Afternoon all,

Seems we are missing a section in our wiki explaining commonly used abbreviations in the forum?

On the most part I'm sure it's accidental but heavy use of abbreviations in this user help forum is counter productive. It is needlessly increasing the level of things you need to learn when we are already dealing with a complex subject matter. In extreme cases it can come across as a bit clicky and seem like the forums own language making it difficult for new folk to join in. The abbreviations are often not easy to seach for on the internet either.

I'd suggest something that searches posts for abbreviations and adds links to explanations, much like you see for many forums/sites where keywords are detected and highlighted for adverts. I do however appreciate that unless this is available as an off the shelf add on there isn't the time to develop that right now.

I have done it in my posts, but particularly when we are starting to delve into linux and single board computer lingos lets either try to keep it to a minimum, or add a page to the wiki. I will do this myself if it's not there by the time I've sorted my current headache - sorry, machine build!

Please can you pin the most recent Firmware and DWC threads to make them easier to find? Quite often older firmware threads will appear above the most recent ones due to ongoing conversations and that's a little confusing.

Yes, to find the most recent releases I look at the github.

No category tag for threads about general website (.duet3d.com/) issues or suggestions?

Morning,

The default colour for links on user bio pages and last posts should be revisited for the dark colir scheme. They are difficult to read.

Thanks!

Your spread sheet is very similar to mine! Spookily so.

https://forum.duet3d.com/topic/5588/non-linear-extrusion/6

I fell into the same trap trying to verify the non linear extrusion. Some where in the gcode references or non linear extrusion guides it says that it is not applied to non travel moves.

Also, don't do what I did next which has to do a high feedrate test with a fraction of a mm move in one of the axis! If you do that the feed rate is only relating to the motion axis and not the extrude and so it attempts to cram 100mm of filament through the hot end in a fraction of a second! Life is generally slow and quiet while calibrating extruder. Suddenly this makes it loud and fast!

I requested a feature to allow us to allow non-linear extrusion while stationary but don't think there was much comment.

Reading up on this and I can see it is a mine field with huge learning curves on some bits if you let youself drop too deep!

I had heard in the past about people worrying about the failure mode of switched power supplies causing mains on the DC out but this appears to be mitigated by an internal fuse. Float between multiple PSUs would not be great but I've commoned my PSUs.

So the earth tie must be to try and stop the DC floating higher than the expected 'safe' 24VDC, but this is also potentially done internally by a similar RC circuit to that between your ethernet socket and -VIN? Tieing the earth to the -DC does potentially reintroduce the noise kicked out of the non-linear switch mode supplies to your -VDC and so screwing with your temp readings.

Are there no standards that we can look for that indicate the earth is tied with an appropriate RC circuit internally and there is no additional need to tie? Is this a product of buying cheap PSUs? Or perhaps is it my PSU which needs an RC circuit because of the noise that it kicks back to the mains that is confusing me?

Edit: I've made a few references to RC circuits (resistor capacitor), but apprecieate the internal circuits in the power supplies may be more complicated with over voltage shunts and transient suppression to mention a couple of things. I guess the main question is if it is safer to tie why don't the PSUs all arrive like that?

You said you know your system can do better. Do you mean you have seen others do better or your system did better pre Duet? With age thing do wear out, and the duet has come second! May be worth a nozzle swap if it is old.

At least 4 areas for optimisation being investigated here and improving one can easily make another look worse when all thats really happening is you are reducing a randomising effect by curing one issue which was hiding the magnitude of a second. Very frustrating, but it's the nature of the beast.

In no specific order:

• Ringing/ghosting - you've mentioned your stock pully a few times. Assume this is at most £5 to swap out with a gates model? Some other threads indicated that gates may not be the best, but at least they are respected for quality control. Creality not so much. Also try to get the belt running without scuffing the sides of the pulleys. That said as the pitch changed with speed I think this maybe a mode of oscillation in your frame somewhere. I've no experience with enders, is it easy to brace the top on the z-axis back down to the front or rear of the machine? To treat the oscillation you need to make the thing that's exciting it vibrate differently (less or at a different frequency; quieter stepper setup, belt/pulley choice, belt tension, etc), or change the natural frequencies in your frame by making it stiffer or more compliant. More drive current on the motor will likely reduce the amplitude of oscillation of the motor around it's target position. Keep check on how warm the steppers are. Ok you don't want to run higher currents than specced, but a stone cold stepper would indicate it's not working particually hard.

