Stefan Gruenendahl , Feb 01, 2012; 01:08 p.m.
I notched the aperture ring and glued a metering chip to a micro-nikkor 55/3.5 (S/N 199392, silver barrel) so that it can be used on a D70.
A few notes from the conversion:
'AI' notch: The aperture ring can be unscrewed after removing the single screw visible from the outside that links it to the inner aperture gear. It is not necessary to remove the lens mount. As the aperture moves a bit with focus distance on this lens (CRC correction, I assume), it is important to reassemble with both aperture and focus ring at the same setting as during disassembly. Otherwise the slot for the aperture ring screw might not be accessible.
A notch needs to be filed into the aperture ring to accommodate the '8 o'clock tab' on the D70. This tab is a switch signaling that the lens is set to its minimum aperture. I used an AI lens to copy the general features of the notch. The exact begin and end of the notch were marked while holding the ring against the D70 mount.
The minimum aperture signaling end of the notch needs to just move the tab when the aperture ring is rotated to minimum aperture. The other end of the notch needs to be cut far enough to clear the tab when mounting and dismounting the lens. On this lens the minimum width extends a few millimeters beyond the aperture ring screw location. (For mounting clearance at all aperture settings a wider notch would be needed.) I filed the notch to the minimum depth needed to clear the tab.
Metering chip: I obtained a metering chip ('AF Confirm Metering Emulator Chip') from an ebay source ('tagotech' in Singapore: $50 for two chips, including shipping; shipping took 30 days). I selected the 55/3.5 lens for my first conversion because no mechanical modifications were needed to mount the chip itself.
The chip, a black plastic ring segment housing the circuitry and the spring loaded contacts, glues right into place. I copied the chip location from an AF lens and glued the chip to the inner baffle with cyanoacrylate 'super glue'. For better visibility during gluing I did not use the template provided with the chips. The chip sits tight on the inner baffle and touches the mount flange on the inside edge, but has a slight gap at its outside edge (see picture).
I managed to get some cyanoacrylate on the contact side of the chip; I was able to remove most of it with a cotton swab soaked in isopropyl alcohol. One of the contacts got stuck in a slightly retracted position and lost maybe a third of its travel. The tolerances apparently are big enough that the chip makes good contact and
achieved full functionality after a gentle 'sanding' of the contacts with an Arkansas stone, to remove traces of the glue on the contact surface.
The chip programmed as described in the instructions. After each step (max. aperture; focal length; min. aperture) powercycling the camera got the chip ready for the next sequence.
The chip seems to work flawlessly, and makes using the lens much more convenient and faster than the old trial and error method using the histogram. In 'P' mode the rear command wheel cycles through the different time/aperture combinations. Since the aperture gear in this lens is not linear (not an AIS lens) there are slight variations in exposure between the different combinations. I haven't been able to achieve correct flash exposure yet. The last picture was taken with the lens after the conversion; I left the image exposure as it came out of the camera.
Ariel S 
, Feb 01, 2012; 01:34 p.m.
Interesting experiment, but that seems like a pretty high time/effort investment to make for one lens. Considering the hassle, wouldn't it have just been easier to sell your D70 and pick up a D200, which is about the same size but better in every way, and will meter with non-chipped lenses? They're selling for pennies these days. Heck, I got mine by buying a trashed body for $150 and getting it repaired by Nikon for $250. Brand new camera, as far as I'm concerned, and good for the next 100,000+ photos.
Stefan Gruenendahl , Feb 01, 2012; 02:29 p.m.
Interesting point about the D200. Upgrading the 55/3.5 took actually less time then writing up the note. Maybe it has a bit too much detail? :-)
I thought about upgrading to a camera with better available light capability, but that's in the far future, because to make it worthwhile (>2 stops better than my D70) I have concluded I'd need at least the high ISO performance of a D7000. It will be a while before those become affordable. And even then having to pick each non-chipped lens manually from a list stored in the camera seems like an unnecessary hassle.
Short term I am planning to upgrade my collection of manual Nikkors one by one, starting with the easier ones (105/4 micro, 35/1.4 AIS ?), and hopefully eventually figuring out a way to chip the 85/1.4 and the 50/1.2. The latter one might require sanding down the chip.
Mihai Ciuca , Feb 01, 2012; 03:15 p.m.
Hi Stefan,
Thanks for sharing your experience. Please keep us informed with your plan to chip your other lenses. I'm also intending to start chip my MF Nikkors... I use a D7000 but I want to have the right EXIF info without to change the menu every time... I'm getting old :) and merely forget to do the change in menu...
Stefan Gruenendahl , Feb 01, 2012; 11:21 p.m.
Chip & lens dimensions for 85/1.4, 35/1.4 & 50/1.2
Chip & lens dimensions:
The metering chip is 2.75mm thick (radially), plus 1.1mm for the contacts when fully extended. It is 3.20mm high, and at the outer radius 19.2mm wide. On an 18-70 kit lens the contact mount has the same thickness as the aftermarket chip, and is mounted with a setback of 2.15mm from the outer edge of the bayonet. Hence mounting the chip needs a total of 4.9mm from the outer edge of the bayonet to the inside radius of the chip.
