Suggest A Quality, Hackable, 3D Printer for Lab Research

[LaserSteve]

I work at a University. A professor requested that I look into acquiring and modifying a 3D printer for her. Goal is to print various UV cured biomaterials.
As the 3D printer market is crowded and confusing, I figured I'd ask here first. Please only suggest machines that you have seen, used, borrowed, or otherwise have actual experience with.

Requirements:

Budget 1K to 2K for core hardware, and higher is better, we need not be cheap

Easily modified hardware/software. Preferably open source software

High quality printer with easy attachment of a print head, so when I build her the new print head, we do not have issues with mass/attachment /interface etc

High quality, good motor drives, microstepping,  encoder feedback on Z if possible, encoders on XY would be delicious.

Space to attach moving UV curing lamp/fiber optic, or high power leds and heat sink

Prefer double or color changing print head

Understandable Windows Software environment, Linux only OK if the machine is a VERY special  unicorn otherwise.

Lack of useless gimmicks

  Good quality in bearings, leadscrews, high stiffness etc.

100 uM accuracy would be fine

Mainstream manufacturer that will be around a while, no fly by night garbage.

Material will be delivered by a external precision pump assembly to a syringe type nozzle, nozzle to be kept at 37' C for bio reasons

Ability to adjust system timing of everything, for obvious reasons

As much metal in the machine as possible, to aid in modification, ie no 3D printed parts in print head area if possible.

User / Target audience are  Chemical Engineering Graduate Students,  not an  EE community

Suggestions?   Material is a monomer with a gel like consistency that will not kill live cells. Monomer cure time is fairly long, so need to be able to print slowly. Because we may print live cells, the monomer needs very little UV to initiate a cure, by design.    Objects to be printed are small, 4x4x 2 cm max.  I have all the usual university tools, ie machine shop, electronics shop, CNC,  everything but a dedicated programmer.  I have 15 years in the research support business, so "complicated" is the norm in my life.  We do have some experience in 3D printing in the department, but nobody is going to let me Mod/Hack the Stratasys down the hall.


Steve

[Mechatrommer]

Easily modified hardware/software. Preferably open source software

Mainstream manufacturer that will be around a while, no fly by night garbage.

Suggestions?

you are dreaming. the other specs will just increase the magnitude of the dream. the only way to realize this dreaming is ask stratasys to make one special for you since you put some leeway in the... "Budget 1K to 2K for core hardware, and higher is better, we need not be cheap", prepare to pay 5 digits cheapest.

[frozenfrogz]

Since you mentioned Stratasys you seem to be thinking polyjet technology. That is not going to happen in the DIY / open source sector any time soon.
You might look at building your own SLA printer, or check out the Moai SLA kit that seems promising. If you look on Youtube, I remember having seen a real-life review of one the first prototypes.

[LaserSteve]

Actually, I'm not dreaming, and I'm looking for a gantry FDM machine as a platform. Even if its a kit.  My co-workers agree I'm not dreaming, either. We know a good platform is out there. We'll add the UV source, and make a new print head to do our own squirting.   When very basic toys are starting at 375$, two grand or so  should buy something fairly nice.  While I can use some ancient,  professional motion stages we have laying around,  my bet is there is a good commercial platform available, mass produced.

 If  Little Machine Shop and other vendors can crank out a complete Sieg  X2 CNC machine for 3900$, a simple, quality,  FDM  gantry capable of a few thousandths of  of an inch resolution with marginal repeatability  is out there for 2000$.

There are commercial machines doing this at 10,000$ base cost. Someone started with a common platform for making those machines. I just need to find out what it is.

Steve

[LaserSteve]

Moai would be wonderful if I could use SLA, as I work with Galvo scanning for a living.  In fact I used to be the caretaker on a early galvo based commercial SLA.

 However I need to squirt a thick gel  from a blunt syringe needle with a fixed curing light  source aimed at just past  the needle.  Most other  published prototypes in this area use 365 nm air cooled  LEDs in flood mode. A few use a 405 nm laser diode.  All the published work uses a stepper driven  syringe pump or a very carefully balanced air pressure delivery system to pressurize the fluid.

