Best Auto level sensor & Extruder?

Quote
SupraGuy

I'm willing to accept that the heated bed must remain heated for the duration of the print. I do that anyways for consistency.

I as well.

Quote
Ed3D
^{(Probably a repeat thread so apologies if it is!)}



Also what do you think is the best option for an extruder? Ive been looking at things like the nimble or the titan. Any thoughts? Anything open source?

Myself, I decided to go with a Bondtech BMG.

I kind of got turned off of polycarbonate used in 3D printers after E3D ran into problems with the original titans cracking. It ended up being the oil used in the bearings would cause the plastic to craze. They fixed the problem, so I'm sure it's not an issue if you want a titan.

Red Herring as in all of a sudden people are talking about granite beds, some may feel they are missing out on something.
Some may experiment because of it, I just keep moving it around till it fits, and right now it fits as a base(i'll post pic when I have one).
I have no reason to use Granite(assuming it is some composite that noone knows how heavy it is or isnt) when I have a sheet of glass the same size which is lighter but can you really tell from the pic?

Perhaps a Red Herring dedicated thread would be helpful, those little extra's that just turn out to be a diversion.

Edited 2 time(s). Last edit at 11/17/2018 04:13PM by MechaBits.

With regards the original post I've tried the standard ormerod and mendel extruders, E3D titan, and Bondtech.the printed extruders worked well but didn't last well. I never got along with the Titan. Found it fiddly to set up, lacked power and ground filament when there were loads of retracts. That said it was consistant once setup. The bondtech is my favourite but I don't think it is the only double driven wheel setup. I guess there could be lessons to learn from mig wire drives. Think a few of them have gone to two sets of driven wheels, and I'd also like to try the belt extruders.

With regards to where the post is going I wonder how consistant the filament diameter would have to be, or how tighter control over the polymer melt temperature (not hot end temp) would be needed in order to realise the potential benfits that some of these design tweaks are likely to achieve? All to make a squishy plastic part that is likely to change in both shape and mechanical properties after manufacture due to exposure to UV, chemicals, water, and heat.

Great engineering is about balance and not getting drawn into developer gold plating on aspects which are likely to have negligable effect on the end product. It also of course involves working to a budget and allowing for what materials are easily available. That and if your machine is to remain multimaterial then the range of thermal conditions that it must work well with makes designing to low thermal expansion for the whole range very challenging.

All that said the discussion is interesting and raises a few thoughts. I'll be watching this with interest.

My two pence worth is that going too far down this route will likely end up with a machine running on one material from one supplier in one colour. If you take more advantage of electronic corrections the machine is likely to be ultimately more capable. I'd guess with microstepping and potentially encoder feedback on the drives it will quickly become cheaper to hold the tollerances electronically rather than with tight tollerance design. Electronics will however struggle to correct for pooly design practices such as four leadscrews on a bed etc.

Quote
WesBrooks
With regards the original post I've tried the standard ormerod and mendel extruders, E3D titan, and Bondtech.the printed extruders worked well but didn't last well.

I have had excellent results (better than my Titan) using the spring-loaded variant of the Ormerod extruder - the further variant with a PTFE tube lining the filament input path.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Yes, considering the Ormerod extruder had no pressure adjustment iy seemed fire and forget. The only issues I've had have been the bowden clip wobbling reduding retract effectiveness and standard PLA being a bit crap long term and cracking up and letting go.

Quote
Number_5
The only flaw, or perhaps I should say compromise with Digital Dentist's build plate, is that the plate is indexed on one side. This should not be a problem if you leave the bed heated for the whole print. However if you only want to heat the bed for the first couple of layers, the bed will shift. How bad will depend on the location on the bed, the side opposite the indexing pin will see the full expansion of 0.5mm or 0.8mm, or whatever it is.
It may even move more then a "regular" heat bed fixed at 4 corners or 3 points, at least that may stay relatively centered.

Heating the bed is done to ensure that the print stays attached to it. You don't heat the bed for the first couple layers then turn it off. If you do that the print is almost guaranteed to come off the bed before it has finished. It is common practice to heat the bed to one temperature for the first layer, then drop a few degrees for the remainder of the print. That few degree temperature drop will allow the bed to contract slightly, but that will happen whether the bed is indexed or not. There is a potential for that temperature change to leave a visible flaw in the print, but the temperature change is so slow that it doesn't usually produce a sudden discontinuity in the print surface (it can under some conditions)- it is usually a gradual change. If the bed isn't indexed you still get the same change in size/position with the resulting effect on the print. Having the bed indexed and on a kinematic mount means that if you stop a print and resume later, such as after a short duration power failure, once reheated, the kinematic mount guarantees that the bed will return to the exact same size and position. If the bed isn't indexed, you never know where it will be positioned after a temperature change. If the power is off so long that the bed cools, it may let go of the print and then it won't matter if you have a kinematic mount, the print has failed.

