Its name is Agony, it took one year of my life but it's finally working as I'd hoped.

This is it, there are many like it, but this is mine.

Still beta of course, need better cable management.

It's a relatively large printer. 610x610x610 frame. about 400x400 print area. That was the main challenge. I find that anything larger than about 300x300 becomes exponentially more difficult to build.

Anyway it can happily print at 200mm/s without too much distortion. The following are printed at 150mm/s external perimeter, with a 0.5mm nozzle (my new favorite nozzle size).

Why

It started because I couldn't find a printer with decent print area and reasonable speed. I also wanted to keep it super simple with few easy to print pieces and easy to source off-the-shelf hardware.

All extrusions are 550mm, all rails are 500mm, you don't need special tools to cut/drill/tap, everything can be easily sourced.

If I had to start over today I would probably build an Annex K2... but it is what it is and learned a lot in the process.

A bumpy road

Agony was originally a CoreXY (named Corezilla), but I couldn't solve some resonance issues I had and that generated heavy vibrations at certain speeds. So I converted to a cross-rail cartesian. I then found out that the resonance was due to the belt tensioner design, but it was too late but I'm happy I moved to a cartesian system anyway.

Another issue I had was with z-banding. Apparently it's impossible to get decent leadscrews (or ballscrews) and anyway it's incredibly hard to align 3 of them. The revelation arrived when I saw somewhere a sort of magnetic cushion to be placed on the Z-carriages.

Basically the leadscrew wobbliness is not transmitted to the carriage and now the Z is just perfect.

Finding the right motors was another adventure. Long story short: 0.9° on the XY and 1.8° on the Z. I tried quite a few motors from steppersonline and LDO but ultimately the best I've found are the Moons MS17HA6P4200.

One suggestion, if you need a big print area but you don't need a lot of height, just make the printer as short as possible. That helps a lot with stability. If I had to redo it, I'd probably limit it to like 250 on the Z.

Lastly the gantries.

I tried with the bare rail and with a 15x15 backer. While the bare-rail solution is better in many ways, at the end I had to use supports for the rails. That was a game changer in terms of stability. I was using the rails vertically and at high speed they tend to "flap" at sudden moves change. So I put them on 15x15 extrusions and placed them flat.

Hotend + Extruder

It's basically a custom extruder based on the now famous small LDO NEMA14 and the guts of a bondtech extruder. The hotend is an original Mosquito (the standard one) and from my initial tests everything's fine up to 200mm/s, over that I start getting extrusion problems on acceleration... but haven't tested it carefully yet.

Despite being so small the 45mm Delta blowers push a lot of air, so at the moment I'm happy with them but I might switch to bigger fans if speed requires it.

Duet & RRF

Now the hard part.

I wanted to use a Duet3 mainly because of the toolboard. I jumped in a little too early unfortunately and the experience was really painful. I was able to fix most of the issues but at the end I hit my head on a firmware bug that produced inconsistent extrusion. I understand that bug is now fixed (and a new toolboard is out), but it was the last straw so to speak so I removed everything and redesigned the printer for the Duet2+Duex5 (that I already had on another printer) but this time I wiped RRF and put Klipper on it.

Now. I owe 90% of what I know about 3d printing to this forum and the help of the generous community and I do love the "everything is gcode" RRF philosophy... but Klipper really made my printer fly.

It was painful at the beginning and the community is much less supportive, you are mostly on your own, but the documentation is good. It's in a constant beta(alpha?)-stage but at the end it worked wonderfully and I feel that any problem I have is now caused 100% by me.

The Duet is a wonderful piece of hardware, something that most of the chinese boards can only dream of BUT my suggestion is: if you need a quick replacement for your printer totally get a duet2/3 with RRF; if you want something complex and custom... in my experience a fysetc board and Klipper is a better option. I hope you'll forgive my candor. I'll still be using the Duet boards I have for other projects but I guess to each its own.

