Homemade 35mm Adapter

[Bob Hart]

Here's some useless information regarding the AGUS35 - Australian Plumber's Version - Image Erector.

Compared to the SG35, Letus35, Redrock, etc, it is inefficently large because of the CD sized disk. Except for being dropped or run over, I expect it to last a long time as the only fast moving wearing parts are in the motor which runs at about 1500 rpm.

This version does not permit integration of rod style supports through to the camcorder as the disk enclosure gets in the way. The enclosure is also too large to move on top as it then touches parts of the camcorders.

It could be redesigned to work off to the left or right side but would be a bitch to operate handheld because the weight would be way off to one side or the other.

The side view has altered since I began fitting up the FX1 type cameras and I dont have a current accurate view.

The image tube has become shorter in order for the FX1 to "see" more of the groundglass without the prism edges getting in the way. It still works for VX2000/PD150 in this shorter version, in fact better than the original.

From lens hood mount rear face to the AGUS35 Nikon Mount front face is 111mm. The whole thing, including motor, glass disk and 1.5v battery, without lens attached weighs 874grams.

The optical centres as originally illustrated are pretty much as built for the PD150, HDRFX1 and for the the DVX100 they are fine when matched up hand held. The current .pdf files in media/hart are no longer correct for the design and should be ignored.

I have a profile for the DVX100 lens hood bayonet mount but I don't have the correct orientation so will have to inspect a camera again or get somebody at dvinfo to scan or photocopy the hood from the camcorder side to see the correct orientation of the bayonet lugs relative to the set screw. - Any volunteers??

[Bob Hart]

I have tried uploading 5 very short uncompressed .avi clips to the following address.

http://www.putfile.com/bh_107f

They apparently cannot be downloaded and must be viewed off the site . I've tried working them and they don't seem to work but it might just be my mismanagement.

Maybe somebody can give me some advice. All I am doing is becoming frustrated and getting the red mist, which is not good for stress mitigation.

The clips are one from night-vision into PD150 and the rest, tests into HDRFX1. The titles are bh-AR001 to bh_AR005.

Whilst the performance of my version into the HDRFX1 cold be described as adequate, there are other versions which do better.

Regards all.

[Bob Hart]

The clips I sent up which are mentioned in the previous post do not work. It seems I have mismanged them so I shall try again sometime soon.

Today, I took an FX1 attached to an AGUS, down to one of our local TV broadcasters who kindly allowed me to use their Lemac chart.

I put all the Nikon mount prime lenses up at two aperture settings, wide open and f.5.6. Also I tested the 12mm -24mm Nikon digital zoom wide open only and found with careful attention to backfocus and actual focus, this lens can perform as the equal of the others. The f2.8 135mm Auto-Tamron also held up as sharply. This is an old lens of the Adapt-a-matic series not Adaptall.

Tje 12mm - 24mm digital zoom surprised me as my impression was it was a distinctly inferior lens in this application. Against my expectations the 50mm lens at f1.8 suffered the most for being wide open. So much for one's assumptions based on casual observation.

After running the test, I then found, that the vertical and horizontal bar blocks are marked as not resolveable by video which left only the Siemens star focus indicators, the four "eyes" in the corners. These suggest there is little difference between the best resolution off the disk and the camera's own view. However, their appearance fluctuated in a regular periodic beat which altered when the disk motor was shut down.

The hoizontal and vertical bar blocks seem as visible to the camera through the disk image as they are directly. The "B" and "G" blocks yield a moire pattern which suggests they are being resolved. This leaves only the finest "A" and "F" blocks as not being resolvable both via the disk image and direct to camera.

The fluctuation on the Siemens indicators suggests the disk is still running out sufficiently to affect backfocus at HD resolutions. The disk seems to run true by eye match so the issue may instead be end float which in my device is contolled only by the magnetic field of the motor.

There remains an added slightly warm greyish cast to the colour compared to the camera view under the same lighting. It can be white-balanced out.

