3D Printer Extruder Motor Issues

[jmeosbn]

Hello hackers,

TL;DR - We likely have a fault with the Pololu board for the extruder - does anybody have either a A4983 or A4988 Stepper Motor Driver Carrier that we can test with?

Extruder Issues

For some time the extruder stepper on the 3D printer at the hackspace has had issues with not feeding the material evenly.  This has been most noticeable as "banding" due to regions with little or no material for parts of the tool path and, most recently, failure of a print job due to material no longer being fed at all.  It could also be heard as an occasionally ticking/knocking of the stepper motor indicating missed steps, or a constant ticking with the motor incrementing back and forth.

At our last hackspace, Steve saw that part way through a print, the stepper started to reverse constantly (no stuttering as before).  The day after, the next print started okay but had this same occurrence near completion of a test part - with the stepper reversing the filament completely out of the extruder.  Since that print, the stepper always runs in reverse, regardless of the intended direction.

Likely Cause

Rob noted that this fault had happened occasionally in the past and seemed to come and go based on the wiring being moved (so, an intermittent fault caused by a broken wire or connection). I have also had this fault appear and resolve itself after lifting the printer in and out of the cupboard.

The only connection I can think of that would result in this behaviour, is a break to the E0-DIR input of the Pololu Stepper Motor Driver Carrier.  This would cause the stepper driver to drive in reverse [1] despite the actual direction wanted or, as in the previous case, reverse intermittently when being driven forward, depending on whether the E0-DIR connection is broken or not.

Unfortunately, the E0-DIR connection is made from the Arduino MEGA, through the RAMPS header, and to finally to the small Pololu board for the extruder.  As such, there is no wiring that could be replaced.

Possible Fixes

I've tried reseating the the Pololu and RAMPS boards, in case of a poor connection.  I also remade some of the extruder stepper motor connections that looked damaged or had poor contact with the header pins - just in case.

Most likely, there is a fault or broken connection on the Pololu board itself - it's also possible (though less likely) that the connection from the RAMPS or MEGA is broken.

Therefore, I believe the next step is to try a different Pololu board to drive the extruder motor.  I spoke to Steve about this briefly on Wednesday and we were both reluctant to swap it with one of the other working boards.

Next Steps

 

The pololu boards we have are marked Pololu md09b, and are likely the A4983 Stepper Motor Driver Carrier (since replaced with the A4988).

Before we order a replacement board, does anybody have one of these boards spare that we can try - or have? ;)

[1] From the firmware on github, we see in Marlin/stepper.h:

#define REV_E_DIR() WRITE(E0_DIR_PIN, INVERT_E0_DIR)

and then in Marlin/Configuration.h:

// for direct drive extruder v9 set to true, for geared extruder set to false

#define INVERT_E0_DIR false

[Malcolm Pett]

Yep I do Wednesday be ok?

[jmeosbn .]

Excellent, Wednesday would be great!

Jamie Osborne

--
You received this message because you are subscribed to the Google Groups "swindon-hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swindon-hacksp...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jamie Osborne]

P.S. I'm interested to hear from anyone on whether my assessment is reasonable or not - if there could be another cause of this issue, and if I  should I test for anything else?

Jamie

[Jamie Osborne]

One thing I could test, is to confirm the fault is on the Pololu board by using the E0-STEP signal to also provide E0-DIR. On a working Pololu, this should result in the stepper always driving forward (on our configuration that sets EO-DIR to true to drive forward).

Another test, is to confirm the stepper can drive forward ok, by reversing the direction driven via its wiring.

Jamie

[Robert Longbottom]

I always thought it was an issue with the wiring in the crimp plug, or the crimp wasn't quite connecting with the pins on the driver, hence why wiggling it fixed the problem.  But that wouldnt explain why it always goes backwards now, unless the plug has been reversed.

You could try in no particular order

1. see if the motor is okay by plugging it into the xaxis connector.
2. See if the driver is okay by using it to drive the (working) xaxis motor. ( what was the resin for not wanting to try the driver board in a different port on the ramps as a test?)
3. Drive the stepper from a totally different driver - I have a motor shield for an arduino if you want to try that
4. Drive a totally different motor using the polu - ie. a random one recovered from some other electronics - isn't there one in the grey cupboard???