The custom delivery hardware we can do in house, at very low cost.  What I'm looking for a is the platform, basic motion hardware, and basic software.
More of less to avoid re-inventing the wheel and to save time.

I may be aiming high in the first post, but there is a nice compromise out there.   Google "3D printer" for yourselves and get flooded with results. I'm trying to filter those results by depending on EEVBlog's collective experience.

Steve

[CM800]

Here you go:

http://www.ebay.com/itm/Zeiss-Numerex-CMM-3D-Coordinate-Measuring-Machine-for-PARTS-broken-glass-scale-/162353372295?hash=item25cd03dc87:g:WU0AAOSwEzxYdQKt

Buy this, then spend some good money on a proper motion controller, ACS, Delta Tau, Elmo Motion Control, OMS Motion.

 I envy you having access to the US's ebay, they have some great stuff you can't get in the UK often.

[LaserSteve]

I'll look for a dead  CMM locally, that might just do it.  Gantry, nice hardware, and drivers are cheap.
Still would like to consider a cheap printer.
Ebay is not available to us at a State ran facility, which is one of the reasons for looking for a commercial item.

Steve

[CM800]

I'll look for a dead  CMM locally, that might just do it.  Gantry, nice hardware, and drivers are cheap.
Still would like to consider a cheap printer.
Ebay is not available to us at a State ran facility, which is one of the reasons for looking for a commercial item.

Steve

Find a friend who works at a machine shop or something similar, make a deal so he buys stuff off ebay for you and slaps a small fee on top so you can get your facility to purchase it

[691175002]

Most hobby-level 3d printers would be questionable in this application because they are very lightweight and have extremely poor rigidity.  They work around mechanical limitations by using very light printheads (ex: bowden).  You need a machine that can move more than 50g of printhead, and ideally something rigid and robust enough to ensure your performance is never limited by motion quality.

Most printers can technically repeat to <0.1mm in ideal conditions, but if you rest a cup on the bed or lightly push on the printhead everything stars to flex.  Printers with poor rigidity will also lose alignment when moved.  There are probably 3d printer kits that can do what you want but I don't have any firsthand experience with specific models.

Pretty much any used CMM, liquid handling robot, pick-and-place, or similar would be suitable for this application but you could go overbudget depending on surprises and retrofit complexity.

If you don't need high speeds or accelerations I think a taig/sherline cnc mill would be ideal.  You might even be able to buy one without the spindle to save some dollars.  They will repeat to <25um and could probably survive getting hit by a hammer.

[691175002]

Actually just buy one of these: http://sherline.com/product/5000-cnc5100-cnc-standard-mill/

The primary disadvantages of using a mill for additive processes is that screws are slower than belts, and the high mass of a milling machine limits your ability to make high acceleration direction changes.  Neither limitation seems to be a problem for your application.

You will have to supply your own steppers and electronics.  Since the mill will not be fighting against cutting forces you can use low-cost 3d printer electronics and steppers, at the low end it could be running for <$100 (note you need Nema23 steppers).  You could run servos and a proper motion controller if desired, but I don't think the cost/benefit tradeoff makes sense.

[kkessler]

I recently did some printer shopping in about that price range, and the nicest I found where the Lulzbot Taz 6 and the Makergear M2 (I got the M2 and am happy with it).  If these do not meet your needs, you probably have to kick up the price range.

[LaserSteve]

Thanks KKessler

The M2 is made forty-five minutes  from campus and looks like it has enough open space to meet our  needs.  That is a good start, I'll see how "open" the software is.

Steve

[Kilrah]

I'd also suggest looking at CNC mills rather than 3D printers for their rigidity, it's completely open mechanically so you've got ample space, and you can mount 2kg of stuff on that and be laughing, they're built like tanks in comparison for obvious reasons.
The typical 3020 CNC has about 60mm Z travel which is sometimes limiting in a CNC application, but if your target is 2cm no problem.