Constrast seems to be reduced.

The two grey scale blocks at the darker left side of the bar were less separable than in the direct camera view.

View was via the FX1 LCD flipout screen. I have no HD capture facility. The broadcaster did not either as they use their HVRZIUs (similar to HDRFX1) for news in the DVCAM mode.

Lighting was by one 500watt flood which their camera department had just finished using for their own test.

The camera was left in full automatic mode fir all functions.

The test suggests there is room to allow slightly less resolution in favour of a more filmlike image off the groundglass and yet find an acceptable performance into a HD capable camera. This is consistent with Ben Gurvich's observations on the SD footage I have sent to him for comment.

The problem with falloff due to the proximity of one edge of the combined prism path remains for the FX1 which uses more of the area off the groundglass than my own camera, the PD150. I hope to resolve this by using longer but also thinner prisms which should fit into the same workspace as the 40 x 40 x 56mm x 40mm common thickness set I now use.

Images as .jpg files are going to be of limited use as I can only create them via the MiniDV codec. If anyone wants to see these I can post them at putfile if they are requested here.

The news cameraman I spoke to was interested in terms of being able to put long lenses on the HVRZIU.

The above information is relevent only for versions which use the 24mm x 18mm image frame off the groundglass. The FX1 is actually taking a little more width than that and slightly less height at full zoom-in via a +7 acromatic dioptre.

[Bob Hart]

If anone is still building spinners, there is a part which may interest you in terms of setting backfocus and focal plane alignment.

There is a generation of very low cost DVD players which are being found on roadside rubbish collections (in Australia at least) or discards from repair shops as being uneconomic to repair.

Some of the slimline styles with centre trays use a transport, the SHINWA SHD-2502. These have analogue motors for spindle, tracker and loader. The spindle motors have a short shaft, which enables a more compact device. The motors are smaller in profile than older CD player motors but have the same face mount system and centres. The build precision of the shaft to bearing fit is quite amazing.

A bonus is that they have a three axis adjustment on the spindle motor mount similar to the arrangement for analogue tape recorder heads, ie., one fixed screw on a pillar and two adjustment screws which bear against two small but strong compression screws beneath the motor mount plate and secure this to a subchassis.

By cutting away a small portion of the subchassis to carry the motor mount assembly, a precision three axis backfocus and focal adjustment, an adjustable motor mount, ready-made becomes available and requires only simple fitting to a less sophisticated structure. - Epoxy glue even.

The downside will be that being mounted against spring pressure, there is potential for movement if the device is knocked or bumped. However if the DVD players are capable of surviving longdistance transport, I doubt careful use of an AGUS device will make any more demands than this motor mount arrangement can deal with.

These analogue motors may not continue to be available as there seems to be a newer generation of DVD players with brushless DC motors which require a digitally controlled pulse power supply like CR / DVD recorders, burners and scale aircraft electric motors. There remains one analogue motor for the load tray. Removing the spindle from some brushless motors of the outrunner style has proven impossible.

The shaft diameter of the brushless and analogue motors appears to be the same.

I find the availability of spindle hubs a greater difficulty as you only get one for every three motors you salvage.

Good luck folks.

[Bob Hart]

I have again attempted to post short clips in H264 Quciktime this time at this address :-

http://www.putfile.com/bh_107f

The stupid part is I cannot download them back to my email computer as it is a dog. So I can't view them off the web. If ayone actually gets them to work, I would appreciate the advice that they were viewable.

Here follows some furthur useless information for ayone who may be trying to build a plumbers version.

It is a grab from a recent email I sent.

""""""""""""""""""""""""""""""""""""""""""""""""""""""
Hello again.

My diagrams don't show one or two later developments.

