M6 tapped hole completed successfully   Took it slowly, a couple of turns cutting at a time so lots of goes at it.  Swarf cleaned off and re-lubricated each time.  Hole drilled with 3mm then 4mm then 5mm Bosch titanium drill bits, lubricated.  Tap only lightly tightened in drill chuck to limit torque.  All operations using drill stand with mains drill with limit set to a low speed.  Those spiral fluted taps are really sharp!  Went through the aluminium like a knife through butter   What a difference using the right, high quality tools makes

Heatsink block now finished.  That's the M6 tapped hole, counter drilled 10mm in top to take the Bowden coupler such that the end touched the heatbreak.  The the 40mm length cut off the 100mm billet.

Been working on the heater block, some small modifications and drilled a hole for the thermistor.  Now all assembled with a pair of 24v 60W cartridge heaters and thermistor with thermal paste for heaters and thermistor.  Heater block now clamped to water block, again with thermal paste between.

Here's a photo of the complete X carriage.

X carriage now installed on X rail.  XY carriage installed in main frame and levelled.  Very nearly ready to install Duet board and wire up.  Just need to attach top acrylic sheet and extruder motor.  Then the wiring can be sorted out and the XY cords threaded round and attached to the X carriage.  (The Z drive is already done except for the wiring to the Duet board.)

Now I have the bones of the machine assembled I've been thinking about the possibility of boxing it in to provide a fume cabinet and also to enable a heated enclosure to help when printing ABS and any other type of filament that benefits from keeping warm.  The external dimensions of the frame are 720mm x 610mm, with the motors on the XY frame sticking out back and front.  In view of this I think one of the shorter sides would be best for the door.  It probably isn't necessary to have more than one side transparent (the door) and make the other sides from lengths of timber and plywood.  I plan to include a webcam somewhere to provide remote viewing of the printing process.

1 hour ago, Gina said:

Now I have the bones of the machine assembled I've been thinking about the possibility of boxing it in to provide a fume cabinet and also to enable a heated enclosure to help when printing ABS and any other type of filament that benefits from keeping warm.  The external dimensions of the frame are 720mm x 610mm, with the motors on the XY frame sticking out back and front.  In view of this I think one of the shorter sides would be best for the door.  It probably isn't necessary to have more than one side transparent (the door) and make the other sides from lengths of timber and plywood.  I plan to include a webcam somewhere to provide remote viewing of the printing process.

Here is a good tutorial on using a Raspberry Pi and MotionEyeOS to control a webcam and integrate it into the Duet web interface.

It describes using a Pi Zero which I tested and it's adequate with a Raspi cam and takes up littlemspace.

I plan to use a Pi3 so I can attach multiple cameras. E.g. one watching the bed, one on the spool. Only one will be visible via the Duet but the Pi has its own web interface for monitoring all the attached cameras.

Now sorting out all the cabling...

6 hours ago, tekkydave said:

Here is a good tutorial on using a Raspberry Pi and MotionEyeOS to control a webcam and integrate it into the Duet web interface.

It describes using a Pi Zero which I tested and it's adequate with a Raspi cam and takes up littlemspace.

I plan to use a Pi3 so I can attach multiple cameras. E.g. one watching the bed, one on the spool. Only one will be visible via the Duet but the Pi has its own web interface for monitoring all the attached cameras.

Good point about multiple cameras Dave   I've found there is no one place that's right for a camera so multiple cameras should solve the problem.

Lots of wiring up completed and extruder motor mounted.  Used spiral wrap to attach the wiring for the hotend and parts cooling fan to the Bowden tube to keep them tidy and prevent tangling up in moving parts.  I shall also attach the cooling water tubes to the bundle lightly (to prevent squashing as these are very flexible).  Duet board attached to one of the frame rails.  Job progressing slowly but steadily

Soon be time to sort out the power supplies.  Here is the power supply wiring diagram.  Main power ON/OFF control will be provided by a PC type ATX PSU with the +5v Standby supplying 5v for the Duet logic.  The main printer power of 24v will be supplied by a separate 24v PSU with its mains supply controlled by the standard +5v rail of the ATX PSU and SSR (Solid State Relay).  The ATX is controlled by the PS-ON input from the Duet board.  This allows for automatic power off after the print run has finished or in an emergency.  Remote power control is recommended by the makers of the Duet WiFi 3D printer control board for safety. 

The circulating pump for the water cooling is 12v and is powered by the +12v rail of the ATX PSU and controlled by the Duet.  The Duet also controls another SSR which is used to switch the mains powered print bed heater.  With the printer being remote controlled via its web interface this means standby power for the Duet logic can be on continuously and the main power remotely controlled.

Now to the current required.  The Duet can source up to 2.4A per motor but I doubt I'll go above 2A.  The hotend heater is a total of 120W at 24v which is 5A so at the maximum rating I might want 4x2.4 + 5 = 14.6A.  At 2A per motor this would reduce to 13A so a 15A PSU would be wanted is either case.

With the extruder mounted and the Bowden tube connected up I find the tube tends to catch on various things when the XY carriage is high up.  This will mean the printing height will be restricted to something like 600-700mm or maybe less.  It's not certain that I shall keep to a Bowden feed - in some ways direct feed would be easier though adds considerably to the mass of the X carriage assembly.  Anyway, I'll just see how it goes.

There are lots of discussions going on whether to use a bowden or direct drive.
I never used a Bowden, my pal Guy does use Bowden. We did compare our prints very often, and the result : there's absolutely no difference in perimeter quality at all.
In the end, when we were comparing very detailed prints, mine were a little bit better, hardly visible but there it was. These prints had 0.1mm layers and we even tried 0.05mm layers. These tests took us weeks, lots of time was spent tuning our slicer. Very detailed prints were rather difficult with the Guy's bowden. Most of the problems had to do with retraction not doing it's job.

