Hi,
Just finished building motorised microscope with low-cost optics. I used 0.8:1 ratio gears for speed. Also since I didn’t have sangaboard and illumination pcb, I used alternatives, and used white teflon tape instead of pmma lens (I recall that i read it somewhere, but i’m sure i’m not right because live picture seems too grainy). What can i do to correct shaky movement of stage (i added a video link, and sometimes it doesn’t move at all). I also added a captured image, microscope seems not focus on left upper side and i see a red spot when looking at stream at empty stage.
For slide scan, i recently watched tech demo on youtube and there is slide scan option with live stiching preview. I also recall seeing it Rwanda post. Is there a way to obtain slidescan option?
Thanks in advance!
Stage movement: Watch WhatsApp Video 2024-07-24 saat 22.19.09_b1220a07 | Streamable
1 Like
Well done on your build @zealkarel.
There are several points in your post, I shall try to address some of them.
A scatterer, like white PTFE tape, instead of the condenser lens at the position of the lens should give a large illuminated area with a reasonably large numerical aperture. The downside would be that much less light would make it into the camera lens. A grainy image might be that the gain has to be very high on the camera. With the low cost optics the illumination NA required is quite small, so a condenser lens with longer focal length should be OK, if that is easier for you to find.
The motion you show is diagonal, but seems to travel in a bit of a staircase trajectory. What happens if you move in only x or only y? Is that smooth or jerky? What are the occasions where it sometimes does not move at all? Overall the motion should be fairly smooth but not perfect. Your symptoms may be that one of the actuators is partly broken.
A coloured spot when you have no slide in place is the white balance and flat field correction not quite working properly. In the settings tab you can re-do that calibration. I would first disable the flat field correction to see the uncorrected illuminaton, you may then need to run the auto gain and shutter speed. Then check that the illumination is reasonably good over the entire area, with no really dark patches. It does not need to be fully uniform, but there does need to be at least some light in all areas. Then run the calibration again.
Poor focus in one corner could be from several different sources. The stage might be not quite flat - if you pack up one side of the slide with paper can you make it better? The lens may not be quite straight in the lens spacer part. The camera may be at a small angle relative to the camera board.
The stitching is discussed in another thread.
Good luck with further testing on your microscope.
1 Like
Hi again, thanks for reply @WilliamW,
I bought some lenses from aliexpress with specifications diameter 13, focal length 8. But it was no edge lens, so it didn’t fit. I purchased different lenses again, but didn’t receive them yet. I’ll try to update when shipment arrives. Can i try to fix lens with some colorless tape?
I tried to move only x or y this moment, x axis movement seems normal, but y axis sometimes doesn’t move or too shaky. Don’t know what to do, i had so many failures to this point. This is 4th printed main body.
Video: Watch ScreenRecorderProject1 | Streamable (sorry for watermark).
Coloured spot and focus doesn’t seem to fixed with re-calibration and packed up paper, respectively.
Again thanks in advance.
Starting with the colour. If what you are talking about is the slight ‘blotchy’ pattern in your lower image, then I think that is from low overall light level. The flat field overall looks to be well corrected, and the patch size is uniform across the image and about the size of the correction cells in the algorithm. It is also pinker in some cells but greener in other cells, which looks like noise. Low light means that noise in the calibration images makes a difference to the calibration, and the next image has different noise.
For the condenser lens, you could glue or tape a lens in the condenser body. It is not a precise illuminator.
Finally jerky motion on y not x unfortunately does suggest damage. What were the symptoms of you previous main body failures? Have you got any pictures? It can really help the project if we can understand failure modes. If we know what is happening we can think how to change the design or instructions.
1 Like
It’s worth mentioning that sticky stage movement can result from missing out the washers in the actuator assembly, or from not having lubricated the long M3 screw that goes into the brass nut. It’s often tempting to skip those steps, but they make a big difference.
Various people have also reported issues with the motors - if you think one motor is better than the other, it’s easy to check by swapping them.
1 Like
Hi again, thanks for replying @WilliamW and @r.w.bowman,
For colour as i mentioned i was using white teflon tape it seems it doesn’t allow enought light for illumination. I’ll try with again condenser lens and post pictures again. I also plan to build with a 40x objective (actually two low-cost, and two with a 40x objective). Will condenser lens with focal length 8 result in good image with 40x objective?
