i just installed a PT1000 from https://www.precisionpiezo.co.uk/ yesterday. works as you'd expect, fits in the E3d block nicely and its 3 x 20mm off top of my head

@dc42

That's exactly what it was! It was indeed something very basic, stupid I didn't notice it. Thanks DC42! 🙂

Michiel de Bie

The only one of these I have on a heated bed at the moment is labeled "40A", which probably means it is 16.

The bed is about 1400W, on 110/120V (US standard). So about 12A. Seems OK. Also, it has been in operative use for some months, and I've applied the highly scientific "put your hand on the case" temperature reading. It is never even warm, in an average room (about 21 to 22C).

The usual cause is poor WiFi signal strength and/or interference. The firmware attempts 2 retries when the connection to the access point is lost. This will be increased in a new firmware version soon.

Meanwhile, if you have a PanelDue on your system, you can send M552 S1 to reconnect. It's also possible to set up a push button connected to a spare endstop switch to run a macro that sends this command.

Many thanks, I have been wading through the duet documentation but not read that bit.

This was my attempt to use 10x8x500mm CF rods roll wrapped tubes.
https://www.youtube.com/watch?v=y-h5bUulqL4

Here are the tubes.
https://smile.amazon.com/gp/product/B00RWYANKY

Tolerances DO vary. One batch I got from a particular manufacturer was so bad I returned them. The ones I got from ARRIS though required only minimal work on 2 of the 4 tubes with 2000 grit sandpaper and a quick polish and they were smooth end-to-end in the Igus bearings.

When I decided to upscale to a 500x500mm print area I tried 12x8x1000mm tubes but the tolerances were so bad I gave up and went with the Igus 17mm rails.

alt text

V20 worked great! It sent the print, and I didn't see any errors in the log. I also tried the upload and simulate options and didn't have problems with any of them. Thanks.

@cata On the quad, there was no motor or any other mechanism that would indicate some means of actively mixing the filaments together. Also, the filament path to and through the nozzle was too short to indicate that there is any effective passive mixing. So as far as I have been able to observe, and from the very limited literature available, I would say that as far as mixing is concerned, the quad has the same limitations as a Diamond hot end. That is to say, the filament coming out of the nozzle is akin to stripey toothpaste. That will be the same, no matter what material goes into the input.

So the colour of any printed object will vary around it's perimeter. It would not be possible to make say "Fire Engine Red" which in CYMK colour space uses proportions of roughly 0% Cyan, 84% Magenta, 80% Yellow and 19% Black because unless the filaments are truly mixed, those proportions will vary throughout the cross section of the filament being extruded. For sure, one face would have a heavy Magenta bias, another would have a heavy Yellow bias. Other facets of the print would vary. Maybe if the printed object was cylindrical, at some point around the perimeter it might actually be "Fire Engine Red" (perhaps).

The demonstration unit that I saw at the TCT show, was printing with transparent filament. This is a trick that I use with the Diamond which disguises the fact that the filaments are not being fully mixed. It's OK and works reasonable well but of course, not every object lends itself to being printed with transparent filament.

The unit was very small and compact, I'll say that. It was running very slowly though. If I was considering making a purchase, I would want to know what the reason for that was. If it has something to do with limited melt rate or thermal management (there was no heat break that I could see), then that would be a potential reason for me not to make a purchase.

But, other than a brief view at the TCT show, I have no experience of this thing. What I have written above is just my own opinion based on limited observation.

I don't have a drawing for the IR sensor but the critical dimensions are given in the Appendix at https://miscsolutions.wordpress.com/mini-height-sensor-board/, just before the comments start.

Excellent, that's pretty much ideal behaviour for my purposes.

Idris

@tjb1 because I want to wait at least 5 minutes with the fan on to cool the part and bed. Depending on the print temp the hotend may have cooled off by then or not.

ah! Thank you. So, given the example of a 100mm cube, if I set M376 Hn where n is <100, then I'll get a perfectly flat top, and if n >100 then I'll get a wavy top.

I wonder, then, how the tapering works. If I have M376 H90, and y "units" of compensation are needed, doesn't it apply 1/90th unit of compensation per layer for 90 layers? I guess it would have to....

Add in all the other magic that the duet performs (and the complications of a delta printer) and it really makes for an impressive package!

@dc42 I think, it's the other way round. I split z to Z and U and hide the latter (P3) by

M584 X0 Y1 Z2:3 U3 E5:6:7:8:9 P3

Or maybe the U3 is messing things up?

Sidenote: hiding axis(es?) is causing trouble elsewhere too. Can't find the post right now, but I know someone else mentioned it. Will post back ASAIK.

Thanks - I'll try that the next time it happens.

@briskspirit said in Rotor inertia:

But what about rotor inertia? 34mm - 35g/cm, 40mm - 54g/cm. Is it a big difference? Is it critical?

To answer this initial question: If you are e.g. using 16tooth GT2 pulleys on the rotor this will add an artificial mass of of 130 or 210g respectively to the mass of your Y axis and will slightly reduce your maximum acceleration. Unless you have a featherweight Y axis these extra masses due to rotor inertia don't have much influence.

For details see Maximum Acceleration Calculator.

@billsrf https://github.com/dc42/RepRapFirmware/releases

It's part of the latest release candidate.

Correct, not needed.

In face, occasionally, the printer will drop off the network for some reason or another. (Router reboot, whatever). It just keeps printing...

Well I connected to the Duet with micro USB and used the Adafruit TTL on it and was able to communicate over the PanelDue connection so I placed the Teensy on and then the commands started showing up in the Duet and Teensy was receiving information so I removed the micro USB and went back to wall power only and now the Teensy is receiving information? I don't understand why it works now but it does? Thank you David.

@barracuda72 said in Building dc42's Kossel?:

beside....1350€ is alot...have a similar build and all together its about 700-800€

Corner

Rails

Extrusions from Motedis
Smart effecor...duet Wifi..
Stepper from stepperonline

ok..many of the small parts like srews etc i had already at home. Beside its quite nice to improve the design...for example build it upside down with steppers on top so you can build it with a smaller footprint

I would be surprised if you can can reproduce that build kit for that price. One major advantage of the kit is a custom wire harness with actual copper wire. Chinese parts come with cheap wire that has caused me all sorts of issues. It's a very well thought out BOM. If you compromise quality though I'm sure you can reduce the cost greatly. Always the balancing act.

But I can say that the reprap.me one works I have a couple of them and there fine you do need to solder the connector in place tho guess thats so you have a choice to hardwire if your wish