Geek Articles

Disassembly of the Canon RF 50mm f1.2L

Published December 12, 2018

It is the best of lenses, it is the worst of lenses, it is the lens of perfection, it is the lens of complexity, it is the lens of new technology, it may be the lens of non-repairability. 

A little while ago, Canon released the computer generated MTF charts for the Canon RF 50mm f/1.2 lens. Despite my ambivalence about the Canon EOS-R camera, this lens, and this lens alone, made me lust after the system. (For those of you keeping score, I have managed to go almost 3 months now without buying into any mirrorless system. I am very proud of myself.)

So when we got them in stock I managed to get one shuttled over to the repair department so we could take a look inside – as long as we promised to have it back working fine and ready to rent in a couple of hours. We’ve done this a hundred times or more so we weren’t worried about it. (Narrator: “But they should have been worried about it.”)

Usually, I start tear down posts with a joke about ‘those of you who are following along by disassembling your own lens at home’. Well, no joke today; this is not a home disassembly project. I’m not really sure it’s even a Lensrentals disassembly project. But we got out tools out and boldly went where we probably shouldn’t have gone., 2018


Starting with a False Start

Aaron decided this lens looked like it would best be opened from the front, so the makeup ring came off., 2018


Which let us see a very odd screw arrangement in the front, with multiple screws off-center from the slots and angled through some plastic wedges. I’ll be honest, we had never seen an arrangement quite like this before but we figured it was some kind of binary code for ‘Do Not Enter’, so we didn’t enter., 2018


We beat a hasty retreat, flipped that thing over, and started disassembling from the back. Things looked fairly routine back here, except that Canon used some side screws to hold the rear baffle in place (many other manufacturers do that, but Canon usually has a snap-in rear baffle).  We’ll fondly look back on this part of the disassembly, when things looked fairly routine. Good times.

Anyway, a few screw removals and the rear baffle comes out. Then we remove a set of nice, heavy-duty screws to remove the bayonet mount., 2018, 2018


The smaller screws in the image below are the ones attaching the electronic connector and baffle to the bayonet mount, the larger screws attach the bayonet to the lens. As always with Canon lenses, there’s that feeling that strength is engineered in: larger and longer screws; and plenty of them., 2018


Something we haven’t seen often in Canon lenses, there’s a grounding or heat transfer tape coming from inside the lens back to the bayonet mount., 2018


After lifting the tape up, we can remove the rear spacer., 2018


The spacer is labeled for thickness so different spacer thickness is probably used instead of shimming for infinity focus. Again, this is a fairly routine Canon method., 2018


Something else we don’t usually see in Canon lenses, there’s some electrical shielding above the PCB. Whether this is because the RF lens mount is closer to the sensor or because of the additional electronics, or both, I don’t know., 2018


Looking at the PCB does seem to confirm that there are more electrons running around in this lens, though. The board is pretty densely packed, there’s a lot of flex connectors, and Holy Nikon, Batman, there’s some wires going into the PCB!!!!! Since Canon strongly prefers flex connections, we assume there’s more current going through here than the legacy Canon lenses we’re used to. (It could also be to lower resistance over a longer wiring run.), 2018


Pulling up the piece of tape shows that the wires are going into a plug, not soldered to the PCB, which makes us happier because soldering and desoldering is time-consuming. And you can burn your fingers and such., 2018


After disconnecting all the flexes, the PCB comes out in the usual fashion. Then there’s a little twist we see every so often – one flex wraps around the rear inner barrel and is stuck down with tape. The PCB area is packed tight and this flex has gone around to pass down on the far side along the barrel., 2018


There’s also a nice weather seal ring inside here (although given it’s location, probably more about dust than moisture). I should mention that we saw good foam at each of the barrel joints and the usual rubber seal under the bayonet mount., 2018


One more set of screws is removed and then the rear outer barrel can slide off., 2018


Underneath is the large version of Canon’s ring USM motor (this is the same motor that is in the new Canon 400mm f/2.8 IS III – hint for those of you wondering what the next tear down will be). The Canon RF 50mm f/1.2 is said to be a quick focusing lens. Given the amount of glass that moves with focusing (we’ll get to that in a minute) a big motor is probably necessary. There’s an interesting tensioning spring on the side of the motor, something we’ve not seen before. At this point, we have no idea what it is for but assume we’ll figure it out later. (Narrator: “But they would not figure it out later.”), 2018


