Is there any news or blog on the Duet3

dc42

Are absolute encoders for linear position actually available, at affordable prices?

Mabover

@dc42 the CUI AMT20, AMT21 and AMT23 sentries of encoders offer absolute output in spi, rs485 or ssi respectively and support software zero position setting. They can be had for about 50USD from digikey.

They also have incremental encoders and commutation encoders for BLDC motors that don't already have commutation sensors, they set the commutation position relative to the motor pole position via software in a guide they provide.

https://www.cui.com/amt-modular-encoders

JoergS5

This post is deleted!

kuhnikuehnast

As the TCT starts today, is there any chance to get an info / Look about new products on the Roadmap? Would really like to come over, but from germany it is quiet far...

Phaedrux

@kuhnikuehnast https://forum.duet3d.com/topic/7010/duet-3-demo-at-tct/12

OG3D

@dc42 Yes they aren't cheap. I have been think about DRO for quit time similar to this.
https://www.igagingstore.com/6-Absolute-Digital-Readout-DRO-Stainless-Steel-Su-p/205470.htm

I know the DROs aren't fast enough to keep up with the motion but it could be used a reference to pull the positions when requested, like during boot up. What do you think?

DocTrucker

Any updates? Still on track for Q2 this year as mentioned on another thread? Higher VIN limit? Presumably targeting a higher price?

dc42

@doctrucker said in Is there any news or blog on the Duet3:

Any updates? Still on track for Q2 this year as mentioned on another thread? Higher VIN limit? Presumably targeting a higher price?

A batch of Duet 3 pre-production main boards is scheduled to be assembled this week or next. Following testing and some firmware work on them, we expect to provide some to key customers in February. We won't know how many working ones we have until we have tested them all. Many of the prototype boards failed testing because of a manufacturing difficulty related to the new stepper drivers, but we think that should be solved in this iteration.

The price will certainly be higher, because the stepper drivers will cost more and there will be 6 of them. The higher specification processor and the CAN bus interface also contribute to the higher cost. So we expect the Duet 2 WiFi/Ethernet to continue to be mainstream for most applications.

The maximum VIN voltage will be higher, probably around 35V-40V.

DocTrucker

Thanks, thats great.

Coincidentally has anyone shared or blogged about the relevant regulations that need to be met for shipped kits and prebuilts? I'm aware documentation should be in place for CE marking but it's largely the manufacturer's responsibility to ensure it is complete, unless it is called for then there is a fixed time in which it has to be delivered. Outside the remit of the machinery directive?

Grogyan

I somehow missed the news on the upcoming Duet3

Have been rather slow with my projects, one of which requires higher power stepper drivers than what is commonly in use, and was looking at getting Duet2 for this.

I really like the direction of Duet3
Happy to see someone formally bring in CAN bus to 3D printing.
two small issues I see, are no mention of if the process has an inbuilt FPU.
And
Not using the USB-C connector for interfacing.

T3P3Tony

@grogyan

In built FPU: the SAM 4E MCU in the Duet 2 have FPUs.

USB-C: this is still too expensive to make it worth it for a connector that is only required for initial setup, if at all. Micro USB is ubiquitous and cheap (I bet nearly everyone has one of these cables in a draw).

@t3p3tony said in Is there any news or blog on the Duet3:

@grogyan

In built FPU: the SAM 4E MCU in the Duet 2 have FPUs.

USB-C: this is still too expensive to make it worth it for a connector that is only required for initial setup, if at all. Micro USB is ubiquitous and cheap (I bet nearly everyone has one of these cables in a draw).

To keep cost at a reasonable level I think what you pointed out for the usb is right. For a connector only used for setup (in best case only 1 time in a lifetime) the cheapest but still reliable (THROUGHHOLE of course, NOT those only one side surface-mount-=)§)(/)=§("!!! ) usb-connector would do. ( I know you guys have and will choose only throughhole on every possible moving-cable since you did that in the past. Just wrote that because I stumbled in my so far young life enough times over consumer-electronics where they took the cheaper only smd-usb-connector and that for sure will come of the pcb after a few times un-/plugging the cable)

XTL

That looks like RJ-45 on the CAN? Is there a reason for such a large connector? I've used 4-pin RJ-11 in the past and that's already more than enough pins for CAN and would even carry a pair plus voltage and even ground if needed.

