Hello RepRappers - I am building my first 3D printer and am curious about the drilling for the plate.

The mounting system I will be using is standard; a four-point M3 countersink screw and spring setup. What is not standard is the massive quarter inch 310x310mm Mic6 plate.
Should I be drilling holes, or slots to account for (slight?) expansion at 110 Celsius?
If holes, the plate is so thick that I can either drill it clean or drill it and tap it. Does it matter?

Also, the real question: knowing that the screws are M3 countersink and the plate is quarter inch Mic6 how far out should I be comfortable drilling the holes without worrying about cracking?
I have not been around enough commercial 3D printers to steal their potentially well thought through dimensioning choices, and I cannot find hole dimensions anywhere. The screw heads are 5.75mm in diameter. I want the screws to be out of the way of the nichrome. My original idea was center of holes 5mm from each side on the corners but now I am not so sure.

[www.eng-tips.com]

You don't have to worry about the plate cracking when you drill it, and you won't be putting a lot of mechanical force on it when you use it as a printer bed plate, so no worries at all there.

What kind of heater and how much power are you planning to use? This site will give you a pretty accurate estimate of the time required to heat the bed: [jscalc.io]

What will the bed support structure be like?

Will the bed be moving in Y, Z, or fixed?

In general, for a plate that size it is better to use 3 screws than 4. 3 will make it easy to level the bed and keep it stable. Using 4 screws will result in instability and/or difficulty leveling. See [drmrehorst.blogspot.com]

Here's how I built the bed for my corexy printer (moves in Z): [drmrehorst.blogspot.com]

Here's the same stuff applied to a Y axis moving bed: [drmrehorst.blogspot.com]

---

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

Custom 480W wall power nichrome/kapton loop switched with a SSR. I ran the specific heat/heat up time and wire calculations myself.

It is a modified Hypercube Evolution so sitting on a 2020 square frame.

CoreXY camp because I value the limited footprint and enclose-ability given recent patent news.

I am aware of the four screw problem, but the plate is so heavy (1.6kg) that I would rather (for now, until I at least get something that works) run four screws, especially because we are talking a cheap screw kit with springs made from solid Chinesium. This is probably an elaborate excuse to stick to the standard build plan. Is using a four-point design that flawed? I see you have experience with this nature of plate. Do you think I should convert to three screws, and if so why would you suggest that they be placed at the midpoints of three sides of the plate rather than two adjacent corners and the opposite side's midpoint?

Is that to say that I should just drill holes, not slots, and not worry about tapping them?

Quote
ders
.. This is probably an elaborate excuse to stick to the standard build plan. Is using a four-point design that flawed? I see you have experience with this nature of plate. Do you think I should convert to three screws, and if so why would you suggest that they be placed at the midpoints of three sides of the plate rather than two adjacent corners and the opposite side's midpoint?

Is that to say that I should just drill holes, not slots, and not worry about tapping them?

It is NOT STANDARD, it is POPULAR because most people lacking basic mechanical knowledge keep propagating bad ideas despite being told otherwise by a MINORITY.

Most often, the best solution is the simpler one proven over decades (mechanic is rather "mature"), it just needs proper quality parts.

As for the potential issues, they pop up here regularly. Ex: [reprap.org]

Anyway ....

BTW, there are other bolts than M3 !

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

If you've ever had a geometry class, one of the first things you learned is that it takes exactly three points to define a plane. You also know this from practical experience in cafes and restaurants: a three legged stool/table is stable on any surface, a four legged stool/chair/table is not. That's the difference between three leveling screws and four. The bed plate is (supposed to be) a plane that is parallel to the printer's XY plane defined by the X and Y axis guide rails. You make the bed plate parallel to the printer's XY plane by using the leveling screws. It doesn't really matter where you put the leveling screws, but it's most convenient if two of them lie in a line parallel to either the X or Y axis. I like to put the leveling screws close to the lifting elements- lead screws or belts- for stability. Keep the lifting elements close to the Z axis guides.

The mass of the bed has nothing to do with the number of screws you use to level it. A very large or a non rigid plate may require more leveling screws for support, but 1/4" MIC6 at 300x300 mm works fine on 3 screws.

