Posts made by jwilo

@achrn OK thanks, good to know I'm at least not chasing something that doesn't exist.

Next up then, is how to even configure a rotary axis? At the moment, my rotary axis is successfully configured as a linear axis, and the microstepping configure along with the steps/mm, such that if I try to move '1mm', I get 1 full rotation.

M569 P1 D3 S1                   		; physical drive A axis goes clockwise
M584 A1					; set drive mapping for rotational nozzle axes
M92 A800

But, that's not what I'd like. I want to be able to command the axis by degrees, is this possible? Adding R1 to M584 should apparently make the axis rotational, but it doesn't change behaviour at all, as far as I can tell.

@handyandy you can configure it directly, rather than tricking Duet into thinking it is a fan.

See my thread here, where I did exactly this to control some linear pull solenoids, and an LED ring light for a vision camera.

This way you'll be able to control your light using

M42 Pn S1 ; turn on your light
M42 Pn S0 ; turn off

Where n is the number you assigned to your light, via the M codes in my thread linked above.

In my case, both the ring light and solenoid required 24V so I used OUT5 and OUT6, with the voltage select jumper set to VFUSED, as I too am supplying the board with 24V.

Alternatively you can configure the jumper to supply 12V on these pins, if you're within current limits.

Finally, I'd need to check the schematic to see if the outputs are just low side MOSFETS, in which case you likely can externally supply whatever voltage you need (within the drain-source breakdown limit of the FETs) and just use the ground side of the output pins, but take this last paragraph with a pinch of salt, as I've not done that check.

@dc42 thanks David, does that apply to any axis, not just Z? I may be reading too much into the meaning of the "Z move comp" response...

@BlitzReg have you configured microstepping?

Steppers will be their very nature be incredibly... steppy, if driven in full steps.

I'm trying to configure a rotational axis, for the nozzle rotation motors of an OpenPnP build (this thread title might sound similar to another thread I just opened here, but it's very much a different end goal.

Whilst I can configure the axis to accept degrees, how do I get rid of the axis limit values?

For example, I might need to rotate a nozzle +180degrees, 1000 times in a row.

Presumably I could just set a very high axis travel limit, but that'd still only a band aid. Presumably there's an M code I'm missing to continue a rotational axis with no limit?

If I wanted to control something like a PCB conveyor motor using Duet, how would I go about doing it?

From a high level, the control approach would be to start the conveyor stepper rotating, and continue until a light gate is broken and pulls a digital input low.

That sounds like an endstop, but it's not really, that light gate would typically be for example the light gate in the middle of a machine.

I would then want to start running the conveyor again, but that light gate would be pulling the input down still until the board that triggered it passes by.

Then I would want the conveyor to stop again when it triggers a second light gate at the end of the conveyor.

This would be trivial with a PLC or custom hardware and firmware, and I'm sure it probably is too, but I'm not sure how to go about configuring it.

Essentially, I'd like to run said axis I'm velocity control mode, rather than position control.

I ran some tests tonight, rather it appears to be simpler than this.

Combining M400 onto the same line, delays the return of "ok", i.e.

G1 X100 Y100 F1000 M400

However, what if you want to receive the "ok" as confirmation of receipt, followed by "done". The following appears to do it, but if you don't send the M400 before the move finished, it doesn't work, which for very short moves is quite likely...

G1 X100 Y100 F1000
M400 M117 P4 S"Done" L0

Hypothetical one here, not for printing or CNC use etc, just a 'what if' in my mind.

Say for example we're controlling a cartesian machine via Telnet, each time we issue a G1, G2, G3 etc - "ok" is returned as soon as the message is decoded, and the move is added to the motion queue.

If the machine sending commands to the controller needs to wait, to know when those moves are completed, before it does something else, how would we achieve that?

I see "M400: Wait for current moves to finish", but that does not control the return of "ok", does it?