Since I fitted up a HDRFX1, I have shortened the length of the image tube to compensate for the thickness of a proper Nikon mount on the front cap and for the FX1 to be able to focus on a wider piece of the groundglass. This involved cutting about 10mm off the front which leaves about 8mm from front of case to front of tube and about 14mm off the back to enable closer coupling of the FX1. The pipe caps themselves now have to be trimmed down as well so they don't fetch up short against the case.

This also involved moving the motor mount closer to the camera by adjusting it with the adjusting bolts to within about 14mm of the front face of the rear prism which was another reason to shorten the image tube at the front. To keep spring pressure I had to put some spacers under them between the front case and the motor mount.

You also need to have a restraining device to stop the disk from moving up and striking the front prism which overhangs it. This can happen if the device gets bumped hard or dropped.

At the Cunderdin Airshow last March, the device dropped off the front of the camera and hit the ground hard while running and shattered the disk and chipped the front prism. I had quickly assembled it with glue tacking, meaning to make permanent screw fitting later and forgot once I had the thing working the way I wanted it. The glue softened in the hot sun because I had the device painted flat black.

There should be a little hook bolt (a cut-down and re-threaded bicycle spoke) from a hole in the motor mount plate below the battery holder and a little to left of the bottom motor mount adjustment stud, - through a hole in the bottom of the case and a small nut or spoke nut to pull up the slack and restrain the motor mount and the disk downwards to keep prism to disk clearance at about 1.5mm.

In the end I got lazy and just glued the adjustment nuts and exposed areas of the studs at the motor mount with lots of silicone bathroom sealer and put a little piece of cardboard across the bottom face of the prism after I set backfocus.

Feeding the bicycle spoke through whilst trying to fit up the motor mount is a real pain.The nut and the spoke end sticking out the bottom is untidy. It also adds complication to the backfocus adjustment and could put it off if the spoke was knocked during use.

Backfocus is easy but has to be very precisely done, otherwise the wide lenses like 24mm, 28mm or zooms will not focus sharply.

All three motor mount nuts are used and interact for backfocus and alignment. The bottom one allows vertical alignment with the lens centre-axes as well as backfocus. . Each side one allows horizontal alignment with the lens centre axes as well as backfocus.

Because the side adjusters are below the image area on the upper half of the disk, screwing the bottom adjustment in (frontwards) actually causes the upper half of the disk to move backwards.This adjustment equally affects the adjustment of both of the other two side adjusters simultaneously.

I set initial backfocus by using the depth gauge of a vernier caliper and measuring from the Nikon mount front face to the front of the glass disk. Because I grind the rear face of the glass closest to the camcorder, the measurement is 46.5mm less the disk thickness which is 0.9mm for my glass home-made disk. So it is about 45.5mm.

I find it helpful when setting each adjuster, to gently press against spring pressure, the motor mount near the nut being adjusted to pop the mount slightly off focus at that point and allowing it slowly to come back onto focus, then snugging the nut up against the plate then letting go. You have to apply pressure very near to the nut as the mount plate can flex if you are not over the spring and the adjustment is then no good.

You need to have a distant defined object like a radio tower on a hilltop and a good tripod for a successful backfocus adjustment, though I have also fluked it by measuring to an object on a wall from the groundglass, then setting the lens to that distance by the number on the lens ring by the same method of focussing the edges of the image.

You check the Nikon lens groundglass image at each edge left and right, top and bottom by using pan and tilt on the tripod to place the same distant object at each of the edges in turn and adjusting the nuts for sharpest image at that point. When you have the edges all the same sharpness, the centre should be correct. When setting this up you need to have the Nikon lens on infinity focus.

If you are in a hurry, then just sharpen the two sides and the bottom edge of the image as you see it in the camera LCD. this should be close eough but not always as good as it can be.

With the FX1, and Century Optics 7+ achromatic dioptre, the groundglass image seems best when it focusses sharp to the camcorder when the camcorder's distance scale reads 1.4 metres in the LCD viewfinder.