Are you planning to somehow apply nichrome wire to the bed plate yourself? How?

---

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

Quote
MKSA
BTW, there are other bolts than M3 !

Thanks, I always overlook the obvious. Will examine this further.

I define the four-point mount as standard because it is used by the designer and most implementations, thus being used or accepted as normal or average. This is not an average printer, nor am I trying to be average. Thank you guys for highlighting this flaw. I appreciate being pressed on the design because nobody else has had the experience to question it.. I should have joined here earlier.

I will choose between finer adjustment (two adjacent corners and opposite midpoint) versus stability and quicker adjustment (three midpoints). Still wondering are slots necessary? The primary goal being warping avoidance.

The bed was definitely what I was losing sleep on while trying to consider. It is a good idea to talk about it as a whole. Plans are vague around the attaching. I will be using straight up nichrome double wrapped in kapton tape, kapton taped to the plate? Shielded with a similarly sized cork board. I have no plans on thermal standoffs and the screw mounts are ABS.. HM. I have to look at whether my SSR provides optical isolation in order to determine what is relevant to ground in case of a short, which running this setup is undoubtedly more likely. The kapton and nichrome were cheap so if this is a bad idea I do not mind balling out and going the silicone route.. I guess. As mentioned earlier I value the enclose-ability and will be enclosing in the future, running a filter setup, and a heated chamber (might as well at that point). I would appreciate the heater-building experience for when I have to build a rather extreme air heater.

On top of the plate I am running a removable 3mm glass bed (binder clips). I am avoiding adhesive sheets (PEI, kapton) to move on to a magnetic spring steel bed in the future.

4 screw leveling is a mistake that goes back to some of the earliest 3D printers and persists because printer "designers" have been copying the error without thinking ever since. The bed has 4 corners, so you need 4 leveling screws, right? Duh! Actually 4 leveling screws might make sense when everything is flexible, like the type of construction you find in a lot of cheesy printers. But if you are using solid construction techniques and materials, and the bed plate starts off flat and stays that way, it makes no sense to use 4 leveling screws.

If you want to make a kinematic mount, one slot is needed. It doesn't have to be milled into the plate- it can be formed by screwing a couple small pieces of metal to the plate. You also need one hole, optionally chamfered using a countersink bit in a drill.

The point of the kinematic mount is stability. It allows the bed to expand and contract with temperature changes without forcing anything to flex or causing anything to lift. The kinematic mount guarantees that the bed will always be in the same lateral location at any given temperature. If you want to make a printer that doesn't require frequent releveling of the bed, the kinematic mount will help you get there. OTOH, if you're planning on active leveling and unflatness compensation, you probably don't need to bother with it, though starting off with a flat, level, stable bed is still a good idea in those machines, too.

If all this leveling stuff is too much of a PITA, you can make a printer whose bed does not have to be leveled at all, and doesn't need a bed heater, either. You can use a piece of PIR foam for the bed, and just replace it after several prints. Prints will stick without heat. The foam is cheap- you can buy a 4' x 8' sheet of it for $15 at Home Depot stores and cut it to the size you need with a razor knife. See: [vimeo.com] and [vimeo.com]
Don't try to use the pink, blue or yellow stuff for this! PIR doesn't break down and produce toxic fumes at nozzle temperatures.

Making your own bed heater is risky business. If the wire gets too hot it will burn through the kapton. The adhesive on the kapton may not be able to withstand the peak temperatures reached. Good luck.

Edited 2 time(s). Last edit at 08/20/2019 07:26AM by the_digital_dentist.

---

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

People fall into the trap of assuming that because some good printers, like the genuine Prusa for example, use a four-point system that copying it will give good results - but it is not that simple. A good four-point mounting is a matter of making the best compromises to get acceptable printing quality, ease of adjustment and yes, cost as well. If you have the mechanical knowledge, computer tools and the time to try out different geometries then maybe go for it.

A three-point system is actually the simpler option to get good results. In the self-built environment is also not significantly more complex to make.