The only way to have a bed that doesn't change size with temperature is to use a material like Schott glass that is formulated for nearly zero thermal expansion. If the glass is clamped to a surface that expands and contracts with temperature changes, and you don't somehow index the glass, you don't know where it will end up after a temperature change. Other problems include heating it evenly, getting prints to stick to it, and the potential for breakage if you drop a tool or a spool of filament on it. People often use glue on glass beds to get prints to stick and have problems with the glue pulling chips out of the glass surface. Many people have worked out acceptable solutions to some of these problems and successfully print on glass every day.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
WesBrooks
With regards the original post I've tried the standard ormerod and mendel extruders, E3D titan, and Bondtech.the printed extruders worked well but didn't last well. I never got along with the Titan. Found it fiddly to set up, lacked power and ground filament when there were loads of retracts. That said it was consistant once setup. The bondtech is my favourite but I don't think it is the only double driven wheel setup. I guess there could be lessons to learn from mig wire drives. Think a few of them have gone to two sets of driven wheels, and I'd also like to try the belt extruders.

The bondtech BMG is a neat looking extruder. I might actually give that one a go. I was initially set on the Nimble but its so expensive for what it is and i'd be very concerned about that flexible shaft being somewhat elastic.

Quote
the_digital_dentist
Many people have worked out acceptable solutions to some of these problems and successfully print on glass every day.

I successfully print on glass every day across multiple machines. Its ghetto but I put the glass atop a cheap heated bed and it stays reasonably flat! (I dont have any easy way to measure it but haven't ever had flatness issues).

For this new machine I'm building I really want a more engineered solution so thats the reason for the post. I noticed your post and how your bed sits on a kinematic feature and thats really the only 'perfect' way of doing it that I can think of but I can think of a few easier ways of making the kinematic that I might try. I like the spring loading to hold the plate down too. The kinematic could be integrated in a way that makes taking the bed off really easy. for those difficult to remove prints.

One thought I have been thinking over recently is the idea of putting a silicone heater directly onto a glass sheet forgoing the aluminium heated beds. The glass really has much better thermal expansion properties than the aluminium and some quick simulations show massive reductions in distortion on even a badly mounted bed. I can't think of any reason it wouldn't work but I thought I'd hear what you guys thought of the idea! (Mounting might be tricky) I really wanted to avoid mains powered silicone heaters though so unless I can find a 24V heater of some sorts it might be a bust for this build.

This post has gone off topic but is probably more interesting! It's a shame I cant rename this thread to "Bed discussion"

Slicers have the option to heat for the first layer for adhesion, so it seems a reasonable assumption, whether or not you actually do it. The point being it would be better indexed in the center, for purely arguments sake. I do like that you have this reference point for the bed.

"Heating the bed then dropping a few degrees" is semantics for warming the bed for the first few layers then shutting it off. It's the same thing. Nothing wrong with warming the bed above ambient to reduce thermal shock on the first few layers. There are a lot a variables at play to make any almost guarantees.

The potential issue for Schott glass is that unless you are going to go thick enough (as mentioned earlier with the 1:6 aspect ratio Zerodur that stays optically flat without support), it still needs to be supported and heated. It expands less then regular glass, but not so appreciably that it mitigates all these other issues that are there with most other print surfaces IMO.

Masking tape has a similar CTE as Schott glass, yet I wouldn't say masking tape is the only way to get a flat bed with temperature change, you need the qualifier to make Schott stay flat, which would need to be about 2" thick for a 12" build plate. (50mm x 300mm) Assuming you have money for 2" x 12" Schott, you may have money for Zerodur, in which case there would be no point in regular Schott glass because Schott glass moves quite a bit and Zerodur is the most temperature stable material on the planet.

One time I placed 1/4" glass on a 12" build plate to do just as you suggest, Schott or nott, you can see the sag with the naked eye so what is the point? (Though I can see it's potential on 6" and 8" printers, it kinda almost works as a passable build plate.)
I suppose the confusion comes in comparing build surfaces with build plates. The build surface is an optional intermediate layer between the filament and the build plate to aid in adhesion. The build plate is the flat foundation that optional surface may be applied to.