The polypropolene sheet (sometimes used as whuiteboard material) is less acoustically live than harder plastics and less likely to transfer motor noise conductively to the camcorder body. It is softer and may change shape over time which means periodic checking of the backfocus may remain necessary.

I try to reduce this tendency by fitting a flat washer on both faces of the motor mount, ie., the spring side and the adjusting nut side. So far, the backfocus has held true during transport and use. It is more likely to go off during a long period of storage. I recommend storage of the device in the upright position so the weight of the motor and disk bears directly along the plate and not frontwards or rearwards through it.

When I update the design I will email the revised pages.


Regards

Bob Hart

"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""

What I forgot to add was that the relay focus has to also be evenly sharp across the image frame. I achieve this alignment by making the three screw holes in the sides of the caps into slots, not round holes, adjusting the cap for best sharpness across the gg after Nikon backfocus has been completed, then snugging down the screws firmly so the rear adaptor which fits up to the lens hood bayonet fitting remains square-on relative to the groundglass.

[Bob Hart]

An update on the image tube/prism block.

The design has been altered to move the horizontal (front) prism apex 3mm to the left as viewed from the rear. The apex of the vertical prism remains the same. The effect of this is to move the falloff area attributable to the horizontal prism apex and vertical prism edge out of the left camera frame as viewed to permit a larger 16:9 area off the groundglass.

This has a side-effect of shifting the frame centre relative to the radius of the disk and may require the image tube centre to be moved about 5mm outwards relative to the disk enclosure.

This is to allow the prism block to be turned anti-clockwise as viewed from the rear, about 7 degrees in the image tube to restore the correct tangential relationship between the lower side of the horizontal (front) prism relative to the groundglass disk.

The two original 40mm x 40mm x 56mm x 40mm common thickness prisms are replaced with two 45mm x 45mm x 65mm x 32mm common thickness prisms, which more closely resemble porro prisms.

To compensate for the added thickness, apex to apex of the two larger prisms, the two prisms are to be installed face to face and fixed to each other by UV curable optical grade adhesive. This will also eliminate a narrow space between the two prisms which is a dust trap and impossible to clean except by dismantling the prism block.

The glass disk I am currently using has a 5 micron finish with a slight backpolish, made using the same method as the original disk which was broken. This disk had been originally rejected as there were defects in the groundglass finish. The defects were larger pits remaining in the finish resulting from the original cutoff process which had not been dressed out by the first coarse grit run. I could be dreaming but the existence of this coarser defect in the finer grade finish seems to produce a slightly better image than my original disk.

Finally, totaly off-topic. The cat (the ginger one in some of the test images) decided my carefully marked out mount mount plates waiting to be cut out from a polypropolene sheet was a ideal place to deposit the biggest furball in the whole entire universe. I didn't see it for two days by which time the biro mark out lines on the polypropolene had become bleached off.

[Alexandre Lucena]

Go ahead and write a novel on this subject.

[Wayne Kinney]

I have been watching your progress in this, thank for sharing!

[Bob Hart]

For those who might be making erecting (flip) versions out of PVC pipe and caps and finding great difficulty in getting the centre axes right. -------------

If you are measuring from the front face of the front cap and the rear face of the rear cap to get them parallel to get the centre axis of the SLR lens, the centre axes of the prism path and the centre axis of the video camera lens all parallel and finding that despite your very precise and meticulous efforts it doesn't hang together right and you get a big soft edge to your images, very likely it is not your fault.

PVC pipe caps like many such injection moulded products have a shape memory. The plastic wants to go back to that it once was, a blob. There are latent stresses within the material and when you cut a hole through it, the cap will deform slightly, most times the centre becomes bowed inwards.

With a non-erecting version, the lens mount will most likely be placed dead centre in the front cap, the camera dead centre in the rear cap and any concave deformation in the cap will not cause a problem as the deformation is centred on the hole.

