X pins not working? on commissioned O2

So,

I've finally gotten my printer assembled, and on commissioning the X axis wouldn't work. (heated bed, hotend, Z and Y all worked fine) I tried reversing the Z and X cables, and the X motor ran fine off the Z pins, and then the Z motor didn't off of the X pins. I checked the X wiring loom with a multimeter and it looks fine, there's no crossover in the cable either.

I'm running DC42 1.00 O

M111 S2 returns nothing, and it also seems like the Z probe isn't working all it returns is 0 (1023), and it's not under any strong lights where it should be maxed out or anything.

I'm thinking somethings up with the duet.....

Going to contact support, but wondering what you guys think

If you have checked continuity on the loom and the motor then I would also say that there is something wrong with the Duet.

With the Z-probe I would perform some continuity checks first as well.

Quote
stuart576
With the Z-probe I would perform some continuity checks first as well.

I didn't really look too much into the Z probe issue, mainly because I'm not sure what 0 (1023) is indicative of.... I think that the second number is the probes 'off' reading, and the first number is the Delta-IR with modulation, so the 'true' IR reading compensating for background.....
So, am I correct thinking the duet thinks it's reading 1023 with the IR LED off and 1023 with the IR LED on?
I'm not entirely clear on how the Z probe works.

I know there is a section in troubleshooting about this, so I'll have to go back and try it, best to be sure the Z-probe is working before I get a warranty order shipped for the duet...

A reading of 0 (1023) suggests to me that either the proximity sensor has been miswired or that the ground connection to it is open circuit.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Ok, so I’ve gone through the troubleshooting for the z-probe
I checked the cable and all the correct wires are connected at each end with no crossover.

Voltage between yellow and green pins 3.28 Volts
Shorting Red pin to ground, duet gives 0 (0) for probe reading
Jumper between red pin and blue pin, duet gives 1024 (1024) for probe reading

Measurements with probe plugged in with black lead on ground and red on pin at connector

Probe far away from surface
Red = 0.016v
Green = 3.26v
Blue = 1.648v
Yellow = 0.005v

Probe close to white surface
Red = 0.021v
Green = 3.26v
Blue = 0.011v
Yellow = 0.006v

Does this indicate an issue with duet, wires, or z-probe?

Also sent this info to RRP support, haven't heard back about any of this yet. But I know they are very busy right now, and running behind on support.

Can you post photos of both ends of the Z probe loom, showing clearly which colour wires connecy to which pins?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I'll post pictures when I get home tonight (~ +8 hrs from now), but I think I just realized that Yellow (gnd) and Green (+3.3v) might be reversed on the Z probe end of the loom compared to the picture in the install guide.


I know I've triple checked that the duet side wires colors are going to the right pins, and that it's wired up according to the diagram, but I'll post a picture regardless, I could just be missing something.


Support got back to me today,

Sounds like RRP isn't 100% sure how your firmware handles the Z-probe, but I thought as long as M558 P2 was in the config.g your firmware handled the default probe the same way as stock.
They're thinking that it's probably something with the wiring loom, or the board.

We're also continuing to investigate the duet X stepper driver, going to get some pictures of the duet, and loom to support as well.

Quote
shadow651
Sounds like RRP isn't 100% sure how your firmware handles the Z-probe, but I thought as long as M558 P2 was in the config.g your firmware handled the default probe the same way as stock.
They're thinking that it's probably something with the wiring loom, or the board.

That is correct. My guess is that either two of the wires are swapped at the probe end of the loom, or there is a bad crimp connection at one end of the ground wire in that loom. A bad joint on a resistor on the proximity sensor board would also cause that problem.

Edited 1 time(s). Last edit at 03/17/2015 09:19AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Looks like the wires are reversed on the z end of the probe loom, green and yellow like I thought. Not sure if there is a way to get the wires out of the connector to fix it myself.





