Duet3D

engikeneer

@sd_matt your tool 0 is currently off (it says under the tool name) so is not selected. Maybe add 'T0' to the very end of your config.

Also, unless you disable it, you will likely need to heat up your tool before it will let you extrude edit: M302 P1 allows cold extrusion

engikeneer

@tenaja That's down to how the firmware interprets the commands:

G1 E1 F5

Will extrude 1mm of filament at 5mm/min (RRF uses mm/min for speeds). As you are only commanding an extruder move, that sets the speed. So with F5, it will take ~12s!

G1 Z0.242 E44 F5

Will extrude 44mm of filament whilst the Z axis moves at a commanded speed of 5mm/min over a distance of 0.242mm. That will take ~3s, and you will be extruding 44mm of filament, so your extruder would be moving at ~15mm/s (that's over 900mm/min so a lot faster than the first one!)

All of this is also bound by what speed limits, accelerations and jerk settings you have for all of the moving axes (X, Y, Z, E etc) in your config, so your actual moves might be a bit slower/take a little longer depending on your settings.

engikeneer

@gnydick I think you're confused between Pressure Advance and Dynamic Acceleration Adjustment.

Pressure advance aims to compensate for the elasticity of the filament and the extruder system which can cause under-extrusion whenever the extrusion rate is increasing, for example at the start of a straight line when the nozzle has to accelerate from zero or near-zero speed, or over extrusion when the extrusion rate is decreasing.
https://duet3d.dozuki.com/Wiki/Pressure_advance
It is mostly effective at stopping under/over extrusion blobs at sharp corners. If you want to calibrate this, I'd recommend the steps at the bottom of the wiki page above.

Dynamic acceleration adjustment is used to help ringing/ghosting in your prints by tuning out the frequency you set. I think this is what you want
https://duet3d.dozuki.com/Wiki/Gcode#Section_M593_Configure_Dynamic_Acceleration_Adjustment

I'm guessing you've already chased after all the mechanical things to aid ringing (belt tension, making sure your pulley mounts/axles are solid, reducing mass on your moving gantries etc). If after all of this, you find you can't get raise your printing accel as much as you want, you can always use your slicer to vary it by feature type (I know PusaSlicer, Cura and IdeaMaker all have this at least). That way, you can set much higher accelerations for infill, but drop it for internal/external perimeters to maintain quality. In Cura & IdeaMaker, you can also do the same with Jerk.

engikeneer

@Piet said in Using bed leveling screws with independent z motors:

For example in an extreme case where one point of the bed is 10mm off. If I were to level that initially using the ABL it would torque the entire Z-axis assembly and would put a lot of strain and resistance on the leadscrews.

Whether you're using manual levelling assistant or auto bed levelling, the process is taking the measurements from probing the bed surface.

I think your concern is that the bed is not level with the z-gantry (e.g. because one of the manual adjustment screws is way off). So you want to get the bed close to level to the z-gantry (and so level with the lead screws) so that they aren't kinked over later. The trouble is the Duet has no idea where the lead screws are as it is still only probing the bed surface. Only exception/caveat is if you are using independent endstops on them as well (and somehow worked out the offset for them all).

One option could be to remove the bed and try probing the z-gantry to get that near level (so the leadscrew are all level) using auto bed levelling. Then put the bed back on and use the manual assistant to get the manual screw right, then use auto levelling again to actually get it all tuned in fine (assuming you didn't with manual). Might work, but probably will end in tears trying to probe aluminium extrusions....

Alternatively, just try making sure your bed is level to your z-gantry when building the printer. Simplest way is to use fixed spacers (i.e. no manual adjustment). Or just a spirit level/your eyes? The lead screws should take a little misalignment. Unless you have a massive bed, if you build your printer 10mm out and don't see it, you're probably gonna have a bunch of other problems in there too...

engikeneer

@Raniot Hello and welcome to the Duet world!

I'm guessing you have already read up on what ringing is and things you can do to change/improve it. In general it comes to trying to push a heavy print head round a corner too quickly, meaning the flex in the system bounces a bit and causes the artefacts you're seeing.

As well as changing the belt tension, you might want to double check all the joints/bolts/connections on the gantries are tightened - In the past I have found a screw loosened in the printhead which gave me some ringing.

After that you're mostly onto tuning the printer to try and slow it down round the corners. For this you can play with speed, acceleration, and jerk, and you can adjust them in your printer config, and/or in your slicer. General advice is to set sensible limits for the machine in your config file (i.e. things that will mean it won't skip steps, break itself etc), then set more detailed/controlled limits using your slicer. In the Slicer you can usually set different limits for different move types, so you can have fast, high acceleration moves forthe infil where ringing/quality is less important, but slower and more controlled moves on the external perimeters where quality is paramount. On an 800mm tall print like yours can do, this can make a world of difference on print times!

Regarding the firmware flavour, you should definitely set it to Reprap. Reprap and Marlin are very similar, but some of the commands are slightly different and use different units (e.g speed & jerk are in mm/min in Reprap but mm/s in Marlin).