An erecting version is another matter. The lens mount is likely to be fitted over a hole cut well off centre in the pipe cap. In this instance the centre of the lens mount and the centre of the concave deformation in the cap do not coincide. The lens mount will not be correctly aligned even if the front outer rim of the pipe cap is.

The solution to making sure the lens mount is parallel to the rear cap, is to place a straight edge across the mount face itself and measure back to the rear cap from that. An alternative is to place the device face-down onto a plane surface such as a sheet of glass and with the lens mount face firm on the glass, measure back from the surface of the glass to the rear face of the rear cap.

The caps themselves will be unlikely to be parallel to each other once the face of the lens mount is made parallel to the rear cap.

The rear cap will also present the same problem but the hole for the front-end of the camcorder is much larger and the effect of the deformation less evident. This also only has an effect where the device is directly mounted to the camera via the filter mount of the lens hood bayonet fitting.

Okay! The next question is how do you adjust for the deformation. Fortunately, the deformation across the front face of the caps the caps after the hole is cut is much less than the angle of taper inside the caps where the fit is over the tube. It is a simple matter of skewing the cap slightly on the tube until the parallel adjustment of the front face of the lens mount to the rear cap is achieved. Once the alignment is right, the screwholes are drilled and the screws fitted.

This method is also only known to be valid for Nikon mounts. There may be difficulty with Canon mounts because of the collar on front of those mounts.

Like the rest of the appliance the method is not every elegant but it does work.

A better method of course would be to machine the front face of the cap but then the construction method would have to move from the kitchen table to the machine shop, which rather defeats the purpose of this exercise.

[Dennis Wood]

Bob, someone should be giving you an award. Your posts both historical and current are a huge asset/help to the adapter world. Cheers!

[Jim Lafferty]

Hmmm.... history...

[Agus Casse]

i just love the cigarrette video... i remember that it was like 5am when i recorded it...

[Jim Lafferty]

I think I can safely say many of us here have a fond memory of that video, too :) That's why I've kept them all... maybe soon I'll do a "history of the 35mm adapter" video for the wiki entry and just for kicks on my site.

Let me know who I'm forgetting, because this is the timeline as I remember it:

1) Agus Casse introduces the idea; made from about $20 and assorted parts from around the house, glued together. Over time, many people contribute to refining this design concept over a myriad of implementations...

2) Alain Aldus moves to the first static design, using aluminum oxide ground glass -- the Aldu35.

3) Brett Erskine schools everyone on the value of condensors, achromats, etc.

4) What is to eventually become the RedrockMicro implementation of the "spinner" styled adapter makes its debut, bringing on the first of the commercial endeavors of many to come...

5) Frank Ladner is the first to successfully employ microcrystalline wax and shows what is in my opinion the best static DIY adapter footage evar.

6) Quyen Le develops, along with Brett Erskine and others at the same time, the first practical oscillating GG device.

7) ...

[Quyen Le]

Agus Casse

Happy to see you around. Do you have anything in your sleeve to influence people here? Thanks for sharing your technics and designs.

Jim Lafferty

Don't forget Dan's works, he has contributed a lot. From number 7 and on, it's hard to come up with anything new unless we have some breakthru technics. All the flipping, relay ... have been on the market and are not new. Hope some of us can someday comes up with something significant, thanks.

Quyen

[Bob Hart]

Agus.

Long time no see.

I seem to recall you were working at a TV station. How is that going? Have you progressed your own device or still using what works best for you?


Quyen.

I think I might have a handle on the ghost image thing some operators are experiencing.

I was doing some outdoors tests with wide lenses trying to wring the last bit of sharpness from them. I happened to pick up the sun through tree leaves and there was its ghost image, shimmering slightly in harmony with a slight runout I am getting with my current disk.

I think it is occurring, on my specimen at least, between the shiny side of the groundglass which faces the front and the optics in the SLR lens.

This might be a case for reversing a popular preference for placing the groundglass surface closer to the camcorder.