geert_2

I usually print my medical models with 100% infill, which by nature very much reduces the possibilities of holes and leaks, althoug of course there are always thin "canals" of entrapped air in-between the extruded sausages, that are not filled. You can not avoid this. Printing slow also helps, and with enough extrusion: better a little bit over- than underextrusion. Also, user cloakfiend's acetone smoothing works very well: this tends to fill tiny gaps. Search for acetone smoothing on this forum. Further options: spraying a thick varnisch? Or dipping the model in it? However, for cell cultures, I would rather try to find commercial injection moulded containers in PP or PE (polypropylene, polyethylene): these repell water, and are chemically quite resistant. Or some other injection moulded containers that can be autoclaved. I wouldn't like the idea of cells, bacteria and chemicals getting into the little holes in the plastic of 3D-printed models, and contaminating everything. If not commercially available, and you need enough copies of the exact same part (let's say a few thousand), it is worth looking into custom small-scale injection moulding via prototyping companies like Shapeways. They can make injection moulds in aluminum, for up to 10 000 pieces, and run the production. Might be way cheaper than 3D-printing. You make the design (according to injection moulding rules, thus with draft and equal wall thicknesses), you make a few 3D-prints to verify if it works well (very important), and then send the design online to them and they do the rest.

No shame on you. We do not always have the time to fully explain things. Further, I am working in an educational institution (university), where we are by law required to: "provide education, research and service to the community". Often I also have to provide explanations and guidance to collegues and PhD students about computers, software, and laboratory equipment. So I am used to it, and it fits within my job. Also, I enjoy sharing knowledge, so I don't mind doing this at all. :-)

If you use a 9V battery, and you put the LEDs in series, it will work with the same resistor value, but just give a bit less light. But usually I would prefer to recalculate the resistor: voltage over resistor = battery voltage minus first LED voltage minus second LED voltage. Vr = Vs - Vled1 - Vled2 For an educated guess, that would be: 9V - 2V - 2V = 5V over the resistor (as a crude order of magnitude). Use two identical LEDs. And then calculate the resistor, based on the recommended current through the LED. If it is a high-efficiency LED with low power-consumption, the current could be 1mA. Then the resistor would be: R = Vr / I = 5V / 1mA = 5KOhm If it is a medium efficiency LED, with a current of 5mA, it would be: R = 5V / 5mA = 1KOhm For a LED of 10mA: R = 5V / 10mA = 0.5KOhm = 500 Ohm. This value does not exist, so we take a nearby very common existing value: 470 Ohm, or 510 Ohm. For a very old LED, or a brighter LED that requires a bit more current of 20mA: R = 5V / 20mA = 0.25 KOhm = 250 Ohm. This does not exist, so we would take 240 Ohm or 270 Ohm. For the existing resistor values, Google for: E24 resistor series Among the images in Google, you will then also find the color codes. If you are not familiar with electronics, avoid very high power LEDs like those used in spots or in traffic- or billboard signs. These may get very hot and require special cooling and mounting features. Tiny low power LEDs like in keyboards, stereos, etc..., don't get warm if a correct resistor is used to limit the current to the recommended value. If you want to buy new LEDs, search for low power high efficiency LEDs, because: 10x less current = 10x longer battery life. Search for 1mA or 2mA LEDs, provided they give enough light for your purpose. This may sometimes be very hard to guestimate from the specs, so you may want to try various types. It is a long time ago since I bought LEDs myself, so I can't say what is on the market today. About 20 years ago a typical low power high efficiency LED of 3mm diameter consumed ca. 2mA, but that is 20 years ago... And LED-voltage was 1.65V for red LEDs, 1.9V for yellow, 2.2V for green. Hence my 2V guestimate, which is usually okay for red, orange, yellow, yellowish-green. However, blue, white, and "traffic green" LEDs (=blue chip with phospor on top) usually are around 3...3.5V. But different values exist. Look them up in specs of distributers like RS-components, Farnell, etc... Mr. Google is very helpfull today (credits to the original photographers): However, solder the wires, instead of wrapping. Or try these boards for experimenting: they are very handy, but watch out for short-circuiting wires: With this board you can try lots of different resistor values and LEDs in a short time. I used them a lot. Note: LEDs have virtually no internal current-limiting features: if you apply a too high voltage without external resistor, or if you short-circuit the resistor by accident, the current can get very high and immediately burn out the LED. Don't ask how I know... :-) So, *always* use a separate resistor for current limiting, and never rely on the very unpredictable internal resistance of batteries, power supplies, or LEDs. Further, when plying the leads, use a plier to grip the wire close to the LED, and bend it around the plier, on the side away from the LED. So don't put too much mechanical stress on the plastic housing of the LED. Pulling hard on the leads of tiny 1mm LEDs could cause them to break, since the plastic is not very strong. Here too, don't ask how I know... :-) These are a couple of very good educational Youtube videos on this subject: - https://www.youtube.com/watch?v=Bozb8t6d1Xk - https://www.youtube.com/watch?v=Yo6JI_bzUzo - https://www.youtube.com/watch?v=NfcgA1axPLo - https://www.youtube.com/watch?v=VSpB3HivkhY This reply is a bit longer than I intended, but fortunately I can type very fast. Also, I realise this is a bit off-topic concerning 3D-printing, but I think it is close enough. It may also be usefull for people who want to modify their 3D-printers to mount some indicator LEDs in it. For example you could mount a LED on the bed heater or on the (cold side of) the print head and nozzle, to see when it is on. Carefully calculate resistor-values (Ohms) and power-ratings (Watt).

