Toolboard vs filament sensor

Chriss

Hi Duet guys....

Happy new Year to all of you!

You wrote https://duet3d.dozuki.com/Wiki/Duet_3_firmware_configuration_limitations#Section_Permanent_limitations that the filament sensor must be connected to the same board as the driver board for E:

I'm not very happy with that decision to be honest and I kindly ask you to re think about that decision again please. I guess your motivation for that behaviour is a work around the latency on the can bus, isn't it? You guys try to stop E right at the exact same moment the sensor recognize out of filament situation and want E to extrude more filament than available.
But why is that a permanent limitation here and not a decision of the user? You accept a z endstop/sensor on the toolboard while Z is driven by the mini or HC6. Here is a real risk to damage the printers hardware if the latency kicks in, but you do not allow me to stop E for a filament runout? I have at least 80cm of reverse bowden between E and my filament sensor, so what risk is here when the filament runout detection stop the E stepper 1ms "to late" for a CAN bus latency?
That is not very logical to me to be honest.
You want me to route at least one cable all the way from the lower back of my Voron, trough 3 dragchaines up to the toolhead on the far top front only for a filament sensor while I do not need to do that for the Z Sensor? Do you get my confusion?

Could you please let the user make the decision whether he want to accept the risk of the latency or not? I bet that this would be very, very helpful for more people than only me.

Cheers, Chriss

dc42

@chriss the primary types of filament monitor supported by RRF are ones that correlate measured filament movement with commanded extruder movement, in order to detect filament slipping and not just filament runout. For that to work well, the filament monitor code needs to access the extrusion commanded in the middle of a move with very low latency.

These considerations do not apply to simple filament runout sensors, and there is already a pending feature request to support filament runout sensors that are not connected to the same board as the extruder. Meanwhile, you can configure your filament runout sensor to use a M581 trigger instead of configuring it using M591.

Chriss

@dc42

Thanks David, I will give M581 a try. Let's see how that will work for me and my sensor.

I fully understand the problem with the latency and the sensor but I do not understand why that is a problem here and not a problem at Z. You write in the wiki that this is a permanent limitation, that sounds to me more like: We will not do that
Than is there a pending feature request you did not refuse, did you?
It would be very nice if you guys could allow the user to decide here themself.
I'm more than willing to accept a "to late" e stepper stop if I would not need to route the cables trough all of the chains on more than one printer. That sounds to me like a better solution than having no sensor at all, which is my current work around.

Cheers, Chriss