Looking at the duet for the stepper issue, looks like some solder is shorting out the X stepper driver (U5)

Quote
shadow651
Looks like the wires are reversed on the z end of the probe loom, green and yellow like I thought. Not sure if there is a way to get the wires out of the connector to fix it myself.

I can't confirm that they are swapped as I have an Ormerod 1. But if you are certain, you can easily swap them. Press a small point into the rectangular hole on the front of the connector to disengage the barb, then you can pull on the wire to remove the pin.

Quote
shadow651
Looking at the duet for the stepper issue, looks like some solder is shorting out the X stepper driver (U5)

I am truly astonished that even after all these months, RepRapPro and their suppliers are still not testing Duet boards adequately prior to shipping them. Even a visual inspection should have picked that up. The hot end boards and PanelDue board I ship are fully tested prior to shipment (I even built an automatic test rig for the hot end boards), and I wouldn't dream of doing anything different.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
I can't confirm that they are swapped as I have an Ormerod 1. But if you are certain, you can easily swap them. Press a small point into the rectangular hole on the front of the connector to disengage the barb, then you can pull on the wire to remove the pin.

Swapping green and yellow wires worked, the duet is now reporting probe values as expected!

For the record the correct order for the wired on z-probe end is red-yellow-blue-green

Quote
dc42
I am truly astonished that even after all these months, RepRapPro and their suppliers are still not testing Duet boards adequately prior to shipping them. Even a visual inspection should have picked that up. The hot end boards and PanelDue board I ship are fully tested prior to shipment (I even built an automatic test rig for the hot end boards), and I wouldn't dream of doing anything different.

It is a bit irritating I must admit, even more so with me in the US. It makes email a 1 day affair to answer a simple question since if I'm at work I can't look at the machine...... and don't even get me started about the speed of the post office....... Hopefully duet's are expensive enough parts to be sent UPS... or I might just pay to have it sent that way.

I don't think it'd be that hard to rig up something with 3 ribbon cables and some other connections to power the board and test all of it's functions.... Does also kinda make you wonder if some of the pins you don't use right away might not work later if you decide to use them


Good to know that you check your stuff, I'll probably be getting something from you at some point. But I'd like to get mine running first, save up a bit, and then I'll have to decide if I find the Z-probe/always on fan more irritating, or if I want the convenience of the touch screen

Quote
dc42

Quote
shadow651
Looking at the duet for the stepper issue, looks like some solder is shorting out the X stepper driver (U5)

I am truly astonished that even after all these months, RepRapPro and their suppliers are still not testing Duet boards adequately prior to shipping them. Even a visual inspection should have picked that up. The hot end boards and PanelDue board I ship are fully tested prior to shipment (I even built an automatic test rig for the hot end boards), and I wouldn't dream of doing anything different.

Looking at that photo the U5 Stepper chip has obviously been placed skew wiff on the board prior to soldering not an easy thing to fix without the correct SMD re-work tools and I certainly couldn't do it now my eyes ain't what they used to be.

Grounds for warranty replacement I reckon

Doug

Edited 1 time(s). Last edit at 03/18/2015 04:15AM by dougal1957.

shadow651: I'll order a warranty replacement for you. I'll email you regarding this. I forgot to add the instructions to swap the crimps over to my email, though you've obviously been able to correct this:
"If you look at the housing, as in this picture [reprappro.com] (top right), you can see the crimps. Each one is held in by a small plastic tab that is part of the housing. Using a thin, sharp point (like the tip of of a craft knife), lift the tab a mm (but no more, or it will snap off!), just enough to allow the crimp and wire to pull out. You can then swap the position of the wires, and test the probe."

dc42: As 'truly astonished' as you are, it's just not the way the world works, is it? We expect our suppliers to test the boards for us, and they promise us that they do. However, clearly they don't, or at least not 100%. We are in constant contact with them, continually trying to get them to improve their processes. We have a testing rig here, but again it takes time, and not inconsiderable cost, to test every Duet (rather more involved than your hot end board, or LCD panel), and/or every other component of the kit. So to keep prices less than extortionate, we accept a level of failures, and have a very reasonable attitude to warranties. As a diy-built kit, the electronics are always going to be the most vulnerable part. We are slowly learning how to improve this.
Also "Press a small point into the rectangular hole on the front of the connector to disengage the barb, then you can pull on the wire to remove the pin." That's okay for the 100kk crimps, but not the 1800T crimps on this connector - you'll crush the crimp if you do that. You have to lift the tab in the housing.

