An interesting innovation.

Sigma has just announced an 18-35mm zoom for APS-C sensor DSLRs with a fixed maximum aperture of a speedy f/1.8. It works out to 27-53mm full frame equivalent.

It’s no lightweight at 29 ounces – 5 ounces heavier than the stellar 35/1.4 – but the finish appears to be the same, meaning excellent. Like that prime, the new optic will be programmable using Sigma’s dock. I would guess pricing at $900 for the lens and under $100 for the dock. The lens comes in Sigma, Nikon and Canon mounts. I’m sure my D2x would love one, but I am very happy with the inexpensive 35/1.8 Nikkor G prime.

It’s good that an independent maker can challenge the big boys on both quality and price. It seems that true innovation is mostly coming from Sigma and Fuji today.

Sample images appear here.

A qualified recommendation.

Background:

See my earlier Sigma 35/1.4 review comments based on the faulty first sample of the Sigma 35mm f/1.4 DG HSM A1 for Nikon bodies.

For the street snapper using an FF sensor body (me!), arguably no lens is more important than the 35mm. Trained in the classic Leica rangefinder tradition the 35mm was the lens I most often used in the days of film. When Leica’s bodies fell behind I moved on, first to a Canon 5D, then to a Nikon D700, and now a D3x. Thus it’s worth sweating the details in getting the best 35mm optic, especially if low light use is contemplated.

I would preface what follows with the fact that I am brand agnostic – camera and lens. Whatever works. Until now I have used only Nikkors on my bodies because I know little of aftermarket lenses and the price differentials are mostly immaterial for a long-term user.

After I complained to Sigma USA about the faulty loaner, they loaned me a brand new copy of Sigma’s 35/1.4 lens to try. They volunteered this. I did not ask. After some quick snaps at home of my long-suffering test target, Bert the Border Terrier, all seemed well, so I hit the streets. I did give him a cookie first, though.

The first lens I had borrowed from BorrowLenses.com had a random AF error, sometimes front- sometimes back-focusing. Heck, sometimes it was right, too! And when it was right it was beyond compare. So trying again with another sample was worth the effort.

I had explained to Sigma that I would be more than pleased to buy the loaned lens given its stellar performance, provided AF worked every time. f/1.4 is nice, but I already own the f/2 pre-Ai MF Nikkor, so spending all that money for one more stop and unreliable AF does not solve.

Results:

“What are you looking for?” At f/1.4.

Well, things worked out great and the second loaner was perfect in every regard. I ordinarily use point/lock focus/recompose with AF lenses, as it’s fast and consonant with my snapping theme which is mostly street work. Further I do not need focus tracking as my subjects are not moving fast and I certainly do not trust 51-point autofocus to decide where my key area of sharp focus resides. How on earth can it know? At very short subject-to-lens distances I will compose first, then change from the central focus point to one over the subject, thus obviating the change in subject distance occasioned by the ‘recompose’ step. However, I generally dislike using the adjusters on the Nikon’s back as they are slow and clunky. By the time I have dialed in the focus rectangle of choice the subject has moved on, as often as not.

Triple Rock bar, Berkeley. At f/2.

So good is the AF in this sample that my D3x body needed no AF fine tuning and that is just as well, for the APS-C sensor D2x in my other Nikon body provides no such capability and it’s nice to be able to use the Sigma on it with confidence. Sigma’s promised USB attachment which may make AF fine tuning possible in the lens might address issues for bodies without the fine tune capability, but that will have to wait determination until the device becomes available. One thing it will do is make it possible to do firmware upgrades as these come along, as possibly required by new bodies down the road.

Why buy it?

It bears repeating that there is very little point in buying this lens if most of your snaps are not taken at f/1.4, f/2 or f/2.8. There are any number of excellent Nikkors in both MF and AF guises which perform every bit as well from f/4 down and weigh a fraction of the Sigma. They are also far more compact. Indeed, so large is the lens hood Sigma provides with their optic that I do not use it as it simply sticks out too far, and I find I am constantly whacking it against something with the camera over my shoulder. I just opt for a protective (67mm) UV filter and have at it. The lens is very heavy and you are not going to enjoy carrying much more than one big body with this lens and maybe a medium telephoto like the 85/1.8. Much more than that and you will be hurting before long. Pain is not conducive to happy picture taking. Even on the heavy and large Nikon D2x and D3x bodies I use, the lens makes the combination top heavy and the kit will not stand upright on these cameras’ broad bases, preferring to tip forward. However the large size and excellent ergonomics more than counter this front heavy design.

Saloon. At f/8.

AF versus MF. No test charts, honest!

I tested the lens on a tripod at typical snapping distances of 5 and 10 feet at f/1.4, taking sets of three snaps:

• Lens set at infinity, using AF
• Lens set at the closest focus distance, using AF
• Lens set on MF, using the focus confirmation light in the finder of the D3x

I did this both by daylight and by incandescent light at f/1.4.