Inside the rear barrel that electrical connecting tape that we took off of the PCB comes to an end. It doesn’t have an obvious connection or sit on top of any obvious electronics within the lens, so our best guess is it’s for static discharge. Or perhaps it’s decorative. I’d love to think of a bunch of Canon engineers, sitting around drinking sake, and one goes ‘Hey, we’ve got lots of electrical discharge tape. Let’s put a piece in the rear barrel that goes nowhere. Roger and Aaron will take it apart and have no idea what it’s for’. Mission accomplished, Canonista., 2018


Here’s another view of the lens. Notice there is more sealing felt between the motor and outer barrel (red line). You can see a multi-flex connector on the left, and in the slot of the focus ring assembly, some position sensors are barely visible (green line)., 2018


Another set of 6 screws holds the focus motor assembly in place. After removing those the assembly comes right off.  Because someone will ask, gloves are needed to keep finger oils off certain electronics. They aren’t needed for taking off barrels and such., 2018


Underneath the flex enlargements are a couple of the focus ring position sensors. In case anyone is not aware, this is a focus-by-wire lens, but these days that’s at least (and perhaps more) accurate than mechanical focus. This isn’t your grandfather’s electronic focus., 2018


There are more absolute position/position change sensors within the focus assembly. We put off disassembling the focus motor assembly since we had promised this lens back in a couple of hours. (Narrator: “Yeah, you got it. They did not come back to disassemble the focus motor.”), 2018


At the bottom of the assembly is a very large, robust focusing key., 2018


This inserts into a slot in the mid barrel. The focus motor moves the key, which moves the slot, which turns the focusing barrel to adjust focus., 2018


Let’s take a look at the remaining lens core, now that we have the AF motor assembly off. You can see the rear group sitting on top. (They’re out of focus, but you can see one of the pairs of screws that attach it to the lens.) Then there are slots in the barrel with lots of air inside there. If you look through the slots, you can see the back of the focusing group down at the bottom. In other focus positions, this group slides back into the air, closer to the rear element., 2018


One other new thing: you can see a number of springs attached to the inner barrel and disappearing into the front barrel. At this point, we have no idea what those are doing. Our usual thought in this situation is that they’re holding tension on a plate containing a zillion ball bearings just waiting for their chance to escape.

One more look at just how empty most of the lens is when the focusing group is moved forward., 2018


For orientation purposes, let’s take a look at the lens diagram. Over on the right side, you can see the 4 elements making up that rear group bolted on the back of the lens. Everything else, all the elements on the left side of the diagram, move when focusing. So now we see why, since Canon wanted fast focusing, they slapped the big-boy USM motor in this lens. That’s a lot of glass to move around.

Courtesy, Canon USA


Once we’d gotten to this point, we were starting to feel a little lost with the Canon RF 50mm f/1.2L. Things were vaguely familiar, but not quite like any other Canon lens we’d taken apart before.

Looking at the other side of the inner barrel, we get another look at the Springs of Unknown Purpose, the Shielded Flexes of Forevermore, the Borrowed Wires of Nikon, and another Ring-Around-the-Barrel Tape (already pulled back in the image)., 2018, 2018


With the flexes out of the way, let’s move that big front focusing group down into the air space where we can see it. One thing we were impressed with is how much movement occurs. A long focusing throw like this should mean very accurate focusing., 2018


The second thing to note is there are some really robust eccentric adjustment collars in the back of this group. As Canon does these days, they adjust them at the factory and then put silicone glue in place to make sure they don’t move. Or, if you want to adjust them yourselves, make sure it’s a pain to get the silicone glue out before doing the adjustment. If you look back up at the lens diagram, though, you’ll see this adjustment is for the group containing a ground aspheric element. Those are very delicate to adjust; I can understand why Canon doesn’t want it touched., 2018


There are 6 eccentrics at this level, three for tilt, three for centering. Each eccentric is paired with a white nylon post as you see above.