Want more?

I would never ever suggest anyone to build Agony, but if you want to look around, head to onshape and search "Agony" in the public projects. You'll find it, it's the only one... but be warned it's a very messy project.

Thanks, sorry for the long post.

I have a slice thermistor too and my educated guess is that it's not reliable at (very) low temps. It makes sense being a thermistor for high temps. I added H6 option to add 6 degrees in the low end, but it's purely cosmetic.

I checked with a thermal camera and I'd say as soon as you reach 100C the reading is accurate.

the returned M307 won't probably work, at least it didn't for me and I had to manually tune it to avoid heater errors. I'll have to find a way to embed a thermistor into the bed itself instead of relying on the one with the heater.

What you need to know about tweaking spreadCycle/stealthChop is a little all over the places. Here below a list of all the docs you need (I read them all!):

• Tuning Stepper Motor Drivers
• Stall detection and sensorless homing
• Reducing Stand-Still Noise
• TMC reference (pdf)
• Parameterization of spreadCycle (pdf)
• Help needed with Duet 3 6hc TMC5160 StealthChop config (forum)
• Duet 3 StealthChop 2 Tuning for Nema23 (forum)

My issue is as follow.

Custom CoreXY. XY are 0.9° steppers. I get very bad vibrations at about 60mm/s (resonance?). It happens only if both motors are moving, diagonals are fine.

It's not a noise issue, I don't care about that. The problem is that the printer vibrates A LOT.

My motors are rated 2A and I run them at 1.4A. If I drastically lower the current the problem alleviates (≤1A).

I tried tweaking SpreadCycle but nothing seems to be helping. Increasing TOFF over 6 makes it noticeably worse.

I was able to enable and configure StealthChop and that seems to be solving the problem completely, but I have to set it at about 100mm/s otherwise I get "false positives" (ie: spreadCycle triggers too early) and I get a lot of jerks during printing.

So my only option is to always keep stealthChop but both @dc42 and @Phaedrux advise against that over and over again (I print pretty fast... not warp speed but still).

I tried with different motors from different companies (LDO and moons) and they have more or less the same issue at slightly different speeds. To be fair the two sets of motors are very close in specs.

Dampening the motors makes no difference. I also tried different kind of brackets and idlers.

I can't be the only one with this issue... any suggestion or I can finally tableflip this thing? Thanks!

I have 2 part fans connected together and 1 thermally controlled tool fan, they are all 12v connected to the duex5. I'm in a corexy setup and thinking of reducing the number of wires that go from the head to the duetwifi/duex5.

It is my understanding that the positive lead is shared with all fans. My idea was to use just 1 pos and 2 negs. On the duet side I'm connecting the 2 part fans together to one header (+ and -) and the tool fan to another header (just the negative since the positive is shared on the other header). Does it make sense? Does that impact on the total max amperage? Just wanted to check this config with you before burning my duet Thanks!

PS: I wanted to add that both this forum and the wiki are an incredible source of knowledge, I was able to go from zero to full functioning corexy by just RTFM so thanks guys!

okay I was ready to swap the wifi module but before that I tried to just "extend" the antenna... believe it or not it worked. I hadn't a single disconnect since

Basically cut the onboard antenna and soldered a new one. The antenna must be of a very specific length (check 2.4 antenna length) but apart from that it's a 5 minutes hack.

Now, I don't suggest to anyone doing that... but it worked for me.

So I got this texture on a corexy I've built. It's 3x leadscrews+rails.

The pattern looked very regular, so I thought z wobble.

Changed z motor mounts, leadscrews, couplers aaaand... same result. Nothing changed at ALL.

Pulled the trigger and bought a digital microscope.

I printed a 1-wall cube. This is the result of two opposite sides (sections).

What the hell is going on? The wobble in just on one side of the wall?! Help me Duet3d Forum, you are my only hope!

