Widely Regarded: Nikkor 20mm f/4 AI

The Nikkor 20mm f/4 AI (and its near identical predecessor, the K) was the optical successor to the very well regarded Nikkor-UD 20mm f/3.5 Auto. The K was produced from 1974 to 1977 and the AI released for a short period after til the start of 1978, at which point the Nikkor 20mm f/3.5 AI had been released. Both the f/4 AI and f/3.5 AI lenses had good reputations but how are they really?


^{Nikkor 20mm f/4 K with factory AI conversion aperture ring}

Physically, the f/4 is small - the f/3.5 itself is already small but this is even shorter, with it just a little larger then the Nikon 50mm pancakes. Focus throw is longer at 140 degrees compared to 100 degreesfor more minor focus corrections and the f/4 being about 20g ligther.


<sup>Nikkor 20mm f/4 AI height vs Nikkor 20mm f/3.5 AI


Nikkor 20mm f/4 AI height vs Nikon 50mm f/1.8 pancake</sup>

Optically, the f/4 has a 10 element/8 group design and its housed - as most smaller AI lens - in a single objective, whilst the f/3.5 AI was a more advanced design with a 11 element/8 group optical construction.


^{(C) Nikon - Nikkor 20mm f/4 AI optical cross section}

The lens has a unique slip on lens hood, HK-3 that was released for the AI lens - you can see the rubber grip has the pattern of the AI/AI-S as apposed to thoe narrower K design language.


^{20mm f/4 K | K with AI conversion ring and HK-3 lens hood}

Both lenses are comfortable shooting into the sun, with very a ghosts and very flare resistant - however the f/4 ghosts a more easily in similar conditions but is not overly problematic.



The 20mm f/3.5 AI performs a touch better in similar conditions.



Ghosts are worse/larger at wider apertures.




^{Nikkor 20mm f/4 AI @ f/11}

Sharpness is pleasant and sufficient - not biting sharp but good enough; between the two 20mms, I'd say they're pretty close and difficult to see one over the other but perhaps the f/4 edges it. Both are centre crops wide open.


^{f/4 at MFD}



Bjorn Rorslett posted the following observations:

All the f/3.5 and f/4 versions of the 20 mm Nikkor are good, but not without some restrictions.

The 72mm f/3.5 UD has some barrel distortion and moderate contrast. Its workmanship is beyond reproach, though and it handles beatifully. I mainly use it for IR for which it gives excellent results. Lots of flare of course in Ir as usual with wide-angles since their coatings aren't optimised for IR. Doesn't bother me at all once you are acquainted with this behaviour. I consider it not optimal for landscapes except the nearer ones.

The successor [ 20mm f/4 AI ] is much smaller and due to size does not handle as well as the 72 UD. It is sharper and have better image contrast though. [..]

The 52mm [ 20mm f/3.5 AI or AI-S ] follow-up to the f/4 has phenomenal performance for shooting into strong light sources and is superb with the narrow K-1 ring. Its corner performance can be slightly soft for landscapes unless you stop down to f/8 or beyond.

The 20/2.8 AIS has some barrel distortion and impressively good performance up close (goes to 0.25m on its own). It flares much more than the f/3.5 though. I use it mainly for underwater photography.

Lets look at the f/4 lens through f/4 - f/8, after which it starts to deteriorate.


<sup>full size crop


full size crop


full size crop</sup>

Comparing the f/3.5 and the f/4 directly, we see at f/5.6 the f/4 is sharper which was a little surprising given the older optics - this at a focus distance of 0.4m on both lenses.


^{full size crop}

At further distances f/4 clearly edges it - note the left side yellow bricks are sharper with the f/4 and the buildings on the horizon line


<sup>full size centre crop

full size centre crop</sup>

For far objects, the lens the details f/5.6 the details are good but optimal performance seems to be at f/8.


^{full size centre crop}

Distortion is there and seems better than the f/3.5.

General Use

Most shot on DX at or close to infinity at f/8.












^{sun stars at f/16 with some minor ghosting}

Servicing

This is where this lens hurts - as with a number of early AI lenses, there was a tendancy to apply a lot of adhesive seals. The front name plate has a set screw that reveals a retaining ring sitting on the focus ring - this was impossible to remove, even with heat. The lens objective itself has the bulk of its lens elements (6) housed in a unit that has to be unscrewed from the front - there are lens spanner slots but I don't know why they bothered - the technician who assembled this at the factory had thread-locked the front optical housing so good that I bent one of my thinner spanner tools trying to unseat it. There is a "locking" hole on the side of the objectived where we would add acetone/IPA to disolve any seal.

Optics

The clearance to get to the slots is insanely tight - to get a set of pointed slots to fit I had to file down the sides my (cheap) spanner tips to fit: ensure that the tips DO fit otherwise it IS possible to deform the objective wall and that can lead to heavy focus feeling (see later). Leave the lens objective in the barrel housing so you can get good purchase to apply (a reasonably) considerable torque to unseat the front optical section:



If you are able to open the front optical block you will have access to elements as laid out:



The front portion of the optical elements live in this block - if you can get to this point, servicing the rest of the optics is easy - there are other seals that need to IPA but minor inconvience. What is not clearly shown below is the objective has alignment marks for each seperatable housing - when reassembling ensure we get back to this tollerance; if there are no alignment marks, make them yourself before unscrewing.