The first two were always perfectly focused. The third was seldom correct, reflecting the too-broad range of the focus confirmation LED in the Nikon. AF beats MF every time at f/1.4. If your eyesight, like mine, is not the greatest, you propose to use f/1.4 and you find you opt for LED confirmation MF rather than screen MF, then this finding should give you pause in deciding whether an MF f/1.4 lens (Zeiss/Cosina, Nikon MF, Samyang) is for you. The AF choices – Nikon G or Sigma – appear the best option in this case. There are alternative focusing screens for certain Nikon bodies which might help, but I have no data. I got like AF results on both FF (D3x) and APS-C (D2x) sensor bodies.

For a whole bunch of AF examples at f/1.4, look here – I do not photograph test charts. These were taken under incandescent and fluorescent light. More wide aperture snaps accompany the article you are reading.

Buying the lens:

The owner and the loaner. Two 35mm f/1.4 Sigma lenses in Nikon mount.

I had made it clear to Sigma USA that I was a buyer, not a hack reviewer trolling for clicks and freebies, I told them I was pleased with the loaner and asked to buy it. But no. They wanted the loaner back. When I complained that made no sense they offered to sell me a new one from their stock, and I quote from their email:

The writer meant ‘courtesy’ but ‘curiosity’ seems equally appropriate.

Bottom line? The loaner I received was pre-screened. Rational extrapolation? I would guess that all review samples are cherry picked before being sent for review.

Well, this raises some questions as you only screen review samples if you are worried about your QC, but I reckoned I might as well make the process a ‘win-win’, so I replied thus:

The one I eventually bought was every bit as good as it should be. It, too, was cherry picked, as the above emails disclose. Here is the serial number of the one I bought:

The lens I purchased.

Off-center focusing:

The Nikon D3/D3x/D3s/D700/D800/D800e/D4 all share the same AF module, the Multi-CAM 3500FX. Many users of the D800/800e have reported focus problems when using other than the center focus rectangle for AF. I tested this at 5′ distance and f/1.4 with the Sigma on my D3x focusing by using the extreme left and then right AF sensors at f/1.4. Focus was perfect for each, so it may be a D800/800e build or Nikon QC issue which is involved here. Once again, this is based on a sample of one, so treat this information with a pinch of salt.

What I saw:

Based on the two samples from Sigma and the loaner from Borrowlenses.com, this is the highest resolution 35mm lens, at wide apertures, I have used. It is also the bulkiest and heaviest. The Sigma easily out-resolves the 35/1.4 Nikkor G which BorrowLenses.com loaned me when I returned the first faulty Sigma, and is equal to or better at f/2 than the 35mm f/2 Asph Summicron for the Leica M which I owned for many years before giving up on Leica’s dated bodies. It also renders out of focus areas better at f/2 than the Asph Summicron, which tends to harshness. There’s more to life than raw resolution and the Sigma does not disappoint when it comes to color rendering where it easily matches the best I have owned from Nikon and Leica.

Images downloaded into Lightroom pop on the display. Colors have a natural quality commonly seen when Leica or older MF single-coated Nikkors are used.

And while pixel peeping will show that f/2 is better than f/1.4, f/1.4 is fully usable at all times and I find that I never hesitate to go full bore, gaining shorter shutter speeds in the process. F/1.4 with a fast shutter speed and no blur beats the alternative. As f/1.4 means backgrounds will be blurred, it bears adding that the out-of-focus bits are pleasantly rendered by this optic. You do not need a single test chart to tell you all of this. Just use your eyes.

Ethnic pride in North Beach. At f/2.

Sample variations:

• The first loaner from Borrowlenses.com had perfect focus collar resistance but the AF was faulty. When focused correctly resolution was the best I had ever seen at f/1.4 to f/4, equalled by some other makers’ lenses at smaller apertures. Serial number 50004444.
• The second loaner, from Sigma USA, had a slightly too stiff focus collar, perfect AF and resolution to match the first loaner above. AF nailed focus every time. S/N 50004693.
• The one I eventually purchased from Sigma USA had perfect focus collar resistance, perfect AF and resolution to match the above two. An occasional squeak can be heard from the AF mechanism, audible to the operator only. I expect it will go away with use. S/N 50022095.

So there are sample-to-sample variations but the one constant was the high resolution of all three samples. Let’s be fair. Even Leica, Nikon and Canon have sample-to-sample variations in their premium lenses. Just check Roger Cicala’s blog to confirm this.

Other considerations:

What else is there not to like? The lens is a tad sensitive to flare when the light source is really bright and close to the axis. I suppose using the hood might help here but that’s not something I am willing to do. So, maybe not a fair test, but one consonant with my working method.

