There are two popular ways to get plastic pushed into a hotend, bowden and direct extruders:

Bowden: while a “light extruder” might initially seem like a good idea in rl it plays out exactly the other way around:

  • Its almost impossible to dial in pressure advance with a bowden system which is a basic need for going fast with a printer.
  • Ultra high retracts in the 3mm range severly slow down print speeds as a lot of time is spent on waiting for retracts.
  • From 100mm/s onwards print defects like blobbing and inconsistent extrusion in corners start to really show.

Direct Extruders:

  • Get a light, geared extruder like the E3d Titan for example, the gearing allows for small “pancake” style nemas, reducing overshoot on x significantly.
  • With a light direct extruder and a light bed its possible to reach equilibrium between them, resulting in an easy to tune / fast to drive motion system
  • A direct extruder allows for ultra fast / ultra small retracts ( 0.4mm / 2k accel / 75mms is enough with pressure advance ) significantly reducing print times again.



The bed is a rather simple affair as there are only very few options to build it:

Top Sheet:

Phenolic resin bound sheet at 1mm, common product names are: Pertinax, Hartpapier etc.

As we are printing on the phenolic resin surface the exact type of fibre bound within doesnt really matter.

Its compatible with pretty much any material extrudable, Dry sand the surface with 400grid, attach it to the aluminium bed with 3M 468MP.

Bottom insulation:

Again there are no other options than exhaust insulation, comes in slightly under a mm, one side aluminium foil, sandwiched glassfibre in the middle, self adhesive high temp resisting sheet on the other.

Bearing blocks:

Dont use aluminium ones, way to heavy, printed holder are perfectly fine here.


Trapezoidal rods

There are a lot of different ways one can set up his z rods on an I3 style printer, only few are feasible for fast printing though, before installing the z rods one needs to make absolutely sure they are dead straight ( Roll on a flat surface ) If they are not theres pretty much no remedy, so exchange them for straight ones.

Stepper – Rod coupling:

Either get rigid couplings or stepper with inbuild TR Rods, reason is simple: we need to kill any play possible in the z rods to get the extrusions as evenly layered as possible.

Trapezoidal Nut:

Either get pretensioned nuts like shown in the picture or make sure theres absolutely no play in any direction. Upward play is bad as during fast printing there are numerous situations in which the extruder gets pushed up more than his own weight pushes down / the hotend tip melts in short timeframes. ( Extrusional force, crossing slightly overextruded print parts, step upward bending overhangs etc etc )

Top loose bearing:

As we are moving fast the top of the rod need an radial support, otherwise it starts to severely bounce around, a simple remedy are 8mm bore skate bearings ( 608 ) in an printed holder.

In depth summary and Powermod for the KLD-2150 K1

– Out of the box the cover was cracked ( to be expected the way its packaged )

– Spindle / stepper connector misaligned and not secured

– Power button loose

– Backside PSU connector cold solder spot came of after repeatedly disassembling the printer

Problems with the working unit:

– Raspberry PI3 oiverheating, needs an additional fan so it doesn’t go into thermal protection and downclocks.

– Powder coated build surface: while being able to directly print to it and having sufficient adhesion, sometimes big parts curled up Fix sanded the plate first with 80 grit on an orbital sander to get the powder coat of and afterwards 40 grit manually to get really deep scratches in it and now everything sticks perfectly.

– Distance between Led / cooling array and metal light guide to high leaking a lot of uv – Fix and Part for fix in the comments.

– Rotational sway between mounting block and build plate, check comments for a Fix

– 12v 6v Buck converter only 32W same as in its smaller brothers YHD-101 / KLD 2150, despite 6x 10W LEDs build in.
Fix https://plus.google.com/+lenne0815td/posts/haQQCMvFnqN

– Thermal grease for UV LEDs badly aplied

Wiring them in series has the added benefit of no brightness deviation which is a problem with parallel wired LEDs.

– Aluminium posts connecting top and bottom plate misaligned.

– Cover not really fitting the unit ( not enough height to allow Zs full range etc )

– Psu only rated at 12v 5a barely enough with a proper buck converter Fix bought a generic 10a 12v unit to replace it

– The protective film wasnt removed from the lcd and the flap to remove it hidden under the black tape. Fix I removed the black tape completely and changed it to a much thinner electrical insulation tape as the thick black one gave me huge problems aligning the build plate first.

– The right calibrating space between display and build plate is three layers of standard office printing paper, it needs to get stuck ever so slightly on all 4 corners of the display ( Check one after another )

– Metal light funnel distorted, Fix unscrew from below and align properly.