In the beginning I was a bit concerned about that stepper on the XY-carriage also.  The first thing I did was change infill pattern. Standard setting when installing a Slicer is honeycomb pattern.   The main reason to change infill pattern in the first place to 'Line pattern' was to have it printed quieter, but most of all not to have the printhead violently shaking back and forth all the time because of that  silly 'honeycomb pattern'. That pattern is mechanically not correct for the printer.

So next logical step was to test a few infills. I thorough tested different infills. I hung literal 50kg(100lbs) of weight on small pieces of PLA. 'They' say honeycomb is the best/ strongest. Don't believe it. They're not all as strong as honeycomb.
A few are. Line pattern shure is strong, that's the one I always use.
During these tests (except for one) all thin testpieces survived with flying colors, my plastic bucket filled with steel and bronze didn't...

Great research. When I started using PETG I had lots of issues with stringing. This was using a direct extruder too so not a Bowden issue. I found that tuning slicer parameters to minimise retractions got rid of almost all stringing. I found the best infill pattern was concentric as it drastically reduces the number of retractions followed by a move. I stopped using honeycomb when I had my Prusa. It would almost shake itself to bits unless going very slow.

I too found honeycomb unsuitable and having tried all sorts (in Slic3r) went for Rectilinear which is fast and vibration free.  This may be the same as linear in Simplify3D.  It draws lines at 45 degrees in one layer and -45 degrees in the next forming a criss-cross pattern.   I tried concentric and that was good too but I think it took a bit longer to print.  It was a while back and I can't remember now.

As for Bowden feed, I last tried that several years ago and many improvements in both hardware and software/firmware have occurred since then.

As a short cut to providing power I have two standard 12v PSUs (25A and 16.7A) connected in series to provide 24v for the main supply to the Duet and 12v for the water pump.  Later these will be replaced with the ATX and 24v 15A PSU as per the diagram above but for testing I shall have always on power with just the switch on the mains socket to switch on and off.  After testing and when everything is working (hopefully) I shall move the printer out of my living room with the PSUs changed over.  The 24v 15A PSU I plan to use is currently providing power for the print bed on my Titan printer which will be upgraded to a mains powered bed heater, providing faster heating and freeing up the 15A 24v PSU.

I now have power and about to start testing.  I think I need to connect the Duet to USB to determine the IP address after which I can control it from my main desktop and Firefox.

Connected laptop USB to Duet, ran Repetier Host, selected Manual Control tab and sent M552 to the Duet which replied with the IP address and Access Point etc.  Disconnected USB and switch on mains power and now have remote control via Firefox on Linux Mint desktop.  Now for testing...

Test Results :-

Endstops - all working but all inverted logic!  Changed from S1 to S0 to correct X & Y endstops.

M574 X1 Y1 Z1 S0                         ; Set active low endstops

The Z endstop uses a probe and problem cured by adding I1 in M558 command as below.

M558 P4 H10 F120 T6000 I1                ; Set Z probe type to unmodulated and the dive height + speeds

Quote

NPN output normally-open inductive or capacitive sensor

Connect its output wire to the cathode of a diode, and the anode of the diode to the E0 STP pin. Connect the sensor ground wire to a ground pin on the Duet, and the sensor's + power wire to a suitable voltage (typically to VIN because these sensors usually need between 6 and 30V power).

The diode should preferably be a small-signal Schottky diode such as BAT43 or BAT85, but a small signal silicon diode such as 1N4148 works for some people.

Select mode 4 in the M558 command. If using firmware 1.15e or earlier, include M574 E0 S0 in config.g to select active low logic level for the Z probe. If using firmware 1.16 or later, include parameter I1 in the M558 command instead.

Quote

Mode 4

The trigger signal is read from the E0 endstop circuit. You can choose whether the signal is high or low when triggered. In firmware 1.15 and earlier this is done by using M574 E0 S1 to select active high, or M574 E0 S0 to select active low. In firmware 1.16 and later, use I1 in the M558 command to select active low, otherwise it defaults to active high.The device connected to the E0 STP pin must be able to sink 1.5mA for the Duet WiFi, or 2.1mA for the Duet 0.6 or 0.8.5.

Test Results :-

Test temperature sensors at room temperature - tick

Hotend heater - 17-200°C very fast - tick

The new Duet wiki has better information on setting the Z probe height.

Quote

Calibrate the Z probe trigger height

1. Make sure the dynamic test is successful (Z probe stops when it senses the bed) before doing this.
2. Use the X and Y jog buttons to position the nozzle over the centre of the bed
3. Jog the nozzle down until it is just touching the bed or just gripping a sheet of paper.
4. Send command G92 Z0 to tell the firmware that the head is at Z=0
5. Jog the head up by 5 to 10mm
6. If the Z probe needs to be deployed, deploy it
7. Send command G30 S-1. The nozzle will descend or the bed rise until the probe triggers and the Z height at which the probe stopped will be reported. If you are using a nozzle-contact Z probe, the trigger height will be slightly negative. For any other type of Z probe where the probe triggers before the nozzle contacts the bed, it will be positive.
8. Repeat from step 5 two or three times to make sure that the trigger height is consistent.
9. In Duet Web Control, go to Settings -> System Editor and edit the config.g file. Set the Z parameter in the G31 command to the trigger height that was reported. Save the file.

This will be a very useful section in the wiki when it's finished.

Motors are all working.  Increased Z maximum speed to 3000mm/m = 50mm/s and it runs fine   A 100mm step taking just two seconds.  So a 700mm range is moved in 14s.  I checked the calibration with steel rule and it's correct within the tolerance of this method of measurement.  This movement seems easy even with motor current left at the default 800mA.

Water cooling system set up and working.