Flare central – what happens with the sun in the frame or just outside it.

The focus collar turns the way most lenses do, other than Nikkors that is. (Only Zeiss seems to go to the trouble of conforming the rotation direction on its Nikon versions). However, as I use AF with the Sigma all the time – and the AF is faster than that in the competing 35/1.4 Nikkor-G – it makes no practical difference. The focus collar on my loaner from Sigma was stiffer than on the first, the one on the purchased copy is just right.

I also dislike not having a physical aperture ring on the lens but that is the way of the world. That’s a Nikon issue, not a Sigma issue. You have to use the control dial(s) on the body to change apertures and may have to touch the shutter release first if the LCD display has gone to sleep to check your setting, otherwise no amount of control dial twiddling will change aperture. But the stellar performance of this optic makes this a light cross to bear.

North Beach paint job. At f/2.

Best of all, paired with a high pixel count full frame sensor like those found in the D3x, D600 and D800 bodies from Nikon and in the Canon D1x, 1Ds/III, 5D/II and 5D/III, there is high excess detail in files which allows for selective cropping when you could not get as close as you would have liked to your subject. I have cropped to one quarter of the frame and made 24″ prints and the results are stunning. No other word for it. That makes for a very capable body-lens combination and increasingly finds me leaving the medium telephoto at home.

Warmth. At f/2.8. Cropped from one quarter of the D3x 24mp frame.
This prints beautifully at 18″ x 24″.

Qualified recommendation:

The circuitous purchase route I took and the fact that my lens was pre-screened by Sigma USA’s tech people makes it impossible to give a blanket recommendation for this lens. Yes, good ones are as good as it gets. Both the loaner, also pre-screened, and the one I finally purchased direct from Sigma are outstanding. But buy ‘off the shelf’ from B&H or Amazon or whoever, and you may still find you are in a lottery. This may mean returning one or two before lucking out, and a growing body of AF issues is now being reported on chat boards. These also seem to be a problem with Sigma’s 50/1.4 and 85/1.4 lenses with samples requiring substantial focus tuning adjustments. In fairness, chat boards tend to be like hospitals. Only the sick have an axe to grind there. Also, bear in mind that the issue with my first loaner was not one of micro-focus adjustment. The AF was faulty. Period.

All manufacturers will make occasional duds. It’s a toss-up whether Sigma’s QC beats that of the opposition. Only industry insiders know. At this time there is no statistical basis for concluding whether Sigma’s quality control is improved from the bad old days or not. Nor do we know how it compares to the QC at Leica/Zeiss/Canon/Nikon et al. What is troubling is that, as both the Sigma and the 80% costlier Nikon 35/1.4 lenses state ‘Made in Japan’, I struggle to understand how Sigma manages to sell its lens for so much less without some cost cutting along the line. The math does not solve. Would I pay $1,100, $1,200 or even more with assurance that my lens has been subject to rigorous QC or pre-screening? Yes. This optic is easily worth that amount, especially when compared to the competition from Nikon and Zeiss, the latter MF only.

One other unknown must be considered. Nikon lenses are famously long lived. Any number of 40+ year old pre-Ai MF Nikkors I own testify to that fact. Will AF lenses last as long, given the complexities of motors and gears and electronics? Has Sigma cut cost saving corners compared with Nikon in making this optic? I do not know. Only time will tell. Still, if one gets three years of hard use from the lens (which is the Sigma USA warranty period to the original purchaser from an authorized US dealer for its DG non-EX lenses, like this one) then it probably does not matter. Sell at the end of the warranty period and replace if it’s a concern.

Conclusion:

I recommend you buy this lens if it meets your low light needs. You may have to try more than one sample if your first is a bad one. While I was unlucky with my first sample, there is no statistical basis for making any conclusive statements about Sigma’s quality control. A sample of three is not meaningful. One day we will probably see statistically meaningful results from the likes of Roger Cicala at LensRentals.com. Roger tests dozens of samples of each lens and can meaningfully address sample variation. But Roger, please make these tests include finder controlled AF, not LiveView. That’s why you use f/1.4 – street snaps in poor light. LiveView is not a concept here nor are tripods involved.

Blinds. In the tradition of Eugène Atget. At f/8.

USB dongle:

10 pins on the CPU.

Sigma has promised to sell a USB attachment which will allow firmware updates – and maybe other adjustments? – to the lens. As you can see, the lens has 10 contacts on the CPU compared to a maximum of 8 on Nikkors, suggesting the other two are needed for this device.

Sigma’s rumored USB dongle.

Serial number:

This is the s/n on the Sigma USA loaner I had to return.