Hopefully then it will become clear which bit is at fault or if there is a bad connection somewhere.

[jmeosbn .]

Yes, I understood you meant the wiring, but didn't realise you had wiggled the connector specifically.  My own experience is that twice, it has been lifted it into the cupboard not working (once last week, and on Wednesday), only to be working once lifted back out.  Either way, that is why I had remade some of those connections and changed the connector housing - even though the change of direction doesn't really match those four wires - but, (as I said above) just in case...

Unfortunately, the motor always goes backwards (i.e. also in retract) - else reversing the connector would be an easy fix! ;)

Good suggestions there, I may try 4 if I can find the stepper.  3 would be a good test too, though Mal has a spare Pololu board so it'll be quicker to try that (you could bring the shield just in case though - and I'd be interested to play with it anyway).

Steve was hesitant about 2 due to not wanting to (re-)adjust the current limiters to different steppers - and I had not already tried it before then for the same reason.

I kinda confirmed 1 already, but will try your way too when I am in again. :)

Jamie

[jmeosbn .]

Damn it! Missed an obvious chance to say "reversing the polarity"!!!

Jamie Osborne

[jmeosbn .]

Oh, and before I forget: if somebody could bring some crimp pins suitable for that connector (and special tool if anybody has?) I'd like to recrimp them as they've been soldered and that's what prevents them connecting well mechanically.

For now I've made do with de-soldering them enough that the male pins can be fully inserted. :)

Jamie Osborne

[Damian Axford]

If it won't change direction at all, then it may be that the centre two pins have been swapped at some point... Or the driver is knackered. I have pins and a crimp tool... Will bring them on Wednesday

[Jamie Osborne]

Cool, Rob says he can bring some - but if you have the proper crimp tool then that would make a nicer job of it. :)

The colour of the wires seem to match the others; with the exception of one (blue I think) being extended with orange. I had traced them back from the motor as best I could and it does seem to be correct. Rob today had suggested doing a continuity check through each coil - that should confirm the middle two pins as well (so long as the outer two are correct!).

Jamie

> On 14 Dec 2014, at 17:52, Damian Axford <dam...@axford.me.uk> wrote:
>
> If it won't change direction at all, then it may be that the centre two pins have been swapped at some point... Or the driver is knackered. I have pins and a crimp tool... Will bring them on Wednesday
>

[jmeosbn .]

We confirmed during Wednesday's hackspace that our Pololu had a fault by using Mal's purple GE Tech board to run the extruder. Unfortunately, the print quality was not great - this was most likely due to less material being extruded (it appeared to mostly extrude in a dashed - - - pattern.)

Mal did leave us with some StepStick boards that just needed to have pins attached, so I did that on Friday but have been having issues getting the boards to work (drive ran partially - clicking/skipping steps - then not at all).

To rule out any serious issue with the stepper, I tried one of the pololu boards from the X axis and confirmed that works ok in E. (Though that doesn't really prove the stepper doesn't have some minor fault causing it to pull more current than expected.)

Reading the wiki page for the StepSticks, they have a current limit of 1A, where the Pololus are limited to 2A.  This is to compensate for the design having half as much copper in the board and therefore not allowing for as much heat dissipation (and as such, runs up to 20° hotter at just 400mA draw).

Mal had some heatsinks on Wednesday, so I'm afraid I may have toasted the StepSticks by not using them - each one drove the axis initially (but missing steps) then stopped soon after. Apparently there is over current protection but the failure condition for over heating is pretty much what you'd expect: that they run okay initially on power up, but fail once they get hot. I went in today hoping the StepSticks might run the X axis - neither one worked.

I've also tried the purple GE Tech board in the X axis and that only drove half the distance driven via the web interface (50mm instead of 100mm etc.) - this would explain why the prints so far had too little material extruded.

Looking at the main pololu site, official boards are $5.95, and if purchase in a bulk pack of five, can be got for $4.25 - unfortunately I think that takes up just over the limit for import tax and the site isn't showing any sensible shopping option atm (though that seems to be from failure to accept a postcode).  I'd at least want to email them before making an order.

http://www.pololu.com/product/2975

Jamie

[jmeosbn .]

