My future Solidoodle

Wow, ok, more power to you. I am glad you have the vision, ambition and capability to do this. No disrespect intended, massive props. As far as ROI goes, how much better of a printer will it be when you are done and what will that cost? Will the cost be offset? Though I do understand the perfect answer to the question "Why?" is often "Because I can!".

Looks great have you decided on materials yet for the red parts.. Aluminum, delrin, uhmw? Also really like your tensioner idea will you be adding a fixed angle indicator into the spring?

I wnt to offer you a suggestion: Fix that bed too. As it is now, the way the aluminum angles upward and the leveling screws are like three long stilts that just wave all around. One little blob in the middle of a print knocks the entire bed off course. Not only can it shift, it can also twist. If you are going to put linear bearing on the Z axis rods, you need to fix the leveling design also to make it worth while.

My suggestion is to drill out the rivets and flip the aluminum upside down. Make it so that the springs fit between the underside of the bed and the end of the leveling screw. This will make it so that the distance between the thread and the head of the screw is minimal as possible, removing all of that play.

Please see the attached of my highly technical drawing for a better representation of my idea.

To further this and remove more distance between the bad and the frame, you could remove the stock heater and install a PCB heater that would go ontop of the Kapton with either glass or tile final surface:

Frame: Aluminum bed: PCB: Glass or Tile.

Just my 2 cents as I have evaluated the SD3 design myself. Right now, I have yet to make an irreversibe upgrade. I have glass and an enclosure, but I have stopped on when it comes to the Z axis as I have no banding. My circles are as goood as they are going to get. The imperfection has more to do with the controller. Also remember that it can only slice a circle when it is a circle. Most models I have seen others do leave the circle as 24 segment lines. I edit my circles to be 240+ segments, and I get nice (and slow) true circles out of a stock SD3.

I guess my point is, I have no good reason to fix the Z axis since my banding is so small, you can see moire effects instead of banding. And the real problem I have with my SD is that anytime there is problem in the print (ie- overhangs tooo big, brides that dont make it, anything where the nozzle 'bumps' a lifted edge) my bed gets whacked and the entire print actually gets shifted on itself. Not only that, but my level usually needs to be redone right after. I blame all of this on having over an inch of screw that has 0.125" of anchor. I equate it to the entire bed design is simply a bobble head.

I see your point and agree with you. Having the bed so sloppy (from loose bushings, undersized shafts, thin aluminum that flexes, and finally the screws and the springs) gives the nozzle the best chance of taking hits and absorbing vibrations of the fast movements of the machine.

One thing to note is that this thread is about someone who is going to do just that though. He is adding aluminum blocks and linear bearings to remove slop. I would also say that my idea on a stock SD3 would probably be bad since the X carriage would break apart on collision, but this guys design would withstand the impact and quite frankly no harm would be done as it would result in skipped steps, not cracked ABS.

I agree, all of the Mendel and other designs benefit from having the springiness of the bed. Without strengthening the entire design, this idea would be very bad. And from your points on dampening vibration, I have no good argument to support making anything better on the stock SD3. I was just offering that if someone was going to go the distance and remove all of the slop that current design has to not overlook source of the most 'sloppiest' part of the entire machine.

Oh, and I have reconsidered my idea. You were right about 'hard-locking' the aluminum bad to the aluminum frame. I did not finish my though before drawing it up, but the idea was to use nylon lock nuts on the ends of the screws. Its basically reversing where the springs are, but the bed still flexes up and down just like before, just not side-to-side as much.

So now you got me thinking. While making the whole design much more rigid, would this be counterproductive since any small problem would result in skipped steps instead of just the bad wobbling around (I think of it as the print 'fixes' itself). For instance, if the first layer is too close to the bed, the over-extruded plastic will have ridges that the nozzle bumps against until the second layer get put down. As long as extrusion is still calibrated, the second layer lays down the correct width if plastic, and by the third layer the print is laying down as desired (where an over-extruding problem just gets worse). So, any little 'blip' in the printing process will eventually fix itself and is probably unnoticeable in the final print quality. A rigid design would be more prone to skipped steps, which do not fix themselves.

Hmm, this gets me thinking.

Nope. I need people to give me guidance. Another thing with the previous idea is that the threads would need to be drilled out, otherwise the aluminum bad would just drop to the frame.

