6
\$\begingroup\$

What does m% mean?

Like in this datasheet from TI at the gain error graphs: http://www.ti.com/lit/ds/symlink/ina149-ep.pdf?&ts=1589388038122

Is it supposed to mean per-cent-mille (10^-5) or per-mille (10^-3) or ppm (10^-6)?

Update: Here is an other datasheet with the same unit: http://www.ti.com/lit/ds/symlink/ina226.pdf?&ts=1589357516155

Here the gain error is given without a unit in the table section (unlike in the first one where it is %FSR), but the values still don't match up with the graphs.

\$\endgroup\$
0

2 Answers 2

Reset to default
4
\$\begingroup\$

Assuming you're talking about figure 33, looks like they're plotting 1000x the error in %. So milli-percent.

Maximum gain error is 0.047% so that would be 47, which fits with the graph.

enter image description here

\$\endgroup\$
2
  • \$\begingroup\$ 27, 29, 33, and 34. Does 0.047% fit the graphs though? Because that is where I got really uncertain. The maximum in the graph is just 20. That is significantly less than the maximum 47 and a lot more than the typical 5 they specify in the table section (on page 4). Plus the 0.047 is given in %FSR, is the graph in FSR as well or it is showing the real gain error? I honestly am so confused... \$\endgroup\$
    – Cerike
    Commented May 13, 2020 at 17:29
  • \$\begingroup\$ One is typical and the other is guaranteed maximum. Typical is stated as +/-5 which matches Figure 27. It does not match Fig 33, which makes me think that +/-5 is only at 25°C, despite what the datasheet says. Typical numbers are not guaranteed, of course. An almost 10:1 difference between typical and maximum is a bit much in this case. It's not unusual for there to be a large difference between typical and maximum. 2:1 or 3:1 is common. Sometimes, for relatively unimportant things like leakage on digital inputs that are hard to test quickly, the difference might be many orders of magnitude. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented May 13, 2020 at 17:54
6
\$\begingroup\$

It must be a per cent mille (milli per cent) or 10⁻⁵. I have no idea why TI decided to use such an obscure unit in a data sheet. It is occasionally used in other fields of engineering, but I have never seen it used in electronics.

\$\endgroup\$
2
  • \$\begingroup\$ National always had better datasheets - I don't know why TI was unable to utilize these skills when they purchased National Semi \$\endgroup\$
    – D Duck
    Commented May 14, 2020 at 11:31
  • 1
    \$\begingroup\$ Isn't it obvious? They reassigned all the NatSemi technical writers to menial duties or let them go, it's a natural part of the acquisitions process! \$\endgroup\$
    – Doktor J
    Commented May 14, 2020 at 13:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.