Ian
RepRapPro tech support

Quote
droftarts
dc42: As 'truly astonished' as you are, it's just not the way the world works, is it? We expect our suppliers to test the boards for us, and they promise us that they do. However, clearly they don't, or at least not 100%.

However you choose to do it, it's your (RepRapPro's) responsibility to ensure that the boards you ship are fit for purpose. Planning how boards will be tested is an important part of electronic design. If you are relying on your supplier to test the boards, then this is presumably covered in the contract with your supplier. Have you provided your supplier with a test schedule, and a test rig if they need one to execute that test schedule?

Quote
droftarts
Also "Press a small point into the rectangular hole on the front of the connector to disengage the barb, then you can pull on the wire to remove the pin." That's okay for the 100kk crimps, but not the 1800T crimps on this connector - you'll crush the crimp if you do that. You have to lift the tab in the housing.

Sorry, my mistake, I was thinking of the wrong type of connector.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
However you choose to do it, it's your (RepRapPro's) responsibility to ensure that the boards you ship are fit for purpose. Planning how boards will be tested is an important part of electronic design. If you are relying on your supplier to test the boards, then this is presumably covered in the contract with your supplier. Have you provided your supplier with a test schedule, and a test rig if they need one to execute that test schedule?

Of course it is, and we strive for this, and when we fail, we replace under warranty. We do all of what you suggest. But do we expect 100%? No, it's unrealistic. And then there are problems that pass testing anyway, eg poor USB socket soldering, or failures that take time to appear. Then there's handling and shipping damage, and that's before the customer even gets it. If we built fully made printers, rather than kits, it would be easy to test and seal them. They'd be more expensive, but they'd be a known-quantity, and all testing can be done at the end, rather than on individual components, that may fail at a later date, after shipping. It's what 3DSystems, Makerbot et al do. And yet they still have huge warranty issues. But we don't make completed machines, we make kits. It would be great if we had no warranty requests, but realistically, it's never going to happen.

Regarding shadow651's issue, I haven't seen a stepper driver that badly soldered in any of the over 3000 Duets we've shipped. And, in fairness, in every other respect, the board works, so it could have gone through testing, but an unobservant tester may have missed the fact that the X stepper was not moving, though all others were.

Ian
RepRapPro tech support

Edited 1 time(s). Last edit at 03/18/2015 07:37AM by droftarts.

Quote
droftarts
Of course it is, and we strive for this, and when we fail, we replace under warranty. We do all of what you suggest. But do we expect 100%? No, it's unrealistic.

It isn't unrealistic to expect that 100% of the boards that you ship are fully working when you ship them. I seriously suggest that someone at RepRapPro go on a short course in Total Quality Management, or whatever the current buzzphrase for that concept is. I can understand that in the early days of the Ormerod you had issues with USB sockets and the like not being properly soldered on the board. But after well over a year of shipping Duets, those issues should all be ironed out. For example, the quality management procedure could include a visual inspection (by you or your supplier) to check those joints, and any similar issues that automatic test is unable to find.

Quote
droftarts
Regarding shadow651's issue, I haven't seen a stepper driver that badly soldered in any of the over 3000 Duets we've shipped. And, in fairness, in every other respect, the board works, so it could have gone through testing, but an unobservant tester may have missed the fact that the X stepper was not moving, though all others were.

Maybe not that badly soldered, but I recall someone posting a photo of a board with an obvious solder bridge on one of the stepper drivers.

