A rebuildlog for a Delta printer

Thanks, DragonFire. As soon as I saw your posting I checked out the extruder but it doesn't look to have any involvement. I have done many repeats now and it looks like the right-hand tower stepper motor is losing steps suddenly. The problem has become more common since it first appeared and everything has been changed except the stepper motors and the TMC 2225 drivers.
The front-running candidates - disregarding demonic possession for now at least, are the following:-
1) The TMC2225 stepper motor driver for that tower is a bit hinky and goes into thermal limiting for a fraction of a second once in a long while. Irms is set to 0.88 amps (Vref=1.25V, sense resistor 0.11Ohm)
2) The stepper motor itself is having a mechanical problem where the rotor is jamming against the stator - maybe there is something loose in the gap?
I wrote that last night but didn't post it. Since then I have changed the drivers to old A4988 drivers which I set to 1.25A and got a fail in the same right-hand stepper motor after about 10 layers of a small print. The motor feels, sounds, smells, and looks good although all 4 drivers seem to be running quite hot.
Sitting here pondering, weak and weary, I wonder if perhaps my old IsoTech multimeter is reading very low. Maybe it is time to break out the 1960s Avo model 8.

Mike

Might be a failing breaing in that motor then. They do wear and seize eventually.

Or, maybe you got some debris in there jamming it up.

I'd unscrew it from the back and take a look. Generally needs a gentle metal tap to get the aluminium casing bits apart, bui it is all serviceable, and the bearings look identical to skate bearings to me

A spray with a bit of PTFE lube might be all it needs, generally a good last step before you put it back together again. Can get a bit oily to hold though, it's all a bit of a faff first time around taking apart a Nema 17.

EDIT: A loose grub screw on either the motor gear or the return (idler or gear) might be it too. Or, the belt mounting. You did say you had checked everything mechanical but just in case you did not, I thought I'd mention them.

Edited 2 time(s). Last edit at 06/19/2023 04:40PM by DragonFire.

Oh, and yeah, it could be the steppers are set a bit high, but they shouldn't really have changed since you last had it running?

Regret I can't advise on tweaking, I have used both driver types but am switched what it was on the TMC, I run at .86 on the A4988.

Let's be honest here, ALL of your motors are due a service. Even if it doesn't fix, it's about time they had some internal TLC.

EDIT: Ah... I recall sometimes in a ball bearing, the balls are coated with anti-corrosion layer. If that cracks off, you can get flakes floating around the inside. That does happen with linear rail balls inside the mechanism.

You're telling me you can twist freely, but it might be a little bit more subtle than that,

Edited 1 time(s). Last edit at 06/19/2023 06:48PM by DragonFire.

Maybe it's REALLY subtle...

... Dust or debris is more likely to accumulate in the pot trimmer for setting the current. Which would increase the resistance, set a higher current than what it was when you put the printer to one side?

That would affect all the steppers roughly equally, which is why they are all running hot.

If you give each trim pot a little squirt of WD40 switch cleaner, that might get it all running smooth n cool.

Edited 1 time(s). Last edit at 06/19/2023 11:18PM by DragonFire.

Thank you DragonFire, I think I have found the problem: Sometimes several different problems can all give the same symptoms and this appears to be the case here.
1) The multimeter is reading very high so the stepper drivers were set just above their upper limit.
2) The A4988 that I put in the same position on the RAMPS board is likely faulty. They were all set below the driver limit but the heatsink on one seems to run very hot.
I found these by setting up the previously removed boards (P 2B, Arduino Mega 2560, and RAMPS 1.4) on the workbench and using a trusty AVO multimeter to check the Vref and motor current. Working on the bench was much easier than working in the cramped space on Miranda's underside.
So my IsoTech multimeter, a birthday present in the late 90s is now in the bin. Modern instruments are just not made to last.
I will need to recalibrate and do a lengthy test but feel a little more confident. In the longer term, I may do some further rebuilding, including servicing the motors and replacing limit switches; but for the moment I need a rest.

Mike

Edited 1 time(s). Last edit at 06/20/2023 06:18AM by leadinglights.

The auxiliary box with the vacuum and blower pumps is finished bar only some shake-out tests. The black carbuncle on the side contains the vacuum failure switch, the air filter for the cooling pumps, the silencer for the vacuum pump, and a box cooling fan.



The present, and perhaps last, part is a touch probe for bed mapping. Some time ago I decided that underbed sensors were inherently problematic for mapping the whole bed surface. My present approach is to use a single underbed piezoelectric sensor for determining the nozzle Z position, and simultaneously reporting on nozzle cleanliness. A second sensor of the touch type would be used for mapping the surface.

Although a switch type sensor such as a BLTouch or Euclid probe would be satisfactory, I am already using a piezoelectric touch sensor on my Biceps printer so I decided to try something new: A dockable piezoelectric sensor.

The present status of this is that "first light" has been achieved - the sled attaches to the effector and detaches - both smoothly. The piezo senses contact at a very low pressure and what problems are apparent don't seem to be unresolvable. Much more details on this on [reprap.org]

When I finish this I may buy a 1937 SS Jaguar to restore - something nice and easy like that.

Mike

Now working generally O.K., but some work to do. I have put a video on [www.youtube.com]

Here is a gif to be going on with



I have put a bit more for discussion in Tech-Talk [reprap.org]

Mike

At a loose end while waiting for parts I thought I would report on the present status of the rebuild.

The dockable piezo is above expectations and works well with cloakedcodes's dockableprobe Klipper module. Comparing the mesh from two runs, the two meshes are substantially within 0.00625mm at most points. The under-bed probe gives a nice accurate Z datum, again within one step. I am not presently using the "plastic on nozzle" detection as I haven't worked out how to integrate this with Klipper.

The vacuum bed is in a somewhat less happy position. I have known for a long time that printing ABS on a thin film was pretty much a non-starter as ABS's tendency to curl was greater than any vacuum could overcome. I have however been printing with PLA on a 0.15mm polyester film on the vacuum bed with good results so I decided to incorporate a vacuum pump as part of the rebuild.

On trying out my new vacuum system, having previously used my big laboratory vacuum pump, I discovered something about small diaphragm and vacuum pumps - that they leak backward through the valves. The result of this is that the pump can pull a vacuum of 850mb below atmospheric pressure, but even with a reservoir of 350ml the leakage would result in a vacuum loss of about 5mb per second. Trials with a variety of vacuum pumps gave results not a lot better than this.

Edit: The leak above is with the pump switched off i.e., what vacuum the pump will hold when used to evacuate the reservior.

The problem seems to be that the pressure at which the valves in the pump open, known as the "Cracking Pressure", has to be low to get a good vacuum, but this means that the valve and seat must match very well in order not to leak. The poor leaking performance is a result of an engineering/cost compromise.

The parts I am waiting for are a selection of solenoid valves that can be used to stop the loss of vacuum when the pump is not running.

Mike

Edited 1 time(s). Last edit at 10/09/2023 06:12AM by leadinglights.

Quote
moosix
The switch to a 24V main PSU is also a good move, as it can provide more power and potentially improve heating and overall performance. Managing different voltage requirements for fans, Arduino, and Raspberry Pi with buck converters is a practical solution. Speaking of interesting games online. There's information about [casinosanalyzer.com] Google Pay casinos that just can't help but delight. Now I'm many times quicker to fund my casino account and nothing can stop me) The addition of an enclosure for the diaphragm pumps and a dual sensor setup for Z-related tasks shows a dedication to optimizing the printing process and enhancing precision.

Thank you for taking an interest, but wtf with the spam in the middle??

Mike