Shaving the chip: sanding down to 2.6mm is ok; at 2.5mm the back side of the metal contacts starts showing through.
Radial lens clearances, from outer edge of bayonet:
85/1.4: 2.0mm to outer radius of aluminum shield; 3.25mm to inner radius of aluminum shield; 4.65mm to rear element lens mount. So one would need to cut the aluminum shield and to reduce the chip thickness or the clearance by 0.25mm.
35/1.4: 2.15mm to outer radius of blackened shield; 4.15mm to inner radius of blackened shield; 4.60mm to rear element lens mount. Need to cut the blackened shield and to reduce the chip thickness or the clearance by 0.3mm. (If you look down the lens barrel the rear element lens mount seems to extend further outward radially, but this thicker part does not come up above the bayonet flange, where the chip would sit.)
50/1.2: 2.15mm to outer radius of aluminum shield; 3.50mm to outer radius of black lens shroud; 4.54mm to actual rear element lens mount. The black lens shroud seems to directly contact the lens mount; not clear that it can be removed and cut. Cutting only the aluminum shield one would need to reduce the chip thickness or the clearance by 1.4mm.
Mihai Ciuca , Feb 02, 2012; 05:37 a.m.
I wonder if the chip you are using is the Dandelion made by someone in Belarus... If not I wonder if the Dandelion chip has the same dimension like these you ordered from Singapore. Please keep posting your findings. I am interested to chip at least these lenses:
28/2 AI-S
35/2 AI-S
50/1.2 AI-S
58/1.2 AI-S
85/2 AI-S
105/2.5 AI-S
135/2.8 AI-S
Bjorn Rorslett , Feb 02, 2012; 06:26 a.m.
A general point of warning is that the benefit of "chipping" non-AiS lenses is questionable unless (a) you use the lens on a camera which can be set to employ the aperture ring on the lens for metering or (b) you only need the correct EXIF and don't care about the actual exposure. Do remember the non-linear operation of the stopping down action can *only* be properly accounted for when the aperture is set directly on the lens itself. So, the metering with such lenses *will* be correct, but the actual exposure is off by a wide margin (unless the safe action described is followed).
I have CPU-modified nearly all lenses on the list by Mihai (50/1.2 still remains), but not used the Dandelions as they are very fragile, easily breaks off, and most of the units I've purchased have been DOA anyway (indicative of poor quality control). Thus I gave up on this product. I'm now using original Nikon contact blocks with the proper sturdy metal backing and retaining screws for my chips.
Mihai Ciuca , Feb 02, 2012; 07:58 a.m.
Bjorn, my reason is exclusively to have the right info in the EXIF, because I have a good number of MF lenses and I may expand the kit in the future. I merely forget to change them in the menu and this creates a big mess.
I'm disappointed to hear about the weak part of Dandelion chips. In the same time I am very interested if you can share the info about the source and the cost of the original Nikon contact blocks. This may be very helpful to me and probably to others in the same boat.
Stefan Gruenendahl , Feb 02, 2012; 08:46 a.m.
Thanks for the information about the Dandelion chips. I was (and still am) afraid about the DOA part. I was quite relieved when the first chip worked, despite my gluing misadventure. The chips I bought look like the picture on the Dandelion website http://filmprocess.ru/nikon_spec_en.htm, and have the same programming sequence, but I have no idea where and by whom they are manufactured. The contact block in my Nikon 18-70/3.5-4.5G does not have screws. Bjørn, are you able to order these parts (contacts, chips) directly from Nikon, or do you have a helpful dealer?
Non-AiS exposure using camera control for the 55/3.5 micro-Nikkor: with my very limited testing I have seen very close results from f3.5 to f11, maybe 25% underexposure using the dial at f22, almost correct using the lens ring, and both methods 50% underexposed at f32.
Flash exposure also seems to work; I use flash very little, so there might have been some operator error with my initial failure.
Bjorn Rorslett , Feb 02, 2012; 09:08 a.m.
I have a direct source of parts (contact blocks) from an authorised Nikon repair facility. The custom chips are sourced elsewhere. They can be programmed either as type "P" or "G" depending on the actual lens/bellows/extension etc. and how it is to be used. You can fine-tune metering accuracy by specifying the exit pupil as well.
One might of course order Nikon chips, but they are not programmable so not very helpful in this case (unless you only care about the aperture range of course, then matches can be found). They are also way too big to fit inside most candidate hosting lenses.
Here is a photo of two Noct-Nikkors 58/1.2 modified with CPU. Not a task for the faint of heart, true, but doable, yes.
Overall I may have more than 100 CPU-modified Nikkors. It's always funny when I meet Nikon reps (even in the Japanese Nikon factories), ask about such a conversion service, am being told it is impossible, then show one of my converted lenses :)
CPU-modified Noct-Nikkors
