1
\$\begingroup\$

Leaving the package and the resulting thermal characteristics aside, do diodes with the same name from different manufacturers have the same characteristics?

I would think they should, but when comparing a BAT54 from Diodes and NXP in LTSpice, I get different results.

NXP BAT54, SPICE model under "Documentation"

Diodes BAT54, SPICE model listed under "BAT54(Z)"

I simply put the diode in a DC sweep circuit from 0-2V and measured the forwardvoltage and -current.

What I got was:

Diodes BAT54: 200mV@1mA ; 254mV@5mA ; 287mV@10mA

NXP BAT54: 225mV@1mA ; 278mV@5mA ; 309mV@10mA

Is one model more accurate than the other? Or did I mess something up? The Diodes file is a .model, the NXP is a .subckt with an additional resistor.

I copied the spice files (NXP_BAT54.prm and Diodes_BAT54.txt) into the folder with the LTspice circuit, added a standard Schottky, renamed it to BAT54, set the Prefix to D / X and added the files with a ".lib NXP_BAT54.prm" command.

When I put both curves on top of each other, you see the difference: U-I Plot BAT54 Diodes and NXP

\$\endgroup\$
5
  • \$\begingroup\$ Do the datasheets differ between the NXP part and the Diodes, Inc part? It's very possible that these are tuned to the different minutiae of the different manufacturers' offerings. \$\endgroup\$
    – Hearth
    Commented Sep 29, 2019 at 15:10
  • 1
    \$\begingroup\$ Now this has me wondering, what exactly (if anything) are the standard specifications for each of these generic part numbers.... \$\endgroup\$
    – Hearth
    Commented Sep 29, 2019 at 15:11
  • 2
    \$\begingroup\$ All in all those are tiny differences against which any circuit you design should work just the same. This is to take into account normal production dispersion of both suppliers. \$\endgroup\$
    – carloc
    Commented Sep 29, 2019 at 15:14
  • \$\begingroup\$ Welcome to the real life case of using a simulator: it's your duty to verify not only the models, but also their datasheets, and cross-compare. Only after this you can say the models are ready to be used. Yes, it's bothersome, but also yes, it makes the difference between having some graphs as a result, and having a result you can guide yourself after. \$\endgroup\$
    – a concerned citizen
    Commented Sep 29, 2019 at 15:25
  • \$\begingroup\$ If you're optimising a design for a huge production run, then you ought to have 3 separate SPICE models for each manufacturer's component, one at typical performance, one at max and one at min. Note that some manufacturers' SPICE models will produce the typical behaviour, and some the worst case. Hobbyists like typicals, professionals need worst case. \$\endgroup\$
    – Neil_UK
    Commented Apr 29, 2020 at 13:22

1 Answer 1

Reset to default
3
\$\begingroup\$

Models and graphs are given to shown nominal , while specs are common to both and others with ;

240 mV @ 1mA.
320 mV @ 10mA.

The Min/Max ratings at given conditions should be the same.

Diodes Inc have pioneered lower voltages by improving the bulk series resistance by design and may have some advantage. (Hundreds of patents)

\$\endgroup\$
1
  • 2
    \$\begingroup\$ +1 In other words, as long as both manufacturer's devices meet the datasheet their SPICE models can be different. \$\endgroup\$
    – Elliot Alderson
    Commented Sep 29, 2019 at 18:47

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.