-switching circuits).

The resistance is probably measured (using a Kelvin connection) at points on the leads similar to where it would be soldered into a board, but only the manufacturer can tell you that for sure.

If the leads are 0.8mm diameter and if they are made of copper (they may not be!) then they would be about 0.035 ohms/m so a difference of 70mm (35mm on each lead) would change it by 0.5%. +/-6mm (about the led length) would change it by less than 0.1%. That kind of resistor is not accurate or stable enough for that to make much of a difference.

\$\endgroup\$
7
  • 1
    \$\begingroup\$ So, what kind of resistors is preferred to do the job? \$\endgroup\$ – iMohaned Apr 16 '17 at 19:38
  • 1
    \$\begingroup\$ And get it in surface mount with four terminals if possible. \$\endgroup\$ – pjc50 Apr 16 '17 at 19:48
  • 8
    \$\begingroup\$ A resistor that says "current sense" and/or "low inductance" in the description. Here is a very expensive one. Here is a cheaper one. And a still cheaper one here. You can also get surface mount jumper-like structures or through-hole bridges. As pjc50 says, SMT with 4 leads (Kelvin connection) is better if you can keep the power dissipation reasonable. \$\endgroup\$ – Spehro Pefhany Apr 16 '17 at 19:52
  • 4
    \$\begingroup\$ There are much higher harmonics in the PWM, so it's too optimistic. I would definitely spec a low-inductance resistor for your application. \$\endgroup\$ – Spehro Pefhany Apr 16 '17 at 19:56
  • 3
    \$\begingroup\$ @SpehroPefhany DARN. I just realized what was the fault with a power project I did years ago. Thanks for the (late) enlightenment. \$\endgroup\$ – Marcus Müller Apr 16 '17 at 21:00
3
\$\begingroup\$

Looks like 10W 1mm wire which is 18AWG or 20mΩ /m and this is only x cm long.

Only a problem if R< 0.001 Ω but > 0 as 0Ω Resistors are common and readily avail. (jumpers)

Special low inductance types are folded magnet wire then twisted for 100nH/cm with 2mm wire of untwisted wire.

enter image description here

Here is a common current sense shunt for SMPS

enter image description here

Here is a common spec for 0 Ω with a tolerance. enter image description here

To avoid measurement errors Rsense must be low ESL with SMPS at all useful harmonics for ZVS mode.

\$\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.