Microscope v7 stage is jumping / skipping / rewinding

[jyellick]

I just finished assembling a v7 microscope and the assembly ultimately went fairly smoothly. Everything is mostly working, and I’m able to capture some really very nice images of some assorted sample slides, but, the stage movement (and Z actuator) is exhibiting problems.

On any axis, the movement when traversing is generally okay. I see the servo rotate the small gear and engage the large gear. And mostly, I see the sample move in the appropriate direction. But occasionally the actuator seems to jump / slip / skip back to a previous position, and visually it looks like the movement is being rewound (quickly). This is a bit annoying, but not awful in the X/Y axis, but, in the Z axis, it seems to cause auto-focus to consistently fail. I initially thought maybe I was overstretching the bands, or running out of thread or something, but, the skipping occurs when traversing in either the positive or negative direction.

My first guess was that the Viton bands had broken (and indeed, I did snap a couple of these in the assembly), but pulling on the feet the bands still seem to be firmly attached. I thought next maybe I hadn’t adequately oiled the bolt threads, but dabbing additional oil onto the bolt through the nut-slot hasn’t seemed to have any impact.

For parts, I sourced the Viton O-rings, hex bolts, washers, and brass nuts as indicated in the parts list from McMaster-Carr – so I don’t think anything should be a knock-off or out of spec. The stepper motors are generic and cheap from Amazon, but, as best as I can tell they are not struggling and things are rotating as they should.

Sometimes near a skip it does seem like the motor begins to stall for an instant almost like something internal is seizing up and then breaking loose. But the stall doesn’t seem to line up precisely with the skip (though the slight lag in the video feed makes it hard to say for sure) and it doesn’t always occur before a skip. I also can hear a slight audible clicking sometimes associated with the stage resetting. If I try to push the stage around manually, other than the plastic flexing a bit, it seems to stay in place. Looking inside the nut-slot of the actuator, I do think I can see a piece of plastic inside that rotates a bit back and forth sometimes as I reverse direction on the stage, but nothing is obviously happening to it when the skip occurs.

Here is a video that shows the camera output on my monitor, along with the movement of the X-axis gear. The behavior is similar along the Y axis. The Z-axis seems to slip more quickly, but it’s also harder to tell visually since it is the focal plane, rather than the sample movement of the X/Y.

I’m hoping that this is a known problem and something with an easy fix, but I suspect that perhaps I’ve broken something inside the main body (or misprinted it) and will need to reprint, possibly with changes. Any insights or help debugging would be appreciated!

[WilliamW]

This is an odd problem, thank you for the very clear video. From what you show in the video it is a nicely made microscope and you have good images.

In the video the motor and the large actuator gear are moving continuously in one direction and continue to move when the skip happens, so it does not seem to be in the motor. It is particularly odd that it happens on all three axes as it does look as though something is broken or the threads are stripped. My first thought is stripped threads on the hex bolts or the brass nuts. You can unwind the bolts and inspect and replace them and the brass nuts without removing the bands from the actuators. The threads can strip very quickly if not oiled, but this would be particularly quick in one day! I can’t quite think of a type of breakage in the plastic mechanism that would give this symptom.

You say that you had breakages when putting in the bands, it is not usual for that to cause breakages of other parts of the mechanism. When inserting the bands it is important to use the nut tool to jam the mechanism so that there is space left at the top to get the bands over. It is also important to get the band even in the tool and the foot at the start, so that there is the same amount of band in all four corners. This keeps the stretch in the band the same in all parts so that no part goes beyond its limit.

[jyellick]

Thanks for the reply @WilliamW – I wasn’t sure if I could safely remove the bolts without damaging things, but I’ve gone and checked them now.

Unfortunately, the bolts seem to be in fine condition.

I checked both the X bolt and the Y bolt, and aside from some dust that had gotten stuck on the oil, things seem to be perfectly true. I tried threading a fresh nut up and down the bolt with success.

Since I had the motors off, I decided to take a couple additional videos where you can see some more detail. Sorry they’re a bit shaky, but hopefully they give a good level of detail.

First, is the stage movement itself. I suspect this will not be unexpected, but, you can visually see the stage jump, especially when watching the sample relative to the objective. The stage jumps twice in this video, and, I’m simply casually turning the large gear with my thumb.

Secondly is a close-up of the actuator through the nut-hole while I turn the large gear. The stage jumps multiple times during this operation. The video shows the traversal in both directions. Aside from the slight twisting of the actuator, as best as I can tell, the actuator itself is not jumping up or down (which surprised me), but somehow the stage is still jumping.