For main body failures, first one stage same symptoms, stage sometimes didn’t move and at some point, i heard a snapping noise and assumed body failed but later realized that didn’t cut ties for rear legs and i think that was the snapping noise. Since i was new to 3d printing (my first print was just_leg for printer test), assumed print failed, and printed new one, same problem stage was not moving as expected, printed new one, same problem. I have bodies 2nd and 3rd print also this one. Will try to dissamble actuators and post pictures.
As for the @r.w.bowman mentioned;
Checked for washers they seem in place. Also lubricated M3 screws but maybe didn’t use enough oil. But since i couldn’t find brass nuts for actuator assembly was using DIN932 A2 grade m3 nuts. Can this be the problem?
I have motors to spare and will try them too and post results again.
Again thanks for replying.
Hi again,
Sorry for replying too late.
Shaky movement was due to ripped viton bands (i think). I disassembled actuators and noticed blotches on the top walls. Also later i figured out my band assembly tool was not symmetric (right side didn’t fit well to holder, printed again now fits well). Actuators seems fine. Now trying to assemble microscope again will post again later how it works. I added photos for top wall and actuators.
1 Like
Hello again,
I apologize for bringing up an old topic, but despite my long-term efforts to resolve these issues, I have unfortunately been unsuccessful. To summarize, I am still experiencing significant problems with stage movement. I would like to explain my current situation in detail and would greatly appreciate any help.
I have built a total of four OpenFlexure microscopes: two “low-cost” versions and two equipped with 40x objectives. For 3D printing, I use a Creality CR-10 Smart Pro. Recently, I’ve been using Creality HyperPLA filament (previously, I used local brands).
Here are the technical details of my builds:
-
Gear Ratio: 1:1.25 on all microscopes.
-
Electronics: I am using a compatible board instead of the original Sangaboard.
-
Power: The board is powered directly from a Raspberry Pi 3A AC adapter.
-
Hardware: During the initial build, I used stainless steel nuts instead of brass nuts for the X, Y, and Z movements.
-
The Nut Issue: I recently sourced brass nuts, but they appear to be thicker and wider than the stainless steel ones. While I managed to force them into a “nut test print,” they simply do not fit into the slots of the main body.
-
Versions: My most recent build uses the Beta 5 and f0799cf main body versions. Three of my microscopes are the Upright version.
Troubleshooting steps I’ve taken:
-
Motors: I tried swapping the motors, but it didn’t solve the issue.
-
Bolts & Lubrication: Thinking they might be worn out, I replaced and re-lubricated the M3x25 bolts. For lubrication, I use standard machine oil (commonly used for sewing machines in Turkey).
-
Scanning Failures: I have attached images from my slide scan attempts using Beta 3 and Beta 4 versions. In today’s scan, the stage stopped moving entirely during the last 4-5 areas, so I had to cancel it. As you can see in the images, although they are labeled as different fields of view, the stage did not actually move.
The stage does not move as intended, which makes automated slide scanning impossible. I suspect that not using brass nuts is a major factor, but the brass nuts I currently have are too large for the printed slots.
Is there a known way to modify the main body or the nuts to make them fit? Or is there a specific recommendation for the “stainless steel vs. brass” issue in this context?
Thank you in advance for your support.
Test scan with beta4 version.
Today’s test scan with beta3 version.
Last 4 images from failed scan.
Brass nuts i managed to source. Dealer mentions 5 mm in outer diameter.
Thank you for your perseverance!
I shall deal with the simple part first: Your stainless steel nuts are made to the DIN934 standard pattern, the brass nuts are not. To fit in the body you will need brass nuts to the DIN metric full nut standard, 5.5mm across flats and 2.5mm thick.
Your other problems are slightly intermittent, which makes them very hard to diagnose - a lot of your scan looks fine. The essence of what you have shown is that the stage clearly does move in x and y, and is able to focus, so it is moving in z as well. However you are seeing motion that is not quite as expected.
