The distortion of image during scanning, Zeiss LSM 710 and Zen2009

[lookin]

Hello everybody! I had a problem with the scanned image using Zen 2009 and LSM 710. The image is vertically shrinked by the factor of ~1.5, as well as the center of the image is shifted to the top edge and therefore exactly not the same as I can see in the oculars. I used another program (Zeiss SMT) with calibration objective and the problem persisted. I believe that this comes from the technical problem, but possible anybody encountered similar issue and solved it. Many thanks in advance!

UPD: the distortion is not linear, and increases from top to bottom, also the distortion is higher for low-magnification objective and decreases with increase of X. And finally the distortion is greater for lower wavelength laser. I guess, the problem is due to the non-centrality of microscope optics and the optics of scanhead. Your thouhgts?

[ShawnPatrickCasey]

Post a screenshot or pictures.

[lookin]

Thank you for reply! Sorry, I cannot provide a self-explaining image. Just imagine an ideal square, then now it is shrinked roughly by 1.5 on average, but the shrinking extent is higher at top then at the bottom. It may come from the situation where the optical centerline of the microscope is not aligned to the optical centerline of the scanhead/detector. That is, the beam from the microscope falls to the curved part of lense away of its center, not to its center as should be in the ideal optics. So an image distorts due to the curvature. Then, if I use higher magnification, I reduce the curved area and the distortion reduces. If I use lower wavelength (higher energy), the beam scatters at higher angle and it makes the distortion higher too.
I also found with an external USB camera that the right sideport of my AxioObserver creates good non-distorted image but the left sideport (where I connect my scanhead) makes somewhat strange, non-schrinked but slightly rotated counterclockwise. This may be initially incorrect optics, which then go to the ideal optics of scanhead and produce actual shrink. And I also found that right and left sideports use different prisms, so the problem may arise from the damaged prism served for left side. I would try to acquire the images from left and right sideports and post them within a few days. Thanks!

[FishCsCells]

When relative sizes go strange, my first thought is the scanning mirrors since it is not uncommon for those to fail after a while. A classic sign of bad scanners is that te Live image kind of “breathes”. If you’re under service contract, I would recommend contacting your local service engineer. If you aren’t under contract, us SMT and the calibration objective to run the Scanner Calibration. When the scanners on my 700 went out, that failed miserably. Good luck!

[FishCsCells]

Oh, it could also be the controlling board for the scanners rather than the scanning mirror unit. For my systems that were under contract they often replaced both the scanning unit and the associated control board at the same time to avoid cross-damage.

[lookin]

Thank you for your reply! My system is not under service contract, so I cannot ask Zeiss for checking within the warrant period. Also, I am unable to use SMT as usual, because I see similar problem with calibration objective (the center of the grid of the objective image is shifted (in SMT - leftwards, in ZEN - upwards) and shrinked) and I cannot therefore locate the control dots before calibration procedure (I will do and post the screenshot later). If I simply omit this, the calibration fails, of course. I checked the detector modules using CanCheck and found they are OK and passed tests (while some nodes said potential conflicts, e.g. SDAL, SDPW etc). But it looks like the motorized things and the associated lenses in the scanhead are all OK (detector, filterwheels etc). At the moment it looks like the trouble arises from the prism/mirror which passes the light from the sample to the left sideport. This prism/mirror serves as the pathway for both fluorescent and T-PMT images, and they both are equally distorted. Unfortunately, the hole size on the right of my AxioObserver is not the same as on the left, so it seems I cannot simply connect the scanhead to the right sideport and check how it works.

[lookin]

To say about the scanning mirrors - the acquired image is not flickering or breathing, everything is OK (regardless I use one-way scan or two-ways scan), at it scans normally at low and high speeds and at low pixel/high pixel frames. The sounds from the scanhead during scanning are OK too. Altogether seems to be not due to the scanhead fail.

[FishCsCells]

OK, if it’s a straight shift, rather than constantly changing, it does sound like some optical component.

Have you taken the stage and nosepiece off the microscope and looked down onto the prism from above to see if there’s anything obviously wrong with the prism surface? (I think to really get at the prisms you have to take the bottom off the stand). Any evidence of immersion oil or others gunk dripping into the microscope (lots of oil on the nosepiece, etc)?

[jmreece]

Perhaps the y-galvo is having an issue. The noise you hear is from the x-galvo.

[lookin]

It is difficult to dismount whole system and open the bottom side of the AxioObserver. So, I looked to the mirror (I guess it is rather mirror, not prism) from the opened frontport and right sideport and found no visible damage or dirty surface. I also tried to slightly adjust the position of the wheel which rotates mirrors. At least some changes I could see.

