Nikon RayFact 3.5x

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Nikon RayFact 3.5x

Post by ray_parkhurst »

OK, I admit it...the seller lowered the price from $2500 to $1755, and I pulled the trigger (and drained my PayPal)! I had not received an answer back from the seller regarding the aperture setting, so I was taking a chance, but the seller offers 30 day returns...turns out there is indeed an intermediate stop between f2.4 and f4, so all is well. I don't know what the setting actually is, so I will just call it f2.8 for now til I can make some physical measurements.

This is a big lens, much bigger than the Printing-Nikkors. It has M67 threads front and rear, and the rear case is 65mm diameter. I also ordered a 65mm-42mm protection circle adapter same time as the lens, but it hasn't arrived yet. After ordering I realized that necking down to M42 would limit me from being able to test the lens wide open, so when the lens arrived I did not wait for the adapter to start testing.

Turns out I had a cheap Canon tripod mount with 65mm ID (with flocking), and it fits and works perfectly.

I ended up using one of my taller stands based on the 4040 column with countertop tile base and 480mm Arca plate. The specs say 70.8mm WD (which is another reason I decided to go for this lens), and 651mm object to image distance. I set up the system with these two parameters, which required me to extend the Arca plate well above the 24" tall column. For interconnect/bellows, I made a tube of black flocking material and attached it to the camera, leaving it mobile around the lens. This is obviously not my preferred way to do this work, but it seems very effective.

I mounted the camera to the column / plate with an Arca clamp and a cheap Fotomate rail going horizontally. I centered the camera on the rail, so I can shift +/-50mm. I mounted the lens to the column directly, using another Fotomate rail going vertically to give some easy WD adjustment.

Manual fine focusing and framing is via a micrometer XYZ.

I'll take a picture of the setup and publish it a bit later. Results first...

My first subject is of course a Lincoln Cent, a 1955-S BIE. The BIE area makes an interesting subject since the metal is often taller there than anywhere else on the coin. The BIE is created by a die chip, which makes a hole in the die such that the corresponding raised area on the coin can be arbitrarily tall.

The published spec numbers were almost dead-on! Just a little adjustment and I was at 3.5x exactly. Zooming in 100% in Live View on the BIE showed a shockingly sharp image...a very promising beginning!

My first shots were an aperture sweep...f2.4, 2.8, 4.0, 5.6. Although there is an intermediate stop between 2.4 and 4.0, there is no such stop between 4.0 and 5.6. To my eyes the f2.4 is slightly sharper than f2.8 (!!!), so my next shots were a coverage sweep to see how far out the lens stays sharp at f2.4. I shifted the camera 21.5mm, 31mm, and 41mm to simulate the corners of FF, and the ends of 62mm and 82mm line sensors. See the images attached below:

Note: these are unprocessed 100% crops of single images, with critical focus on the flat high points on left and right sides of the die chip between the B and E.

Center, f2.4
Image

Center, f2.8
Image

Center, f4.0
Image

Center, f5.6
Image

I see very slightly more detail in the critical focus areas on the f2.4 shot, with slightly more CA vs the f2.8, which has slightly more than the f4.0. I also see false colors on the f2.4 and f2.8, and less on the f4.0 and f5.6. This gave me hope that the f2.4 may be usable across the frame, so I did the camera shift sweep at f2.4.

Edited to add: the below images are native 1024x900 to avoid forum scaling. Aperture sweep images above are still 1200x900 and have been rescaled to 1024x768 by the forum software.

Center, f2.4
Image

21.5mm / FF Corner, f2.4
Image

31mm / 62mm Line sensor end, f2.4
Image

41mm / 82mm Line sensor end, f2.4
Image

Amazingly, the lens holds up well at f2.4 all the way to the end of the 82mm sensor! In fact it looks like the CA improves at the FF corner, and then stays OK all the way to 82mm. This coverage is better than expected, at larger aperture...so far so good!
Last edited by ray_parkhurst on Sat Mar 23, 2019 4:45 am, edited 3 times in total.

Lou Jost
Posts: 5985
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Ray, congratulations! I have to admit that if this lens had stayed up for sale a little longer (like a couple of months longer) I would have bought it. (Needed some time to find the funds....) It looks promising, especially for stitching sensor-shifted images. I look forward to seeing more of your results from it.

