Latest print is fine except for lifting off the bed at the corners - but I wasn't using a brim so no wonder.

Been looking at where there could be problems in the printer that could contribute to errors in printing.  As I posted above, I found a drum that was running askew very slightly.  I thought it was both but it turns out to be only the RH side motor the other is easily within 0.1mm.  I shall see why the errant drum is out of alignment. 

The other problem is the Z carriage (print bed frame) wheels.  ATM I have some 6mm plywood attached to the underside of a pair of rails with wheels on the ends running on the corner upright V-slot rails.  When I built the printer the wheels were right up against the rails but now I find there's some clearance.  I think the problem is that this is relying on the stability of plywood and that is just not sufficient.  Humidity changes cause changes in length and when damp the rails are pushed effectively inwards on the screws, then when it gets drier the wood contracts and leaves a gap.

What seemed a good idea at the time has proved not to be, so I'm going to change the design.  I think the answer is to have wheels either side of the corner rails at the back.  These 4 wheels when mounted on metal plates will be quite adequate to hold the bed in position laterally.  An alternative would be to replace the plywood with two aluminium rails.  Not really sure so I shall sleep on it and decide later.  The suspension cords take care of the Z axis.

Here's a photo of the current arrangement with one wheel on each corner.  The brown envelope was to stop the reflection from the flash in the glass.

Sorted out the drive drum on the RH motor and it now runs true   I've had another look at the print bed frame and by far the simplest solution will be to use aluminium rails.  Changing to two wheels at the back of each rail would mean considerable changes to the way the cords are attached and the vertical adjustment.  Also, this would mean the frame would not be as rigid at the front as it would with wheels at the front.  It could skew slightly using the rear rails as pivots - or put a lot of stress on the wheels by leverage.  I have spare aluminium extrusion.

Regarding extrusion, I have two pieces I can salvage from my Pilot printer that are a little longer than I need, so just a matter of cutting them shorter.

Extrusion cut to length.  So now the printer should be as accurate as it can be hardware-wise.  Next is the check Z calibration I think.  After that it probably orthogonality which the Duet is able to correct in the event of error.  This printer has been designed with a view to getting orthogonality correct straight off.  Actually, I can combine the Z and orthogonality calibration prints as a single print.

Now the dratted WiFi won't work and I haven't done anything that could affect it.  The router is alright as it's connected to my laptop by WiFi.  I want to check the calibration further.

Well, it's no good if it's not going to be reliable   It can't be signal strength, I brought it to less than 6ft from the router and it's been working a lot further away than that.

I dunno!!!  I get the new printer working nicely do a minor mod to the hardware and now it won't work!    I haven't touched either the firmware or any setup files.  In fact I've been nowhere the Duet board since I had it working yesterday.  Guess I'll put it to one side and go back to my Giant printer until someone can suggest something or I can think of something myself.  I dare not try the second Duet on my Titan printer until I'm sure the Mini is working reliably.  Looks like I need your help again Dave @tekkydave, please

The mains voltage bed heater pad has arrived, would you believe - sooner than expected.  Sometimes a slow boat fro China is not so slow   Pity I don't have a working printer to put it on!!!!!

The only thing I can suggest is restarting the router and Duet. I had similar issues at the start but it has been pretty stable since (touches wood).

I did restart the Duet - several times but not the router - didn't really think of that as the other WiFi devices were working.  I did try closing Firefox and restarting but no joy.  Then I thought I would see if the connection via USB still works - it does.  Sent M552 and it said WiFi idle.  Sent M552 S1 and it turned WiFi ON and now WiFi works   But how/why did it get turned off?  The config.g tells it to turn on automatically.  Is this going to happen every now and again?  Surely it shouldn't be necessary to have a USB connection available and suitable computer just to re-set the WiFi?  Maybe it will settle down

I'm now running the calibration test print and it's going well

Photos of the calibration test print run.

Calibration results :- X=100mm, Y=100mm, Z=102mm.  Orthogonality :- XY fine, XZ slightly out, YZ fine.  Print Quality :- slight curling on corners, tiny imperfections on column, poor printing at top of column.  Why it should print well all the way up to around 95mm and then become poor, I don't know.  Of course, it's a good test with a tall narrow column.  This is with cheap ABS filament.  No part cooling.  Extrusion temperature :- 260°C, bed 100°C.  Printing time :- 2hr 4m.

Photos :- On print bed, closer in, sideways lighting showing up more surface detail.

I think I've found the cause of the regular rings round the column in the Z direction - there's a slight problem with a tooth on the Z drive gear and the drive is not perfectly smooth.

Th WiFi has gone again

Spent this afternoon titivating.  Tidied up the gears for the Z drive and a bit of tidying up of the Z carriage, making sure the wheels all made contact with the V-slot rails then re-levelling.  The gears are not well printed so I think I'll either reprint them or replace with timing belt and pulleys.  This printer is already printing better that any other I've had so I'll use it to print new gears for itself

Unfortunately, the Z drive motor pinion became loose on the shaft so I'm having to print a replacement on my Titan printer.  Goodness me what a racket after the almost silent new Mini printer!!  And I wonder if the Titan will not only be quiet but also more precise when I convert it to Duet control.

New pinion printed and put on motor shaft instead of old one.  This one printed better than last time (on the same machine).  Now testing with calibration test print.

I seem to have got the print bed pretty flat as there is no visible Z movement as the nozzle proceeds over the bed.  Been playing with Z offset (Z Baby Stepping) and got the nozzle closer to the bed.  Also, upped the bed temperature for the first layer and increased the extrusion rate (to 1.3).  Bed adhesion seems improved.  I have also increased the printing speed of the infill.

Generally the print is good except for the pillar which has too much extrusion.  Think I'll stick with higher extrusion for the first layer but return to standard for the rest.

Adjusted Slic3r as above plus added a 3mm brim and started another calibration print with a simpler model.  Looking good so far (a couple of cm of column).

Left to run overnight and this morning I found a perfect print on the print bed and power shut down as programmed   There will be no need to replace the Z drive gears - the new pinion has cured that problem.  No lifting off the glass - the brim worked perfectly   As this was a calibration test print I've checked up on that. As before X=100mm, Y=100 and now Z=100.18mm and all three axes are orthogonal as near as I can tell with the corner of a piece of paper

There remains just one problem - the WiFi won't work without the USB connection to laptop.

I'm taking a "sledge hammer" approach to the ultra-weak WiFi signal and ordered a long range WiFi Access Point to connect to my router instead of the internal WiFi.
Ubiquiti Networks UAP-AC-LR 175.7 x 43.2 mm 2.4-5 GHz 802.11ac Dual-Radio Long Range Access Point - White

I wonder if improving the WiFi signal strength and sensitivity with an access point will cure the not-connecting problem.  Nothing like hoping