When the bolt came back clean, I thought perhaps it might be the nut that has somehow been stripped. But, if that were the case it seems like the whole mechanism wouldn’t be working. It wasn’t obvious to me how to un-capture the nut to inspect it. But, the bolts felt like they threaded back in cleanly. Also, given that the actuator column doesn’t seem to be jumping, it seems like it’s probably not the bolt or nut?

I did try to keep the bands relatively even as I installed them. I can’t remember if I ended up snapping a band in each actuator, but it’s possible. It feels odd that I would have broken the exact same component in every actuator this way, but, stranger things have happened.

Anything additional I can do to help debug? If this is just too weird, I can always take the main body to the band saw and post some pictures for a post-mortem. I had my brother do the printing for me, so it will be a week or two before I can print another, but it’s always an option.

[j.stirling]

I think something has snapped inside the actuator. When you turn the gear you can see the actuator column turning through the nut hole, this shouldn’t happen. The actuator column should be rigid and only free to flex at its base relative to the lever arm to the stage.

If you remove the foot quickest way is to cut the band. IT would be good to know if the bottom flexure is broken/damaged. If not whether the actuator is broken higher up.

If you can cut the front off the actuator housing (where the letter is in X/Y) then it might be possible to see where/how it broke.

It could have been a weak layer in the print causing them all to be very fragile, but that isn’t clear on the outside. Other thing it might be is if the bands are the wrong size, or the less stretchy than ours. I am sure that McMaster will have supplied the correct thing (oh how I miss McMaster Carr!!). One thing we have seen is there are lots of different types of viton and the elasticity and elongation can vary, this is something we haven’t found how to spec very reliably yet. Do you have the McMaster Carr product number?

[jyellick]

For the McMaster Carr part, my order says:

Chemical-Resistant Viton Fluoroelastomer O-Ring, 2 Mm Wide, 30 Mm Id
9263K119

I was all ready to cut open the main body to take a look at what was going on, but when I went to stretch the X foot open to cut the band, the actuator came sliding on out. For the Y and Z feet, it took a little more effort, but the same separation between the actuator and the flexure allowed me to remove them without having to physically cut anything other than one of the Viton bands.

The thing that sticks out to me about the actuators is that it seems like the bolt has threaded itself into the bottom of the actuator. Maybe this is a case where I over-tightened the gear into the actuator and basically sheered it off the flexure? In retrospect, I seem to recall the stage being very misaligned with the objective during assembly, and then things ultimately snapping back to center at some point.

During the disassembly I also noticed that the flexure seemed to separate from the very bottom of the print, does this indicate a problem with my print settings?

I was able to get one of the Viton rings out whole. It seems to have stretched a mm or two vs the remaining ones I got in the bag. Not sure if this is expected, or, cause for concern.

[WilliamW]

Oh dear, that is not great. Thank you again for clear pictures.
Going back to your earlier post the actuator screw in the photo in your second post looks very long. The thread length should be 25mm, which is similar to the diameter of the large gears. That one looks to be quite a bit longer, which would then be able to push the bottom of the actuator out. Failure of the layer adhesion is a potential problem in the flexures at the bottom of the actuators, as you have seen. It is not clear to me whether that was the only problem, or whether the root cause is bolts that are too long.

Some permanent elongation of the bands is expected after installation.

[jyellick]

Ah, indeed! I just checked my invoice and the bag of hex head bolts and they both indicate a 35mm length, rather than 25mm. So, this is certainly at least part of (if not the primary source of) my problem. I discussed the print parameters a bit with my brother, and he thinks he has some tweaks in mind to help with the layer adhesion, so perhaps that will help too.

I just re-checked the bill of materials CSV and HTML and both indicate the correct 25mm – I guess I must have misread or mistyped/clicked when ordering. Sorry to have required so much of your time debugging such a simple mistake.

Thank you so much for your assistance! In a few weeks when I have everything re-printed and reassembled I’ll be sure to post a full build report in that section of the forum.

[WilliamW]

I am glad that there is a simple reason. I just had a delivery of DC barrel connectors and found I had carefully ordered the plugs not the sockets.

If the bolts are 10mm too long I expect that that is your only problem and your layer adhesion is fine. You should be able to cut 10mm off a few bolts. Thread a nut on first so that it can re-shape the thread after cutting. If you want to test your printing and actuators, you can print the separate z-axis from the upright version of the microscope. That is quite a quick print and is the same actuator mechanism as on the main body.