I would start by using the values from your config file provided by the printer manufacturer in your slicer. I mostly use PrusaSlicer which only allows you to play with acceleration, so would recommend starting with just that, then you can play with Jerk later if needed. Looking at your config file, your machine has acceleration limits of 1000mm/s^2 in X & Y (from the M201 command), but these are then limited to 500mm/s^2 for printing moves (i.e. moves where the extruder is moving) and 1000mm/s^2 for travel moves (both from the M204 command). So I would put 500mm/s^2 for all the printing moves, and 1000mm/s^2 for travel moves in the Cura controls. Cura will then start inserting M204 lines in the Gcode files where needed to change these limits as needed during the print.

From there on, you mostly need to play with the slicer values on test prints until you get to a result you're happy with. At the same time, you might also consider changing the speeds for different move types in Cura (e.g. slowing down external perimeters a bit). For reference, on my big home made CoreXY, I print external perimeters at about 60% the speed of my infill, and with 30% the acceleration. You may be able to go faster than this, or might need to go slower - only tuning will tell!

Whilst tuning things, you could disable the acceleration control in the slicer and manually send different M204 commands during a print via the web control to see in real time what the results are. It may be a little harder, but will be quicker. There are a few good forum posts out there on the best methods for doing this, e.g.:
https://forum.duet3d.com/topic/14250/tuning-jerk-accel-speed-settings/3?_=1592332339377

Regarding the slicer time estimations, you will probably find these get a lot better when you set up the accceleration control in Cura. I think the acceleration & jerk limits in your config are reasonably low compared to most printers (this is probably to be expected because your machine is so big, it will have a lot of mass to move around!), so will take a bit longer to accelerate/decelerate than Cura might expect by default.

If you're confused/unsure on the commands in the config file, you can always look them up here:
https://duet3d.dozuki.com/Wiki/Gcode
The descriptions are very good and will generally give you a good idea of what is going on, but it can be a bit daunting at first!

One final thought, you might want to consider using 'Dynamic Acceleration Adjustment' - M593:
https://duet3d.dozuki.com/Wiki/Gcode#Section_M593_Configure_Dynamic_Acceleration_Adjustment
This is a useful feature to combat ringing by tuning out the first natural frequency of your motion system. There's some details on what it is,how to use and tune it etc in the link above. First check if this is already in your config file from the printer manufacturer though!

engikeneer

@Nxt-1 I was meaning bolting another aluminium extrusion, either a full 30150 or even just a single 3030 to keep it nice and light. More thinking it might be a lighter weight solution!
One other thing to consider is to do all of the towers slightly differently. That way you'll get some separation of the resonant frequencies so the towers don't interact and further excite each other. I'm guessing you'll get this a bit for free with the concrete though

engikeneer

@jallen810 not too sure on the power supply. Does it give the expected 12/24V out (disconnect from the Duet before measuring)? I'd probably try powering over USB only and see if you get anything with M122 before putting 12/24V back through the Duet. I'd also do a close inspection of the Duet (unwire and take off the printer) to look for any signs of damage front or rear before putting 12/24V back in. I don't know much on SSRs, but how is the light powered? From 5V, VIN, the Duet control signal?

When you do power up the Duet again, does it measure the correct VIN in DWC?

On the fans, I have previously blown 5015 24V fans by wiring them up the wrong way round on my Duet and only powering them for a few seconds. TBH, I was probably lucky not to have damaged the Duet!

Edit: side note - I'd also be careful with the driver heatsinks you've fitted. They won't do much as the heat will go through the board to the back/underside (the top of the chips ar plastic so don't conduct well) and they may hinder airflow more than they help conduction. General advice is to not fit them in case they short something and to fit a fan to cool the board (like you have done).

engikeneer

@alielsayid also sounds like you might need some pressure advance if it's not just on your first layer. (Edit: in your printer config, not the slicer)

engikeneer

Dowel pins are probably the best. Make sure you get parallel (most are) not tapered. I've made that mistake before... the most common parallel ones are DIN7 or ISO 2338 which should be ground to within a few microns. Most are hardened too

engikeneer

@NonnoGio98 You only need to use a single dc supply (usually 24V or 12V on the Duet Wifi). You can supply this direct into the board and the internal regulator will provide the 3.3V & 5V it needs to work. The board then in-turn supplies the 24V to the motors, heater, bed etc.

However, you can deviate from this if needed/desired. Most common deviation is to power the bed separately, either because it is a mains AC bed (as you suggest?), or it is a 12/24V bed that draws more than the Duet's bed heater channel can cope with (I believe it's 15A..?). You then use the bed output to control an SSR (for AC beds) or a larger external mosfet. From what I can see you have got this right. However, I am confused why you have the 600W power supply? If the bed requires 600W AC, you don't need a power supply - mains power comes in from the mains. As it is, the bed wiring will work, but you don't need the 600W PSU and can put the Live staight into the SSR and the Neutral straight to the bed.