[mikeselectricstuff]

I'd also suggest looking at CNC mills rather than 3D printers for their rigidity, it's completely open mechanically so you've got ample space, and you can mount 2kg of stuff on that and be laughing, they're built like tanks in comparison for obvious reasons.
The typical 3020 CNC has about 60mm Z travel which is sometimes limiting in a CNC application, but if your target is 2cm no problem.

I was about to say exactly that -  my reading is that you basically just want an XYZ platform to hack your own head on to - if the limited z height isn't an issue, a cheap Chinese CNC is probably the way to go, using whatever stepper drivers you like - plenty of control options out there. I'd go for the 3040, which has ballscrews. 

 

[Mechatrommer]

so Moai the Kickstarter and that Wang Gong CNC are mainstream not the fly by night garbage? that will survive in generations of lifetime?

Mainstream manufacturer that will be around a while, no fly by night garbage.

we are talking about a mainstream that will hand over their firmware here and willing to support for the education... try adjust the system timing for the OSS reprap alone.

Ability to adjust system timing of everything, for obvious reasons

[Kilrah]

I think everybody has de facto ditched that aspect since it's purely unreasonable given the other constraints OP has given, consumer 3D printer manufacturers that offer products in the budgets he mentioned have at most about 5 year life expectancy.

They can probably find someone who will last and support them but they'll have to add a 0 or 2 to the budget, that solution was mentioned already but was explicitly excluded right in the OP.
Then it makes little sense if they have the resources to build the machine in the first place, supposing it breaks parts are generic enough that they can replace them / replace the entire machine by one that's not even exactly identical (control is standard enough as well).

[Mechatrommer]

Most hobby-level 3d printers would be questionable in this application because they are very lightweight and have extremely poor rigidity.  They work around mechanical limitations by using very light printheads (ex: bowden).  You need a machine that can move more than 50g of printhead, and ideally something rigid and robust enough to ensure your performance is never limited by motion quality.

3d printhead like below is 500g, excluding its rig weight to hold to the gantry. OSS reprap marlin fw and my modded prusa is happy to run it without miss steps (well, almost always). what hobby 3d printer that can only bare 50g of head? i dont know, sounds more like a plastic toy to me.
http://www.ebay.com.my/itm/3D-Printer-MK8-single-Extruder-Print-Head-for-1-75mm-MakerBot-Reprap-Prusa-parts-/131800846468?var=&hash=item1eaff15084:m:mrj_S07A2DKNPz51XYZR9Ng

and bowden head like below weigh 100g excluding tubing and rig... yeah i have both as spare parts..
http://www.ebay.com.my/itm/3D-Printer-Extruder-Metal-J-Head-Extruder-V6-Nozzles-1-75MM-M111-/311809342464?hash=item489948fc00:g:ZH4AAOSw~AVYr0Zh

I have all the usual university tools, ie machine shop, electronics shop, CNC,  everything but a dedicated programmer.

if you are willing to ditch out the "Mainstream manufacturer that will be around a while, no fly by night garbage." requirement...

building gantries is the easiest part, you should be able to make it quality and sturdy with what you have in the university with some reference/layout provided free in the net. or better refer to your mechy expertise in the house. the harder part is try to understand the whole built system if you havent seen one (board,drivers,motors,wiring etc), will take weeks and lot of wasted money if you try to learn from zero. OSS fw difficulty is somewhat in the middle if you are willing to reuse OSS reprap fw, there are many variants in the net, will be the hardest part and probably take months or years to build from zero. probably took weeks trying to understand and fault finding of existing reprap fw.