Since the UK suppliers seem to add a crazy amount of markup to these boards, I checked each of the official UK distributors for a reasonable price. The best I've found is £5.59 each or £4.87 ea. for 5+ (all inc. VAT).  Unfortunately, they are showing as out of stock - I'm sure they had stock yesterday though.

http://www.technobotsonline.com/pololu-a4988-stepper-motor-driver-carrier.html

I'll have a look at unofficial distributors and eBay etc., but this seems a reasonable price, especially when compared to the others so far.

Since I'd like to replace the E driver, have at least one spare, and replace Mal's boards, I propose that we order 5 as that's almost as cheap as getting 4 - especially if we order from the US. (about £14 from pololu.com, sans shipping and tax - or £24 from technobots, shipping should be £2.95.)

I'd like to order these soon in time for our next Hackspace, any chance you could put this order in Rob, once I've found a decent place to order from that has stock?

Jamie Osborne

[jmeosbn .]

Have emailed pololu.com wrt shipping and import tax.  Have referenced the bulk packaged, pins soldered variant as at $5.50 ea. for 5+, these are still cheap while not having to worry about soldering them.

If we have to buy elsewhere, and the StepSticks and clones should work with the heat sinks attached (so long as we won't need > 1A current draw), these "Fulfilled by Amazon" clones aren't too expensive, and eBay (via China) is really cheap.

Jamie Osborne

[jmeosbn .]

If we didn't mind waiting a bit, the purple Drv8825 boards matching Mal's that we tested on Wednesday can be had for £2 (and £1.25 ea on aliexpress). Though we'd have to make sure we can actually use these with our printer; Steve says it should just be the axis_steps_per_unit parameter that needs changing in the firmware.

Atm, I'm thinking the Sintron StepSticks are the best option, especially as they are shipped by Amazon (allowing for easy returns!).

I don't have enough cash in my online account to order these now, but do have a refund with amazon to process, which I can use to buy these if nobody has a better option?

Jamie Osborne

[Robert Longbottom]

Yep, the Sintron ones you found from Amazon look reasonable at £10.95 for 5.  If people agree then I'll order them.

(This looks like a reasonable source if we only wanted one or two, but the Amazon link is cheaper for 5 http://www.ebay.co.uk/itm/3D-Printer-Stepper-Motor-Driver-A4988-Heatsink-Pololu-Reprap-Mendel-Prusa-/201046247659)

Cheers,

Rob.

On Sunday, 21 December 2014 20:43:41 UTC, jmeosbn wrote:

If we didn't mind waiting a bit, the purple Drv8825 boards matching Mal's that we tested on Wednesday can be had for £2 (and £1.25 ea on aliexpress). Though we'd have to make sure we can actually use these with our printer; Steve says it should just be the axis_steps_per_unit parameter that needs changing in the firmware.

Atm, I'm thinking the Sintron StepSticks are the best option, especially as they are shipped by Amazon (allowing for easy returns!).

I don't have enough cash in my online account to order these now, but do have a refund with amazon to process, which I can use to buy these if nobody has a better option?

Jamie Osborne

On Sun, Dec 21, 2014 at 8:09 PM, jmeosbn . wrote:

Have emailed pololu.com wrt shipping and import tax.  Have referenced the bulk packaged, pins soldered variant as at $5.50 ea. for 5+, these are still cheap while not having to worry about soldering them.

If we have to buy elsewhere, and the StepSticks and clones should work with the heat sinks attached (so long as we won't need > 1A current draw), these "Fulfilled by Amazon" clones aren't too expensive, and eBay (via China) is really cheap.

Jamie Osborne

On Sun, Dec 21, 2014 at 7:16 PM, jmeosbn wrote:

Since the UK suppliers seem to add a crazy amount of markup to these boards, I checked each of the official UK distributors for a reasonable price. The best I've found is £5.59 each or £4.87 ea. for 5+ (all inc. VAT).  Unfortunately, they are showing as out of stock - I'm sure they had stock yesterday though.

http://www.technobotsonline.com/pololu-a4988-stepper-motor-driver-carrier.html

I'll have a look at unofficial distributors and eBay etc., but this seems a reasonable price, especially when compared to the others so far.