To separate the various elements that have their individual housings, you'll want to grip the front portion in a rubber cup and use the various spanner slots. The front element has housing with circular slots but partially obscured by the above housing so you may have to take apart more items to get to your desired housing.


^{front element's retaining ring removed (and front element getting a mini IPA and H₂O₂ bath)}

Removing the 6th element exposes the rear of the couplet (4th/5th element) - the couplet can be removed from its housing but it has its own (brass) retaining ring



The rear of the 3rd element is seen on the right (its the rest of the optical block sitting on the nameplate to avoid scratching the front element) with its own rear retaining ring



On the rear lens objective, we can get to the rear 4 elements by unscrewing the rear collar that houses the last 3 groups/4 lens elements. The rear element's housing takes up a lot of space and if you don't need to remove any other elements you could get to this by removing just the rear mount to give you more leverage to spin off the housing. This will expose 2 pairs of spanner slots to get to the 7-8th couplet and the 9th element.


^{rear element in housing removed to allow access - notice the factory made alignment marks}

Helicoids

To overcome the barrel's glued-unremovable focus ring collar, which overhangs and obstructs the focus ring's securing screws, we can file an angled groove that is wide and high enough to allow a driver onto the screw head and sufficient clearance for the screw to clear its slot. Filing the groove at an angle will allow for an unmarred collar when re-assembled. Once removed we expose the focus limiter ring screws



The focus limiter ring on this copy was stuck even after removing the delicate slotted screws - there is a slot at the side of the ring that I applied accetone and IPA that helped free it.



To strip down to access the helicoid key, the rest of the action is at the rear. The usual rear mount and aperture ring come off first and then we need to remove the aperture ring coupling screw along with the ident spring. This will give us the remaining chrome grip that is glued with the distance scale.

The helicoid key has a lip that faces towards the front of the lens, which is unlike similar contemporary construction lenses such as 20mm f/3.5 AI, 35mm f/2 AI and 50mm f/2 AI which have a near identical key where the lip faces the rear of the lens.



Separating the outer helicoid is standard fare but care must be taken with separating inner and central helicoid. The limiting block on the inner helicoid has its screws entirely obstructed by the focus limiting ring. This is unlike other latter AI lenses where the top of limiting block can be exposed. To separate the inner helicoid I found that with the limiting ring freely moving, we can carefullly turn both helicoid and the limiting ring as to remove from the front. There isn't a lot of turns before it separates so be careful - I did find that there are only 2 mating positions so if, as I did, the inner helicoid separates inadvertantly you won't suffer too badly: the remated helicoid will must be able to sit the holes aligned to secure the helicoid key.



Normally we'd have a very thin layer of a NLGI #00 grease on the inner helicoid but I found that with this lens - and its front heavy lens objective - we will want a touch more grease for the inner-central helicoids to avoid a slight scratchy feel when fully assembled with the objective inserted: a thin layer on the inner helicoid and then a very thin layer on the mating surface seems to work well. Remate the inner/central helicoids and turn to ensure its not heavy. Ensure the helicoid key/slot is very lightly greased.

Conclusions

One note - the unserviced lens had a heavy focus and this remained even after new grease applied - with the lens objective removed from the barrel the lens' focus is smooth and lightly damped as we'd expect. Inserting the lens objective is a very snug tight into the barrel and once inserted the focus becomes heavy. There is nothing catching (the aperture fork and activiation key are correctly positioned) and its not apparent what causes this heavy feeling unless its the weight of the lens elements pressing into the inner helicoid. Other users do not report such heavy focus so this might be an factory issue with this particular sample.

After more analysis/swapping lens objective into another 20mm f/4 barrel with the same heavy focus, I found that it was the upper lip of the lens objective that was slightly out of round which (I assume) put additional pressure on the inner helicoids and thus uneven/heavy focus. I resolved this by gently reshaping the lens objective around the slots with a few small light taps of a jewellers/watch makers hammer. This made sense given the barrel itself was smooth without the objective inserted.

So a warning, servicing the optics and/or the helicoids is impossible very hard if the lens was assembled in a similar fashion to this copy - no heat, acetone, IPA nor tapping would break the adhesive bonds and you NEED a lot of torque. This is incredible frustrating as this is lens does not appear so often on the 2nd hand market, with about 33,000 lens made based on Roland Vink's data, and any that do become available will almost certainly require servicing given it will be nearly 50 years old.

This is huge shame as the 20mm f/4 AI is a nice little compact lens with nice optical output. The f/4 is sharper than the f/3.5 but sun flare resistance is not quite as good (a few more ghosts) but its certainly very flare resistant. If you can find a clean f/4 take it since the servicing aspect is very challenging but if not an very suitable legacy 20mm alternative can be found in the Nikkor 20mm f/3.5 AI/AI-S which is quite similar in size and optical performance.