The key to testing electronic circuit boards quickly and accurately is to have a good test rig. You shouldn't need an observer to see whether the axes are all moving, the rig would check for the correct stepper motor drive automatically. The procedure would be to put the board in the test rig, press it down to engage the spring-loaded test connectors with the back of the board, press a button, and wait a few seconds for the test results to appear. The test rig would provide appropriate inputs monitor all the outputs to check for correct functionality. It could even have phototransistors to check that the LEDs are lighting up at the right times. It should take no more then about 30 seconds to insert, test, and remove each board. See below for the test rig I use for my hot end boards.

Only a few days ago there was a report of a Duet that was obviously missing some components. Given these issues and others reported in the past, I find it hard to believe that the stepper motor and heater outputs are tested at all.

Replacing boards under warranty is all very well, but it doesn't help your reputation when customers receive boards with such obvious faults on them. When shipping to far-away places such as the USA or Australia, you are faced with either paying for the replacement to be sent by courier, or annoying the customer further by making him wait while the package is set by ordinary airmail.

I like the Duet and I have been recommending Duet boards on the Delta printers forum, but I'm wondering whether I should continue to do this unless the quality improves. I've yet to read of anyone having to return a Smoothieboard.

btw will there ever be distributors for the Duet in the US and Australia? I have been asked more than once "where in the USA can I buy a Duet?"

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

CPU boards are generally easy and quick to functionally test if you have designed and made a suitable test rig (which often takes as long or longer than the original board design). It is not practical to make such a test jig for small quantities, but by now I should think the volumes are enough to warrant such a development. With an intelligent test rig you simply plug the Duet into the test fixture (all the pins being catered for, including those presently unused), connect USB and Ethernet cables and run special test firmware in the Duet & test fixture, which will then test the functionality of every pin and port on the Duet (and shorts between them) within seconds - including the A/D calibration, voltage on all power rails and current consumption. You also work out any paths that are *not* tested by the fixture, and either test those separately or subject them to additional visual scrutiny.

No, it will not pick up faults that develop after testing, such as poor solder joints or transit damage, but will prevent most of the board faults I have seen on this forum, and as the Duet has no BGA components a good visual inspection should pick up the remainder.

You cannot expect a board fabricator to functionally test a product unless they have been given a test fixture and/or a comprehensive written test schedule (and the charge for such an additional service is generally greater than it would cost to test in-house). The faults being reported would tend to suggest that the Duets receive either a very cursory or no functional testing whatsoever prior to shipment, leaving the customer to be the first person to test that the board is fully functional. Now that you have a reasonable volume of international sales, that model is not going to work very well, and RRP really needs to bite the bullet and develop a satisfactory test process. Like a child, bad behaviour is tolerated in the infancy of a product, but there comes an age when people no longer smile and shake their heads.

Dave

Well,

I certainly didn't mean to start a flame war over QC practices last night....

I can see both sides of the issue. On the one hand, it's very easy for us the consumer to sit around and point at a company and say 'well, I wouldn't do it that way' or 'you should've known better' or whatever. But at the same time, we want all of these goods at a prices that's as low as possible. Sadly we can't have our cake and eat it too. I think we all love how affordable RRP kits are compared to some of the other systems out there, it's basically a selling point. If the kits cost more, we'd be guaranteed that everything would work, but would they lose sales because of the price increase? and thus make less?

I don't know how common duet problems really are, but looking at the forums is going to give you an exaggerated view because we see everyone that has the tiniest problem, but we probably never see some people who build the machine with no issues.

All that said, I don't think it would be impossible to design and build a rig that would automatically test the boards, but then that's a new boring job for someone to do (plus the cost of designing and making a test rig). And at some point it's almost inevitable that someone would assume so-and-so tested this board, and pack it up for shipment before it was actually checked (which could be what happened here)
As long as humans are involved, mistakes will happen no matter what, and I typically have bad enough luck to encounter them all.


