I really don’t see the problem with the maths. We already have it working for two motors so it’s just the same calculation process twice more. And once it’s done it’s done.
Sounds like a worth while project that we sort the maths out now and include it in the firmware as an option so the community could then start experimenting with the hardware.
are you talking about a 4 motor version?
I really don’t see the problem with the maths. We already have it working for two motors so it’s just the same calculation process twice more. And once it’s done it’s done.
right now, any errors in the chain lengths result in the sled being in slightly
the wrong place, if you overconstrain the sled position with motors on all four
corners, when the chain lengths are slightly off, something will have to
bend/stretch/break.
Sounds like a worth while project that we sort the maths out now and include it in the firmware as an option so the community could then start experimenting with the hardware.
it’s not as simple as ‘include it in the firmware’, due to hardware limitations.
Each motor currently needs 5 GPIOs (two for the encoder, one for motor enable
and two for motor direction), and there just aren’t enough GPIOs left to have
two more motors on the current system (let alone the problem that the driver
board only supports 3 motors and 5 would be needed)
My 2 cents - https://www.adafruit.com/product/1749
6 axis controller but it’s pricey - a 4 motor variant might be a fork of the Maslow opposed to an evolution. I rather like the idea of counter tension at the bottom corners. I’m involved in several types of 3D printing. There is a open frame delta - I call a flying Delta design that seeks to balance tenson
I prefer the idea of using a 3rd motor to take up the slack at the bottom 2 corners. These are just my thoughts.
I might have more to say once my Maslow is up and running.
Thx
a 3 motor approach actually also would make the linkage obsolete.
if the sled has the 3 chains mounted on an equilateral triangle then the routerbit is ALWAYS in the center no matter what.
the drawback would be a reduction of the workspace in the lower corners (right?)
Would be interesting to see what the effective workspace will be in this kind of configuration.
And from there decide if this makes sense
Almost forgot to post it: I drew a 3 motor config idea / effective work area diagram
Motor positions are A,B and C
the brown areas are 4’x8’ sheets in portrait or landscape orientation
The circles are there to get an idea how the sled would sit between the motors.
imagine a chain from A B and C to the sled. (I forgot to draw the actual chains to the sled)
to be clear:
in this config the sled can never move out side of the triangle
the reason i moved A and B to the bottom is that C takes control of all gravity issues, so possibly less chain-lag in this config (?)
The over all height of the entire machine would be A LOT higher then the original Maslow
(i could never fit this in my little shed) But I find this a very interesting option non the less.
The third motor doesn’t have to be the same distance from the center of the work area as the others. Just need to be able to manage the portion of chain on the slack side. Using suitable turning sprockets, it could be off to the side. It all adds expense, though.
My 2 cents - TinyG CNC Controller Board v8 : ID 1749 : $165.00 : Adafruit Industries, Unique & fun DIY electronics and kits
6 axis controller but it’s pricey - a 4 motor variant might be a fork of the Maslow opposed to an evolution. I rather like the idea of counter tension at the bottom corners. I’m involved in several types of 3D printing. There is a open frame delta - I call a flying Delta design that seeks to balance tenson
5 axis.
it would be nice, but if you look at the kit, doubling the motors, chains,
brackets, and controller channels pretty much doubles the kit. The only think
that isn’t duplicated is the Z axis stuff and the arduino.
is this really enough of an advantage to double the cost? I don’t think so.
- The Hang Printer
I prefer the idea of using a 3rd motor to take up the slack at the bottom 2 corners. These are just my thoughts.
if you follow the development, they are having a lot of trouble keeping all the
lines at the right distance (they either get too tight and stretch/break/jam the
machine, or they get loose and accuracy suffers)
Hmm but then the triangle gets smaller…?
I need to figure that one out.
With a rectangle or trapezium i could agree, but still in doubt how this would work in a ‘lower’ triangle
You have created a Delta CNC router! If you flip through the Hang Printer Development they demonstrate how the system adjusts itself. However I was really suggesting a lower system using 1 motor and 2 pulleys to adjust the bottom half of the Maslow. This lends itself to the rule of planes any 3 points in a plane is stable. 4 motors makes me think of a medieval torture device.