so my advice, you may buy a cheap hobby printer like this... and then build a new aluminium body and suitable head/rig for it. strip out motors, board, necessary ee, mechanical parts (usually they are decent quality) etc and put in your new body. learn how to modify and upload marlin FW using arduino the ide and off you go, quick.
http://www.ebay.com.my/itm/Tronxy-Upgraded-DIY-3D-Printer-High-Precision-Reprap-Prusa-i3-Free-Shipping-8GB-/322426940695?hash=item4b1224ad17:g:nZEAAOSwCU1Yp5d4

attached is my reprap/marlin compatible build, using mdf wood + some acrylic, not sturdy enough for your need but enough for my need (making pcb etch resist and drilling holes). if only i have the usual tools in your university i can make it better. i even recycled inkjet printer's smaller stepper motors for it as you can see in the pic. next plan is to add solder paste dispenser and pick n place head when i have time. all control will be from gcode command file and my homebuild pc sw... fwiw..

[LaserSteve]

       We have a call into Makergear. We're going up for a tour,  taking a long hard look at the M2, as a starter kit. They were more then friendly, and have often supported customization. We'll look at the Lulz machine and possiby the Chinese CNC as well. We have ancient  Sherline in another building, but it's in use.

  At the same time, I'll order some driver boards to close the loop on some Aerotech precision motion stages that I pulled out of the campus disposal chain. The motors and encoders are there, and the home CNC market has driven the cost of a DC servo driver  down to less then 150$.  As these were designed for massive loads and long travel,  they will be overkill. But they are ballscrew based. The four racks of 19" drivers are easily replaced by small boards today. That is the long term machine.

Funding for basic research is tight these days. If I went to the right  office and asked for an extra "zero" that could take one to two years to obtain, if at all.. How grant money can be spent  has changed drastically in the last twelve years, and is more student focused, with less emphasis on hardware.    I'm forever mindful to spend corporate donations and public money very  carefully.  That often means modifying COTS products, where before I would build something for long term use.

I know what my last in house , comparable motion system cost to build,  and while it had to run while charged at 30 Kv,  the core parts ran about twice what a Lulzbot or M2 would cost. Not to mention the time needed to purchase parts from multiple vendors and to CAD it for the machinist.

  Another consideration is always lab space, which is a precious commodity. Finding space that can support a 1400 pound gantry from a CMC is difficult. 

Thanks Ladies and Gentlemen,  this thread allowed us to converge on a "good enough"  solution rapidly.

Steve

[sokoloff]

Might also be worth looking at Prusa, which are fully open source. It does use 3D printed parts extensively for the printer, but that's more of a feature than a bug to me, given your desire to modify the thing. You could easily get 2 Prusa i3 Mk2 kits, keep one as a working 3D printer, and use it to build any modified parts you need along the way with the second kit.

I have a clone of a FlashForge Creator Pro (which is also fully open source, hence how easily and often it's cloned), but at this point if I had it to do over, I'd buy the Prusa. (In fact, I may very well buy one anyway, but the existing printer is chugging along quite nicely still.) I do agree with prior posters that printers in this price range are several year, but not decade lifespan. That said, they're cheap as hell, so it doesn't matter all that much. You cost as much in a half week (or less) as a researcher.

[691175002]

What hobby 3d printer that can only bare 50g of head? i dont know, sounds more like a plastic toy to me.

They aren't missing steps, but they are overshooting and producing ringing artifacts.  Just looking at the extruded plastic makes it clear that a reprap style printer is not holding within +-0.1mm of the commanded path while in motion.

Maybe thats fine in this application, but in my book any platform that needs a bed realignment every time it is bumped or moved is not ideal.

I'll again point out that a base Sherline is $900 and comes made in the USA from a company that has been around for fifty years.  A cheap chinese CNC will also provide high quality motion and the ballscrews should have zero backlash as well.

[MDM3D]

My recommendation if you want open source software would be to use linux cnc with the 3040 cnc it will hold the tolerances and with proper acceleration settings ringing and deflection will be negligible the only thing I have not actually done with Linux cnc is is coordination the extruder and cure leds with the 3 axis motion but it should be possible by modifying the rigid tapping code. Otherwise use repetier host and firmware with slic3r or simplify 3d to generate your tool paths.