At the end of the day for a business, it's all a balance of the cost to replace boards under warranty vs the cost of trying to find every defective board before shipment; and it could be the board supplier who's eating the cost of new boards, and then it's more up to them what QC policy to use.


At least RRP honors their warranties (at least so far, I've had excellent service), there are other companies I've dealt with who just ignore you, or blame your incompetence for whatever issues your having.

Quote
shadow651
I can see both sides of the issue. On the one hand, it's very easy for us the consumer to sit around and point at a company and say 'well, I wouldn't do it that way' or 'you should've known better' or whatever. But at the same time, we want all of these goods at a prices that's as low as possible. Sadly we can't have our cake and eat it too. I think we all love how affordable RRP kits are compared to some of the other systems out there, it's basically a selling point. If the kits cost more, we'd be guaranteed that everything would work, but would they lose sales because of the price increase? and thus make less?

If the Duet were a low-cost board like Arduino/RAMPS, then it might be easier to justify the faulty-on-arrival rate. But the Duet is not a cheap board - although at least it is not as expensive as the Smoothieboard (which IMO is seriously overpriced).

Quote
shadow651
I don't know how common duet problems really are, but looking at the forums is going to give you an exaggerated view because we see everyone that has the tiniest problem, but we probably never see some people who build the machine with no issues.

I'm not talking about tiny problems here, I am talking about stepper drivers not working at all, which seems to be a regular occurrence. I've not raised this as an issue before, because it's often hard to know whether such a fault has been caused by the user connecting/disconnecting motors when the board is powered up. But there have been enough failures due to obvious manufacturing faults that it seems likely that the boards are not being tested adequately.

Quote
shadow651
As long as humans are involved, mistakes will happen no matter what, and I typically have bad enough luck to encounter them all.

One of the main purposes of a quality management system is to make sure that human error does not lead to customers receiving faulty or incorrect goods.

FWIW I have done the provisional design for a low-cost 32-bit controller board to run RepRapFirmware (not really a competitor to the Duet or Smoothie because it doesn't have Ethernet, it's intended more as a RAMPS replacement). I am considering having it manufactured. If I go ahead, I will most certainly build an automatic test rig for it so that I can ship fully-tested boards without spending more than a minute per board on testing.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

It's not a flame war by any means, but constructive suggestions to RRP. Several of us have gone through the pain of growing a small startup manufacturing company from scratch and are more than familiar with the problems and compromises. It is important to know but often far from obvious as to what is unimportant and can be neglected, and what seems OK but will bite you in the bum later unless you address it early. I think RRP is great and would hate to see them succumb to the problems of up-scaling that has seen the demise of so many promising companies. The RRP team I have encountered all seem capable, enthusiastic, hard-working and dedicated and I'd like to see them all able to retire wealthy in due course. The fact is that what works for 5 people hand-assembling in a kitchen-table type operation does not work when your shipping figures and number of employees start increasing. Formal Q.A. is a PITA but a very necessary evil. Yes, we do not see posts from the many people who have had a kit that worked first time, but we also do not see posts from people who gave up because their kit did not work and they assumed it was due to their own failure to build or understand it correctly. But far more important for RRP are the number of customers who will have complained directly to RS and demanded their money back rather than battle on with the help of the community, which will risk RRP losing a very juicy plum that they must have worked very hard to get. RS may be understanding at first, but will not put up with a high-maintenance supplier for long.

Yes, inspection and testing can be a boring job and so prone to human error, which is exactly why I suggest a process that is far, far less likely to result in faulty product reaching the shipping line than the very high possibility of a bored operator failing to spot a hair-thin solder short between pins or even a missing component. It also reduces costs in the medium term rather than increasing costs. Increasing the number of manual/visual inspections is not usually all that effective, because each inspector is less likely to be diligent, figuring that (1) anything they miss will be picked up by the next inspector and (2) if faulty product gets shipped, the blame will be shared (diluted) between several people.

Dave