Seriously you will need a elastic to deal with the timing difference, the real answer if either a soft bottom 3 motor or a much longer chain manipulating 2 points at a time from 1 motor each. I really do buy into the Maslow Design - I could make this a big X/Y plotter but the same issues would remain. Let’s back up the cart and look at it in a very simplistic way, the rotational force of the routers torque is potential, it is moving in 1 or more directions Right now we rely on the angel and weight of the sled. If you want to strap the router down you will have to over build a structure strong enough to overcome any potential raised by the 3-4 motors we have in motion now, the router counts as one. I think a up take motor working on tension with no connection to the controller , and independent microcontroller could work.
My drawing will suck - But I will put one up in a short while.
I love drawings that suck, misunderstandings usually trigger neat new idea’s
I looked at the hangprinter though the 3rd dimension still boggles my mind on thatone…
I find it a wonderfull printer, though it would be cooler with servo’s in the corners instead of heavy steppers on the effector 
looking forward to your brainwaves!
I agree on thatone. the original maslow is far more cost effective.
I hope we find a simple solution with all these little sidesteps.
for lack of a better term I don’t want to push on one side and pull on the other at the same time because that is bound to have a timing issue. I want a loose tension system to act as an anti Z backlash system by holding X/Y taught and fighting quick changes. Am I making any sense ?
Interesting drawing! that option seems to even out the gravity in a nicer way!
Just use a weight hanging on a pulley instead of the
take up motor
That would work too. I’d need a few more pulleys to move it around back and set up the counter weight. My way is a microcontroller and few pulleys a stepper and 2 microswitches. The thing I dislike about mine over yours is I need more wires and electricity.
the problem you will run into is that there is a limit to how flat the angles
can get between two adjacent chains before you loose the accuracy and force to
move the router. Currently the Maslow lets things get to ~140 degrees IIRC
If you plot this out, I you will find that you loose access to most of the
current work area (you would need to have the third motor several feet below the
center of the work area)
The maslow is only ~6’ tall, if we enlarge it to 6.5’ (the largest that will fit
through a standard door.
Put a motor on the floor under the center of the area, then move the outer
motors so they only have a 7.5’ spacing instead of the current 9’+ spacing.
no motor can get closer than ~16 inches to the sides of this triangle in the
center. This leaves a very odd three legged shape in the center with arms that
are only about a foot thick as they spread out from the main body. I’ll have to
try and draw this out and take pictures of the result.
David Lang
it’s worse than that, the sled can’t get close to the edge of the triangle as the angle between the chains gets too wide
If the distance between A and C is 7.5’ (equilateral triangle 6.5’ tall, just short enough to fit through doors), keeping a similar angle to what the maslow is limited to now (140 degrees between chains IIRC) have to stay at least 16" away from the sides.
I you could get to 150 degrees that lets you get down to 1’ away from the edges and lets you pull the sled almost all the way to the motors.
no, chain sag isn’t going to change with the direction. It doesn’t really matter the orientation, when you start moving away from being directly under C it’s now working at an angle and starts to run into limits
Your thinking much faster than me. That is probably a good thing. It dawns on me my thought process in this is based in sailing rigging. I think yours is much more math.
I like the cut of your jib.
Thx
@Brandon, a lot of it is that I’ve fielded the same or similar questions before
There have been a lot of people who post quick “if they did this it would be
better”, but there are only a few ideas that hold up to being explored in depth.
We’ve all agreed that a triangulation approach would help, but we didn’t have a
reasonable way to implement it until the ‘throwing my hat in the ring’ thread
and the ring and linkage approaches. Those are SO much better that I really hope
that the 3rd batch of kits ships with one of them instead of the stock design.
I think the top-beam frame approach is another significant change, but I suspect
that we have had so many other accuracy problems that the flexing of the machine
hasn’t been the biggest source of error for anyone yet.
David Lang
Building on your diagram, instead of a third motor, take each of the lower two lines all the way across the bottom and up to a multi-purchase pulley system working against the chain take-up sprocket on that side. That way for instance when the left motor feeds out chain and winds in on its take-up sprocket, the tackle attached to that sprocket pulls in on the right lower cable. Inserting a stretchy link into that line, maybe at the router end, would provide the actual tension, the tackle provides adjustment of that tension relative to the chain length and the router’s position. It might take some playing around to decide how much purchase the tackle needs; the sprocket moves one inch for every two inches of chain movement, so using a gun tackle arrangement would move double that half inch so that the tension cable would move one inch for each inch the chain moved. That may or may not be the right amount, it would depend on the tension of the stretch link.
I know, Rube Goldberg would be jealous.