It probably works in this way: the post holds the lens group in place but has some play that the eccentric is used to fine tune, then the glue is applied. Until we unglue one and learn how to make the adjustments ourselves that’s just an educated guess, though, and that task will wait for another day. The takeaway, though, is Canon continues to have more potential optical corrections than almost anyone, and they’re making the adjustment mechanisms both more accurate and more robust.

At this point, we had decided that today was not the day to pull out the glue and learn eccentric adjustments. Removing the rear group really wouldn’t reveal any new information. So we took another look at the front ring (Remember, the front? That’s where we started this.).

We could now look from the back and tell which screws would remove the front barrel. How is this done, you ask? It’s pretty simple; you turn a screw in the front, look at what’s coming out of the back, and go ‘no, not that one’ half-a-dozen times., 2018


Finally, we removed the six long screws that let us slide the front barrel off., 2018


Inside the front barrel, you can see several things. First, the Flexes of Forevermore are traveling up inside the front barrel between inner and outer plastic rings. That makes sense, if the Control Ring can control almost anything, it has to be connected to almost everything. The Control Ring is a feature I really like in the short time I’ve used RF lenses. I hadn’t really thought about how much engineering goes into it. I had thought this feature might be a ‘trial balloon’ of sorts, but I’m now realizing this was a huge engineering commitment on the part of Canon., 2018


You can see in the image above that we can now see the underside screws that would let us disassemble this ring further. There will obviously be a bunch of position sensors underneath there, but we’d already answered one of our big questions: can we replace the filter ring (a part that breaks fairly frequently) easily? It looks like the answer is no, and that makes us sad; replacing the filter ring will mean a major disassembly (if the filter ring is sold as a part) or replacing the entire front barrel assembly (which may be the case). Either way, it’s not the quick and inexpensive repair we would hope for.

With all the outer barrels and electronics removed, we now come to the pure core of optical goodness that is the Canon RF 50mm f/1.2 L lens. Like a shaved cat, it’s always kind of shocking how small the core of the thing is., 2018


As we’ve seen throughout this lens, there is another thick foam weather resistant gasket around the front element., 2018


Time pressures (and fear, there was also fear involved) kept us from disassembling the inner focusing barrel, but there were a couple of interesting things to see. The barrel itself moves on paired rollers, which you can see to some degree in the out-of-focus image below. Note also the springs and the white bands or collars around the focusing ring, front, and back. There appear to be sets of small ball bearings under each band with the springs supplying tension across them to give a smooth resistance and nice feel to the focusing ring. It’s complex, but I will say the focusing throw is as smooth as could be, so apparently, it works well., 2018


And there is another set of eccentric adjustment collars way down deep (you can just barely see the screw head, you can’t make out the collar). So we have adjustable groups on both sides of the aperture, as expected in a Canon lens., 2018


So, What Did We Learn Today?

I wish we had enough time (and the guts) to disassemble further into the front internal and external rings, but we weren’t pushing our luck today. We already suffer from PBBA-SD (Post Ball Bearing Accident Stress Disorder) and have experienced a couple of episodes of TSAKI (touched sensor and killed it), so we’re a little nervous about going further in this lens. At the time of this disassembly, it was not replaceable. We’ll have to come back and do that at a later time; when it might be relatively OK if we mess a lens up.

What we did see, though, is the R lenses are not only entirely new optics, they are also largely new electrical and mechanical systems. There are a lot of different things in here that we haven’t seen in any Canon EF lenses. Some of them we should have expected, like the increased electronics going to the control ring. Others we don’t really understand yet, like the tension spring in the ring USM motor or the increased electrical shielding.

Some days we show off our amazing disassembly skills doing tear-downs. Some days, we get overconfident, decide we can do anything in two hours, look inside, get overwhelmed, and beat a fairly hasty retreat. This was one of those latter days. We’ll be back on a day when there are plenty in stock and we can set aside 4 or 5 hours to explore.