All in all no deal breakers, generally the unit has been built way better than expected:

– Proper z linear rail

– Ballnut screw with flexible connector to the stepper, loose and fixed bearings.

– Solid ! all aluminium build ( results in a really heavy and stable machine )

Display Specs:

192mm ( 2560px / 75my / 338,6^ ppi / 13,3^ ppcm ) x 112mm ( 1600px / 75my / 338,6^ ppi / 13,3^ ppcm )

Onto the software:

Photonic3D on a octopi distro ( no remote access as far as i can tell beside photonic ) / MOYIU “My Power XYZ” with Sprinter preinstalled and working ( as in able to talk to the machine 😉 )

Display properly set in the settings.txt file, working at full resolution right out of the box.

To get Photonic in a usable state the following changes had to be done:

;** Header **

G21 ;Set units to be mm
G91 ;Relative Positioning
G28 ; Home Printer
M17 ;Enable motors
;<Delay> 3000

;** Lift Sequence **

G1 Z${ZLiftDist} F25
G1 Z${(-ZLiftDist+LayerThickness)} F170.0

And nothing in the shutter section.

Later down this comment train a lot more sophisticated settings for photonic are available.

Heres a link to a pi image which should work straight out of the box and employs tweaked dynamic Lift cycles.



Latest Version 1556 ( 03.07.2017 ) in a complete package including +Bud Hammerton s latest firmware for the ramps board and a quick readme / install instruction.


Updated Base Resin profile:


NanoDLP wont work without a flashed ramps, so dont try to laze out of it.

DruckWege Resin Type D Transparent / Blue / White – A comparison.

Safety: Type D resins do not contain any volatile organic components (VOC) Safety Datasheet https://www.druckwege.com/app/download/14373847724/1310_GB-en.pdf?t=1496675081

Initially ive been underwhelmed by DWs mainly epoxy based resin, right of the printer it could match FTDs resin in nearly every aspect but there didnt seem a lot more to it so the question i asked myself was what would make me choose it over a different fast curing acrylic resin, what would even justify a higher price ( which only applies to colored versions, clear is just the same price as FTD ind. ) than a classical well dialed in acrylic resin like FTD ind. ?

Im gonna answer that most definitely.

Smell: Muted, like the last batches of FTD ind. i got. When heated theres a discernible epoxy smell to it and something else which i couldnt pinpoint, pe maybe. Certainly mute enough to use in an apartement, given the usual precautions of ventilating the room well. Im a bit biased there though because i like epoxys smell 😉

Curing times: especially for the lower powered uv printer crowd the most important factor, with my 60w UV 8.9inch printer:

– FTD ind unpigmented: 2,8s

– DW Typ D Unpigmented: 3,5s

– DW Typ D Blue ~RAL 5012 Light blue: 3,5s

Another point worth mentioning is base Layers need a fair bit more cure time to stick properly, i could get FTD clr to stick @ 15s ( when it started sticking further base layer curing just increased time, not adhesion ) DW needs atleast 30s, after that it sticks very well ( as described below )

Shrinkage: Again similar to FTD but it seems to highly depend on the models geometry aswell, its certainly is one of the resins which dont shrink much.

Print Bed adhesion: DW sticks well to a roughly sanded aluminium plate ( 40 grit manual as the last step ) in fact it sticks a lot better than FTD (which itself already sticks well especially on aluminium ), this allowed me to raise lift speeds up to ~F175 ( retract F700 ) a speed where i had many failures with FTD before, the parts just wiggled loose until they completely dropped from the build plate.

Since then i backed of lifting speeds a fair bit ( F145 now max ) but in turn could hugely reduce lift heights ( 1mm lowest / 2mm heighest / 2,5mm for burn in / small parts ) as DW doesnt seem to stick much to the fep at all.

Again a discernible difference but nothing major either, certainly not enough to sway me one way or another, especially as i can drive FTD ind clr down to 4.9s cycle times @ 0.05 which is imo insanely fast for a uv / lcd printer.

Light bleed Again comparable but in certain situations DW fared better, ive never had big problems with FTD but vertically aligned channels just come out a bit better with DW, the base layers dont grow as much either.

Details: I didnt look much at the details the prints gave me as my main goal was to evaluate the brittleness, seemed comparable / slightly better than FTD though, when i have pigmented DW ill look more closely at it.

Heat sensitivity: DW is more sensitive to heating resin, first few layers had definitively less fep adhesion than later ones, together with good print bed adhesion this didnt mean much, in the future i hope that we can use it to reduce cure times.

Onto the reasons why you will never buy another all purpose resin again

Some generic blabla here: Brittleness of acrylic resins is really high, some go up to a surface hardness in the regions of shoreD 90 which is really hard but worth absolutely nothing in any real world application because they remain brittle, ive only ever used a single FTD part in my quadcopter which exploded on the first light crash.