• Z wobble. In one of your pictures there appears to be heavy scoring on your right hand inner delrin wheel. That shouldn't happen. Check the extrusions to ensure the v-slot is very smooth. If anything that the wheels run on can scratch or be felt as uneven by your finger then it needs to be fettled or replaced. Check all other wheels for wear and that the bearings run smooth (remove wheel to check). With no z axis screws connected the axis should move up and down easy with no tight spots. The wheels should be tight enough to the rail to limit wobble, but not so tight as to overload the bearings. There is a big difference between quality bearings and wide tollerance cheap bearings. I have found heavy duty wheels (polycarbonate I think) harder wearing. Linear rail z maybe asking for trouble with an extrusion based machine as you are likely to end up spending an age trying to align the two rails close enough to each other to get the axis running smoothly. A bit like the issue Tom3D discussed on YouTube when trying a twin linear rail y-axis setup on another bed slinger. You may also need to (as far as possible) de-couple (in x/y-plane) the z nuts from the motor and idler ends of the x-axis. A popular mod on P3Steel machines. My next step is belted z-axis with a worm reduction drive to stop axis drop on power down.

• Your "///\" pattern is often referred to as "Vertical Fine Artifacts" or VFA. Somewhere in this forum there is a link to a long thread on a Prusa forum about trying to tackle them. From memory specific motors (despite similar spec sheets) do make a difference. There seem to be a few quality stepper motor brands rising up at the moment. LRO are making noise on twitter (and I think they supply E3D) but I have no experience of them. I'd be a little cautious about definitively saying it was vibrations from the fan causing issues. I'd wonder if for example it was more electrical noise? Can you check for electrical noise and would you know what to look for? Ensuring there is good continuity from your stepper motors (and maybe hotend) to the mains protective earth would be wise first principles step. likewise on the noise front can you lock your toolchanger and run with the magnets off in a trial?

• Irregular extrusion. Direct is easier as you are about to find out! Have you run the calcs to check the stepper can run at the commanded speeds? I suspect the loud odd noises you mention on some retracts are skipped steps. Have you already checked e-steps carefully (slow extruder drive with nothing after extruder), linear & non-linear extrusion correction, before doing detailed pressure advance and retraction calibration? Ideally the retracts should be minimal but slightly excessive to avoid occasion stringing.

Final point... Plan your experiments or you will spend way over the odds on this. Try to find ways of quantifying/measuring the issues you can see. Tackle maintenance issues (scored wheels, bearings etc) first and always check system for loose bolts, couplers, and signs of wear. I have - as most of us have - got history of missing a simple problem while searching for a complex one. Be warey of compairing your prints too much to others. Some are dishonest with lighting to hide artefacts etc, but to be fair I've not seen too much of that on this forum. A walk around TCT Show last year and I saw many worse examples of ringing than what I thought was critical on my machines on both demo machines and parts trying to sell commercial level services.

Some swear by the direct link. There doesn't appear to be anything easily to hand that confirms things definitively. In my recent searches I found that this is prescribed by code for some types of devices but not others - including some scientific instruments.

Checking schematics helps. If they have shorted the pads around screw holes to ground then you may as well do it at the psu. The Ormerod 2's external 5V regulator did this, meaning that without a ground and dc -v link any DC +v short to the chassis/ground (such as heater plate shorting to aluminimum heat spreader, through the bolts and bearings to the chassis) would then try to make it's way back to the psu terminal through the external regulator board. (Not necessarily the reason this connection was toasted on a board I recently spoke about on the forum.)

I always ground as much as possible of the chassis to earth which also includes earth bonding the stepper motors. I struggled to get or test a good bond with aluminium extrusion but realise normal multimeters aren't a great test for this.

If you do bond -v and earth use the largest connections on the earth bonding as practical so the majority of noise and fault currents go direct back to the psu rather than effecting the electronics.

Edit: The other big ones from the safety point of view is fusing and earthing. Where practical add fuses (or reset-able circuit breakers/trips). Main candidates are on the +V feeds to the hot end heater cartridge and heated bed. If possible check to ensure you have a good earth connection for your machine. If your in an old house that has never/not recently been safety checked that may be worth getting that done, and avoid long extensions or daisy chaining extension leads where possible. I run my machines after a mains plug in surge protector and RCD trip.

Edit 2: This is no religious matter where by I will stick to no link in the face of thorough evidence to the contrary. I'm not saying that I will always avoid a tie between the -V and ground (and will add it in situations like Ormerod 2) and may well someday understand it better!

People need to tread with extreme caution on this topic. As this is a Duet3D hosted site I feel the admins have the right to dictate owners of clone Duets are not welcome to seek support here, fair enough.

