Pentax 645 300 mm ED f/4 lens

[Astrofriend (Lars)]

I have noticed that I sometimes I get a gradient on my dark images. I found a light leakage in the Pentax 645 lens system.

Some photos and info how I found it and solved it:
http://www.astrofriend.eu/astronomy/...645-300ed.html

/Lars

[Astrofriend (Lars)]

Now I have been up and running the Pentax 645 300 mm lens. Looks to do well even if it's 35 years old.

When looking on the stars they are a bit distorted, it's even all over the frame. With CCD'Inspector I also have the information that the sensor is tilted.

Have a look here:

http://www.astrofriend.eu/astronomy/...645-300ed.html

What do you think, can the tilt also be the problem of the triangular stars ? Or what cause this kind of distortion of the star profile ?

/Lars

[gregbradley]

Hi Lars,

The bottom half of the stars look pretty round and good with a little tail on some (possibly guide errors?).

But the top half is showing bad coma so that is tilt.

Your sensor is not square to the lens. The lens is quite heavy so perhaps its putting pressure on your adapter.

Regardless the solution is to pack out the offending side top or bottom of the sensor depending on whether your image is reversed from the physical sensor.

I do a test to determine how the image corresponds to the sensor.

Take a 10 second exposure and use a dimmed down torch (with a cloth or something over it) and expose only the top half of the lens.

Look at the resulting image, is it bright on the top half or the bottom half of the image.

So now you know which side of the sensor - top or bottom, matches the image.

Now get some thin packing material - spark plug gauges or thin sheets of metal. I got some thin brass sheets from a steel company. They are thin like paper.

Cut a small strip and pack it under the middle of the side of the camera that corresponded with the top of the image you posted.

Refocus and take a test image and see if the seagull looking stars are now round or not.

You may need to repeat this several times until you get it right.

One tip, if the stars are getting worse after packing then its either the wrong side or too much packing. Usually the wrong side of the sensor is being packed. It should improve not get worse.

I have used this lens and got round stars to the corners of a 16803 sensor which is larger but a bit forgiving compared to CMOS sensors with their small pixels that show up defects rather easily.

Greg.

[Astrofriend (Lars)]

Hi Greg,
Thanks for long detailed answer.

It's more the camera hanging on the lens. The camera is atached with an adapter to the lens. Each has some play, and even the lens have a play. Three different places, but I have a spring that press the camera against the lens to help get rid of the play. But even so there is a tilt.

You can see here:

http://www.astrofriend.eu/astronomy/...645-300ed.html

I will try to put in something on one side of the bayonet to find out how much shims I need, maybe steps of 0.1 mm, or ? But I can also 3D-Print something that hold it together with correct angle and have it adjustable.

I think the lens is good enough to spend that time on it. Or wath do you say ?

My plan is to buy a mirrorless medium format camera in the future. Then there isn't so much optics to choose from.

Lars

[Astrofriend (Lars)]

Now I'm taking the first steps to eliminate the tilt. Use aluminum foil to shim the bayonet. I start with 0.16 mm at the 1st quadrant.

Look like this:
http://www.astrofriend.eu/astronomy/...645-300ed.html

Thanks Greg for the tip where to put the shim, get it much easier.

/Lars

[Astrofriend (Lars)]

Hi Greg,
Tonight we got a clear sky for some hours and I could do some test with a shim to get rid of the tilt. Not so easy because it isn't stable enough when the camera is balancing on a shim.

This is the data I got from my measurements:

http://www.astrofriend.eu/astronomy/...645-300ed.html

Now I know something about the error, the thickness and direction too I think. But clouds stopped me to proceed this night.

The center star is Mizar, the double star in Big Bear.

/Lars

[gregbradley]

Steps of 0.1mm would work. Add one and recheck with a focus test shot. Check all 4 corners of the image. Then repeat until the elongated stars are round. The image takes on a definite sharper overall look when you get it right.

Yes the lens is worth it. If you plan on using a medium format camera you may need the Pentax 300mm 67 F4 EDIF lens instead. I tried my 645 lens on a FLI Proline 16803 and it gave poor stars in the corners so I used a smaller camera - FLI Microline 16200 APSh sensor and stars were good in the corners.

