Chipping a Nikkor - A Walk-through

[Bjørn Rørslett]

[Posted 24 March 2010:  Article and image copyright Bjørn Rørslett ]

This is a short illustration of the process of CPU-converting ("chipping") a Nikkor lens. The example is the 18 mm f/3.5 Nikkor AIS. The 18 mm is not a very difficult lens to modify, but it does demonstrate some of the minor niggles often encountered when you set out to implant a CPU. Let's remember that these lenses never were designated to receive an internal CPU, so that it usually is possible might come as a surprise.

The first step is assembling the pieces.

Here is the lens, together with the CPU print, 5 signal pins, and the contact block.



Assembling the contact block is not rocket science, more like playing with Lego bricks.

[Bjørn Rørslett]

Here I have removed the rear light baffle (always black, it usually is secured by 3 small screws). We're looking straight into the internals of the lens, and immediately note two problem areas.

1. There is not sufficient space for the contact block, so material must be removed from inside of the bayonet mount.




2. The aperture coupling slider has two screws the heads of which potentially will interfere with the print. These heads need to be sanded down to be flush against the surface of the slider. An easy task suitable for any keen Dremel user.

[Bjørn Rørslett]

Now, time to drill the two holes for the retaining screws for the contact block. An 1.25(-1.5) mm bit is suitable. I frequently use a jig so it's easy to drill the holes in a correct position. Pin number 4 of the block has to line up precisely with the centre of the topmost bayonet screw (the one adjacent the the index mark on the lens). This correct position will be evident on the next pictures.

After drilling, the holes must be countersunk so the screws come flush to the surface. Otherwise, the lens will jam when it is attached to the camera. The material of the mount rim must be thinned on the inside, if necessary, to keep its thickness to 0.8-0.9 mm in the area where the contact block is to be inserted. Older lenses may have the required thickness already, whilst most AIS lenses need to be trimmed. Only do the removal of material on the inside of the mount.



Here is the contact block in the correct orientation, and it's easy to see that we now have plenty of space for it in the bayonet mount. The mount has been trimmed on the inside to give the appropriate thickness (0.8-0.9mm) as well.



The contact block, with the CPU print, is put into its place and secured by the retaining screws.

[Bjørn Rørslett]

Now, the bayonet is put back onto the lens and the print dressed so it doesn't snag on any moving parts inside. It's beneficial to make a paper dummy if the placement of the print is deemed difficult. The print is fairly delicate and should not be bent back and forth excessively.



Make a cut-out in the rear light baffle to accomodate the contact block, and fasten the baffle with at least 2 screws. Sometimes, the locking screw of the mount is one of these and this has always to be present. In the case of the 18 mm, we managed to save all three holes, so all three screws can be used (and the locking screw didn't need to be touched at all). This will vary from lens to lens, though.

[Bjørn Rørslett]

Here is the completed and rejuvenated 18 mm f/3.5 S Ai-P lens. It is now compatible with each and every Nikon "F" mount camera ever made, and will meter on all of them. You can dial in the apertures from the camera, or use the aperture ring (if the camera allows this operational mode). All metering modes are available, and you can use exposure program modes A, M, P, and S.

(by the way, all illustrations are hand-held snapshots with the D3S/125 APO-Lanthar, 12800 ISO).



That's all, folks !!

(note to self: dust off lens before taking the product shots. The CV 125 APO is darned sharp so all dust specks are mercilessly shown. And let Erik drill the mounting holes, he does a much neater job than me )

[Bjørn Rørslett]

It is worth noting that most medium or longer focal length lenses don't need the cut-out in the bayonet mount itself. The case of the 18 mm exemplifies the typical shape of the bayonet for Nikkors such as 20, 24, 28 mm etc. These shorter lenses also requires a short (5-pin) contact block from an AF/AFD/IX-Nikkor or third-party lens, whilst most other lenses are well served by any long block (7-10 pins; of which only 5 are used) found on AFS lenses.

A cut-out of the rear light baffle, on the other hand, always is necessary, otherwise there will be no space in which to seat the contact block.

The typical layout of the CPU print is the "L" shape shown in this walk-through example. The "L" is a general-purpose form factor that can be used with most lenses.  There are other specialised shapes as well, and they can  facilitate the modification of some otherwise tricky lenses. The 35/1.4, 16/3.5, and 85/1.4, to name some, can only be modified with CPU print with alternate print layouts.

The time necessary for the modification ranges from 15 minutes for the easiest ones to a whole work day for the really difficult challenges.

[Bjørn Rørslett]

In answer to the following

" I am just curious how would this chipping allow the dial on the camera to set the aperture? "

It does, believe me. The technical explanation is rather long, so I'll skip it. But there is no magic involved.

You  can also realise this is happening since the "P" and "S" modes are available and fully functional; they both work by changing the actual aperture on the lens.