In 'pseudo actions', this is what I'm imagining:

1. PC sends G1 move
2. controller sends 'ok' back to confirm decode of message
3. controller forms move, move finishes
4. controller sends back something that says "I'm done"

Is this possible?

A possible solution I can think of is to:

G1 X100 Y100 F1000 ; Issue move command
M400 ; Wait for it to finish
M117 P4 S"Done" L0 ; Send message via Telnet, which is the P4 argument. S is the message. L0 means don't log this message.

@droftarts No problem, thank you!

@dc42 Perfect! Thank you, that's the key I was missing. I thought the P value referred to which output you're using, and the name was a name you wanted to assign to the pin.

I see now that P refers to the output you're creating, i.e. P1 is some solenoid, P2 might be a light etc, and the name is the name of the pin you want to map it to, great - all working now, thanks!

@droftarts Thanks Ian, not sure I understand your reply. The bit I don't understand, if "solenoid" is not a valid pin name, why does your example use it too?

I've read the docs for M950, but must not be wrapping my head around it.

I need to control a simple low current solenoid on/off, without PWM control with a Duet 3 Mini 5+.

I have connected it to OUT_5, and configured OUT_5 using

M950 P5 C"solenoid"

And attempted to switch it on/off using

M42 P5 S1

and

M42 P5 S0

But the solenoid does not energise. I can confirm 24V (my VIN voltage) is present on OUT5+, and that the solenoid works when powered with 24V on the bench, so it must just be something to do with my configuration.

Hope somebody can help, many thanks.

@gloomyandy said in How are multiple moves 'blended' together?:

It would probably help if you describe what it is you are trying to do and why you need to use H2 moves, it may be possible to achieve what you want in a different way.

I'm not actually trying to achieve anything per se with this question, I Just noticed this behaviour, realised I didn't know the answer and thought I probably should.

@dc42 said in How are multiple moves 'blended' together?:

it's not worth adding code to blend them.

Yes I can't see why you'd need to, it's certain only in homing files that I'm using them.

Thanks all.

So in the example G-code I ran, this was not a true replication of what was causing the stops between G1s.

I have noticed, when using the H2 flag with a G1, that moves appear not to blend together, otherwise yes - I'm able to see many many lines of G code be queued up.

What is the reasoning behind not queuing H2 flagged moves?

Ah, 'command queuing' on this page appears to be what I'm looking for, and M595 to set the queue depth.

However, I noticed when running some simple G1 linear back and forth motion

G91
G1 X100 F5000
G1 X-200 F5000
G1 X100 F5000
G1 X100 F5000

That on the second pass, the motion is stopped between the two subsequent X100 incremental moves.

Is the default queue depth zero?

A conceptual question that I can't seem to find an answer for, probably because I don't know the term to be searching for! When Duet/RFF is controlling say a curve made up of multiple segments, how does one ensure these lines of G code are 'blended' together.

For example, when looking at the hundreds of lines of G code that might make up a single slice in a 3D print, if those lines were given to the Duet via DWC or the terminal, 1 by 1, the printer would obviously make each movement, then stop, then perform the next movement.

Clearly that's not how many motion control systems work, so presumably each line of G code is linked together using some sort of look ahead planner?

@dc42 Thanks! I've only been using Duet for a couple of weeks now, but I'm really impressed with what yourself and the team have put together.

I posted this over on the OpenPnP Google group, which some of you may find interesting, if anybody else is trying to drive opto-isolated integrated steppers, and needs a suitable expansion board to do so.

@wayneosdias said in Reading general-purpose inputs in Duet Mini 5+ returns null?:

@jwilo
You should prolly add the config to see what youre working with.

For reference, for anybody interested - @wayneodias and I resolved this over at https://groups.google.com/g/openpnp/c/qxq_8FXAPTI whilst I was waiting for my account to become fully active

@droftarts Thanks Ian, this is really good information - I'll make sure to add that to my reading list.