Extruder is bondtech. Extruder motor I tried with a 1.8 moons and a 0.9 LDO. Nozzle is brand new high quality. If I change layer height the "wobble" seems to be the same (that's why I thought it was caused by the z axis).

just power the rpi separately, if the duet3 is already powered via 12-36v you don't need to send the 5v back from the rpi to the duet.

I haven't looked into it, but I'd think you need a common ground for the main vin and the rpi 5v. I bet either the forum or the wiki have the answer to that.

this is very interesting, following this thread.

I'll have a look at your code soon and possibly compile it on Linux.

Keep up with the good work

@arhi said in I thought it was z wobble...:

I can't see from the images, is the issue parallel to the XY plane or is it under angle)

they look pretty much parallel.

@mrehorstdmd said in I thought it was z wobble...:

So the line width is varying but only to one side...
Is the hot-end perpendicular to the bed?

yes but the width variation is not always on the same side. if the head wasn't leveled shouldn't the wavy lines be always on the same side?

but I'll double check that. Thanks!

How tightly screwed down is the drive gear adjustment screw on the extruder?

I tried to make it very loose and very tight but I can't say I noticed much difference.

Are the drive gears clean?

Brand new.

How consistent is the diameter of the filament you're using?

I'll double check that but it is supposed to be good quality.

Are you using any pressure advance or nonlinear extrusion?

No, I disabled everything while testing.

PS: just wanted to thank you @mrehorstdmd for your great blog, I've built my printer mostly thank to your posts

@lael hey sorry for the late reply. the forum didn't notify your reply.

They are FYSETC S6 V2.0 today I would probably get a Spider.

@zapta OHMYGOD I need it!

@jens55 there's no such thing as "fast enough" in 3d printing

@chriss it seems it is banging the head twice against the frame... I'd say you need to go a little slower....

@martin7404 said in I present to you my Agony:

I remember you from Peopoly forum with the moai battle

we don't talk about that....

@fma said in I present to you my Agony:

BTW, is this geometry more rigid than a CoreXY (all things equal)? With 2 carriages at 90° for the tool head

I find it to be better in every possible way. More precise, sturdier and stronger (let's not forget there are 4 motors for XY).

The downside is that 1) you need more components 2) everything needs to be perefectly square so a little more time spent assembling it up.

If you need to put a lot of weight on the gantry you can use bigger backers for the rails, but of course you are adding weight.

@slimshader said in I present to you my Agony:

for the next one

it's funny because it's true I'm already thinking of building a very small one...

Its name is Agony, it took one year of my life but it's finally working as I'd hoped.

This is it, there are many like it, but this is mine.

Still beta of course, need better cable management.

It's a relatively large printer. 610x610x610 frame. about 400x400 print area. That was the main challenge. I find that anything larger than about 300x300 becomes exponentially more difficult to build.

Anyway it can happily print at 200mm/s without too much distortion. The following are printed at 150mm/s external perimeter, with a 0.5mm nozzle (my new favorite nozzle size).

Why

It started because I couldn't find a printer with decent print area and reasonable speed. I also wanted to keep it super simple with few easy to print pieces and easy to source off-the-shelf hardware.

All extrusions are 550mm, all rails are 500mm, you don't need special tools to cut/drill/tap, everything can be easily sourced.

If I had to start over today I would probably build an Annex K2... but it is what it is and learned a lot in the process.

A bumpy road

Agony was originally a CoreXY (named Corezilla), but I couldn't solve some resonance issues I had and that generated heavy vibrations at certain speeds. So I converted to a cross-rail cartesian. I then found out that the resonance was due to the belt tensioner design, but it was too late but I'm happy I moved to a cartesian system anyway.

Another issue I had was with z-banding. Apparently it's impossible to get decent leadscrews (or ballscrews) and anyway it's incredibly hard to align 3 of them. The revelation arrived when I saw somewhere a sort of magnetic cushion to be placed on the Z-carriages.