The serial number on the Sigma is very hard to make out. It’s screen printed in faint type atop the lens, diagonally up and to the right of the (useless) distance indicator window. My previous loaner was 5000444 – that was the one with the faulty AF.

Both the above lens serial number images were made with the excellent Lumin app on the iPhone 5.

Comparison with the ancient Nikkor MF pre-Ai 35mm f/2 Nikkor-O:

The Nikkor is over 40 years old so this is a brutal comparison.

In each case I use the lens correction profile I have made using Adobe’s Lens Profile Creator application. It’s more necessary at large apertures in the case of the Sigma where significant vignetting (corner shading) at f/1.4 and f/2 is removed. You can expect Adobe to come out with their own version of this profile soon. I also apply Sharpening=66 for the Nikon D3x, which is what I find to be optimal to overcome the anti-aliasing filter located in front of the 24MP sensor in the camera.

Center performance of the Sigma is clearly better in 5 foot wide prints at all apertures down to f/4. The resolution of the Sigma is truly outstanding from f/2 down. Will the Nikkor yield good 5′ prints at f/2 in the center? Absolutely. Will the Sigma appear sharper? Yes. Enough that you will notice on critical inspection.

Boozers. At f/4.

Edge performance is a different story. The Sigma is greatly superior at all apertures, but especially in the range f/2 though f/8. There is no comparison. The Nikkor is a fine lens but it shows its age here. For the street snapper edges do not matter much. For the landscape artist there’s a big difference. The Sigma is in a different league.

The old Nikkor is marginally less prone to flare into a light source.

Color rendering for both lenses is outstanding. The Sigma is multi-coated, the Nikkor single-coated. Whatever.

Comparison with the current Nikkor 35mm f/1.4 AF-S G:

The Nikkor AF-S G is outperformed by the Sigma at center and edge through and including f/2.8. After that I cannot tell the difference. The Nikkor’s AF is slower but no big deal. The current Nikkor f/1.4 is equal to the old MF Nikkor f/2 in the center at all apertures but clearly superior at the edges at all apertures.

The current Nikkor is marginally less prone to flare into a light source and vignettes less, before correction, at f/1.4 and f/2.

The Nikkor is lighter – nice. The Nikkor costs 78% more than the Sigma in the US, ex-tax. Not nice. Used Nikkors are coming to market at $1250, probably as word gets out about the Sigma.

Lunch. At f/1.4.

Adobe’s lens correction profile:

The just announced Beta version of Lightroom 4.4 now includes an Adobe profile for the Sigma.

Lens correction profiles added in LR 4.4 Beta.

I have tested the Adobe profile against mine and there is no detectable difference at any aperture between the two. If you are using my profile there is no need to rush into the Beta 4.4 update of Lightroom. Let it mature and migrate when Adobe says it’s final. If you want to be sure Lightroom is using Adobe’s profile, do erase mine if you have it installed, as it otherwise takes precedence over Adobe’s.

Full disclosure: For reasons I do not understand, Sigma sold me the lens at a discounted price. I did not ask. They volunteered the discount out of the blue, even though I told them my income from photography is zero and that I am an amateur. While the discount I received was not monetarily significant to me, I reciprocated by sending them two nice 18″ x 24″ prints taken with the lens.

Maybe the most significant Nikkor ever.

Shown with the period HN-7 screw-in hood.

If ever a lens deserves the appellation ‘famous’, it’s Nikon’s 85mm Nikkor-H f/1.8, first made in 1964. Nikon has made many f/2, f/1.8 and f/1.4 lenses of 85mm focal length since then, and while the current AF-S f/1.4 probably delivers the highest resolution of the lot, few would place the Nikkor-H far behind optically, and only non-users would argue that the current lens is better made. Much as with its contemporaries, the 20/3.5UD, 24/2.8, 28/2, 35/2, 50/1.4, 50/2, 105/2.5 and the 200/4, the construction quality, fit and finish of this lens have not been improved since.

None of the manual focus f/1.8 optics came with the Ai modification which permits mounting on later film cameras and most DSLRs. Nikon made (ugly and no longer available) Ai ‘kits’ which comprise a replacement aperture ring with the wrong surface finish, so if you want an Ai version of this lens you have three choices. Buy one which has already been converted, send it out for machining, or do it yourself. I did it myself and the task is simple. See this piece for details of how to do it yourself.

But ‘famous’? Why yes. Because this is the lens David Hemmings used in the studio scenes in Michelangelo Antonioni’s ‘Blow-Up‘ to photograph Veruschka while she writhed below him. His Nikon F with the Nikkor-H mounted can be seen in the contemporary poster.

Fame.

I have long lusted after this optic but good ones go for high prices and cheap ones are usually beaten up, the anodizing on the scalloped metal focus barrel worn away and ugly. Well, I finally snapped one up in near mint condition for $210 delivered, and when it arrived the other day I immediately set to machining the aperture ring to permit mounting of the lens on my D2x and D3x Nikons. I would love to have used it on a Nikon F but no digital version of that classic was ever made, and I no longer use film.