In the meantime, we’ll have to settle for seeing the same ‘best in field’ optical adjustability we expect in Canon lenses. We also saw lots of new stuff we don’t completely understand yet and a level of complexity we weren’t expecting. Complexity and our lack of familiarity with the lens make things seem a bit more chaotic than they really are, but certainly, this isn’t quite as modular as most new Canon lenses are.

But one thing that is very clear: the RF lenses contain some new technology they haven’t used before. There’s a lot of engineering that’s gone into these. Things are different inside here. As we’ll see in the next teardown we do, some of that is carrying over to at least some EF lenses. What does this mean? It means Canon has invested very heavily into developing the lenses of the R system. This level of engineering didn’t all happen in the last year, they’ve been working on this for quite a while.

They step back so that they may leap further.  African proverb


Submitted most humbly by

Roger Cicala and Aaron Closz

December, 2018

Author: Roger Cicala

I’m Roger and I am the founder of Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in Geek Articles
  • Not THAT Ross Cameron

    “Like a shaved cat…” ROFLMAO!
    Many thanks for the peak behind the curtain. Looking forward to a similar tear-down of a new Nikon lens – there’s hope that they’ve finally run out of sticky tape.
    Wondering what to call such an ‘initial’ tear-down by LR? A teaser/skimpy? A reccy (reconnaissance) or a peak? Maybe even a modest one?

  • AE-1Burnham

    A comment to add no value, only praise:
    Your humour (@roger_cicala:disqus ), not-so-deeply-buried within technical descriptions, is amazing! A few laudable examples: “the Springs of Unknown Purpose, the Shielded Flexes of Forevermore, the Borrowed Wires of Nikon, and another Ring-Around-the-Barrel Tape”; “Like a shaved cat, it’s always kind of shocking how small the core of the thing is”; and “PBBA-SD (Post Ball Bearing Accident Stress Disorder)”.

    Thank you to the guys at Lens Rentals. As always, waiting patiently for the next deep dive into this universe!

  • Class A

    Great disassembly and write up! I’m very proud of you that you have withstood the temptation to go mirrorless as well! Be kind to your eyes. Treat them with optical viewfinders only. You know the “E” in EVF is for “EVIL”. 🙂 (Joking but not joking.)

  • dsut4393

    That mystery spring wouldn’t be some sort of tuned mass damping to compensate for the torque required to move the larger than usual focusing group would it? Or even simply to reduce nasty ‘clunk’ feelings when the focus movements start/stop?

  • Carleton Foxx

    You’re such a trendsetter! Did you notice in the DC Watch article that your Japanese doppelgängers have built Son of OLAF (which they named OPTIA)?

  • BigEater

    You could create a fine side business if you started selling those amazing looking tweezers. I’ve never seen anything like them in the makeup aisle at my drugstore…..or anywhere for that matter.

  • Stefanie Daniella

    Amazing Canon’s EOS R’s new RF Lenses allow lots of room for FAST (higher-powered) high precision AF as its bigger Super-Tele-Lenses newer DUAL-SPEEDMODE siblings:

    EF70-200mm F2.8L IS III USM Zoom lens (released just before EOS R)

    EF400mm F2.8L IS III USM Prime lens (featured with EOS R)
    EF600mm F4L IS III USM lens (featured with EOS R)
    The two new tele-primes already featured on EOS R, most missed, was featured separately, of course, on 1DXMkII. (If you missed it, too bad!)

    EOS R LiveView AF has already been known to run 40x faster than on EOS 5DMkIV using newer DUAL-SPEEDMODE EF Lenses. (mentioned above)
    There have been 1DXMkII users also discovering EOS R can AF faster than 1DXMkII on their own EF Lenses (even if their lenses are not newer-generation DUAL-SPEEDMODE Lenses)

    RF Lenses not meant for non-EOS R bodies; so obviously, AF comparisons between EOS R vs 5DMkIV bodies were done on newer DUAL-SPEEDMODE EF Lenses.

    Right now, no other FF Mirrorless camera + their OEM Lenses have AF that can keep up, and rely heavily on “slider-mechanisms” to accommodate typically annoying “hunt-back-n-forth” AF.