One of the better results:

https://pbase.com/gregbradley/image/167787458/large

Greg.

Originally Posted by Astrofriend

Hi Greg,
Thanks for long detailed answer.

It's more the camera hanging on the lens. The camera is atached with an adapter to the lens. Each has some play, and even the lens have a play. Three different places, but I have a spring that press the camera against the lens to help get rid of the play. But even so there is a tilt.

You can see here:

http://www.astrofriend.eu/astronomy/...645-300ed.html

I will try to put in something on one side of the bayonet to find out how much shims I need, maybe steps of 0.1 mm, or ? But I can also 3D-Print something that hold it together with correct angle and have it adjustable.

I think the lens is good enough to spend that time on it. Or wath do you say ?

My plan is to buy a mirrorless medium format camera in the future. Then there isn't so much optics to choose from.

Lars

[Astrofriend (Lars)]

Hi Greg,
My first idea was to purchace a Pentax 67 400 mm ED lens, but more then five times more expensive compare to the Pentax 645 300 mm ED I have. But even worse, the few reviews of it told that it's softer then the 645 lens.

One more clear night and I think I have the tilt problem fixed.

Do you use your Pentax 645 300 ED wide open ? I use mine at f/4 not as sharp as it can be at f/5.6, but with the bad atmospheric condition here it's enough.

Mamiya 67 also have some APO lenses, but difficult to use because the focus mechanics is hold in the camera house. But the quality are fine what I have read, but no astrophoto reviews.

/Lars

[gregbradley]

Originally Posted by Astrofriend

Hi Greg,
Thanks for long detailed answer.

It's more the camera hanging on the lens. The camera is atached with an adapter to the lens. Each has some play, and even the lens have a play. Three different places, but I have a spring that press the camera against the lens to help get rid of the play. But even so there is a tilt.

You can see here:

http://www.astrofriend.eu/astronomy/...645-300ed.html

I will try to put in something on one side of the bayonet to find out how much shims I need, maybe steps of 0.1 mm, or ? But I can also 3D-Print something that hold it together with correct angle and have it adjustable.

I think the lens is good enough to spend that time on it. Or wath do you say ?

My plan is to buy a mirrorless medium format camera in the future. Then there isn't so much optics to choose from.

Lars

Originally Posted by Astrofriend

Hi Greg,
My first idea was to purchace a Pentax 67 400 mm ED lens, but more then five times more expensive compare to the Pentax 645 300 mm ED I have. But even worse, the few reviews of it told that it's softer then the 645 lens.

One more clear night and I think I have the tilt problem fixed.

Do you use your Pentax 645 300 ED wide open ? I use mine at f/4 not as sharp as it can be at f/5.6, but with the bad atmospheric condition here it's enough.

Mamiya 67 also have some APO lenses, but difficult to use because the focus mechanics is hold in the camera house. But the quality are fine what I have read, but no astrophoto reviews.

/Lars

I used it mostly at F5.6. As I recall I got a stop down ring as that does not give the artifacts around brighter stars. It was cheap off ebay.

Richard Crisp used a Pentax 67 400 F4 EDIF so you can look at his images and decide if its ok or not.

The main advantage of these Pentax 67 lenses is the large 88mm backfocus. One of the longest of any lens. That's handy for attaching a camera with a filter wheel. You don't need as much for your DSLR imaging.

You get a better result getting a ZWO or QHY camera than a 400mm lens in my opinion. The smaller pixels and better sensitivity, lower noise will give you better resolution, dynamic range and the ability to shoot narrowband.

An APSc sensor like the ASI2800 is very popular or the 294 for micro 4/3rd sized sensors.

It will give you more magnification than 400mm.

400mm and I would start looking at the many 75mm aperture APOs around that are low cost.

Greg.

[Astrofriend (Lars)]

I didn't remeber who Richard Crisp was until I visited his homepage again. That was a page I visited very often in the beginning of 2010s, he inspired me to look for medium format lenses. At that time he didn't own a Pentax 6x7 400 mm ED IF.