The CPU chips can be programmed to work both in "P" ( the AI-P designation originally used by Nikon for 45/2.8, 500/4, and 1200-1700/5.6-8)  or "G" mode, which encompass virtually all AF lenses made the last 30 years. All AF and AFS lenses also are AI-S, meaning they have a linear aperture mechanism. To complicate the matter further, Ai-P are also Ai-S. Nikon never made their nomenclature crystalline clear. The problem is mainly the earlier classes of lenses, pre-AI (before 1977) or AI (1977-ca. 1981). The pre-AI class can be converted to AI (often designated AI'd) so I treat both under the AI category.

Initially, there is no difference as both kinds will allow the aperture to be set from the camera side and on the surface, all exposure programs (A, M, S, P) will be functional if a CPU is added to them. However, although the read-out metering value is correct in both cases, the actual exposure can differ substantially. This is because the camera calculates the amount to stop down the lens (actually, the travel of the stop-down plunger inside the camera) based on a linear model, but the aperture movement of the AI/AI-modied Nikkors is non-linear. So for example, the meter reading says f/8, 1/250 (and these are the settings you get with an AIS lens), the actual settings at the time of the exposure might be f/11, 1/250 for the AI due to the non-linearity. Thus, one easily gets underexposure with the AI kind.

However, if the aperture is set directly on the lens, using the aperture ring, both meter reading and the actual exposure will coincide, and there is no difference between AI/AI'd or AIS lenses.
 
The contact block acts as a control interface and simply relays the electric signals from the CPU print. In addition, there is the CPU, that does the real stuff.

If you employ a factory CPU unit scavenged from a broken lens, its signature cannot be changed. In some cases, using a unit from a zoom lens, you can fake a lens signature by short-circuiting the section that reports the focal length as this will alter the maximum aperture sent to the camera. However, such factory prints are large and bulky and very few lenses have the required space to host them. The customised CPUs shown in this thread are always the better and easier approach.

[richardHaw]

Hello, is this the daisy chip that I have been reading a lot about? Thanks!

[Bjørn Rørslett]

Naah. You are thinking of the 'Dandelion' chip that is a low quality product and quite likely to get damaged in use. (Seen some of those failures over the years).

My CPU chips are custom made and hand assembled. Each is carefully tested before shipping. They are programmed to the exact specifications of the hosting lens, and that includes EXIF showing the correct zooming range for a zoom lens. Depending on the actual lens to be used with, I program them either as "P" or "G" mode of operation. I recommend they be mounted with factory contact blocks so as to provide maximum reliability and robustness.

[richardHaw]

Naah. You are thinking of the 'Dandelion' chip that is a low quality product and quite likely to get damaged in use. (Seen some of those failures over the years).

My CPU chips are custom made and hand assembled. Each is carefully tested before shipping. They are programmed to the exact specifications of the hosting lens, and that includes EXIF showing the correct zooming range for a zoom lens. Depending on the actual lens to be used with, I program them either as "P" or "G" mode of operation. I recommend they be mounted with factory contact blocks so as to provide maximum reliability and robustness.

Thank you! i have always wanted to chip some of my lenses but i just cannot source cheap ones here in Japan 

[Erik Lund]

I don't think Bjørn's CPU's are expensive, especially since they are same quality/identical to Nikon parts
They come highly recommended!

Dandelion CPU's are very fragile, and really the only application where I can say they work on a heavily used lens is for the 50mm 1.2 Ais, where it can be integrated into the F-Mount.

[richardHaw]

Thanks! yes, dandelion! all i know is that it was named after a flower starting with the letter d.

i am hoping to chip my lenses without any modification or at best minimal...

now, since nobody is offering this service here in japan, maybe i can start offering the service myself...

[Bjørn Rørslett]

I am luckily in a position where I can order Nikon parts such as the factory contact assemblies and their specialised screws. Or in principle any other part in their vast inventory. The pricing policy of such items is 'strange' to put it mildly and most stuff can be frightfully expensive, so if you build an entire new camera from spare parts it'll probably cost 10x or more the street price. The contact block parts come in a middle position and getting the blocks as factory spares can be cheaper than buying junked lenses to scavenge parts from. Plus most junk lenses are cheap consumer 'kit' zooms that use inferior components and their contact assemblies are often plastic-based, not metal as the genuine stuff.

[richardHaw]

not sure if i can do that myself! 
in the past 10 years of my career, i actually do more programming and technical stuff than creative things but i still lack the practical electrical engineering skills do produce these chips! (apart from photo-etching,etc)

how much do these chips cost? the dandelion ones cost around JPY3000 each I think that is way beyond my budget 

[Bjørn Rørslett]

I usually sell them for USD 45-50* each plus shipping and handling. Thus the unit price depends on the number ordered.* recent prices have been higher as parts cost apparently increase exponentially over time