Now this only applies to FULLY CURED PARTS, dont make the mistake i did and compare DW right of the build plate with anything else, the changes taking place while fully curing are huge, certainly a lot more important than with other solely / mainly acrylic based resins.

DW is strong. I dont mean “yeah a bit stronger but i can get FTD f.e. from my local dealer why should i care” strong. I mean seriously strong, so strong in fact that i cant physically break the part from the picture.

Im a strong guy, i do physical labor every day ( Not true sometimes i take days of and pretend to be god almighty on internet forums 😉 )

I just cant break the two bigger posts. I can grab them perfectly, i can wedge my fingers in between to load them even higher than i could pull with my arms but they wont break, i stopped trying after a while because i got scared im gonna severely cut my fingers if i finally succeeded.

But wait, theres more…

Aftercuring / Cleaning: Mentioned before, to reach its full potential DW needs to fully cure ( Typical UV Nail station 2+ Hrs for tiny parts )

When cleaning another property of DW becomes apparent, FTD ind parts never fully cure on the outside as oxygen creates an inhibition layer which cant cure, we can counteract that by washing them very thoroughly with IPA which isnt perfect either as that might damage the surface to a certain extend, DW just fully cures and thats it, no more messy parts even though they sat for hours under uv cure.
As of now ive tested this multiple times, DW fully cures exposed to air, ftd and most other resins always stay sticky, so underwater curing becomes a thing of the past with DW.

( Which in turn makes me doubt that continous printing is going anywhere because most of these systems would be restricted to purely acrylic resins due to the way they use oxygen as an curing inhibitor )

Second part of this comparison:

DruckWege Resin Type D pigmented – A comparison.

Now that ive established the basics and changed everything a bit around to reflect more precisely on my findings im having a closer look at the pigmented version of the resin:

Blue ~RAL 5012 Light blue

Pigment saturation: Very low, parts come out translucent / blueish, dont expect results like FTD regular snow white or Photocentric3d Hard Grey which both contain a lot more pigments for the same volume. This has obviously some drawbacks, but in turn some benefits as well.

DW RAL 5012 Light blue strongly reminds me of translucent filament prints in its appearance.

Pigment retention: I couldnt make out any settled pigments after two full days in the vat, so i guess this wont be a problem even for the longest of prints, will add more info once it sat for longer in the Vat.

Pigmentation effect on cure times: none. ( The fastest pigmented resin i tried to date )

Pigmentation effect on post cure handling: it certainly extends the post curing time but due to its low pigment saturation i guess classical uv curing is still possible ( UV light can still penetrate the parts so they can still be post cured throughout )

Pigmentation effect on part strength: None.

Details: Again its hard for me to tell because i have a low resolution printer ( 75my pixel pitch ) without AA the parts look pixel perfect to me but maybe someone with a 5,5′ 2k screen can chime in and offer his thoughts on it.

Color post cure: Same yellowing like the unpigmented version, turns greenish after a 20min / 20min cure on a UV Nail curing station

All in all nothing really spectacular came to light while testing the light blue resin, it just keeps all the great properties of DW unpigmented and adds a blue tint to them.

I think the focus when developing this resin was to keep the curing / part handling similar to unpigmented and in that regard DW clearly succeded.

The amount of pigment is enough to mask the yellowish appearance of unpigmented fully cured resins, so especially functional parts dont look like your granny printed them in her early days.

Pictures of parts / raw resin in the comments.

White ~RAL 1015 Light ivory

Pigment saturation: Medium, parts come out almost fully opaque, this in turn makes working with the resin more difficult opposed to Unpigmented / Blue, expect this resin to only stick properly to a perfectly dialed in printbed, it takes no hostages on the slightest of discrepancies.

Pigment retention: I couldnt make out any settled pigments after a full day in the vat and by now ive pretty much run out of white while testing.

Pigmentation effect on cure times: Fair – i had to raise my base layers to 45s and the normal layers to 6+ seconds ( I couldnt finalize testing as ive ran out of white, i guess finally i would have arrived in the 7,5s range )

Pigmentation effect on post cure handling: Doesnt seem to cure fully throughout, ill report back once i got a part under UV cure for longer.

Pigmentation effect on part strength: Still strong, a bit less than unpigmented / Blue

Details: Again its hard for me to tell because i have a low resolution printer ( 75my pixel pitch ) without AA the parts look pixel perfect to me but maybe someone with a 5,5′ 2k screen can chime in and offer his thoughts on it.