I looked at the photos now but I see most of them are cropped so it doesn't tell how the lens behave at the edges.

Earlier I owned a Pentax 6x7 165 mm f/2.8 lens, almost no vignetting at all even at f/2.8 when used on a full frame camera. But it had some cromatich aberration problem. The only of my lenses I could use at f/2.8.

http://www.astrofriend.eu/astronomy/...vs-pentax.html

Today I made a new shim and placed it in the baynet:

http://www.astrofriend.eu/astronomy/...645-300ed.html

See at bottom of page. If I'm lucky the tilt is now eliminated, maybe there will be a clear sky tonight so I can do a test.

/Lars

[JA]

Originally Posted by Astrofriend

What do you think, can the tilt also be the problem of the triangular stars ? Or what cause this kind of distortion of the star profile ?
/Lars

Hi Lars,

That sort of distortion looks like it is or might partly be due to astigmatism, perhaps with a touch of coma. Both aberrations / Astigmatism should get better at smaller apertures and away from the corners, so perhaps run some tests imaging at f/4, 5.6, 8, 11 and compare the various corners/centres to help with some conclusions. It can be painful to have to wait until nightime to run your tests so you can try a test target. I will add a link to one I think might be useful.

Best
JA

EDIT: Aberration Test (Down towards bottom of Article) .....
https://www.lonelyspeck.com/a-practi...erration-test/
The only short coming of the test is that ideally you'd want to do it at infinity for your lens since that's where we use them. Still it may give you a guide to the performance of the lens. Alternatively rather than covering the whole sensor with the whole image at a close distance, you could more further away and just image the whole target using the corners of the sensor, and one at the centre for comparison, thereby allowing testing at infinity focus (probably ~>10-20m, depending on lens).

[bojan]

Lars,

I found that all commercial/terrestrial lenses work better on stars when closed a step or two.

Very often (my experience with Canon tele lenses above 135 mm FL) external diaphragm is better than built-in iris.

So, general rule: if you want f/4, use f/3.5 or f/2.8 lens and stop it down to f/4.

400 mm:
https://www.iceinspace.com.au/forum/...d.php?t=138076

200 mm:
https://www.iceinspace.com.au/forum/...ht=Canon+200mm

100 mm:
https://www.iceinspace.com.au/forum/...ad.php?t=58629

50 mm
https://www.iceinspace.com.au/forum/...ad.php?t=39293

[Astrofriend (Lars)]

Hi JA,

Great article about aberrations, from that I can more precise found what kind of problem the lens has. I have stopped it down to about f/5.6 earlier and it gave much better quality on the stars. But I shall try to have it at f/4 for a while. Color aberrations I can correct in software and some of the other aberrations. The Mickey Mouse stars looks to be the same all over the field and make it easier to correct. Partly it looks to some degree be tangential aberration problem as you noticed also. I had a Rubinar tlelephoto lens long time ago that show the same problem. In that case the problem was that the plastic holder for the main mirror squezed to hard on it, after a fix it gave much better stars. If I can find something simular with the Pentax 300 mm lens I maybe can have it perform better.

Thanks a lot for the link to the article, it will help a lot !



Hi Bojan,
You are wonderful, what lens project you handle Canon 400 mm f/2.8. I almost bought a used (very used) Canon 400 mm f/2.8, it had the bayonet smashed and the wires haning out but the glasses looked good. I offered 1200 Euros but some other bid 1600 Euros, saved me a lot of problem.

The reason I use a medium format lens is to reduce the vignetting. A 50% loss of photos at the corners can be corrected with flats, but a 50% loss of signal decrease the S/N at the corner, it's like they only have half the exposure. My plan is also to have a better camera then my Canon 6D, a mirrorless medium format. Even if they are medium format the sensors are only 44x33 mm compare to what the lens is designed for, 60x45 mm.

But of course a 35 years old meadium format Pentax 645 300 mm ED lens is far from perfect, even the Pentax 6x7 400 ED have problems and is much more expensive.

I don't know about any f/2,8 lens of longer focal lengths. And if there are one I'm sure they are far away from my budget. Once I had the Pentax 6x7 165 mm f/2.8, very good but not the color aberration.

