How to calculate the effect of a resistor's TCR?

[John]

I'm trying to determine the power rating I need for a current sense
resistor and am having trouble calculating the effect on the
resistance due to the resistor's temperature coefficient. I want to
minimize the temperature's effect on the resistance as much as
possible (without having to go to a huge resistor, that is) to
decrease any drift in my current reading as the resistor heats up.

If a resistor has a TCR of +/-20PPM/degree-C., is the effect
calculated as follows?

- Assume worst-case 100C temp rise.
- 1PPM for a .010-ohm resistor = .010/1,000,000 = .01 ohm
- 20PPM would then = .2 ohm
- (100C rise) x (.2 ohm/C) = .00002 ohm rise in resistance
- Resistance at 125C (assuming 25C ambient) = .010 + .00002 = .01002
ohms

At 10A, that would only change the sense resistor's voltage from an
"ideal" of 0.1000V to 0.1002, a change of .2mV.

Is the proper way to calculate this?

Thanks for any help you can give me!
John

 

[John Larkin]

On Sat, 15 Dec 2007 21:26:32 -0500, John
<jmuchow@SPAMMENOTcamlight.com> wrote:

I'm trying to determine the power rating I need for a current sense
resistor and am having trouble calculating the effect on the
resistance due to the resistor's temperature coefficient. I want to
minimize the temperature's effect on the resistance as much as
possible (without having to go to a huge resistor, that is) to
decrease any drift in my current reading as the resistor heats up.

If a resistor has a TCR of +/-20PPM/degree-C., is the effect
calculated as follows?

- Assume worst-case 100C temp rise.
- 1PPM for a .010-ohm resistor = .010/1,000,000 = .01 ohm

Micro-ohms

- 20PPM would then = .2 ohm

Ditto

- (100C rise) x (.2 ohm/C) = .00002 ohm rise in resistance
- Resistance at 125C (assuming 25C ambient) = .010 + .00002 = .01002
ohms

At 10A, that would only change the sense resistor's voltage from an
"ideal" of 0.1000V to 0.1002, a change of .2mV.

Is the proper way to calculate this?

Yes. But it's a lot easier to just say: 20 PPM per degree C times 100
is 2000 PPM, which is a 0.2 per cent error.


Current shunts are usually made of manganin. Manganin has an
inverted-parabolic resistance versus temperature curve. "Resistance
wire" manganin peaks near 25C, and "shunt manganin" peaks around 60C.
Either is intended to have a flat, near zero tc, near its nominal
operating temperature. So don't assume that the tc is linear; total
error is usually less than 2000 PPM. Within +-15 deg c of the sweet
spot, tc will typically be under 5.

I have some graphs and stuff if you're really interested.

John

Thanks for any help you can give me!
John

 

[John]

- Assume worst-case 100C temp rise.
- 1PPM for a .010-ohm resistor = .010/1,000,000 = .01 ohm

Micro-ohms

- 20PPM would then = .2 ohm

Ditto

LOL! Thanks John, I missed that. I mistakenly said "change" instead
of "ignore" when the spell-checker flagged "uohm" as being
incorrect...ended up deleting the "u".

Yes. But it's a lot easier to just say: 20 PPM per degree C times 100
is 2000 PPM, which is a 0.2 per cent error.

Definitely easier to remember it that way!

I have some graphs and stuff if you're really interested.

I'd be very interested, thanks!
John

 

[whit3rd]

On Dec 15, 6:26 pm, John <jmuc...@SPAMMENOTcamlight.com> wrote:

I'm trying to determine the power rating I need for a current sense
resistor


If a resistor has a TCR of +/-20PPM/degree-C., is the effect
calculated as follows?

- Assume worst-case 100C temp rise.
- 1PPM for a .010-ohm resistor

The resistor is the first thing to worry about; you also have
mechanical stresses at attachment points to worry about (and
even thermocouple effects) that won't necessarily be negligible.
The mechanical mounting of your resistor should be symmetric
(both ends at nearly same temperature) and some flexibility
should be arranged for the connecting wiring (because the
tempco doesn't include any thermal-stress allowance, and
strain does raise resistance).

Finally, note that the sense resistor also has some series inductance,
and your low resistance means even a low inductance can
be significant. Calculate it, and think about the implications.

 

[John Larkin]

On Sun, 16 Dec 2007 00:12:23 -0500, John
<jmuchow@SPAMMENOTcamlight.com> wrote:

- Assume worst-case 100C temp rise.
- 1PPM for a .010-ohm resistor = .010/1,000,000 = .01 ohm

Micro-ohms

- 20PPM would then = .2 ohm

Ditto

LOL! Thanks John, I missed that. I mistakenly said "change" instead
of "ignore" when the spell-checker flagged "uohm" as being
incorrect...ended up deleting the "u".


Yes. But it's a lot easier to just say: 20 PPM per degree C times 100
is 2000 PPM, which is a 0.2 per cent error.

Definitely easier to remember it that way!



I have some graphs and stuff if you're really interested.

I'd be very interested, thanks!
John

Here's some badly-photographed specs and test data on the two kinds of
manganin.

ftp://66.117.156.8/Manganin.zip

Shunts in the milli-ohm range are usually manganin, although some
low-ohm wirewound resistors may be nichrome sort of stuff, with
different behavior.

The really good stuff is zeranin, but it's hard to come by.

John

 

[John]

Here's some badly-photographed specs and test data on the two kinds of
manganin.

Thanks John! At first glance, I'm able to read all of them.
Time to do some learnin'.

John

 

[John]

Finally, note that the sense resistor also has some series inductance,
and your low resistance means even a low inductance can
be significant. Calculate it, and think about the implications.

Thanks for the tips! I'll definitely pay attention to the mounting
and wiring.

The resistors are being used in a very low-frequency application
(current measurement for the output of a DC power supply) so I believe
their inductance won't be a problem. Though, I guess at turn-off any
inductance that is present will cause a voltage spike, though probably
very low.

John