Now that you say this, I remember: in the beginning I used to wipe the nozzle with a tissue wetted with silicon oil, and PTFE-oil: sometimes the first, sometimes the second. These sprays can be found in car shops. This also reduced accumulation. It seems that after some time the nozzle gets a coating of PTFE and/or silicone, and it gets less sticky. There is still some build-up of goo, but less. So, now I don't need to wipe them with oil anymore; I just clean them immediately after each print.

Google for "simple led circuits" and then select "images". This shows the setup. Always keep in mind: LEDs do need a resistor to limit current, otherwise they burn out! Usually the voltage over a LED is between 1.6V (old red LED) and 2.5...3V (blue and white LEDs). The recommended current for a nice illumination can go from 1mA to 10mA usually, depending on the LED. Don't come near the maximum current through the LED, always stay well below 50% of the maximum. So you need to look up the specs of your LED, or measure them: - normal voltage over the LED= Vled = ? - recommended current through the LED= I = ? What battery or charger are you going to use (I would recommend a 5V or 9V charger): - sourcevoltage = Vs = ? And then calculate the resistor as follows: 1) resistorvoltage = sourcevoltage minus LEDvoltage = Vr = Vs - Vled 2) resistor = resistorvoltage divided by LED current = R = Vr / I 3) power dissipation in the resistor = current multiplied by voltage over resistor = P = Vr x I Example: Imagine this are the specs: - Vled = 2.2V (=voltage over LED, from the specs of the LED) - Vs = 5V (source voltage, as usually found in chargers for charging USB devices or smartphones) - I = 5mA (=recommended current through LED in the specs) Then: 1) Vr = Vs - Vled = 5 V - 2.2 V = 2.8V 2) R = Vr / I = 2.8 V / 5mA = 0.56 kOhm = 560 ohm (take the closest available standard value) 3) P = Vr x I = 2.8 V x 5mA = 14mW (then add some spare: triple this value and take the next higher available resistor series, so it does not get hot: for example take a resistor of 250mW, a very common series) 4) add an on-off switch. That is all. You values may be somewhat different, but this is the principle. Basic scheme But do recalculate the resistor value according to the specs of your LED and your sourcevoltage or battery voltage!!! It may differ. Usually the long pin of the LED is the plus-terminal. And the pin connected to the "dish" inside the bulb is the minus-terminal. Usually, but check it. It only works if you connect the plus-terminal of the LED to the plus-terminal of the battery or source, not vice-versa. Plastic LEDs like these can be grinded or reshaped with a Dremel and cutting disk, as long as you don't hit the wires and chips (also not the very thin wire on top of the chip). But they do get fragile. I used to do that in model trains and cars, to make them fit. But don't cut/drill into modern white LEDs. Typical resistors. The color bands indicate the resistor value. Google for it. That is all there is to it. Use a battery charger with short-circuit protection. And/or add a mini fuse yourself. (All pictures via: "Google --> Images". Credits to the original photographers/designers.)

The PET I used (which I think is a sort of polyester, just like CPE?) has a tendency to accumulate on the nozzle while printing, and then this accumulation discolors into a thick light brown goo, which eventually sags and gets deposited on the print as big brown blobs. Also, while traveling through space, the nozzle has a tendency to leak, and then upon arriving on the next wall, a blob is deposited on the side. Especially when printing fast, due to the nozzle-pressure not reducing immediately. In my UM2, printing slow reduces the effect, but does not eliminate it. I don't know if this is the cause of your phenomena, there could be other reasons, but maybe you could keep watching while printing a test piece? If it would be the above problems, you can easily see it happening.