Taking a quick look at the four-point mounting with screwjacks in the corners of a rectangular build plate. The screw jacks are mounted on a carriage with springs but this does not mean than the plate is sprung: as long as the downward force on the plate is less than the upward spring pressure then the plate is mounted solidly. If both the build plate and the carriage that the screwjacks are mounted on are rigid then one diagonally opposite pair of jacks is likely to be tighter on the plate than the other pair - the looser pair now suspended on their springs and can wobble.
The answer to this is to have the screwjacks themselves mounted on semi-rigid mountings - the legs of the "squashed frog". When this is done then there can be a fair range of acceptable adjustments. This solution works even with acceleration of the print bed in the XY plane - ut it is not so good with tall prints where the XY movements become rotational forces about the XY axes.
Even if you get a good compromise for the flexibility of the legs of the squashed frog, you still have to determine how great the adjustment range and the spring rate to stop those legs from being plastically deformed during adjustment.

As I said, it is not that simple.

Mike

Well, in basic geometry terms, 3 points define a plane, so anything more than 3 points introduces potential for errors.

As human beings though we seem to like 4 legs. That's how we generally build our chairs and tables, and of course many of those wobble as a result. It does make some sense when you're trying to get the maximum stability from a square or rectangular shape, which we also seem to like as a part of human nature, though it's not necessarily the most efficient. When stability really counts though, we do know enough to use a tripod, and this is the most inherently stable configuration.

For the Prusa printers, they use a PCB as a build platform, which is somewhat flexible, and a 4 supports from the corners makes some sense. I have a heat bed that's billed as an upgrade, and has an aluminum surface, but is compatible with the Prusa PCB heated build platforms, so it also features 4 mounting posts. Since these are what I intended to use, I also designed my printer to use this. I would prefer a 3 point system, as it's much easier to adjust, but this is what it is.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

Quote
the_digital_dentist
If you want to make a kinematic mount, one slot is needed.
[...]
Making your own bed heater is risky business. If the wire gets too hot it will burn through the kapton. The adhesive on the kapton may not be able to withstand the peak temperatures reached. Good luck.

One slot? But the plate will expand in (primarily) two axes?

Thanks for the insight with the kapton approach. And the luck. I will undoubtedly need it..

I believe that DD is using a Kelvin kinematic mount. Both this and the maxwell type are shown in [en.wikipedia.org] and both have their own advantages.

I use a Maxwell coupling which should expand equally from the centre. Picture below is the underside of my mount (Arrow pointing to one of the slots) which does return to its starting position after a mechanical disturbance. Thermal disturbance is presently a problem as heating from 20°C to 100°C is accompanied by ticking sounds and the centre of the plate rising and falling by about 50µm. This is caused by "slick-slip" behaviour of the conical screw heads against the chamfered slots. Work is ongoing to minimise this behaviour.



Mike

The Kelvin type kinematic mount I used has a reference screw that does not move- the plate expands away from it. The plate is free to expand in X and Y because the pitch adjuster is located in a slot and the roll adjuster is just touching the flat bottom side of the plate. There's a good explanation of the type of mount I used here: [www.edmundoptics.com]

When the plate expands, it expands in every direction. If you pick a reference point at one corner, the dimensions of the plate get larger relative to that point. I used a reference point along the middle of one side of the plate. Every point on the plate moves away from that point (because every point on the plate moves away from every other point as it expands). The pitch adjuster is on a direct line from the reference in the direction that the plate expands and the slot is set in that direction. The roll adjuster simply contacts the flat bottom of the plate so the plate is free to slide on it in any direction as it expands. It's easy to make because there's only one slot and it is located along the X or Y axis of the plate which is easily aligned with the X or Y axis of a mill when cutting the slot.

You can pick any point as the reference, but it's most convenient to use one corner or the middle of one side of the plate, with the pitch slot aligned along one axis of the printer, because that makes leveling the plate very intuitive- you adjust pitch first, then roll, and you're done. The roll adjustment doesn't affect the pitch of the plate at all (when you turn a leveling screw, the plate pivots on a line between the other two screws- think about it).