However in terms of the effect of clones on the original company what is the difference between me buying second hand rather than new compared to me buying a clone? For instance in the cold light of day I would have to say the logical conclusion is likely to be I am causing just as much damage to Duet3D buying second hand and asking questions on here as I would be if I had bought a clone. I would like to think I am able to pitch in and help every now and again though balancing my value to this forum.

If we support this user in their infancy in the hobby now and lightly enthuse about the benefits of original products they are far more likely to remain, become active contributors and replace their sloppy cloned parts with genuine when they can afford to do so.

In terms of intellectual property we also need to be careful. Our hobby and in some cases our livelihoods were made possible by the expiration of a patent. Is the predominant failures of the cloners is to add any innovation to the product? Cost reduction could be classed as their innovation, they after all have tooled up for injection molding rather than plastic sintering? But yes without diving too deep into politics the cloners location and legal isolation/insulation leaves our manufacturers unable to compete on a cost basis. So really you are just highlighting the failings of free market economy and blaming the end consumer who has exercised their freedom of choice rather then the governments for failing to balance this inequality. For our companies it forces rapid innovation in order to still attract people to purchase direct, and ensures they do not sit for too long. Alas without further market control the sweat shops of the world have taken the cherry from the cake for product production. Development is tough, support is tough, and selling a stable product is supposed to be the easier money bit that has now less profitable.

Bondtech essentially adapted a type of mig wire feed to a polymer filament feed. If the E3Ds Hermes (that name tickles me!) is little more than using a no-idle-roller bontech design in a re-engineered housing (the tooling for injection molding thet the cloners have done is, after all a not insignificant amount of engineering effort) are we going to be as negative about people buying E3D Hermes over Bontech?

Edit: the cherry on the cake is lost by going open source, that is the nature of the beast. But the company benefits be developing a loyal user base that in turn supports other users. When this is established the original company benefits from this free support and development that would have otherwise need to be paid for. So the cherry is not given away for free, there is a benefit.

Edit 2: This is naturally a topic that will stoke many peoples emotions and provoke strong responses. I have been wanting to raise my opinion on the matter but I do not wish to make this personal. We are free to hold whatever view on the subject that we want to! In a strange sense this is both not the individuals fault, but all of ours.

@t3p3tony could do, would a page on the wiki be easier for all to update? Link to the how to ask for help page?

The format could be similar to the gcode reference but much less wordy so that it is easier and quicker to update. Something like the following table but with the abbreviations linkable:

Common Forum Abbreviations

Additive Manufacture

Duet Specific Abbreviations

Computing

Obscure Chat

@deckingman Morning Ian! I'm avoiding the tempation to go down discussing experimental plan in too much detail as that's a big can o' worms. Bare with me here as I am writing it up, and will share in due time. If I try and defend/discuss the work off the cuff I won't do myself or the work justice.

These are preliminary tests that I have started simple and am gradually increasing adapting the test geometry, and procedure in order to reduce measurement errors as far as practical.

In the first post I mentioned about forcing the positioning if the parts to be integer multiples of the full step size. Hopefully this will reduce the effect of microstep inaccuracies across one build. But unless I queue build after build without resetting the machine I can't escape needing to home which may result in a specific co-ordinate of the bed being in a subtly different absolute angle of the stepper than a subsequent build. One valid response would be to improve the limit sensing if that is the case, and the other is a cheaper route which is forcing a specific co-ordinate to be at a whole step position for the motors.

That said first test this morning will be testing my limits to determine how repeatable they are!

Other factors that are on-the-shelf for consideration in this body of work are:

• Local bed temperature variations caused by heater or the use of glass bed (/ another) surface
• Differences in x and y axis - ie mass.
• Belt tensions and inconsistencies between axis.
• Sticking points on the axis.
• Variations in filament diameter - ie manufactures specified tollerance.

That's just the tip of the potential causes. The stepping is the one I'm trying to investigate right now.

Ultimate aim of work? Develop a simple procedure that can optimise achievable tollerance and accuracy for a machine. Not as trivial as you might expect as part geometry, cooling rate, and even extrusion velocity is likely to have an effect on shrinkage - polymers are weird...

@theolodian good point, I'll read up on the coordinate system method another time. The printer is running in FFF rather than CNC mode which may complicate things a little, but can't see why you'd limit that feature to CNC - even if the request came from there.

As it happens I'm just aiming for the middle of a circular-ish feature, so rotation not an issue this time thankfully.

