Trinocular output very poor

crystallattice · Post by **crystallattice** » Mon Oct 02, 2017 5:51 am

Hi,

I recently bought a Chinese (Meiji copy/clone) stereo trinocular simulfocal zoom microscope. I also bought a 0.5x camera adapter and 16MP c-mount camera from the same supplier. I am quite happy with the microscope but very disappointed with the camera output. Googling tells me it is a combination of chromatic aberration and various other artifacts, none of which is visible through the eyepieces.

I'm now trying to figure out which of the three components are the cause (or, rather the biggest contributor, after all, everything was made in china and cheaper than name brands...)

Supplied adapter: SZMCTV1/2 (0.5x)
Camera sensor size: 1/2.33" (Panasonic?)

It is extremely difficult to get even a fairly nice picture on the camera, it is just not possible to focus and the DOF is virtually nothing. Looking for solutions, I found this adapter is often sold for CCTV camera application, so not sure about microscope suitability!

Therefore I'm thinking that the adapter is ill suited for this application since the eyepiece view is great, at least to me! I'm now looking at adapting an eyepiece c-mount camera adapter to work with the trinocular port.

It would appear that the eyepiece camera adapters are longer and use smaller lenses compared to the adapter I'm currently using. A bit of research shows that the the aperture size affects DOF so I'm guessing that the eyepiece adapter will solve that issue and also enable it to parfocal, which at the moment is near impossible.

The microscope does not have a trinocular tube as such, it just has a short base with female threads into which the SZMCTV1/2 threads. I have checked and the trinocular produces a real image on a piece of paper held a small distance above it, about 20-30mm, it is however very difficult to determine the "quality" of this image as it is so small and difficult to hold the paper at the exact focal distance.

So at this point it is very difficult to determine if the trinocular port needs any correction and if the current adapter has it and it interferes or lacks it.

I don't have much experience with optics and also have no other adapters or microscopes lying around to try and eliminate the cause. I also don't want to buy more adapters to try and experiment as this can get expensive. I'm hoping to get the right input and buy once! (Keep in mind I don't want to spend more for an adapter than I did for the microscope or camera!) I'm able to make tubes/mounts and fittings to adapt the eyepiece c-mount adapter if that will solve the issues.

I'll upload some pictures later to show the image quality or rather the lack thereof!

Some advice would be greatly appreciated!!

Thanks!

ray_parkhurst · Post by **ray_parkhurst** » Mon Oct 02, 2017 6:53 am

You might try kludging a paper tube to place an eyepiece on the trinoc at the parfocal distance of the eyepieces to check the quality of the image there.

Does the scope have a beamsplitter or an "in/out" mirror for the trinoc port? Could be the mirror is messed up. Putting the eyepiece at the viewing location will tell you if the scope is the problem or the 0.5x reducer/camera.

crystallattice · Post by **crystallattice** » Tue Oct 03, 2017 3:33 am

Thanks Ray,

This was exactly my next experiment I had planned and did. Eyepiece held in hand above the trinocular port the image looks as good as the view through eyepieces normally does! obviously it is quite small and you have to "scan" the trincoular port with the eyepiece, a bit difficult to do and maintain focus but I don't see any obvious artifacts. Might have to do the cardboard tube trick and try again.

I also went a step further and also held the camera directly above the trinocular port (the scope was bundled with a 1x adapter but I exchanged it for a discount on the 0.5x when I made the purchase.)

The camera view is a lot better than the with the adapter but obviously very sensitive to being kept level for proper focus. It does appear to have some CA but not so much as before and I would say, it is quite usable, if not perfect once mounted properly.

If you look at the attached pictures you'll see the image was "more focused" in the center. I'm hoping this was all due to the adapter lens and not an issue with the camera/microscope.

It is simulfocal so I would imaging no mirrors just a prism beamsplitter.

So now the questions:

Does the adapter need/have CA correction (which the eyepiece probably has)

Or is the adapter just crap/not right for the application?

Will an eyepiece c-mount adapter jury rigged to fit the trinocular port solve the problems?

I'm thinking(hoping) the adapter is crap and that a cheap ebay 0.5x eyepiece c-mount adapter will solve the issues!

Something like this:

http://www.ebay.com/itm/Microscope-0-5X ... SwHjNV-njJ

Also, is there any practical way of measuring what the best reduction factor is for the new adapter. 0.5x still seems too big based on the camera view. I've read a 0.45x can work but also a 0.4x or even a 0.35x

I'm just scared that the image quality at the "boundary/periphery" of the trinocular port is bad, like it currently shows on the camera view and that by going smaller than 0.5x is actually worse!

ray_parkhurst · Post by **ray_parkhurst** » Tue Oct 03, 2017 6:53 am

When you were viewing through the eyepiece, did you see poor image quality around the edges like you see with the adapter/camera? If so then it may be the objective optics in the scope limiting you. The reason for different magnification adapters is to match the FOV of the camera with the FOV of the eyepieces. Was the field of view with the eyepiece the same as you see with the adapter/camera? If you go to a lower magnification adapter, you will see a wider FOV, and if the edge quality is due to the microscope optics, it will probably be worse.

crystallattice · Post by **crystallattice** » Tue Oct 03, 2017 7:26 am

Looking through the trinocular port using an eyepiece in my hand I have to "look around" inside the port to see the whole picture, if that makes sense. What I can see looking around looks good, but obviously I'm not sure if I'm "adjusting" my focus as I'm looking around and therefore it does not look blurry on the edges.

Normal view through the binoculars with the eyepieces in looks very nice, with no issues right up to the edges. I therefore suspect it is the adapter that is causing the issue. The camera view, with the camera held over the trinocular port also looks good although difficult to maintain focus by hand.