The Maxwell type of kinematic mount uses the center of the three radial slots as the implicit reference point. It's a little harder to make because of the angles of the slots (in theory, exactly 120 degrees apart). If you have a turntable for the mill it's pretty easy to make this type, too, but it's easiest to do if the bed is circular so the levelers can sit at the edges of the bed and you can use an off-the-shelf circular silicone heater. If you want to use the Maxwell mount on a square plate it needs to be larger than the heater or you have to come up with a custom heater and some relatively complex hardware like photo in leadinglight's post. Notice that the way the slots are arranged in the photo, none of the lines between the levelers are along the X or Y axis (parallel to the edges of the plate). That means there's no simple pitch/roll relationship- adjusting any screw will affect both pitch and roll, so the leveling process will be a little more complicated than the Maxwell mount I used. The Maxwell mount can be made with a simple pitch/roll relationship, but for some reason, whoever built that plate chose not to..

The main differences I can see between the two is that the Kelvin mount is easier to make.

Edited 2 time(s). Last edit at 08/23/2019 07:58AM by the_digital_dentist.

---

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

In order to level the bed, it is only necessary to set the distance in the Z direction from the bed to the nozzle above each of the mounting points - it is not necessary either to have a reference datum nor to set the other mounts on the X or Y axes relative to the reference.

Edit. As long as the bed starts less than a couple on mm off level it is only necessary to adjust the screw to set the nozzle to bed gap just once at each of the three positions - this will not affect the other adjustments.

The slightly peculiar position of the mounts was enforced by engineering compromises such as not being in the same position as the bearing pillow blocks or interfering with the XY rails.

The final position of the mounts differed from being 120° apart by less than 2° and the radial distance from the common centre by less than 8mm. The effect of these, both theoretical and practical, is insignificant.

Mike.

Edited 2 time(s). Last edit at 08/24/2019 05:36AM by leadinglights.

Quote
leadinglights
In order to level the bed, it is only necessary to set the distance in the Z direction from the bed to the nozzle above each of the mounting points - it is not necessary either to have a reference datum nor to set the other mounts on the X or Y axes relative to the reference.

Edit. As long as the bed starts less than a couple on mm off level it is only necessary to adjust the screw to set the nozzle to bed gap just once at each of the three positions - this will not affect the other adjustments.

All true. Leveling is easy enough with either type.

Quote
leadinglights
The final position of the mounts differed from being 120° apart by less than 2° and the radial distance from the common centre by less than 8mm. The effect of these, both theoretical and practical, is insignificant.

The radial distance shouldn't matter at all as long as the angles are right. I wonder if the 2 degree eror isn't the cause of the lifting of the plate as the bed heats. If the three levelers are fixed and the plate expands, the slots (and the plate they are in) will have to find new positions that keep the leveling screws in the slots. That means either the plate will have to rotate and shift a little or it will lift. Is the 50 um lift at one screw or the entire plate? Always at the same screw? To me that sounds exactly like the sort of behavior I'm trying to eliminate by using a kinematic mount.

In the Kelvin mount as I have built it, the one slot is parallel to the edge of the plate which is easy to set up on a mill- you just put the edge of the plate against a stop on the mill table. Once that's set up, you drill one hole and then mill one slot and they will be in line. You can get fancy and chamfer them if you have the tools, but it isn't necessary for the mount to work. I used the middle of one edge of the plate, but the reference hole and pitch slot can be located at the corners of the plate instead (maybe a little better than the way I did it). The whole thing could also be built without a mill by screwing a couple small parts to the underside of the plate. There are no holes in the plate in the print area so no need for a custom heater, and the entire bed plate surface can be printed upon. As I built it, nothing stands up over the print surface so there's no danger of the extruder nozzle banging into anything.

I built the same type of mount for my corexy machine with the bed moving in Z and for another machine with the bed moving in Y. It's working perfectly in both and I don't have to relevel either bed unless I take something apart to make modifications.

This is the support for the bed in the printer with the bed moving in Y:



I mounted the leveling screws as directly as possible above the bearing blocks in the linear guides for maximum stability. The reference adjuster had to be shifted out a bit from the bearing block, but it's still pretty solid.

