First, I’m not an expert user of CodeV; However, I checked the reported refractive index for N-BK7 in both Zemax and CodeV at the same wavelength with and without the environmental factors on at the reference temperature for N-BK7 (20 Deg). My assumption is that this value can be checked against the Sellmeier formula used for calculating the index, i.e, that the change in index will be 0 at the reference wavelength. Zemax reports the same index of refraction to arbitrary precision. CodeV deviates. If I change the evaluated temperature by 2 Deg up in CodeV (from 20 Deg to 22 Deg) I can recover the correct answer. I don’t know why CodeV does this or if it is a bug or something subtle in how to implement the reference temperature etc.; however, I do think that Zemax is reliable in this case.

Hi Michael, thanks for having a look at it.

For N-BK7, Zemax is indeed exactly identical for the nominal temperature (20 °C). Code V shows a difference of ≃ 2·10⁻⁶ compared to the computation.

However, when the temperature changes, Zemax deviates much more. @-40 °C it shows a difference of more than 10⁻⁴! Code V is still in the 10⁻⁶ range.

N-BK7

n(2,1 µm/-40 °C) :

1.4929312 (equation)
1.4928285 (Zemax δ ≃ 103·10⁻⁶)
1.4929364 (Code V δ ≃ 5·10⁻⁶)

n(2,1 µm/80 °C) :

1.4930318 (equation)
1.4930981 (Zemax δ ≃ 66·10⁻⁶)
1.4930310 (Code V δ ≃ 1·10⁻⁶)

n(2,1 µm/20 °C) :

1.4929616 (equation)
1.4929616 (Zemax δ = 0)
1.4929640 (Code V δ ≃ 2·10⁻⁶)

Across the spectrum the computation from Zemax is OK, but when the temperature changes it starts to deviate.

Any clue from anyone ?

Thanks.

Hi John,

I will try and take another look at this.

I assume you’re using the following equation from the Zemax manual,

For additional reference, here is Schott’s discussion of dn/dT.

Thanks!

Indeed, you assume correctly.

Anyone has found somehting?

Thanks.

Hey John,

Are you also asking on CodeV’s forum?  I don’t really know how CodeV implements it, but Zemax is actually more than just the 3 equations listed in the Help File. 

And the steps are: 

1. Scale the wavelength to air at the reference temperature of the glass and a pressure of 1.0 atmospheres.
2. Compute the relative index of the glass at the reference temperature from the dispersion formula.
3. Compute the index of air at the reference temperature of the glass.
4. Compute the absolute index of the glass (relative to vacuum) at the reference temperature of the glass.
5. Compute the change in absolute index of refraction of the glass at the surface temperature.
6. Compute the index of air at the system temperature and pressure
7. Compute the index of the glass relative to the air at the system temperature and pressure

It’s a pretty complex algorithm to convert from relative index (air = 1.0) to absolute index (air = 1.0003) in order to use Schott’s formula.  When you go through all the equations in detail, Zemax is correct.

 From my understanding with my colleagues who primarily design in CodeV, they always rely on the thermal analysis of Zemax.  It’s a small population sample, but from a handful of people who use both, Zemax is more trusted.  

Thanks Michael for your feedback.

You’re right, following the 7 steps gives the right numbers.

Then, Code V deviates more than Zemax.