I don't know about the UMO, but on the UM2 the bushings definitely need oil. Oil does not only reduce friction, thus it prevents metal on metal wear, but it also allows trapped dust to be removed, thereby again reducing wear. Officially the UM2 rods need thin sewing machine oil. But I found that this dries too quickly (may depend on oil brand and composition), so now I use a high grade hydraulic oil, also used in industrial applications like hydraulic test benches, tractors, bulldozers,... This oil does not dry out at all, it lubricates well, and it contains anti-corrosion and anti-foam additives. Not sure if this is the best solution, but it works for me. One of the other reasons why I use this, is because I have a lot of spare of it. Concerning bent rods: maybe you could see if that shows up if you print a thin test layer of 0.1mm? It should cause a wave-pattern in the thickness, of the same distace as the circumference of the rods?

Can't you tighten the clips carefully with a plier? But I agree, a sort of quick lock mechanism like in camera-equipment would be good, similar to this (the one on top). Although I don't know how well this holds under repeated vibrations for days, which we have in 3D-printing. At least, it would require a strong spring, not just mechanical friction, to hold the lock.

I print most models in PLA (95%), and they are functional: good enough for use in the hospital, and good enough for mould making. But they can't stand the heat of a car interior, even not in mild spring or autumn weather: then they will deform. And of course desinfection has to be done chemically, not by autoclave. Also, PLA is very hard to drill into, or tap threads into, because it melts. Functional items that need flexing, like carabiners or snap-fit locks, will break over time: PLA is not flexible enough, and it hardens over time. Creep is also a factor in PLA: if tightened fast, it will permanently deform over time. This is also why threads in PLA don't work well: they tend to dislodge. Now I use nylon nuts and screws, or glue, to fix PLA models. If more flexibility is required (e.g. for snap-fit locks or carabiners), or for in the car, I use PET. I haven't tried nylon, PU, or PC yet. The cream carabiners below are PLA/PHA (colorFabb) after some time of daily use and flexing. You see where they crack and deform. The green one is PET, which is flexible enough for this application. However, when pulling very hard, the PET snaps before the PLA. So it all depends on the exact application.

Yes, then he would better concentrate on studying the methods to grow his plants well, how to prune (trim) the leaves and bunches of grapes, how to making the grapes taste well, make the wine taste well, and make it conserve well without turning into vinegar. This is not simple and there is a lot of knowledge involved: grapes need special soil, climate, and treatment. The wine-production itself requires even more knowledge, to prevent it from rotting. Making vinegar or rotten fruit juice is easy, making good wine is not. :-) We had grapes, although we never made wine. But in our neighborhood a couple of guys started wine-production from scratch, from bare soil. If I remember well they had to follow specialised evening courses for a couple of years to get the knowledge to make it really good and to cover all aspects. But they made it, so it is definitely possible; a beautiful hobby.

It does exist: in my old Cura that function is called "Print one at a time" (versus the standard "Print all at once"). I don't know the name in newer Cura-versions. However: this needs more room around the piece, to prevent the print head from crashing into already printed items. And if the models are very small, you will run into cooling problems: the nozzle staying on top of the model at +200°C, radiating heat, prevents the model from solidifying and cooling down. Increasing minimum layer time does not help in this case, since the nozzle just stays on top of the print. Moving it away is explicitly what you wanted to avoid. You might get this effect: the left models are printed without dummy cooling tower; the right are printed with dummy cooling tower. That dummy is only to allow the cone to cool down. But even in the white dummy tower, you can see that it had not enough cooling at the top.

For PET from the brand ICE, I usually print between 215°C and 225°C, slow at 25...30mm/s, and with a bed temperature of 80...90°C. No fan if the model allows it. If I need to use a fan for overhangs, then a higher nozzle temp works better. If I need a fan, then I need to use glue to prevent warping, otherwise I print on bare glass. Most of the time I use the "salt method" (=wiping the glass with salt water, and let that dry into an almost invisible mist of salt stuck to the glass), which slightly reduces bonding for PET, but it makes it much easier to remove the models afterwards, with less risk of chipping the glass. (While for PLA, the salt method increases bonding.) PET goo accumulating around the nozzle, and then sagging and leaving brown blobs on the print, is something I also see. The effect lessens when printing slower and cooler.

Depending on the software you used to design the tubes, "deleting" the hollow insides (thus making them solid) might be as simple as just clicking the inside area once, and then press the delete button. Takes a few seconds per tube. In some programs you might even be able to select "all similar hollows" (or something equivalent) after clicking the first inside, and delete all in one shot. It might be worth looking into that?