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

I seem to be having an issue with the images. The 100% crops are 1200x900, which I believe are within the forum limits, but the software is downsizing them to 1024x768. This may be fine, but then the software is displaying them at larger than 1200x900, so they look quite fuzzy.

Rik, any idea why this is happening? My first uploads of these pics were larger than 300k, so I thought that might be the problem, but I re-uploaded versions that were smaller than 300k, and the same thing happened.

ChrisR
Site Admin
Posts: 8671
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

The forum limit is still 1024 square, Ray.
Chris R

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

ChrisR wrote:The forum limit is still 1024 square, Ray.
Hmm, for some reason I thought it was 1200... I will upload new ones tomorrow.

rjlittlefield
Site Admin
Posts: 23599
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

ray_parkhurst wrote:The 100% crops are 1200x900, which I believe are within the forum limits, but the software is downsizing them to 1024x768.
As ChrisR says, the forum limits are still 1024 square. Images larger than that are downscaled to fit. This guarantees some loss of quality due to interpolation of pixel values.
but then the software is displaying them at larger than 1200x900,
Yep, and that software is your browser, probably in response to having a Windows display scaling of 125% or some similar number.

What the forum software does is to serve back exactly what was uploaded if that was within limits, or to serve back exactly the results of downsizing. In either case, the image as it comes from the forum will be 1024 or smaller, at this time.

But then if the operating system's display scaling is set to something other than 100%, the browser will resize all images according to the scaling. If your display scaling is 125%, then 1024 pixels become 1280, again with loss of quality due to interpolation.

This will happen to all images, so I suppose it's easy to get the idea that the forum limits have been changed.

My display scaling happens to be 175%, so for me an image that is nominally 1024 pixels becomes 1792 pixels on screen.

To see images at actual size, I have to pull them into Photoshop. But then the pixels are typically too small to study, because my monitors are about 185 pixels per inch, so I end up setting 200% zoom (or higher).

Sometimes I long for the good old days, when a pixel was a pixel was a pixel. But on the other hand, high resolution displays are Very nice, and those imply display scaling to get results that look reasonable.

The challenge is learning to live with them, when you really care about pixel-level detail.

--Rik

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

OK, all makes sense. Multiple problems. My browser was indeed set for 150%. Thanks...Ray

ChrisR
Site Admin
Posts: 8671
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

Browser scaling, set?
If that's what's changed by zoom in/zoom out, changeable by Ctrl-mousewheel, mine's never set. I seem to be changing it all the time. And I see no display of the factor?
Chris R

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

ChrisR wrote:Browser scaling, set?
If that's what's changed by zoom in/zoom out, changeable by Ctrl-mousewheel, mine's never set. I seem to be changing it all the time. And I see no display of the factor?
I'm using Firefox and it has a scaling slider in the one of the settings menus.

I just uploaded new sensor-offset images at 1024 wide and they look better. Will leave the aperture images as they are.

RobertOToole
Posts: 2627
Joined: Thu Jan 17, 2013 9:34 pm
Location: United States
Contact:

Re: Nikon RayFact 3.5x

Post by RobertOToole »

ray_parkhurst wrote:OK, I admit it...the seller lowered the price from $2500 to $1755, and I pulled the trigger (and drained my PayPal)!
Excellent decision Ray, I had a sneaking suspicion that was you.
Amazingly, the lens holds up well at f2.4 all the way to the end of the 82mm sensor! In fact it looks like the CA improves at the FF corner, and then stays OK all the way to 82mm. This coverage is better than expected, at larger aperture...so far so good!
Rayfact only lists the 1x and 2x as APO, the 7x and 3.5x are not so I was curious about CAs. Glad to hear that its ok in that respect.

Robert

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

OK, I installed the Rodenstock/Linos 105mm f4 Inspec.x L 3.5x (105IXL3p5, hereafter 105IXL) lens into the extended setup so I could compare against the Nikon RayFact 3.5x. The 105IXL takes quite a bit less extension (~584 vs 651mm object to sensor) and interestingly has a bit more working distance (~77 vs 71mm). Otherwise it was a fairly simple swap since the clamp I have for the 105IXL has the same height (clamping surface to optical axis) as the tripod mount for the RayFact. Serendipity.

I did not do an aperture sweep since I already know the 105IXL is best wide open at the center. I do plan to do a sensor-pan sweep to check the 105IXL across its image circle.