Amazingly, the filter size is the same small 52mm common to most Nikkors of that era and the HN-7 hood screws in. No clip-on nonsense waiting to be knocked off.

Comparison with the AF-D version:

I set to making a quick test using the local back yard utility pole as my test target. My test was against the ‘plastic fantastic’ 85mm f/1.8 AF-D lens, set in the most ghastly plastic mount imaginable, but of outstanding performance. This lens is still available new in the US for $460. How Nikon manage to make a bitingly sharp optic encased in ductile materials and cheese beats me, so I confess that I ran my simple test with some trepidation. Could any lens significantly improve on the AF-D or, for that matter, even match it? My Nikkor-H is positively geriatric, having been made in February, 1969, meaning 44 years ago. It cost me $212.

I should have known better than to worry. Central definition and overall contrast of the old lens easily beat those of the newer one from f/1.8 to f/4, and vignetting is identical, disappearing by f/2.8 in the old, f/3.5 in the new. Edge resolution in the newer lens is superior through f/4 after which they are identical. In a lens of this focal length, the ideal portrait lens on full frame, central resolution is what counts.

The focus throw on this lens is very long – fully half a circle from infinity to 3 feet, making accurate focus on close subjects easy. Depth of field is very limited at wider apertures so a slow focus ring is actually an asset.

Here’s my thrilling test target, what passes as utility service in the SF Bay area, allegedly part of the most powerful nation on earth. The old MF lens is the left hand image – focal length and maximum aperture are not reported by Lightroom (in the parentheses) as I had yet to install a CPU when this was taken:

Even viewed via this blog on the 11″ display of my MacBook Air the difference is obvious at f/2, remaining so at f/1.8, f/2.8 and f/4. The enlargement above is from 5 foot 90dpi equivalent print sizes.

Chipped and ready to go, one of Nikon’s best ever.

The CPU confers a host of benefits, described here and installation on the 85mm f/1.8 Nikkor-H is a simple glue-on job. One key advantage is that you can pass aperture control from the control wheel on the camera’s body to the aperture ring on the lens, sidestepping the non-linearity issue I explain here. Your kit will handle better too, allowing apertures to be changed by supporting the lens properly from below with the left hand. $29 for the CPU, 5 minutes to glue it in place and another 15 minutes to program it, as described here.

Can a lens make a better photographer? When it comes with a heritage like this and when the user revels in the operational feel and the results it yields …. well, you can make your own mind up. Me? I’m going to track down Veruschka’s granddaughter.

Veruschka writhes as Hemmings snaps.

Lens correction profile:

This lens is so well designed that what optical shortcomings there are – very minor vignetting down to f/3.5 – are easily corrected with my lens correction profile in LR or PS, which you can download here, but of all the pre-Ai MF Nikkors this one arguably needs a profile least. I can detect absolutely no barrel or pincushion distortion. Likewise, chromatic aberration (color fringing) is negligible, though the profile corrects what little there is. My profile was made at f/1.8, f/2.8, f/4 and f/5.6, with the last prevailing at apertures smaller than f/4. It works with both FX (Full Frame) and DX (APS-C) sensors. The profile was made on my D3x but will work on files from any Nikon DSLR.

To get a sense of what this lens can do in the studio, click here. The handling and balance on a modern full size Nikon DSLR are about as close to perfection as these things get. If you can live without AF, search one of these out.

Use on Panny MFT bodies:

A wonderful lens on the Panny G bodies with a $25 adapter, delivering 170mm f/1.8. Very shallow depth of field, and you retain aperture priority automation, the EVF never dims as you stop the lens down (think about that!) and you have a state of the art MF focusing aid which makes dead on focus trivially simple. Read more about the immense capability of the Panny MFT bodies with MF lenses here. As you are really cherry picking the center of the image circle produced by the lens, definition at any aperture all the way to the extreme edges is not an issue.

The Hemmimgs lens on the Panny G3.

A 1969 lens on a 2012 body, and fully functional. Pretty cool, huh?

Updating classic Nikkors.

Why pre-Ai Nikkors?

Few would disagree that the ‘pre-Ai’ Nikkor lenses for the Nikon F and subsequent film camera bodies were mechanically among the best ever made by anyone. Aesthetically they are a joy to look at, with their scalloped metal focus collars, and ergonomically a dream to use with smooth focus actions and beautifully detented aperture clicks. Optically, most – as I have shown here many times – can hold their best with the latest ‘plastic fantastic’ lenses, and while there’s no AF or VR, there are no motors to fail, either. Many of my prints made using these testify to their outstanding optical qualities. Later ‘pre-Ai’ lenses adopted rubberized focus collars, and while their optics are largely unchanged, the look and feel – for this snapper – are just not the same. The common nomenclature is to refer to the single-coated metal scalloped ring optics as the ‘F series’, the later multi-coated ones as the ‘C series’ and the yet later pre-Ai lenses with the rubberized focus rings as the ‘K’ series. All mine are either F or C.