Another deviation which is sometimes done is to use an additional 5V PSU. In this case you would put the 5V supply to the 5V and GND pins next to PS_ON. Note that this is a 5V input. Using the separate 5V supply then means you can turn off the 24V supply to disable the motors, heaters etc (e.g. from an E-stop) whilst still being able to access the Duet to see what went wrong etc. I do this on my printer. I also use the PS_ON pin to drive a relay on the 24V supplies, meaning that the Duet can shut down the power if it detects a fault. I believe this is what you are attempting to do? I'm confused why you'd have a 600W 5Vdc supply? Mine is 7.5W and I use it to power a a couple of other bits at the same time too... However, in principle, if it is a 5V supply, it will work as you currently have it.
Note that the PS_ON doesn't require the separate 5V supply, but you need some way of powering the Duet at boot (usually done by momentary switch), so that it can then send the M80 command to activate the PS_ON pin and turn on the relay/SSR

Hope that helps

engikeneer

@mildw4ve There are three ways to solve noise problems like this:
1 - get rid of the thing that is exciting the system
2 - get rid of the thing that is projecting the noise
3 - change the system so that the excitations don't get to the thing projecting the noise (by damping or by changing the mass/stiffness)

Id certainly start by checing for anything loose in there (there is something that sounds rattly in the video which isn't great). However, 3D printers are complex beasts with lots of things that can resonate and project noise, and there's lots of paths excitations can take to them. You can try playing with damping, adding masses, changing belt tensions etc, but you will probably find it hard to make much of a difference. Really the best thing is to try and solve the root excitation.

On a CoreXY printer, when moving on a diagonal, one of the motors is moving at root-two (~1.41) times the speed both would be moving for a pure X-X or Y-Y move. You say you see the noise at 40mm/s X-X and 60mm/s on a diagonal (i.e. about 1.41 times higher). This suggets that the vibration source is of a similar order to your motors, so it's either your motors or something that rotates at the same speed as them like the belt/pulley meshing. My guess is that at 80mm/s, you are seeing the 2nd harmonic of the same vibration frequency, and I reckon if you did a diagonal move at ~120mm/s, you'd also see the noise. There will be some noise +/- a few mm/s from each of these peaks, and maybe quieter speeds in between. If you want to get fancy/nerdy, you could use a spectrum analyser app on your phone to work out what the frequency if the noise is and it might help narrow things down even further.

First thing I would check are your drive pulleys and your idlers. Cheap clone belts & pulleys don't necessarily mesh as well as genuine ones. running the toothed side of the belt on a smooth idler is also not great. I'd also check your belts aren't rubbing on anything or rubbing up the side of the pulleys.

More likely listening to the video is that it is the motors themselves that are causing the excitation. Then I think you have a two other options. Easiest/cheapest is to get some damper mounts for your stepper motors, either the cork matt ones or the rubber type. This may help a bit, but doesn't really fix the problem. Better might be to get some different motors. Maybe first of all try swapping in your extruder motor to see (assuming it is a different motor?). Different size stepper motors have different electro-mechanical properties and different rotor inertias so will have their own different natural frequencies. I'd recommend going for 0.9deg motors (I think yours are 1.8's from your config?) if you can get away with the speed. More steps/rev means they generally have a smoother torque profile so are in theory less likely to be as noisy. You could also try bigger/smaller nema17s if you have any lying around?

engikeneer

@jrockland you can see everyone is very interested with the amount of people replying!

First thing the developers will need to know to investigate will be some steps to recreate the issue. Eg. Your config file now, and your config file & firmware version from when it worked (3.1.1?).

There are many commands in your config that could impact this (e.g. M143, which IO port you're using etc).

engikeneer

@jrockland maybe also post your config, hardware and firmware versions so the Duet guys have something to go on?

engikeneer

@LB as Ian said, they've already done that!

@droftarts said in All 5 Stepper Motors Grinding/Stalling:

If you have these macros on the SD card, this also means you have an older board, as we changed the SD card contents to remove these macros quite a while ago.

engikeneer

@Jarery "If you configure a Z probe using multiple M558 commands..."

I.e. it is saying if you split the definition of a single Z probe over multiple M558 commands, it is only the first one that needs the P parameter.

You have two Z probes, so the note doesn't apply and you need to specify the pin for both. Maybe it could be a bit clearer...

engikeneer

@Phil333 I think M116 on its own only waits for the bed and active tool. You don't select T1 until after you've already done your M116

Move your T1 up earlier (or add another M116 after T1). Edit: or add in P0/P1 to your M116 so that it knows which tool to wait for (this is what I do for my chimera)

engikeneer

@Raniot in PrusaSlicer, the acceleration control is under print settings / speed, not under the machine settings (though the marlin flavour does have it here too as you've found...). I would recommend using that with Reprap flavour.

The print settings means it puts M204 commands in controlling the print and travel move accelerations, but won't override/overwrite the machine limits set in your M201 in config. I like that because it means the machine limits set in config are always respected, even you you accidentally type a wrong number in the Slicer.

There may be something similar in Cura if you search though. Maybe @Phaedrux or another Cura user can help point you in the right direction? Failing that then maybe just try some different values and see what the commands come out in the Gcode file? I know there are some marlin commands that are compatible with reprap so it might even work without the 60x multiplication in some cases, but I'm not sure.

engikeneer