Here was what I had in the macro:

; https://duet3d.dozuki.com/Wiki/GCode_Meta_Commands#Section_Variable_naming
; https://duet3d.dozuki.com/Wiki/Object_Model_of_RepRapFirmware#Section_Uses
; https://forum.duet3d.com/topic/26368/object-model-report-contents-of-object
; https://forum.duet3d.com/topic/26370/sanity-check-tool-offsets-g10-p0-x

var num_x_pattern = 3
var num_y_pattern = 2
var x_pattern_space = 50.0
var y_pattern_space = 25.0

var tool_num = 0
var int_x = move.axes[0].machinePosition
var int_y = move.axes[1].machinePosition
var initial_x_offset = tools[var.tool_num].offsets[0]
var initial_y_offset = tools[var.tool_num].offsets[0]

M98 P"0:/macros/intro.g"

while true
	if iterations = (var.num_x_pattern * var.num_y_pattern)
		break
	else
		var x_i = mod(iterations, var.num_x_pattern)
		var y_i = ((iterations-var.x_i)/var.num_x_pattern)
		var x_pos = var.int_x + (var.x_i * var.x_pattern_space)
		var y_pos = var.int_y + (var.y_i * var.y_pattern_space)
		G10 P0 X{(-1 * var.x_pos)} Y{(-1 * var.y_pos)}
		M98 P"0:/gcodes/sample_build.g"
		continue 
G10 P0 X{(-1 * var.initial_x_offset)} Y{(-1 * var.initial_y_offset)}

M98 P"0:/macros/extro.g"

@theolodian I've tested a macro that uses the gcode variables and coding to read in a set location and adjust the G10 tool offsets to make the current location 0,0. This did a regular grid of prints, the computer vision will be used to fettle the expected position to account for work piece variance. There may be a more elegant way of doing it with co-ordinate systems too, but I have something functional to work from initially.

@alankilian That'll do nicely thanks.

I'll setup a macro that fires a message to the SBC then that will run a loop of imaging, calculating a required offset, and returning a G1 command until the required offset is within a set parameter. At that point the SBC will trigger a build gcode file and the process repeats a few times to copy a build at a number of semi-regular (hence need for camera) locations.

After a couple of hours hunting I can see the direct method of connecting a Single Board Computer to the Duet 2 is not feasible. It involves removing the ethernet or wifi adaptor and making connections there.

Looking for guidance on controlling a Duet 2 from a Python Script rather than through the DWC, or more accurately alongside a normal DWC connection. It's likely that the build process would be triggered by a Macros.

@dc42 cheers. Wasn't looking to sell it, just would have been easier to buy it than make for the customer! I will read up, looks simple enough.

Thanks.

Is, or has this work going to be wrapped up into a commercial product?

Hi All,

Looking to do some vision work on a Single Board Computer which will use OpenCV to locate a target start point and trigger a repeat print.

This will be a repeated loop; scanning an image for a feature, moving the head to centralise the feature.

What is the current best supported way to achieve the single board computer to duet communications? Just fire network commands as a prototype? The system is using an expansion board.

So this is the test macro (that works):

var tool_num = 0

var num_x_pattern = 3
var num_y_pattern = 2

var x_pattern_space = 50.0
var y_pattern_space = 25.0

var int_x = move.axes[0].machinePosition
var int_y = move.axes[1].machinePosition

var initial_x_offset = tools[var.tool_num].offsets[0]
var initial_y_offset = tools[var.tool_num].offsets[0]

while true
	if iterations = (var.num_x_pattern * var.num_y_pattern)
		break
	else
		var x_i = mod(iterations, var.num_x_pattern)
		var y_i = ((iterations-var.x_i)/var.num_x_pattern)
		var x_pos = var.int_x + (var.x_i * var.x_pattern_space)
		var y_pos = var.int_y + (var.y_i * var.y_pattern_space)
		G10 P0 X{(-1 * var.x_pos)} Y{(-1 * var.y_pos)}
		G1 X0 Y0 F6000
		G1 X10 Y10 F600
		continue 

G10 P0 X{(-1 * var.initial_x_offset)} Y{(-1 * var.initial_y_offset)}
G1 X{var.int_x} Y{var.int_y} F6000

I would be replacing the following lines with a call to run a build file:

		G1 X0 Y0 F6000
		G1 X10 Y10 F600

Does this seem the sensible approach?

Hi All,

Looking to implement a little meta coding object to repeat a specific pattern a number of times from an offset based on the current head position.

Is the following current best practice for that procedure?

• Read current head position from object model.
• Read current G10 P0 offset from object model.
• Apply appropriate offset to the G10 P0 values to make the current position position X=0, Y=0.
• Run print job with M98 P"build-name.gcode" (Is this any worse than using the specific start build gcode?)
  ...Move head, repeat...

I believe that is better than resetting the position using G92 as it doesn't invalidate any bed probing?

Cheers.

@dc42 thank you!