Wow, that red one looks really smooth, with absolutely no layer lines visible. Did you sand it prior to acetoning, or was it straigth out of the printer?

You could try printing a short, straight section of a round tube sideways, for test. Print rather cold, and with max fan, or even with an additional desktop fan in front for fast cooling. The roof-section of the inside might look a bit like it has grapes hanging from the ceiling, but the canals should stay open. Try the concept on a short straight test piece first, to see how much the inner roof sags, and if this is still acceptable for you. PLA can usually bridge gaps quite well. This is a test of a table model, with custom support structure. At the underside of the support structure, you see how much it sagged. The middle section overhang is 30mm long. But you can do better by printing colder. So, if such an amount of sagging would still be acceptable to you, you can print it without internal supports. The external part will need supports and good brim, otherwise it will fall over while printing. So you might consider some support structure similar to the pink support in the blue spring, and then cut that out later.

In my opinion, although I don't drink wine: wine does not need fancy bottles. It needs to taste well, conserve well, and be resealable after opening. There are already a lot of beautiful standard glass bottles for wine, so it is best to stick to proven quality standards, I think. If he wants the bottle to be eye-catching, which of course I understand, maybe he could design a really beautifyl label with some gold-lined edges and logos? And have that printed professionally in a label-printing factory? These companies are used to do gold-lining, relief-printing, self-adhesive labels, custom cutting, barcodes, and all such stuff. This would give a much higher impression of quality than a 3D-printed bottle, which will always look somewhat poor and amateuristic, sort of "stuck in the prototyping phase".

Keep in mind that the salt method works for PLA and PLA/PHA only. It does *not* improve bonding for ABS, PET, and probably most other materials. Even for PLA, use it for low and wide models: this works very well for me. But not for high models like towers and lantern poles: they tend to be knocked off: the salt-bonding obviously can not absorb repeated shocks very well (e.g. the nozzle banging into curled-up overhanging parts, especially on tall models).

I do this manually, for example: "model_v1234_pet.gcode". But on my UM2 the name can be maximum 20 characters; the rest is cut off on the little display. So I often have to abbreviate the model name and version number. Depending on the priority, and if space allows it, sometimes I also include other specs such as layer height (if deviating from my normal 0.1mm), or the amount of models on the build-plate (if more than one). E.g.: "mod1234_02mm_3x_pet.gcode". This comes in handy after a couple of weeks, if I have to print the same model again.

Does it have to be a circular tube? Or would a pentagon-shape also be okay? Or a triangle on top of a square, like a house-symbol? In that case, you could print it sideways, with the flat area towards the bottom, and the "roof" on top. I will try to insert a unicode symbol here, if the system allows it: ⌂

Doesn't coldspray risk fracturing the glass due to thermal shock, or fracturing layer-bondings in the model?

If you want dampers, I would recommend using professional and tested damper mats, of which the damping specs are listed. Each material has a resonance frequency, also dampers, in which they absorb less energy. A wrong "damper" could cause the opposite effect: the thing getting into resonance, like on a spring. So I wouldn't trust a spring-like design like that. Such damping mats should also be antislip, as gr5 says, so the printer can not vibrate off the table (we have seen a couple of these here on the forum...) In the early days of harddisks (late 1980's) I have seen harddisks crash *because* they were mounted on dampers: the at that time very heavy heads swinging around caused resonance. We had to mount the disks directly and firmly on the metal frame. If the effect you notice is "ringing" after a 90° corner in a print, this is due to the head changing direction suddenly. It won't help putting a damper under the printer. Then a solution could be to tighten the mechanical tolerances of the rods and bearings, and mounting the steppers directly on the driving rods, instead of via belts, I think?

For steep overhangs you need to use supports, or redesign the part so it has no steep overhangs. Or split it and glue both halves together, or print it in such a way that there are no overhangs while printing (if the design allows this; probably not here). I don't know your printer, so if it is a single nozzle printer, you would need supports in the same material as the print, and cut them out later. In dual nozzle printers, you can use one nozzle for the support material, and the other for the print, and dissolve the support later. You could design your own custom supports, if the model has special requirements, like I did in the test below. Or use the standard Cura-supports for general models, which is far easier. Try this on small test pieces first, before doing a big print. An example of custom supports, in order to not destroy the text below the bridge (standard supports would go all the way down). For reference: the main plates of this bridge are 1mm thick; text caps height is 3.5mm.