    Rotational+Slider combination is far more AF Done-Once-Just-Right (no hunt-shenanigans waffling of slider-only AF)

    Others (Sony+Nikon) AF often need to turn Aperture Stop-Down (ASD-AF) mode WYSIWYG-LV “OFF” just to ensure AF works better “wide open aperture” (WOA-AF) as light gets too little for ASD-WYSIWYG-LV.

    Fortunately, Canon WYSIWYG-LV AF is ExpSimLV AF (WYSIWYG-LV AF) = can simulate EVERY f-user choice (F1-F90: EOS R) optimal for AF, even down to F11 under lessening light.

    Others (Sony Alphas + Nikon Z’s) lack ExpSimLV AF, are either:
    ?non-optimal ASD-Only-AF (WYSIWYG Live-View) “ON”
    ?non-optimal WOA-Only-AF (No WYSIWYG Live-View) “OFF”
    Those shooting Sony/Nikon mirrorless used to shooting F5-F8 in lessening light, finding AF totally “halted”, are told to turn ASD-AF WYSIWYG-LV “OFF”, where only WOA-AF will work (=No WYSIWYG Live-View)

    whilst folks in USA are slowest to realize benefits of Canon EOS R + ExpSimLV AF … the rest of the world GLOBALLY, are already far more aware of Canon AF superiority; even if they don’t fully understand how or why ExpSimLV + AF are related: they go hand in hand, especially for stills shooting.

  • Yes, very solidly built, solid metal core inside. That being said, I’m really a little concerned about repair costs. The front element is an aspheric, meaning very expensive if you scratch it. The filter / hood barrel doesn’t look to be easy to replace and it may be that you have replace the entire front barrel, which would be very pricey. If I buy one, I’ll be putting a filter on and being careful with it.

  • Well, it’s already somewhere in CA, so no, it won’t. Sadly, the ones in Brentwood haven’t had the benefit of internal inspection yet. 🙂

  • However, the AF of EOS R is soo much better with large apertures and moving subjects, it’s not even funny.

    that’s what I see too. I mostly keep the EOS R in the face priority mode, and it works very well. Occasionally it misses focus on the closest eye though, even if my subjects don’t move much (AI Servo). I hope they will improve the eye AF in the next R model or maybe via a firmware update. I’m not a 50mm shooter, I prefer ether longer or shorter glass, but I’m very curious to try this 50 out. Who knows, maybe I’ll even buy it.

  • Lars Kvinge

    I don’t do any video, so I don’t take that part into account. What I miss about the 1DXII is the optical viewfinder, the handling, the ergonomics and the overall responsiveness. The overall joy and shooting experience with the 1DXII is fantastic, and hard to beat. Maybe it is imaginary, but sometimes I get the feeling that the 1DXII files have a more “organic” look to them, than the 5DIV and EOS R images. If there is anything in this, it is very subtle. I shot large aperture prime lenses most of the time. The 1DXII is significantly better at that than my 5DIV, especially in low light. However, the AF of EOS R is soo much better with large apertures and moving subjects, it’s not even funny. I get way more keepers and shots in focus than before. Having the freedom to focus close to the entire frame also make a significant difference. So yes, I miss a lot of things with the 1DXII, but I don’t regret replacing it with the EOS R and RF 50L. One last thing, I rarely shoot bursts, so I don’t care about fps and buffer depths.

  • Renaud Saada

    Canon engineers interview (from dpreview): “During shooting there’s a massive amount of data going between the lens
    and the camera, so making sure that was smooth was another challenge,” says Kato. “By having that huge amount of data being transferred between the camera and lens, you can improve the general performance in relation to autofocus, metering and image stabilization. And you can also add features such as the control rings on the lenses.”
    That should explain the need for shielding above the PCB and some other changes. There are more data channels and with higher speeds (Canon patent has been shown on that subject).

  • All this fancy electronics is to mine some BTC for Canon while you’re taking pictures 😉

  • Are you missing the 1dx2? I am. The 4k60p video with APS-H crop was amazing… :’-( but I hope a mirrorless replacement is on the horizon.

  • Sold. Just placed my order to pick it up tomorrow in Brentwood. I really hope it won’t be the lens from this blog post.