Wear from the steel nuts and steel screws in the actuators could give these symptoms, but you would expect that to have got worse over time, and you saw problems early. The wear can be very quick if you are scanning and have not got lubrication. You need very little quantity of oil, but it does need to be in the thread where the screw goes into the nut.
When you say that you swapped motors and it did not solve the problem, did you swap for three new motors, or did you keep the same three motors and swap which axis each motor was on?
When you have disassembled bodies, have you noticed any issues at the base of the actuator columns? One of your earlier posts has the columns detached where you took it apart. The base of the column has quite a lot of force on it and relies on good layer adhesion to remain in contact. Others have seen breakages in that area, for example the first image linked in the first post on the thread Does this mean I have to reprint?, and the solution in Microscope v7 stage is jumping / skipping / rewinding was that the hex bolts were too long and had pushed the bottom off the actuators. These symptoms are similar to yours. If you have a body that has had problems could you look at the area at the bottom of the actuator? or remove a foot on your problem microscope and post a photograph? You say that you are new to 3D printing and that there were problems with your tool print etc. Layer adhesion depends on printer setup and tuning. It is not something that I have needed to attend to so I do not have advice on what parameters to change, the printers that I have bought were all well tuned already.
The other question is how much you see this problem when moving the stage with keyboard commands. Is there always a problem, or only sometimes? Does it happen more when the stage is far from centre? Is it a jumping/skipping (like Microscope v7 stage is jumping / skipping / rewinding) or sometimes just not moving? This will give more detail of the issues than the scan failures where there are many things going on at once.
Hi,
Thanks for quick reply. I’ll try to answer all the things one by one.
For brass nuts: I’ll try to source DIN934 standart nuts. In meanwhile if I force nuts i have at hand to main body will it break something structural? I was lubricating screws outside since @r.w.bowman mentioned it can cause sticky movement.
For motors: I swapped all three motors with new ones in one of my microscopes. Since it didn’t solve problem I didn’t swap motors from other microscopes.
When i disassembled bodies there was no structural damage, here is pictures from underneath the microscope and,
Since it is black and i can’t get better pictures (sorry) it is not clear but there is no damage.
And here is length of a screw i used in building,
For microscope control, if i use keyboard to move stage it sometimes doesn’t move, sometimes makes a half move (stuck for like 0.5 seconds then move). It isn’t always stuck and it is not specific areas like far from centre. For slide scanning I’m trying to scan 5x5 mm samples and try to centre slide to camera as much as i can.
Thanks in advance.
Than you for the detail. As you say, being black makes it hard to see from the photos. The weak point is between the flexure that we can see in the photos and the column, which is underneath in the photo.
When the stage does not move for half a second, do the motors / gears start to move at the beginning or are they still for the half second as well?
If you take off the motors and operate the gears by hand, does the sample move smoothly and as expected?
Probably not the issue, but why is this region looking like that? It kinda looks like it wasn’t squished to the bed / barely stayed on in that spot. Maybe something shifted around? Or is this how the entire bottom looks like, it just hit the light differently?
Also, depending on how willing you are to print, maybe you could print the main body and use a steel nut for testing a run or two. Probably won’t work in the long term, but they have the correct dimensions and all. If the issues disappear, you’ll at least know that getting the brass ones fixes it. I’d also take out the motors from the microscope, place them on the table and just try moving it a bunch. If they get stuck, then you might have a power issue or the board doesn’t work well sending the gcode or something.
Hi again,
When stage doesn’t move i hear motors engaging but no movement, it gets stuck sometimes but motors engaging with every input.
I did take off the motors, the motors i’ve tried (attached to one of the microscope) move with inputs without problems, i can move the sample with hand smoothly.
Example of a scan with this microscope.
I suspect movement sometimes get stuck and it confuses the stitching.
To @adamglia, it’s one of my earlier prints. Not squished well but not what like seems in the pics.
I have a spare main body. I’ll try to assemble new microscope with new screws and nuts. I’ll also try all the motors i have before assembling.
Again thanks for replying and in advance. 
Check the Raspberry Pi temperature and the temperature of the motors during / after a long scan session, or when you see the issue occuring. I saw issues with my prototype when it was overheating, but I chalked it up to the control board (even though I had a thermal alert up on the RPi). Maybe it’s the Pi that doesn’t like heat and is losing it