I agree with you that the Plate or Bed needs some attention, that is why I added the stiffeners with linear bearings. when I had my SD3 apart I noticed the the rivets on the Bed were lose, allowing the bed to twist slightly with little force.
Because my design will be much stiffer than the original I want to allow as much spring as possible. Maybe some type of linear sliding posts that only allowed for up and down motion and restricted twist and X Y movement.

The Z axis upgrade is much more accurate than the original and Including anti-backlash nut it only costs 46 USD.

I agree if you stay on top of your SD3 maintenance wise you get nice looking prints but I want better and as maintenance free as possible.

my printer aces the nickle calibration every time, and i run it about once a week. haven't touched the bed leveling screws in a month (but i did do the thumb screw mod, so my bed screws are tightened down quite a bit).

i don't know if im just lucky to have the most accurate SD3 in existence, or if the machine is just set wrong from the factory and i found the right combination to tweak it.... or perhaps im not looking for .00001" accuracy from my circles.

but i don't see the need for all this upgrading. to me it would just make the moving parts heavier and cut down on motor life + increase stresses on all parts while printing, and cause inaccurate perimeters due to increased mass of the carriages resisting direction change. perhaps im over thinking it, but that's just my .02.

why not design your own 3d printer since you have the means to? why stick to the SD design if your replacing all of it? lol

I would like to take note that this forum has a huge emphasis on upgrades and replacing the stock parts with custom modifications, all of which is what this guy is doing except he is using machined aluminum and market bearings instead of 3d printed parts. If you knock his effort, you are also saying that Ian, lawsy, and adrian have been wasting their time here.

I'm not looking for that type of accuracy either, but at its best my machine printed circles that had a difference of about .5 mm to me that was unacceptable. try putting a micrometer to your circles I bet they aren't as good as you think. But if your happy with your SD3 that's all that counts.

When it comes to mass, the majority of it is wrapped up in the steel rods and stepper motor for the extruder. I don't think the few extra grams I add will make a difference. Also since I am going with bearings motor life should be extended. If I really want I could remove enough Aluminium and make them as light as the printed parts, which I may do once I've created the first prototypes.

I like the Idea behind the SD3 because it is kind of modular, you could increase the size of the printing area and reuse most of the parts. that was my idea after I finish making my parts. New rods wire extensions and a carbon fiber cage, job done.

have you followed the steps for x/y calibration? that's what determines circle accuracy, not machine rigidity, for the most part at least. you would be surprised how accurate these little machines can be if you take a few hours to really dig into the settings and get everything 'perfect'.

my comment on designing your own printer was a bit misguided. i guess i should have recommended something like the rostock delta printer design, im currently collecting parts for one myself. benefits are extremely low mass of the print head that allows for much faster print speeds than a SD, or any other xyz style printer.  and almost double the vertical build area (around 16 inches) and the ability to print outside the bounds of the print bed, something xyz printers cant do.

i think the design would lend itself nicely to being machined out of aluminum, as long as you take care to cut out dead weight areas that don't contribute to strength. which will actually be most of the part. haha.

How exactly do GT2 belts fix this ? I run GT2 belts and pulleys... and it doesn't seem to offer any magical 'fix'.. just a different Pitch...

Thats precisely how it is supposed to work - otherwise when you hit a piece of over extruded print on a perimeter halfway up a model, it would just knock the model clear off the print bed. Or crash the head. Or Crash the bed.

Despite appearances - there is very little 'issue' with the bed other than perception of issues... Its a well tried and true method of running printer beds... And it does return to the same spot pretty much - no where near is this enough of an issue to cause ongoing quality concerns...

Anyway.. continue on with the rebuild.. but I think now you could be accused of slightly starting to over think the quality issues... Perception doesn't always match reality when it comes to how printers work, and what provides the 'best' quality.

the bed is IMHO fine the way it is. Ive never had an offset issue caused by the bed being bumped. the important disclaimer here is that about the 3rd thing i printed when i got my new printer was a set of thumbscrews for the bed leveling screws, tightened them down where the insulation on the bottom almost touches the aluminum rail, and leveled from there with glass bed, this is perhaps why i don't have any issues with spring mounted bed: my springs are under good tension already so they can easily return the bed to position if it gets knocked.