You got a spare 29 or 25 anywhere from circa 1990?
Then the coil from it should be able to drop in fit.
Now to ask why the NB is different in today's Cobras - you can see numbers on the side of coil can - yes, but the frequency response to noises like impulse and buzzing - relied more on high frequency Rolloff at the band in question - in the older days.
They purposely used coils that were not HF types, but IF types retuned to another intermediate frequency in-between - 23MHz to tune for noises and RF "images" and apply that at the last stage of the 10.7MHz IF amp.
What Cobra did is use a coil set up for 26MHz or less, (Peaks out around 23MHz to as high as 26MHz or it's tuned frequency of 12 to 13MHz) and took in the RF from the RF amp - directly from the RF amp (1st Stage) - and then passed this thru into the NB filter - just to get a sampling of noises from AROUND the 27MHz - just below it - near the 12M band where the thinking was at the time, if you can get it on AM on 19 - you can then hear something like it on a lower frequency. We have to use a coil that can shunt or tune out other frequencies so we don't get noises from their bands - also that we can tune out the 17MHz IF that being mixed to generate the 10.7MHz IF inside the radio - that way -the 23MHz to 26MHz typical 3MHz wide bandwidth of the 27MHz coil RF Amp would pass this noise (along with any RF) into the NB - but by the NB working to INVERT the impulses and buzzing - the shunting would lessen the noise but not generate a birdie to mix into the IF accidentally.
So, to bring you to the present - the issue is now the quality of the coils they are using is not the same as of the older days - they have a higher Q-factor - meaning they are really tight tuning so to obtain a noise in the general band area, means the coils coupling to too close to obtain a BROADBANDED signal and its noise level to work effectively as the older days.
So sub in the old leaky lossy coil to get that 23MHz bandwidth (or 16MHz bandwidth) they both worked (they are also images of each other) to get a noise into the NB strip - amplified, Rectified, detected and averaged - to obtain a level of pulse that is the DIFFERENCE signal - and apply it to the FET before the 455kHz downmix mixer.