A newly Ai converted lens, in this case the 85mm f/1.8 Nikkor-H.

Incidentally, while the oldest pre-Ai lenses nearly always come with scalloped focusing collars, the design of the aperture ring seems to be without rhyme or reason. Some are scalloped (my 20/3,5 UD, 28/2, 28/3.5, 50/1.4, 105/2.5, 135/3.5 and 200/4) whereas others dispense with scallops and adopt fluting only (24/2.8 and 35/2). No matter. All are a delight to use.

Drawbacks:

The snag with pre-Ai lenses is that, if you can live with manual focusing, they will not fit a modern Nikon film or DSLR body. That’s actually an advantage, as the need to make this modification mostly keeps used prices of ‘pre-Ai’ lenses down, as they cannot be used on newer bodies in the absence of the Ai conversion. The work required to ‘Ai’ the lens (meaning it will fit) is very easy as either a DIY project or sub-contracted to a skilled machinist who will charge you $30 or so.

There are two aspects to updating classic Nikkors to work on modern DSLR and late film Nikon bodies.

One is to add a $29 CPU to allow proper recording of EXIF data, something I described at length here. You do not have to do this but I strongly recommend that you do. If you, like me, are a student of your EXIF data from which much can be learned and if you like matrix metering and such great features like shutter block for action photography, adding a CPU makes sense.

The other, and this has to be done first, is to relieve an arc on the rear of the aperture setting ring on the lens to provide the required clearance for the aperture feeler tab on the body. Later bodies dispensed with the spigot on the body which couples with the claw on the lens, opting for a rotating, sprung feeler, concentric with the bayonet flange. If the aperture ring on a pre-Ai lens is not machined down, it will damage the aperture feeler, rendering the camera inoperable.

Exceptions:

To effect manufacturing economies for their less costly bodies, Nikon deleted the Ai feeler and the screwdrive focus motor coupling for AF-D Nikkors. The good news is that these bodies allow a non-Ai lens to be mounted without modification (no feeler tab to foul) but the trade-off is that metering does not work. Use an iPhone app like Photometer for manual exposure measurement. And forget AF with AF-D lenses.

The bodies thus designed include:

D40, D40x, D60, D3000, D3100, D3200, D3300, D5000, D5100, D5200,D5300, D5500.

Sub-contracting the work:

When I first started buying pre-Ai lenses I sent them for Ai conversion to John White who does an excellent job for modest cost.

Do it yourself:

Later, as I got braver, I started doing the conversions myself and what I describe below is what John White refers to as the Type A conversion on his site. This works with newer film bodies such as the FM, FE, F4, F5, F6, N90, FM10, F100 and FM3A and digital SLRs such as the D1, D2, D3 and D4 series and the D200, D300 series, D600, D700, D7000 and D800. If your camera body is an “Electronic Only” body such as the N80, D100, D70, D70s, D50, D80 and D90 I recommend you send your lens to John White to have the work done. There’s a little more to it and I cannot advise never having done a Type B conversion, as all my bodies are of the ‘non-electronic’ kind.

Now ‘machined down’ is a rather high falutin’ way of saying that you use a small flat file to remove a sliver of alloy from the rear of the aperture ring of the lens. The process is very simple and easily done by all but the truly cack handed. The latter should use one of the commercial machinists like John White to ‘Ai’ the lens.

You will need a jeweler’s screwdriver to remove the five (sometimes four) rear chromed flange retaining screws. These are either slotted or Philips, depending on the lens. You want the screwdriver to be a really tight fit to avoid trashing the screws. Sometimes screws are frozen in place – after all these optics are mostly over 35 years old – so a miniature soldering iron, applied to the screw for a few seconds, can help loosen them. A broken screw head is not going to make your day. The several lenses on which I have removed the rear flange have been all over the place with regard to the use of thread locker on the screws. Some have none, some a light application and some gobs. As a generalization, the older the lens, the more thread locking compound is used. This dictates the use of large handled screwdrivers which permit far more force to be applied when engaging the screw’s head, minimizing the risk of a damaged screw. Those poncy little screwdriver sets passing as jeweler’s screwdrivers are a waste of money. Check the link in the picture below for the real thing, made by Wiha. Curiously, I have never encountered any thread locker on the two small slotted screw retaining the old aperture follower claw. Go figure.