Color post cure: Same yellowing like the unpigmented version, turns yellowish after a 20min / 20min cure on a UV Nail curing station

White seems to go the mid way between typical low pigment saturated resins and highly saturated ones, if thats a reasonble way of doing it stands to question, i would rather either have a low pigment amount to just tint the resin a bit for contrast when painting without affecting curing time to much or an all out saturated resin for final parts of the plate.


For now i can just recommend to everyone whos looking for more durable parts to order atleast a test bottle of DW unpigmented even if it doesnt meet your specific expectations you will certainly get usable parts of it.

I will constantly update this OP if anything worth mention turns up but for now im just happy / exited that i can finally start designing functional parts again just as i did with my fdm printer.


Anton asked me for feedback on how to possibly change the resin: You well exceeded my expectations for a mainly epoxy based resin, leave it as it is, its great !

And just to be clear because this might read as an all out advertisement for DW im in no way affiliated to them, i talked to +Anton G. twice on the phone and he seemed to be a really nice guy but i wouldnt give a flying fuck if his product was bad or just meh and this comparison would have a completely different tone to it.

Driving UV Led arrays with a tinxi® 600W DC 10-60V CV / CC driver

I wanted to give this a bit more visibility as it might be of interest for anyone looking to drive Led arrays, so heres a short description for solving the problem with a modified KLD 2150 as an example.

Out of the factory my KLD 2150 suffered multiple problems concerning the power train, especially the LED array was wired incorrectly and way underpowered, the uv leds needed new thermal paste.

LEDs need to be wired in series and driven by a constant current, wiring them parallel is only an option for larger arrays where its impossible to get the voltage levels required for a complete series wiring, multiple parallel rows in series are an option there.

Wiring Leds in series has one main reason: due to the nature of Leds when wired in parallel they exibit different output levels, something we definitely want to avoid in a UV Lcd printer.

Based on this knowledge i searched for a suitable stepup / cc driver and found this monster on amazon: https://www.amazon.de/gp/product/B01C2JB0TQ/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1and

It should be widely available as it seems to be one of these standard Chinese psus which are distributed worldwide.

Beside the obvious size drawback ( Check dimensions ) it can provide power for most uv led arrays you can possibly think of.

after getting it i started to work on rewiring and resoldering the printer to prepare it for the tinxi:

– Ripped out all cabling from the manufacturer, i suspect the main power switch to be the bottleneck here but i just redid everything to be sure there are no other cabling problems rearing their heads later on.

– Resoldered the uv led array to run in series.

– Put in a 33A 12v psu opposed to the 6A unit that came with the printer.

– Superglued a heatsink to the ramps mosfet to keep it cool.

After that was done the interesting part started, dialing in the tinxi to provide the exact amount of power needed by the Leds:

– When you want to use a constant current driver for leds you start of by setting it to the total voltage your leds add up to and add about 10 to 15%, this gave me about 22V ( 3,2V * 6 * 1,10 )

– The next step is to dial the i-pot all the way counter clockwise to set amps as low as possible.

– now you can power up the tinxi, set a typical multimeter to the 10A mode and wire the multimeter in series between the out side of the tinxi and the uv led array, afterwards you can start to turn up the i-pot until it rises to the desired level ( 3A in my case )

– reconnect everything and you are done.



Forget everything you think you know about 3D printers.

Throw away your shitty bowden setup, it wont go much over 100mm / s without severe blobbing / insanely long retracts holding back total print speed.

Uninstall Arduino IDE, wipe marlin of your hard drive and regret you ever learned about it.

Roll up your sleeves and get to work – awesomeness awaits.





  • 30 V Psu ( Step up can be used from either 12V or 24V PSU as we only want to drive 2 Step sticks of it )
  • Rigid Z couplings
  • Anti backlash TR Nuts
  • Fiberglass reinforced belts
  • 20 Tooth / 5 mm bore pulleys / Idlers
  • Direct Extruder ( Preferably E3D Titan or similar )
  • TMC 2100 / 2130 Step stick x 2
  • Multiple 40mm Fans, one for every stepper mounted to plastic, atleast two for ramps, rumba, whatever.
  • Pertinax ( Hartpapier ) Sheet as printing surface
  • Fiberglass insulation for the bottom of the heatbed
  • Light bed bearing holders
  • Light extruder holder



  • Octoprint
  • S3D


  • Klipper
  • Smoothieware

Klipper pt6: Gcode / console commands

An overview over the already supported G-Codes / Klipper specific console commands.

A full list can be found here:



McodesGcodesKlipper specific console commands
M106 Sxxx
RESTART ( Restarts host software on the pi, do this after every config.txt change to reload it )
SET_PIN PIN=my_pin VALUE=0 ( Replaces M43 )