Does this effect also happen when you print a simple square tower, without any details, and without curves? In single-wall shels, and double wall shells? If yes, that would definitely rule-out model-geometry specific causes. I have no explanation for this effect, since normally PLA is supposed to print better when cool? At least in my experience... Or are there strong and variable air flows in the environment of your printer? Which would cause sudden temp changes of the nozzle? What if you put a desktop fan in front of the printer, to provide plenty of cooling air: does it get worse, or better? Or vibrations that only occur when the nozzle is printing in front?

Okay guys, it's time to get your red/cyan 3D-glasses out! :-) I tried making a few anaglyphs (=red/cyan 3D-pictures) of a 3D-model. These 3D-images were hugely popular in the nineties, but I haven't seen much of them lately. They give a better understanding of the internal structures of a model, like watermarks; and give a better perspective and feel of depth. Also good for microscopy images of cells and fine structures, where the 3D-effect is important for understanding. They are reasonably easy to create in a 3D-editor: - In the 3D-editor, shift the model to the left of the screen, in perspective mode, by using the "Pan" function. This is the view that your right eye would see. Save this view as a picture, and name it "right-eye-view", or so. The 3D-editor should be in a real "architectural perspective" view mode, not a technical "ISO-perspective" with parallel lines but without perspective. It is the real perspective that makes this work. - Shift the model to the right on the screen. This is what your left eye would see. Save this view as picture and name it "left-eye-view". - Do not rotate the picture on-screen, only shift using the Pan-function. - It is best if the background is pure white. - The red glass is on the left, the blue glass is on the right (think of politics: red=left, blue=right). Actually, the "blue" glass should be a intense dark cyan, not real deep blue. And the red glass should be a deep pure red. - Red text or images are invisible through the red glass, because everything is red anyways; you can't see the difference. Only cyan (the opposite color of red on the color circle) is visible as black through the red glass. So, we need the left-eye-view to have a cyan color, to be visible in the red glass. - Idem for cyan: cyan images are invisible through the cyan glass. But red images show up as black. So we need the right-eye-view to be red on the screen, and appear black through the glass. - In an image-editor capable of layer-editing (e.g. Photoshop, GIMP,...), load both images in different layers, in a new picture. - Make the left-eye-view bright cyan, by setting the Green and Blue channel's output levels in the Levels-dialog to maximum, thus: R,G = 255 dec, or FF hex, or 100%, whatever notation your editor uses. Leave the Red channel untouched. - Make the right-eye-view red, by setting the Red channel's output-level to maximum. Leave Green and Blue channels untouched. - Set the top layer of both layers to "Multiply" mode, so the layer below shines through. This gives the composite red/cyan view. - Now shift both images to the center of the picture, and correctly align them for a nice 3D-effect, when seen through the red/cyan glasses. - If the areas that exactly overlap each other are in front of the model, the model will seem to sit behind the screen (see my examples below). - If the areas that exactly overlap are in the back of the picture, the model will appear sit in front of the screen, and you need to focus in front. - So you can move the model from behind the screen to in front of the screen, by shifting the red and cyan layers in respect to each other. - It works best with monochrome images, but it also works with very desaturated colors. Bright colors do also work, but only if they are ca. 90° separated from red and cyan on the color circle, thus yellow-greenisch, and purple-blue. So, landscapes with fresh spring-green and with a desaturated cloudy sky also work. - On a good screen and with good glasses, there should be not too much "halo" images, where the wrong color shines through the glasses. - It works with renderings with and without black edge-lines, and to my surprise it also works in inverse images. - You may need to lean a bit forward or backward (25...40cm?) to get the best effect. See some of the tests: Model behind the screen, opaque. Model in front of the screen. Model behind the screen, transparent. Model behind the screen. This gives a good view of the internal watermarks and rulers, and of other openings. Very pale model, in front of the screen. This works also, to my surprise. Model in front of the screen. Same model, but in inverse colors. This one is more prone to "ghosting". This is just a technical model, a keychain of it. But for really artistic models like statues and buildings the effect should be way more impressive.

What if you print this in PLA? Does it come out right, or does it also have the same visual defect at the same height (although it might not break due to better bonding)? And what if you print this on its side? Does it then have a defect at the same height above the build plate (which could indicate a problem with the Z-axis), or does it have the same defect now running vertically (indicating a model defect)? When moved manually, does the build-plate or Z-axis move smoothly up and down, without hard, stuck points? Is the bottom of this model open, so you can insert the nut from the bottom? Or did you pause the print halfway, move the head away, insert the nut, and then continued printing? If so, I could imagine this has some influence. Do all print-settings (perimeter, infill) have the same speed, temp and other settings? ...Justy trying to narrow-down the cause...