Support forPWM Conversion

o_lampe

@Nightowl
I'd support the idea, but my CNC spindle has it's own closed loop controller I wouldn't want to remove.
It's controlled by a potmeter, which future Duet boards could mimic with a digital potmeter? (same as on early Duet boards for stepper current adjustment)

Having a digital potmeter-addon board with optocoupled output would be useful for other stuff, too. (like adjusting an external DC-DC boost converter for HV stepper drivers or reduced bed heating time)

Nightowl

@o_lampe
I'm guessing your's is a 'pukka' spindle and VFD, rather than a milling motor with a digital interface, but I think the overall effect is the same.

Looking at the Amazon varieties available and how cheap they are, I'm really surprised Duet3D haven't taken this step, but I suppose CNC users aren't as prolific as those using 3D printers.

I've no idea what the development/production costs of an add-on board would be, so maybe that's the limiting factor (which I completely understand, by the way).

o_lampe

@Nightowl said in Support forPWM Conversion:

I'm guessing your's is a 'pukka' spindle and VFD, rather than a milling motor with a digital interface, but I think the overall effect is the same.

I'm not sure what the differences are, but I always thought they all come with an (optional) potmeter-speed dial? So it would be the most common thing to implement.
A DIY-board would be cheap. But implementing in FW has always been the reason, not to start with any HW project.

Nightowl

@o_lampe said in Support forPWM Conversion:

I'm not sure what the differences are, but I always thought they all come with an (optional) potmeter-speed dial?

They do, but my Sorotec milling motor (read Mafell) has a digital interface with a speed dial. Connecting via the digital interface overrides the speed dial of my Sorotec, but I understand not all digital interface milling motors have both.

I try to differentiate (in my own head) that a spindle has a separate VFD.

A palm router (like the Makita) is just that.

dc42

@Nightowl it's a bad idea to use a common ground connection between something that potential generates a lot of nose, such as a VFD driving a spindle motor, and sensitive electronics such as the Duet. So the recommended approach is to use an optically isolated PWM-to-voltage converter that is powered by a 10V or 12V supply that is not taken from the Duet.

TBH I'm amazed that spindle controllers still use something as primitive as a 0-10V analog input to control the motor speed. It would make much more sense for the spindle controller to provide an optically isolated PWM input.

jay_s_uk

@dc42 ideally we'd be using RS485 to control spindle speed but RRF doesn't support that

dc42

@jay_s_uk RS485 is something we may support in future. We're already looking at making a daughterboard to support it.

Nightowl

Sadly, I'm not at the cutting edge of any of these technologies. I just want my machine to work with as few complications as possible - and the PWM converter is presenting those issues, i.e. not really controlling the speed.

As I said in the other thread, I'll check the voltage output from the converter to see what's being provided against what is needed.

I guess the overall answer to my question is a "No", then

dc42

@Nightowl the PWM converter is working well for some other users, so if you persevere you may get it working.

Nightowl

Thanks @dc42

Perseverence, by the way, was my school motto!

o_lampe

@dc42 said in Support forPWM Conversion:

...
TBH I'm amazed that spindle controllers still use something as primitive as a 0-10V analog input to control the motor speed. It would make much more sense for the spindle controller to provide an optically isolated PWM input.

I guess it's the simplicity of analog OP-amps against microcontrollers. They don't need to boot up and init I/O stuff, they work right away.
While the PWM signal would need digital processing, right?

dc42

@o_lampe said in Support forPWM Conversion:

While the PWM signal would need digital processing, right?

They could use a smoothing network, just as the PWM to analog converters do.

timothyz

@dc42
Just some food for your thought, another open source project has implemented RS485 communication to spindles in a CNC controller called the CNC xPro v5. RRF is definitely better developed overall, and it could be very powerful to port that feature. I am guessing your daughter board would be using a MAX485 style uart to RS485? I have one of those modules and could test if you were to implement it.

This is a board I have that I am planning on hooking up to a spare UART: UART to RS485 Board

The CNC xPRO V5 has a list of spindles that they have already integrated and you would be able to use those for references in your implementation (src/spindles folder). Obviously an RRF implementation would allow for the interface of way more than VFDs, and I think if done right could make it a very capable tool for more industrial hardware.

jay_s_uk

@timothyz i definitely agree. I have been asking for this feature for ages!

timothyz

@jay_s_uk It's so close to being capable I can taste it, maybe to start, we could allow M118 to optionally send data instead of pure strings. Though, I really don't grasp the deeper functionality of M118 so I'm not really sure. It may be best to change M260, but dc42 always seems to know the inner details and how to make things work.

dc42

@timothyz M260 would be a better bet for sending binary data.

OwenD

@dc42 said in Support forPWM Conversion:

@jay_s_uk RS485 is something we may support in future. We're already looking at making a daughterboard to support it.

FYI
Supporting RS485 also opens the door to controlling the output of a plasma cutting system from the cnc controller.
Here is the programmers refererence for Hypertherm RS485 implementation