Below are correct-sized center crops of the RayFact at f2.4 and f4.0, and the 105IXL wide open at f4.0. Colors are a bit different, and I did not attempt to re-white balance the 105IXL, please pardon my laziness.

RayFact 3.5x, f2.4
Image

105mm f4 Inspec.x L 3.5x, f4.0
Image

RayFact 3.5x, f4.0
Image

What I see is what I hoped to see, ie that the Rayfact and 105IXL are very similar at f4. This gives me some additional confidence in both results, as it seems they are both diffraction-limited. The RayFact is clearly better at f2.4, so it appears I have made a wise purchase with this lens.

My plan now is:
- re-install the RayFact
- do a LoCA check across the image circle
- measure the exit pupil to verify the aperture at intermediate between 2.4 and 4.0
- do stacked SR composites to compare f2.4 and "f2.8" to determine best working aperture.

Edited to add: fixed the f2.4 image...they should all be 1024 native now
Last edited by ray_parkhurst on Mon Mar 25, 2019 5:57 pm, edited 2 times in total.

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Re: Nikon RayFact 3.5x

Post by ray_parkhurst »

RobertOToole wrote: Rayfact only lists the 1x and 2x as APO, the 7x and 3.5x are not so I was curious about CAs. Glad to hear that its ok in that respect.
Well, the CAs are worse than we'd expect from an apochromatic lens but they don't seem too bad. I shot an Ag coin (1964 Kennedy Half) at f2.4 with critical focus on the top of the U in TRUST, and then set 50um above and below that level to check the LoCA control. Ag is very close to perfect grey so it's easier to discern hue shifts than for a Cu Cent.

Note the effective aperture is large enough to show significant false colors. At f2.4 and 3.5x I'm at f10.8 effective, just beyond the f9.9 effective for the 95PN at f3.3 and 2x, so I expect similar false color levels with similar illumination. In practice I've found that these colors are easily controlled with Chroma Noise suppression methods. Though of course I'd like to not further compromise color resolution, false colors are extremely annoying. No such editing was done for this test.

Here are the LoCA crops:

Focused 50um low:
Image

Critical focus on letter surface:
Image

Focused 50um high
Image

I plan to do a similar test to compare f2.4 with "f2.8" farther out on the image circle as a criterion for deciding best working aperture.

Antal
Posts: 60
Joined: Fri Sep 28, 2012 4:29 am
Contact:

Post by Antal »

ray_parkhurst wrote: What I see is what I hoped to see, ie that the Rayfact and 105IXL are very similar at f4. This gives me some additional confidence in both results, as it seems they are both diffraction-limited. The RayFact is clearly better at f2.4, so it appears I have made a wise purchase with this lens.
T.4he Linos looks sharper than the rayfact at f4.0 to me.
But anyway, you did right, as the rayfact at f2.4 is the clear winner..

Macro_Cosmos
Posts: 1527
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

Man ~$1700? That's still a lot. Should have sent me a PM and I'll get it for you for the original price of 5000 yuan, which is about $750. I can just act as an intermediate translator while you talk to the seller, won't even ask for a tip since we're on the same forums with the same passion.

I didn't buy it because I was put off with the extremely long extension required and yeah, kind of out of funds :P

Anyhow, congrats on the purchase, it's one kind of a lens and it's a heavy beast! I'd love to see some photos of your setup, just in case another pops up, I'll get my hands on it hopefully.

-- MC

ray_parkhurst
Posts: 3431
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

Macro_Cosmos wrote:Man ~$1700? That's still a lot. Should have sent me a PM and I'll get it for you for the original price of 5000 yuan, which is about $750. I can just act as an intermediate translator while you talk to the seller, won't even ask for a tip since we're on the same forums with the same passion.

I didn't buy it because I was put off with the extremely long extension required and yeah, kind of out of funds :P

Anyhow, congrats on the purchase, it's one kind of a lens and it's a heavy beast! I'd love to see some photos of your setup, just in case another pops up, I'll get my hands on it hopefully.

-- MC
I guess it was on a Chinese-language auction site for much less? I would not have even seen it since I would not know where to look nor could I read the description.

So...what else is selling cheap that will eventually make it to eBay? This was the first and only one of this lens I've ever seen available which is why I snapped it up. Plus, it was essentially new, and they are still selling new for $12.5k. Do you see others available now for the $750, because others on this forum might want a copy.

Are there good links to these cheap sites which might be accessible for language-challenged folk such as me?

Post Reply Previous topicNext topic