This is the update of the problem. My image, if scanned from the left sideport, is 1) shifted relatively to the center of the same sample observed by oculars, 2) slightly rotated, 3) substantially shrinked by vertical. The images I upload here are to demonstrate this.

20211111-633-10X-Zoom0.6-MAXVIEW-TESTOFVERTICALINTENSITYFADE.jpg - this is what I can see in ZEN using empty glass as a sample, 10X objective, maximally unzoomed area.

GridScan-10X-488nm.jpg - here the example of normal scan using 10X, zoom = 1.0, two channels (fluorescent and T-PMT). The image is both rotated and shrinked by vertical (the ideal squares must be seen in the correct situation), the distortion is the same for both channels and this may support that the problem is in the common optics (as fluorescent and T-PMT use the same optical path).

grid-via-external-camera-sideportLR.jpg - Left - this is what I see using external camera (no optics with) attached to the empty LEFT sideport (also no optics in), the image is not shrinked but slightly rotated. The left sideport use its own mirror. Right - this is what I see using external camera attached to the RIGHT sideport, the image is not shrinked and not rotated. This port uses another mirror. The magnification difference between this image and the previous may come from the different distances available for the camera attached to R or L sideport.

MirrorPositions.jpg - Left - this is the normal position of mirror, as is. Right - this is the situation where I slightly move the ring with the mirror (manually by the access from right sideport) to remove the rotation effect. As I can affect the angle of the lines, this initially slightly rotated image may be due to the incorrect position of the mirror.

So, as I can see, the mirror can be adjusted by several screws and they are all accessible from the right sideport. The incorrect position of the mirror can provide all the optical distortions I can see, and also it may direct the light out of the center of the optics of scanhead. This may explain why I can see non-shrinked image from the left sideport using external camera with no additional optics - the image is just initially shifted. But then the shifted image comes to the lenses of scanhead and it is shifted from the optical center, so this can produce also shrinking of the scanned image.

But what I cannot understand - why the shrinking is so much!

Sorry for this wordy explanation and many thanks in advance!

[lookin]

[lookin]

[lookin]

[lookin]

[lookin]

Thank you for reply! Do you think the problem is due to the scanning mirrors? I have tested them via CanCheck, they worked out whole range of positions and said no fails. The sounds are typical for scanning, the scanned image is consistent with no empty lines or so on. Is it possible that the fail in y-galvo can produce consistent shrinking of the image with no other visual effects?

[ShawnPatrickCasey]

No, I do not believe that the vignetting is indicative of a problem with the scanner. It seems to be an alignment issue.

Zeiss is usually pretty good with the documentation and assistance for off-warranty products. They left a 900 upright and a 980 inverted for 6 months in our shared lab… just because… well there are not as many in person tradeshows these days… Even if you can just lure a sales rep to come for a visit and have a quick look they might be able to help especially if they think a bit of goodwill may result in future business.

[lookin]

Unfortunately, it depends on the country and personal features of the servicemen. In my situation, less than year ago, our local Zeiss did not help essentially even under paid service. It was the problem with motorized movement of objective (it stopped movement at all, so we were unable even to focus to a specimen), and that Zeiss guy detached all electronic components and requested us to buy the new set (sure, being ordered in Zeiss). Then he installed these new components and went away, and a day after the movement stopped again. I forgot to say that in my country there are just few Zeiss servicemen (but a lot of local representatives and salesmen), so when I asked him to visit again, he said it would be possible only several weeks after. By our own efforts we opened the microscope and inspected the interior by endoscope. Surprisingly, we found a metal sheet, which moves along the photodetectors and serves as a signaller about the movement range, to be curved at one side. This defect physically blocked the sheet between two photodetectors, preventing any movement. And this defect should be solved by the serviceman, but it was not. After few weeks he returned back and repaired it (now I understand well that we could repair it by ourselves), but it was already impossible to force local Zeiss to refund the costs for the accessories. So, now we have unused set of electronic modules. Altogether, for that service (initial checking of the system, diagnosys, repairment, costs for new accessories; by the way, during the first visit the service payment per hour was ~120 Euro and he spent 3 days in total, the second visit was for free) we paid ~5500 Euro, and this time our system has a problem again. Actually, this is why I try initially to find another way to solve the problem, because my organization has no enough money to cover the expenses too often.

By the way, last years I have requested main Zeiss (zeiss.de) for technical assistance several times, using their web form, but no once they responded by e-mail or by phone. >10 years ago the feedback worked much better.