Yes, FOV through camera is still a bit smaller than eyepiece view and I'm also worried that going to 0.4x or to 0.3x might not do any good although my tests thus far are very positive.

My approach to measure the FOV of the eyepieces vs that of the camera was to view a ruler and then work out a factor of de-magnification, if I recall correctly it was around 1/3 needed, but you only seem to easily get a 0.3x or a 0.4x, a 0.35x does not seem to be so popular or very expensive.

So now I have to choose between vignetting or not getting the full picture. Camera has digital zoom which should be able to eliminate the vignetting though, or end up zooming in too much.

any comment on the ebay adapters?

crystallattice · Post by **crystallattice** » Tue Oct 03, 2017 9:30 am

OK, did another FOV test with ruler:

Eyepiece FOV is about 59mm
Trinocular FOV looks to be about 68mm using the eyepiece (x10).

Camera Sensor is 1/2.33"

Using the camera alone to look inside the trinocular I can see about 17mm of the ruler when it is focused.

So of the 59mm (or should I use the 68mm?) I'm seeing 17mm across the screen with the camera.

I did some calcs and I get to a factor of 0.359x using the 59mm FOV

This takes into account that the sensor is not square and that the circle has to cover the whole sensor. (Assuming I did the calcs correctly!)

Using the 68mm it works to a factor of 0.311x

The 68mm is visible inside the trinocular so I suppose I can use it then a 0.35x adapter should be fine with a bit to spare - I hope!

Comments welcome!

ray_parkhurst · Post by **ray_parkhurst** » Tue Oct 03, 2017 3:38 pm

Sounds about right. Now the challenge is to find an adapter that has high enough image quality to not distort the edges. I've never looked for such adapters, so can't give any recommendations.

How did the camera image look without the adapter? I know it's only 17mm FOV, but it might be instructive to look at the quality of the image you can get without the adapter for debugging purposes.

crystallattice · Post by **crystallattice** » Wed Oct 04, 2017 12:48 am

Well, here are some pictures taken with black cardboard tube as an adapter. One zoomed in all the way (4.5(zoom) x 10(eye) x 0.5(aux obj) = 22.5x) and another zoomed all the way out.

Not great, but I would imagine that once the 0.35x adapter(or whatever it should be) is added it should reduce the noise by giving more light and hopefully correct some CA.

Either way I think it is still better than with the current adapter.

Let me know what you think.

[/img]

Pau · Post by **Pau** » Wed Oct 04, 2017 1:35 am

I think that your last images are pretty good for that kind of instrument. Be aware that most stereomicroscopes aren't very good as photography devices because their intrinsic design limitations: small NA, out of axis light path...

At high zoom settings you can expect lack of sharpness because the limited objective resolution (what is called empty magnification)

ray_parkhurst · Post by **ray_parkhurst** » Wed Oct 04, 2017 7:27 am

Those do look pretty good. Most likely adding more glass won't improve the IQ, just expand the FOV, possibly at the expense of corner performance. It's also not likely CA will be improved, only degraded further.

Can the magnification of the scope body be reduced enough to make the camera-only FOV useful to you?

crystallattice · Post by **crystallattice** » Wed Oct 04, 2017 7:58 am

small NA, IQ, corner performance?

Not sure what that means!

I'll be using the scope as is, no further changes except to get the camera working. I was hoping that the camera adapters have some CA correction to improve things. Still need to do some math to figure out what reduction I need and then decide where to buy one from....

I know this is a trick question but will an adapter for another scope work here, bar the mechanical differences? I can get a 0.35x Fischer adapter that needs a mounting ring to work with my scope, which I'll make, not sure if they are scope specific otherwise???

Pau · Post by **Pau** » Wed Oct 04, 2017 9:33 am

NA - Numerical Aperture, it's a function of the angle of the light cone entering the objective. It's the main parameter that controls the resolution of a microscope

IQ - Image Quality

corner performance - Quality of the image corners compared with the center

Likely an adapter lens from another maker will work if you can put it at the right position to have the camera image at the same focus point than the visual eyepieces.
An alternative if you don't really need a very wide field is to just use a lens less adapter and low power zoom settings, as I formerly commented with high zoom magnification you are not going to get a sharp image

crystallattice · Post by **crystallattice** » Tue Oct 10, 2017 4:51 am

Thanks for the explanation.

I've been in contact with the supplier and they claim I can't use the 0.5x aux objective with the 0.5x camera adapter, I have to choose which one I want to use - not an option for me.

I also asked at a local university but they could not help either, they do not have many trinocs and they were supplied complete no aftermarket fiddling required. I did get to see some SEM's and also saw a few gallium arsenide atoms! That was well worth the visit, but sadly did not solve my problems...

I have thus sat and calculated as best I could what adapter would give the max FOV and not sure if it is correct. I'm hoping someone on here has done this calc before or can advise if my method is correct:

The camera has a 1/2.33 sensor which is 6.16mm wide x 4.6mm high.

This gives me a 17mm horizontal width view on a monitor with no adapter.

The height was not measured but should be 17/6.16 x 4.6 = 12.69mm

Using CAD, the smallest circle to cover (on the corners) a rectangle of 17mm x 12.69 = 21.21mm diameter

The original FOV through the trinoc was about 69mm diameter circle using an eyepiece to look through the port.

Thus 21.21mm/69mm = 0.307x So a 0.35x adapter would not give exactly the full FOV but 0.3x will be too small.

Therefore 0.35x should be right?

Corner performance should be the same, as the trinoc image that was viewed with an eyepiece looks good and assuming corner performance is dependent on the available image corner quality, correct?