USB-C is indeed a minor nightmare unless you can get some sort of module for it with an interface chip and that would not be very cheap either way. Also, cables can be quite brittle unless you get very very good ones. Available functionality in full implementations is brilliant, of course.

@XTL

Just because I read in an article about it: Could also be single-pair ethernet, which has only 2 wires twisted, but can use also ethernet-connectors?

On the other hand RJ-45 are available from thin to thick, flat to round, from flame-retardant to halogen-free, from unshielded to shielded, from home-flavour to industrial-moving-application types, costing very little to very much money... lot of reasons to use that is the amount of options you have ready in the warehouses I guess?

Edit: Also there are lot´s of toroidal-transformer for decoupling available ready with ratings from weak to heavy? Paired with a twisted pair cable isn´t that close to differential communication with CAN? (Have some mercy when you answer, CAN is new to me)

@dc42 said in Is there any news or blog on the Duet3:

@doctrucker said in Is there any news or blog on the Duet3:

Any updates? Still on track for Q2 this year as mentioned on another thread? Higher VIN limit? Presumably targeting a higher price?

A batch of Duet 3 pre-production main boards is scheduled to be assembled this week or next. Following testing and some firmware work on them, we expect to provide some to key customers in February. We won't know how many working ones we have until we have tested them all. Many of the prototype boards failed testing because of a manufacturing difficulty related to the new stepper drivers, but we think that should be solved in this iteration.

The price will certainly be higher, because the stepper drivers will cost more and there will be 6 of them. The higher specification processor and the CAN bus interface also contribute to the higher cost. So we expect the Duet 2 WiFi/Ethernet to continue to be mainstream for most applications.

The maximum VIN voltage will be higher, probably around 35V-40V.

Yeah! Does that mean we could possibly squeeze out more speed with our existing steppers by using e.g. 36V?

@dc42 said in Is there any news or blog on the Duet3:

Are absolute encoders for linear position actually available, at affordable prices?

All this is new to me, so have some mercy when answering to the following idea:

Since there is a "permanent storage" possiblity within the duet-pcb, that is the sd-card, the absolut information could be handled in the ram of the processor and in the case of a blackout be stored to the sd-card...? (Maybe the boards supply voltage has to be "prolongued" with some capacitors for making that possible?

If only an absolut-enoder-wheel is needed - without storage - , I guess if I start buying a stepper which simply has its rotor-axis broken out on both sides (which usually costs not much compared to the whole cost of the stepper) + the absolute-wheel that should be the cheapest option?

O.K. and now I come back to my old idea: If within the new trinamic-drivers "micro-plying" (that is whatever stepping resolution is set at least the movements are always smoothed to 1/256) Couldn´t I just do a mechanical gearing from stepper-motor to moving-axis for the high-resolution needed and use full steps with 1/256 microplying (because of 36V I would still get enough speed out of the motor), and so wherever the motor stops since it always "rests" in the full steps this is also a very secure setup, with low possiblity of loosing steps? Because in the "old" info about micro-stepping what you find is the only reason full steps were avoided was because in the time before "micro-plying" of course in full steps the movement behaviour was much more shaky and so was loud and steps were lost in high-speed, etc.... but if full-step movement is controlled and smooth now, I do not see a point in not trying a small setup that way out saving me the money for encoders (and all the cabling and connectors and and and)

EDIT: I think I have to stop drooling around with my fantasy going too wild

dc42

@xtl said in Is there any news or blog on the Duet3:

That looks like RJ-45 on the CAN? Is there a reason for such a large connector? I've used 4-pin RJ-11 in the past and that's already more than enough pins for CAN and would even carry a pair plus voltage and even ground if needed.

The prototypes have RJ12 (6p6c) wide-keyway sockets on the boards and we plug RJ11 cables into them. The production boards will use narrower RJ11 sockets. The CAN bus needs just one twisted pair in the cable.