  • Yup, I’ve been saying it this whole time, ever since I got my hands on it at the press release… That RF 50 1.2 is a dreamy lens. In fact if you’re weak-wristed, you’ll definitely feel the torque of that “big boy” AF motor when you jam the AF-ON button. And boy is it snappy and reliable. It exposes the old EF f/1.2 L’s as the utter slugs that they were/are.

    As a landscape photographer I still don’t have much interest in switching from Nikon, but as a wedding/portrait photographer, I’m definitely keeping an eye out for whichever “5-series” pro version of the EOS R comes out next… If they can do this level of impressiveness with a ~28 and an ~85, and throw in IBIS and dual card slots, I’m sold…

  • Lars Kvinge

    I have been eagerly awaiting this tear down, after selling my 1DXII to get the EOS R and the new RF 50L. It was worth it. The RF 50L isn’t just sharp, it has mutch of the great rendering that you can get from the 85LII. I would trade the EF 35LII, the 50L and 85LII for this lens, if I had to.

    One thing I didn’t find you commenting on, Roger. Is it as solidly built as the 35LII? I know it has a plastic barrel, but does it have a solid metal core like the 35LII? I wasn’t able to tell for certain from the pictures. Thank you!

  • Perry

    can it run crysis?

  • I don’t think a filter will be a problem. Usually these break when they got dropped or banged against something.

  • bdbender4

    Maya, thanks for sharing. I really enjoyed scanning through this. (Although there was a fair bit of them breaking their arms to scratch their own backs, as the saying goes. “Only Canon can do this” sort of thing…) Now to read through the DC Watch stuff you posted below – thanks again.

  • Andreas Werle

    Thanks for granting us a view to the “shaved cat” and its wires, Roger. It is always a pleasure, cannot wait to see the interiors of the new 400/2.8! One question, do you recomment not to use filters on this lens, if a replacement of a broken filter ring can be expensive?

  • Maya

    You have no idea how much I share that sentiment. DC Watch sometimes (too rarely) interviews lens designers but I’m afraid that Google Translate fails to convey everything faithfully. Their interviews of Nikon engineers for the 58mm and 105mm f1.4 are very interesting IMO because they go beyond engineering – of which I wouldn’t understand much anyway – and explore the question of what makes for a great lens and the difficulty in finding the right quantitative balance between various aberrations and resolution :

  • Deanaaargh

    Thank you,
    I look forward to reading it, like I do anything you guys post on this blog!

  • Deanaaargh, we will, but the reality is we have too much equipment these days to do that analysis by hand like we used to. The good news is the programmers are giving us some tools now that should automate things so we hope to have a report in 2019.

  • I think it’s really about moving a lot of electrons a longer distance than usual. Of course, I could be wrong.

  • bdbender4

    Round wires? Solder?! Canon becomes Nikon. And in the Nikon Z7, no round wires. Nikon becomes Canon. Convergent evolution, like mammal eyes and octopus eyes!

    But never mind that, it looks like the round wires go toward the control ring? Perhaps a late design modification that will be addressed with a proper flex in future lenses.

  • Thank you, Maya. I love it when they let engineers show their enthusiasm and talk about things like this without a marketing person ‘interpreting’ for them.

  • Deanaaargh

    Richard and Aaron,
    thank you for taking the time to write this up. The construction supporting the optics is often as interesting as the optical designs, and I like the look of this Canon.
    It will be interesting, I assume, to see to what extent Canon adopts these new designs. But with all new technology I suspect there will be bugs that need to be worked out.
    I have appreciated your appreciate your analysis and caution when it comes to “weather sealing” when tearing down cameras. In the same vein would you ever be comfortable divulging failure rates and modes for types of camera equipment. I feel like there is a dearth of consumer information about the durability of cameras.

    I ask specifically because often cycle with my camera frequently on gravel roads, and I am interested in assessing potential failure points given the resultant vibration. i.e. would I be better off leaving behind cameras with ibis or lenses with vr?

    TL;DR can you make some data available about failure rates?

  • Maya

    P.17 of that pdf interview of RF engineers talks about the 50mm f1.2’s retention mechanism but I’m not sure that it contains any information that you haven’t already guessed :

Follow on Feedly