The problem with the Canon 6D, the shaft down to the sensor is too narrow and the mirror shadow the sensor. It's also very difficult to clean the sensor because of this narrow shaft. I think also that the Sony sensors perform better today.

I can get a Fujifilm GFX 50 used for 2000 Euro, but the the problem is some people say they have this star eater problem. So have to wait until ther is a solution to this or some other camera that doesn't have this. The Pentax 645Z is better without star eater, but it has a huge backfocus.



Thanks a lot to both of you sharing links and interesting info. Step by step I solve the problems and get a better working equipment.


By the way, I have started using Siril, I must replace the old Fitswork and AstroImageJ. Siril is built on the Old Iris I used many years ago. Now Siril has implemented 32 bits format and make it very interesting.


/Lars

[Astrofriend (Lars)]

Here are some photos taken with the Pentax 645 300 mm ED lens, all taken at f/4, full opening. All images are not cropped, all the way out to the edge of the sensor and it's a full frame sensor. Taken at Bortle Class 9 environment so no weak objects.

http://www.astrofriend.eu/astronomy/...per-stars.html

I really like this lens. When I look at the oversaturated stars, looks if something is in the way for the light beam, like an aperture. Have to look into the lens and inspect it carefully.

/Lars

[bojan]

This lens is really good

[JA]

Originally Posted by Astrofriend

I really like this lens. When I look at the oversaturated stars, looks if something is in the way for the light beam, like an aperture. Have to look into the lens and inspect it carefully.

/Lars

Hi Lars,

Yes I think the lens is doing a good job except for some of the star shapes. You are right to suspect the lens aperture blades are a cause, suggested by the 8 points / diffraction spikes on the larger stars and possibly some slight distortion, possibly present as imperceptible spikes(?) on smaller stars.

As a test, you could try to use an external iris, maybe at 80-90% of the front element diameter, to see to what extent this mitigates the issue. If it does then you could possibly put up with using that permanently and loose 20% of the available light or you could consider lens surgery to remove the aperture mechanism. Have a good look inside the lens with a torch from both ends.

Good luck

Best
JA

[bojan]

Originally Posted by JA

... As a test, you could try to use an external iris, ...

Yep, this will most likely sort the issue.

[Astrofriend (Lars)]

Discard 20% of the photons? No I keep all the photons I can :-)

But we will see how I fix that, but it's already good to me. And I haven't got the tilt eliminated yet. Today I made a 80 my shim and installed it. I f also offset the focus position to get rid of a wraparound problem. Jumpning from 65550 to 0.

Have a look at page 8 at bottom:

http://www.astrofriend.eu/astronomy/...pberrypi4.html

/Lars

[MrB (Simon)]

When I look at the oversaturated stars, looks if something is in the way for the light beam, like an aperture. Have to look into the lens and inspect it carefully.

I hate to be the bearer of bad news, but your lens should not be doing that when wide open. I also own this exact lens and it is clean wide open.

[Astrofriend (Lars)]

Hi Simon,
Interesting, maybe you have some photo with lot of stars at full resolution that I can analyze ? My camera is a full frame with 20MPix.

Aperture problem:
I also took a look at the aperture how it behave when rotate the aperture ring. From fully open at f/4 to close down to f/5.6. It doens't always open fully, it have happen before, the aperture ring could move a little bit when it's fully open.

Deformed stars:
Maybe I have found what cause the stars to be that distorted, look at this photo I took just after I bought the Pentax lens:

http://www.astrofriend.eu/astronomy/...645-300ed.html

All star are much better. Could it be my focuser that set some side force of the lens barrel ? I have to have a high tension of the timing belt to not get any backlash.

Tilt problem:
I thought maybe there are something in my lens that are out of adjustment and that cause the tilt ? I did the same inspection of my photos with the TS130 APO refractor. It look almost to be the same tilt, it must be the camera sensor that is out of adjustment. But I think I'm looking after very small tilt error here.

Have a look at the bottom of this page, the last photo:

http://www.astrofriend.eu/astronomy/...645-300ed.html


/Lars