Basically the leadscrew wobbliness is not transmitted to the carriage and now the Z is just perfect.

Finding the right motors was another adventure. Long story short: 0.9° on the XY and 1.8° on the Z. I tried quite a few motors from steppersonline and LDO but ultimately the best I've found are the Moons MS17HA6P4200.

One suggestion, if you need a big print area but you don't need a lot of height, just make the printer as short as possible. That helps a lot with stability. If I had to redo it, I'd probably limit it to like 250 on the Z.

Lastly the gantries.

I tried with the bare rail and with a 15x15 backer. While the bare-rail solution is better in many ways, at the end I had to use supports for the rails. That was a game changer in terms of stability. I was using the rails vertically and at high speed they tend to "flap" at sudden moves change. So I put them on 15x15 extrusions and placed them flat.

Hotend + Extruder

It's basically a custom extruder based on the now famous small LDO NEMA14 and the guts of a bondtech extruder. The hotend is an original Mosquito (the standard one) and from my initial tests everything's fine up to 200mm/s, over that I start getting extrusion problems on acceleration... but haven't tested it carefully yet.

Despite being so small the 45mm Delta blowers push a lot of air, so at the moment I'm happy with them but I might switch to bigger fans if speed requires it.

Duet & RRF

Now the hard part.

I wanted to use a Duet3 mainly because of the toolboard. I jumped in a little too early unfortunately and the experience was really painful. I was able to fix most of the issues but at the end I hit my head on a firmware bug that produced inconsistent extrusion. I understand that bug is now fixed (and a new toolboard is out), but it was the last straw so to speak so I removed everything and redesigned the printer for the Duet2+Duex5 (that I already had on another printer) but this time I wiped RRF and put Klipper on it.

Now. I owe 90% of what I know about 3d printing to this forum and the help of the generous community and I do love the "everything is gcode" RRF philosophy... but Klipper really made my printer fly.

It was painful at the beginning and the community is much less supportive, you are mostly on your own, but the documentation is good. It's in a constant beta(alpha?)-stage but at the end it worked wonderfully and I feel that any problem I have is now caused 100% by me.

The Duet is a wonderful piece of hardware, something that most of the chinese boards can only dream of BUT my suggestion is: if you need a quick replacement for your printer totally get a duet2/3 with RRF; if you want something complex and custom... in my experience a fysetc board and Klipper is a better option. I hope you'll forgive my candor. I'll still be using the Duet boards I have for other projects but I guess to each its own.

Want more?

I would never ever suggest anyone to build Agony, but if you want to look around, head to onshape and search "Agony" in the public projects. You'll find it, it's the only one... but be warned it's a very messy project.

Thanks, sorry for the long post.

@bot https://duet3d.dozuki.com/Wiki/Accelerometers

the accelerometer doesn't need to be fixed on the head, you can put it on only when calibrating (and every time you change something that can vary the machine kinetics)

@t3p3tony sorry if I am a little out of the loop but it's hard to stay up to date with all the changes... is the toolboard issue where the motor was missing steps fixed?

I think I got the heater naming....

exp.heater3 is an alias on !duex.e2heat (note the !). So to invert it you either use !exp.heater3 with an ! or duex.e2heat without it. At least this is my interpretation...

@Phaedrux okay, the behavior was too erratic, I took the time to re-wire everything. On both the duet and duex. What drove me off is that all heaters were shorted when this started happening.

I use CAT6 wires for all not power hungry connections. I realized that they are convenient but also not easy to peel and crimp. They are just one stiff strand of wire. When you peel or crimp they tend to weaken in one tiny spot and break. Everything seems fine at the beginning but they eventually fail if damaged.

Now I've been very careful when peeling and I loosened the crimping tool so it applies less force.

I believe that was the problem. Everything seems to be working now. BTW, it works without the ! in the heater name. That part is a bit confusing in the doc... well at least for a total noob like me.

Thanks everybody for the help... I hope this is fixed...