Once the chromed flange is removed, the aperture ring can be slipped off. A few lenses attach the aperture ring with a short spring to the aperture actuator inside – like the 50mm f/1.4 Nikkor S and SC – but most do not. The 85mm Nikkor-H, illustrated below, retains the aperture ring to the aperture actuator with a small, radial slotted screw which has to be removed to allow the ring to come off. Once the chromed flange is removed, the aperture ring can be slipped off. It has to be replaced in the same orientation as when removed, so do not fiddle with aperture settings when the ring is off. If you replace the ring oriented incorrectly you will lose the ability to change apertures. Don’t panic. Remove and replace correctly. A little experimentation and you will get it. Force is never involved here. Some lenses (like the 50mm f/1.4) use a claw on the removable flange and this must mesh with the aperture tab inside the lens. Miss the tab and you lose aperture control. Others (like the 85mm f/1.8) use a small radial screw to lock the aperture ring in the correct position. The screw must be replaced so that it enters the thread in the lens barrel.

You will need a small, fresh (meaning not blunt from use) flat file with square edges. Triangular ones work well to if you have decent filing skills. I say ‘square edges’ because you really want the slot you file away to have perpendicular sides.

There are three aspects to making the slot.

• How deep it should be
• Where it should start – the left hand end
• Where it should stop – the right hand end

Depth is very easy. Remove the aperture claw (two small slotted screws). You want to file the slot down to a point where the screw holes just remain intact. You do not want to go down so deep that the screw holes are compromised. There’s no loss of function if you do but your work will look ugly.

The starting point means the left end of the slot when the lens is viewed from behind. This one is critical as it tells the camera the maximum aperture of the lens, and it changes with the maximum aperture of the lens you are working on. But there’s an easy way to avoid making a mistake here by the simple expedient of making the slot too short, then gradually working the end point back with subsequent applications of the flat file. Simply refit the aperture ring and the flange (you need only use a couple of the 4 or 5 screws you removed earlier to retain it, temporarily) and see if the LCD display on the camera’s body will reach maximum aperture (camera set to M or A Mode) when you turn the aperture collar. If not, remove and extend the slot a little more on the left hand side.

This table illustrates where the left end of the slot should be relative to the engraved apertures on the aperture ring of the lens which you have just removed.

Nikkor pre-Ai filing limit
for Ai conversion, maximum aperture.

The focal length of the lens is irrelevant. The maximum aperture is the dependent variable here.

All data in this table are based on lenses I have personally converted or were converted for me by John White. All converted lenses work correctly on Nikon D700, D2x and D3x bodies I own or owned. I can also attest that these data are correct for an F6 film body owned by a friend.

How to use this table:

• Determine the maximum aperture of your lens
• File down the rear of the aperture ring to the engraved aperture position in the ‘First try’ column, above
• Extend the length of the slot around the circumference for a length of approx. 1.5″
• Reinstall the aperture ring on the lens
• Fasten the chrome bayonet flange with two opposed screws
• Insert lens in camera and set to maximum aperture
• If maximum aperture is indicated in finder or on the LCD you are done on the left end of the slot; if the LCD shows apertures higher than dialed in on the lens, your left slot side needs to be extended
• Check you can get the finder or LCD to show the lens’s minimum aperture. Once it does you are done with the right end of the slot
• Reinstall and fasten, re-test.

When an entry in the table states + 0.5 it means you are filing to a half-stop to the left of the stop indicated. For example, ‘f/11 + 0.5’ means filing to a position midway between the f/11 and f/16 engraved aperture settings.

The identical data for f/4 and f/4.5 are not typographical errors. These are based on multiple observations of lenses I own.

Here’s my 105mm f/2.5 Nikkor pre-Ai lens with the slot machined (the focal length is irrelevant, only the maximum aperture matters) – the two claw-retaining screws were replaced after the machining was completed:

As you can see the slot extends at the left to a position between f/11 and f/16 (‘f/11 + 0.5’ in the table, above) and extends around the periphery of the aperture ring for approximately 1.5″.

The right hand end of the slot is not critical. The slot must have a certain minimum length but too long on the right hand end does not matter. The right hand stopping point for the slot is non-critical and needs only be far enough to the right to clear the feeler on the camera’s body with the lens at its minimum aperture. Typically 1.5″ around the circumference is fine. Make the slot too short and the lens cannot be stopped down to minimum aperture (or fitted at minimum aperture, for that matter) as the movement of the aperture collar will be constrained (or blocked) by the feeler on the camera’s body. Make it too long and no biggie. You cannot set the lens beyond its minimum aperture owing to the limit stop in the lens, so making too long a slot makes no difference. It just means you are doing too much work. The limit stop in the lens will make sure that the correct minimum aperture is not exceeded with the lens on the body.

Some paint the machined surface black. That serves only cosmetic needs, as the exposed alloy is not ‘seen’ by the film or sensor in the camera’s body.