This is the bed plate - the reference hole is on the left ear and the pitch slot is on the right. The roll adjuster contacts the flat surface of the plate on the ear at the top of the photo. This bed plate is almost identical to the one I used in the corexy printer, the only difference being the location of the screws that hook on the hold-down springs.



Here's the roll adjuster with the bed mounted on the machine:



And here's the pitch or reference adjuster:



This is the corexy bed support- things don't get much simpler:



I used a tee instead of a rectangle for the bed support and put the levelers at the center of the sides of the plate to keep them as close as possible to the bearing blocks on the Z axis for maximum stability (there are two belts that lift the bed structure and two linear guides located on either side of the bed). The teflon blocks that hold the leveling screws have tangs that fit into the t-slot allowing them to be mounted using a single t-nut each. As built, the reference and pitch screws are adjusted from the top side of the plate and the roll from the underside. It would be just as easy to build it so that all adjustments are on the underside of the plate.

This is approximately what the Z axis looks like (there have been some recent changes):



The knobs on the roll adjusters have ridges that make it easy to grip the screws with fingers. I used 16 ridges because the screw pitch is about 800 um per rev of the screw, so each ridge represents a 50 um vertical displacement. I use the manual bed leveling assistant built into reprap firmware to level the bed plates in these machines and it reports the error in um and adjustment required at each screw, so knowing the exact displacement is handy. It would be even better if the pitch screw also had such a thumbwheel. I may convert both of them to under-bed adjustment using thumbscrews. That would require replacing the spherical head screws I used with acorn nuts on the ends of longer screws that go through the teflon blocks like the roll adjuster, probably a 10 minute job to make the changes.

---

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

Hi DD,

Thanks for the full description of your kinematic mount. I think we agree that a kinematic mount which returns the build stage to its original position after a disturbance and minimises the effects of thermal disturbance can only be a good thing.
As regards the bowing of my build surface during heating: The effects of heating are easily cancelled by thwacking the build surface lightly with the blunt end of a screwdriver. Lubrication of the sliding surfaces with an extreme pressure grease also reduces the maximum excursions to less than 10µm - it remains to be seen if this is a long term solution or if I have to find a better way. I think the diagram below shows what is happening although it applies to a beam supported in two places rather than a plate supported in three.



During heating to 100°C the plate will audibly 'tic' some 3 to 5 times and the middle goes up by very approx 40µm and down by perhaps 10µm with each tic.

Mike

@DD "The radial distance shouldn't matter at all as long as the angles are right. I wonder if the 2 degree eror isn't the cause of the lifting of the plate as the bed heats."

In fact neither the radial distance (except 0) or the angles (of course the 3 "slots" meet at one point but can't be // that is meet at infinity) matter to constraint the bed. 120° and same radius are important to keep the "center" centered during expansion. More stable too. And nice looking.


As for binding, the figure shown above explains it. The taller the support, the more binding. When you move a heavy furniture, better push it at the base isn't ?
Better use steel plates and balls as the contact surfaces are small. Bearing surfaces should be polished. Lubrication helps of course.

I use a Kelvin approach too, can be done with limited tools. Which is the purpose and the challenge.

Besides, let's not forget that these machines are just spitting molten plastic. +- 1/20th is good enough.

Edited 1 time(s). Last edit at 08/24/2019 08:45PM by MKSA.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Quote
MKSA
.............
Besides, let's not forget that these machines are just spitting molten plastic. +- 1/20th is good enough.

Very true, but 1/20ths add up so I always target <5µm but take an "Oh what the hell, that will do" if they add up 50µm and I am feeling especially pessimistic.

Mike

Thanks leadinglights for pointing out that there is a more advanced mounting system that I was not envisioning. Thanks digital dentist for explaining that in fact the roll adjuster is not attached to the plate at a fixed point. This all makes much more sense. I am unsure about mill access for now so if drilling a small slot isn't feasible then I will likely just drill two holes and deal with the single-axis warping, to upgrade later after I get things running. Starting with PLA on the Mic6 will be almost negligible warping anyway, I am sure.