The problem with indexing at the center of the bed is that you usually have a heater and/or thermistor there. Some heaters have a hole in the center or you could have a custom heater made that way. You probably wouldn't want to drill through the bed plate for tool access on the top side because slicers tend to put prints at the center of the bed, so you'd have to ensure that you can adjust the reference screw from the bottom side of the bed.

If the bed moves in the Z axis it should be OK to hold it down with magnets, but it would probably be better to do so if the reference and pitch adjusters were located along one edge of the bed instead of at its center, as I did in my printer. If you were going to do that with multiple bed plates you'd have to make sure the holes/slots depth and dimensions match from one plate to the next or you'd have to relevel after changing bed plates, which would probably be pretty difficult. You'd also need to put connectors on the heater and thermistor leads so they could easily be disconnected. It would be a lot easier to just clamp a piece of your favorite print surface to the bed plate with bulldog clips and leave the bed plate in place, or copy the Prusa design and use a magnetic sheet to hold the print surface in place.

When I said "heating the bed then dropping a few degrees". I did not mean warm the bed then turn off the heater. I meant print the first ABS layer at 100C then drop the bed to 90C or so for the rest of the print, or print the first layer of PLA at 70C then drop the bed temperature to 60C for the rest of the print. I don't know of anyone who heats the bed then turns it off during the print (but maybe I do now).

Optical flatness isn't necessary for a printer's bed. +/- 50 um or so across the surface is adequate to keep the print stuck if you're printing the first layer 200 um thick. The leveling system does support the bed plate, so I'd guess that a much thinner piece of glass would be flat enough for a bed plate. But flatness isn't the only requirement. If you want prints to stick, you need even heating. Glass is a thermal insulator. As much as they try to make heaters with wide, even heat distribution, they are not perfect. If you stick a heater to a piece of glass, you will have an unevenly heated surface.

Here's a thermal image of a Taz printer that came from the factory with a heater stuck to its glass bed plate. Notice the 30C+ variation in temperature across its surface. It was very difficult to print on this machine because prints did not want to stick to the plate.

And here's a thermal image of the same heater attached to an aluminum plate that replaced the glass after it broke. The thermally conductive aluminum smooths out the uneven heat provided by the heater.

Masking tape has nothing to do with providing a flat surface. It conforms to whatever surface it's stuck to. For some reason, molten PLA likes to stick to masking tape, even at room temperature. Until now, I've never heard of a coefficient of thermal expansion for masking tape.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
The problem with indexing at the center of the bed is that you usually have a heater and/or thermistor there. Some heaters have a hole in the center or you could have a custom heater made that way. You probably wouldn't want to drill through the bed plate for tool access on the top side because slicers tend to put prints at the center of the bed, so you'd have to ensure that you can adjust the reference screw from the bottom side of the bed.

Yes Keenovo makes custom heaters. I briefly thought of ordering one with a hole in the center, but I was already suffering from design paralysis. I don't think the overall all expansion is enough to worry about if kept reasonably centered and the plate is able to stay flat of it's own accord.

The heat beds you mention with the thermistor hole in the middle aren't even worth mentioning in this context. They aren't flat enough to begin with at this size, they need 5-point leveling.

Quote
the_digital_dentist
When I said "heating the bed then dropping a few degrees". I did not mean warm the bed then turn off the heater. I meant print the first ABS layer at 100C then drop the bed to 90C or so for the rest of the print, or print the first layer of PLA at 70C then drop the bed temperature to 60C for the rest of the print. I don't know of anyone who heats the bed then turns it off during the print (but maybe I do now).

The reason they print at 100, then drop to 90, or 70-60 is for adhesion. If there is no need to heat for main part of the print, a bit of heat can be used just so the first layer isn't printing on ice cold aluminum covered with tape etc. A room temperature glass sheet is a lot colder then the 5th layer of plastic on plastic.
A cold print doesn't suffer from the thermal gradient that a tall print would have on a heated bed. This is because the heat was shut off before the print starts to form. It's only a bit of heat, so it's not like the plate is going to stretch then shrink a 1/4", it will be similar, perhaps even more stable then dropping from 100-90 because you aren't as susceptible to drafts etc at temps that are already close to room temp. As I say, it's semantics really.