[ShawnPatrickCasey]

Sorry to hear that. What I am saying is that it’s optical alignment, and doesn’t seem to be physically damaged. The detectors are point detectors, and by themselves directionally insensitive. Unfortunately, the 1000 plus pages of manuals doesn’t have much in the way of troubleshooting. It’s not something that the Zeiss design engineers expect the user to attempt even if they’ve read the documentation or are familiar with. When these or most other types of commercial equipment are assembled in a factory it’s not as straightforward as when putting something together on a breadboard. There’s certain steps, tools and procedures that are specific to a particular device or subassembly that might not be apparant.

Have a couple more questions that may help you identify the area where the issue is located. First:

If you use a different channel and laser, do you have the same issue? Could you post a screenshot or picture of a sample with say both the red and green and red detectors?

Another thing is that as an example, most contracts in the country that I was looking located in, the administration holds the remaining 10% of the funds until after the end-user has approved or signed-off on an installation or repair. Don’t know what country you are in, terms of the agreement or if you are dealing with Zeiss or a local dealer.

[ShawnPatrickCasey]

Also some of the pictures and explanations are a little confusing. Mostly because there’s dirt/dust. In some of them it moves as one would expect when changing left to right (inverts). In the last two sets of images there’s a disparity.

Is there a DIC analyzer or some other slider mounted in the objective turret?

[lookin]

Dear Shawn, sorry for my late response!

The answer to first question is shown in “20211116-10X-Z0.6-AllLasers.jpg” (all examples are uploaded in the next post), where a sample was scanned with four lasers (this is autofluorescent signal). I see no principal difference between the channels, but my previous calculations indicate that the shrinking is a little bit higher at lower wavelength. To do this calculation, I located an object and scanned it in two positions - to be at the top of the scannning area and at the bottom. Then I measured the vertical size of the object and obtained two values (sizes): the size at bottom was a little bit higher than at the top (so the image is vetrically shrinked with non-linear dependence). This may arise from the more higher distortion at the bottom of the scanning area. As I described in my previous posts, this is a function of objective magnification (higher X covers less physical area, so the distortion is less) and a laser line (the higher energy wavelength will refract higher and the distortion is higher as well).

The answer to last question is YES, my microscope is equipped with DIC analyzer under the objective turret, but I use empty slot and it does not block any part of the light pathway. I also checked all the positions of all optical things and did not found any improper installation or positioning, which could affect this.

I have checked again the LSM modules. Unfortunately, I cannot conclude anything. First, I run CanCheck and found that there is the only problem - the potential conflicts of motorized control for spectral detector components (they are called as SDAL/SDAS/SDPINL/SDPINS/SDPL/SDPS/SDWL/SDWS). Here the explanation of the abbreviations:

SD = Spectral Detector (first two characters)
L = Long wavelength (last character)
S = Short wavelength (last character)
P = prism (third character, if not PIN)
W = wedge (third character)
PIN = pin (third-to-fifth letters)

Please see image “008.jpg” as an example, the red marks show this problem. Actually all these modules say they are in potential conflict with counterparts (e.g. SDAL against SDAS). When I move all to the “Safe position” (using CanCheck), the red marks disappear. However, when I then run a test for, say, SDAL, it passes the test successfully and moves to a new position, but the counterpart (SDAS) now again has a red mark for SDAL. Nevertheless, motorized components of spectral detector all passed the tests run by CanCheck or System Maintenance Tool (SMT, please see image “SMT-MotorTest.jpg”). I also compared the results in CanCheck reports obtained now and an year before, no differences were found between the entries related to the spectral detector (as well as other components).

As the image is out-of-center and shrinked, I am unable to run “Resolution Test”, “Scan Field Test” or “Scanner Calibration” using calibration objective by System Maintenance Tool.

To summarize, I am so wondering how much is the extent of image shrinking, now I do not believe that the problem comes solely from the optics. But I am in a dead end now and don’t know how to proceed…

PS About the contracts between Zeiss and a customer for servicing - in my location there is somewhat another regulation, which anyway assumes that the repairment result has to be approved by a customer. In our case last year, it was approved by us before the problem with Z-movement of objective turret next day, and then Zeiss repaired the problem within the actual contract with no additional costs. Indeed, my Institution takes 10-20% of the grants (it depends on the grantor), but this amount is not so much and it is used for many other purposes, not only for repairment. Also, while there is a Zeiss representative just in my city, the costs are still high because of the currency rate (they work with Euro). I am from Yekaterinburg, Russia.