The 20mm UD Nikkor: This is the odd man out. Rather than relieving an arc of alloy from the rear of the aperture ring, you must fit a small extension piece to effect contact with the aperture feeler on the body. I reverse and cut down the stock aperture claw to do this. Click here for details.

Proof of the pudding:

Here I illustrate the process of adding the Ai slot on my latest ‘pre-Ai’ lens acquisition, the magnificent 85mm f/1.8 Nikkor-H Auto optic, made in the early 1970s.

The alloy of the aperture ring is fairly soft and little time or effort are needed to do the work. I do not advise the use of power tools like a Dremel. They are far too aggressive and you run the risk of removing too much alloy. When testing the limit point of the left hand (the maximum aperture) end of the slot, by refitting the aperture ring and bayonet flange for fitting of the lens to the camera, I make sure to use compressed air to clear any remaining filings from the aperture ring. Getting these into the lens or camera will not make your day.

My first lens took me an hour (nervy-dervy!) but I can crack these out in about 15 minutes now, no stress.

Tools of the trade. Flat file, compressed air,
Wiha miniature screwdriver kit, Nikkor awaiting surgery.
Magnetic screwdrivers sometimes help, but Nikkors do not consistently use magnetized screws.

The screws and the bayonet flange removed.

Some lenses, like the 20mm f/3.5 UD and this 85mm f/1.8, have an aperture ring retaining screw. Remove it.

The aperture ring removed, ready for machining.

With the aperture ring removed, the spring spigot which engages
the detents inside the aperture ring, conferring ‘click stops’ on the action,
is accessible. If you wish to convert the lens for movie use – meaning no noisy
click stops – remove this spigot, retained by two screws.

15 minutes with a flat file and the machining is completed. The aperture ring is
first blasted with compressed air to make sure any trace of metal filings is removed.
The small aperture ring retaining screw is replaced followed by the flange.
The flange screw holes are not evenly spaced so the flange can only be replaced correctly.
Two flange retaining screws are used here to test the fit on the camera body.

When reinstalling the flange retaining screws tighten them in two stages in a criss-cross pattern, much as you would a car’s wheel. Tightening each down first time with massive torque risks distorting the flange, meaning the lens will not sit dead parallel to the body. I do not advocate the use of threadlocker and Nikon’s more recent MF lenses use none, bearing this out. A properly torqued screw will not come out.

The CPU can now be installed if required.

Factory Nikon Ai conversion kits: Nikon used to make ‘plug-and-play’ conversion kits, consisting of a replacement aperture ring. Most are long sold and it’s unlikely you will find the one you need. You can see which ring your lens needs in this database. If you can find them, they generally run $30-35 and invariably have the awful, non-matching diamond machined texture which will not match the look and feel of the original lens, which is why I avoid them. (My 24mm pre-Ai Nikkor came thus converted and it kills me every time I use it).

Claw replacement: When you are done you can replace the old aperture claw if you like the look or if it helps you key the lens to the body. It serves no other useful purpose. I replace the screws but leave the claw off. In the many lenses I have modified, these screws have always responded to a magnetic screwdriver which makes an almost impossible task easy. Hardware stores carry small gadgets to allow a screwdriver’s head to be magnetized or demagnetized. A few bucks and highly recommended. Be sure to demagnetize your screwdriver if you contemplate using it on sensitive electronic components.

Final touch – color mounting index for protection:

I add a spot of fluorescent orange paint between the fourth and fifth CPU pins (very little, to avoid run off into the pins) which I find greatly helps with indexing the lens when mounting it on the camera body. For those lenses which accept a simple glue-on CPU installation, this also lessens the chance of striking the edge of the fairly exposed CPU on the camera’s bayonet throat, something I once did, managing to damage the CPU in the process, necessitating replacement. Only Leica ever got this right with its red, raised plastic buttons on their lenses which permitted mounting by feel only. Later, Pentax with their M series did the same, though the plastic hemisphere was white and hence less visible. It’s so simple, it mystifies me why others have not done this. My paint dot does not help with feel, but it helps greatly when aligning the lens with the camera’s bayonet mount.

The collector’s dilemma: I have little time for gear collectors. If I do not use something I sell it. Nevertheless, there’s a body of believers out there who will tell you that machining classic Nikkors to make them work on modern bodies is some sort of sacrilege as it impugns the ‘collector value’ of the lens. Utter rot. Most Nikkors of all types were made by the tens and hundreds of thousands – not my idea of scarcity value. The only ‘rare’ lens I would not consider converting is the 13mm f/5.6 (one just sold for $100,000 at auction – only 51 were made!). The other rarity, the $4,000 58mm/1.2 Noct came in Ai and Ai-S versions only, so no conversion is needed.