Quote
the_digital_dentist
Optical flatness isn't necessary for a printer's bed. +/- 50 um or so across the surface is adequate to keep the print stuck if you're printing the first layer 200 um thick. The leveling system does support the bed plate, so I'd guess that a much thinner piece of glass would be flat enough for a bed plate. But flatness isn't the only requirement. If you want prints to stick, you need even heating. Glass is a thermal insulator. As much as they try to make heaters with wide, even heat distribution, they are not perfect. If you stick a heater to a piece of glass, you will have an unevenly heated surface.

That wasn't my point. You can make a telescope mirror out of it, send it to space and it would still be optically flat enough for the job. Space is -270C.
Zerodur could be cast hollow with honeycomb or heat chambers close to the surface to have heat strips installed. Its only money, theoretically speaking.
Flatness is a huge priority with a 300mm bed, especially when the lack of ABL is being touted around.

If a build plate is "flat" and stays flat throughout temperature swings, well that's 2 of the 3 major hurdles surmounted. 3 point leveling takes care of the rest. Adhesion is incidental.
Stability is paramount to the ultimate precision. As you say 0.050mm is probably good enough, providing it stays like that.

Quote
the_digital_dentist
Masking tape has nothing to do with providing a flat surface. It conforms to whatever surface it's stuck to. For some reason, molten PLA likes to stick to masking tape, even at room temperature. Until now, I've never heard of a coefficient of thermal expansion for masking tape.

Right, just as a typical sheet of glass does nothing to provide a flat surface, it conforms to what it is clamped to. (Although 1/4 glass may be acceptable on a 6" or 8" printer to compensate for a warped aluminum PCB heat bed)

Since masking tape is an insulator, it reduces the thermal shock on the first layer, giving more time for the molten plastic to grab onto the cellulose material. A warmer layer will have a better bond on the bottom, because the bottom of the first layer hasn't skinned over and contracted, partially releasing itself, like it does printing on a cold glass bed etc.

I can't follow your logic. You need print adhesion to the bed for the duration of the print. That's what the heater is for. You also need a surface that prints will stick to, which may not be the surface of the bed plate itself- people use different things- some print on glass with glue or hairspray, others use PEI or kapton, etc. I thought we were talking about 3D printers. I have no idea what Zerodur and telescope mirrors in space have to do with any of this. Glass is rigid and flat and is often clamped to crappy, unflat bed plates to provide a flat surface specifically because it doesn't conform to the unflat surface under it.

I'm so confused by your comments I have to ask a couple questions: have you ever used a 3D printer? Have you ever built one?

Edited 1 time(s). Last edit at 11/18/2018 10:47AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
I can't follow your logic. You need print adhesion to the bed for the duration of the print. That's what the heater is for. You also need a surface that prints will stick to, which may not be the surface of the bed plate itself- people use different things- some print on glass with glue or hairspray, others use PEI or kapton, etc. I thought we were talking about 3D printers. I have no idea what Zerodur and telescope mirrors in space have to do with any of this. Glass is rigid and flat and is often clamped to crappy, unflat bed plates to provide a flat surface specifically because it doesn't conform to the unflat surface under it.

I'm so confused by your comments I have to ask a couple questions: have you ever used a 3D printer? Have you ever built one?

Why are you so confused? You are the one who keeps moving the target, bringing up things that weren't discussed. I brought up Zerodur because you obviously didn't read the thread before you went off rambling about Schott glass trying to reinvent the wheel.

You don't "need" heat to print everything.
Hairspray, masking tape, Schott glass and whatever else you brought up has nothing to do with the mechanical aspect of the build plate. Glass is not rigid, correction, glass you can afford is not rigid. 300mm glass thin enough to use with a typical inductive sensor will sag far more then a layer height when supported by the edges, more so if attached with binder clips.

Didn't you know glass is a poor conductor of heat?

I'm done.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Anyway..... the_digital_dentist I like your bed

You did manage to link the same picture before though.

The good thing about having a surface plate & support plate replaced by one flat piece of material is less complexity & maybe lighter overall...cant see that granite(or some other composite 5mm thick) would be much different than glass to print on, maybe keep heat more & more even? (though just seen e3d's heatbed)

once that is achieved, not much need for ABL unless it's to iron out any errors, but the distortions from heating wouldnt really be compensated for as it's only probed once at the begining, when I first started out on this journey I wanted all the toys, and especially when I couldn't get the hang of getting things flat...now I have no problem with it, and do have a problem with adding more wires & bits to the hotend...especially after reading people having issues, which means I'm probably bound to have a few.

Despite having 3 heatbeds & one ssr I've never bothered to wire them up...I just cant be doing with the hassle.

Edited 1 time(s). Last edit at 11/18/2018 09:14PM by MechaBits.