Since I'd like to replace the E driver, have at least one spare, and replace Mal's boards, I propose that we order 5 as that's almost as cheap as getting 4 - especially if we order from the US. (about £14 from pololu.com, sans shipping and tax - or £24 from technobots, shipping should be £2.95.)

I'd like to order these soon in time for our next Hackspace, any chance you could put this order in Rob, once I've found a decent place to order from that has stock?

Jamie Osborne

On Sun, Dec 21, 2014 at 7:06 PM, jmeosbn  wrote:

We confirmed during Wednesday's hackspace that our Pololu had a fault by using Mal's purple GE Tech board to run the extruder. Unfortunately, the print quality was not great - this was most likely due to less material being extruded (it appeared to mostly extrude in a dashed - - - pattern.)

Mal did leave us with some StepStick boards that just needed to have pins attached, so I did that on Friday but have been having issues getting the boards to work (drive ran partially - clicking/skipping steps - then not at all).

To rule out any serious issue with the stepper, I tried one of the pololu boards from the X axis and confirmed that works ok in E. (Though that doesn't really prove the stepper doesn't have some minor fault causing it to pull more current than expected.)

Reading the wiki page for the StepSticks, they have a current limit of 1A, where the Pololus are limited to 2A.  This is to compensate for the design having half as much copper in the board and therefore not allowing for as much heat dissipation (and as such, runs up to 20° hotter at just 400mA draw).

Mal had some heatsinks on Wednesday, so I'm afraid I may have toasted the StepSticks by not using them - each one drove the axis initially (but missing steps) then stopped soon after. Apparently there is over current protection but the failure condition for over heating is pretty much what you'd expect: that they run okay initially on power up, but fail once they get hot. I went in today hoping the StepSticks might run the X axis - neither one worked.

I've also tried the purple GE Tech board in the X axis and that only drove half the distance driven via the web interface (50mm instead of 100mm etc.) - this would explain why the prints so far had too little material extruded.

Looking at the main pololu site, official boards are $5.95, and if purchase in a bulk pack of five, can be got for $4.25 - unfortunately I think that takes up just over the limit for import tax and the site isn't showing any sensible shopping option atm (though that seems to be from failure to accept a postcode).  I'd at least want to email them before making an order.

http://www.pololu.com/product/2975

Jamie

On 14 Dec 2014, at 18:45, Jamie Osborne  wrote:

Cool, Rob says he can bring some - but if you have the proper crimp tool then that would make a nicer job of it. :)

The colour of the wires seem to match the others; with the exception of one (blue I think) being extended with orange. I had traced them back from the motor as best I could and it does seem to be correct. Rob today had suggested doing a continuity check through each coil - that should confirm the middle two pins as well (so long as the outer two are correct!).

Jamie

On 14 Dec 2014, at 17:52, Damian Axford wrote:

If it won't change direction at all, then it may be that the centre two pins have been swapped at some point... Or the driver is knackered.  I have pins and a crimp tool... Will bring them on Wednesday

--

You received this message because you are subscribed to the Google Groups "swindon-hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to swindon-hackspace+unsubscribe@googlegroups.com.

[Damian Axford]

the other option to get the purple ones to work is to adjust the micro-stepping jumpers - I believe they're a different mapping than the pololu drivers.

[Damian Axford]

and i vote sintron pack of 5

[Jamie Osborne]

Doubling the steps per unit works, but this driver is only loaned for testing - so only a temporary fix.

I'm not sure if the StepSticks are damaged or not now - I tried one in X again, and it moved while homing (just about 20mm or so) but driving it manually didn't work and adjusting it didn't either.

After that I couldn't get it to work again. It's as if one of the internal protections trip - one of which is thermal cut out IIRC. It wasn't hot at all though, so I'm not sure if they are damaged by overheating (but perhaps now just don't work long enough to get hot).

Concerned that if it wasn't the lack of heat sink, then new ones may not make any difference. That said, these are supposed to work. :/

Rob, if you're ok to order these, then I'll leave it in your hands and have someone else try their luck fitting them! ;)

Jamie

On 22 Dec 2014, at 10:47, Damian Axford <dam...@axford.me.uk> wrote:

the other option to get the purple ones to work is to adjust the micro-stepping jumpers - I believe they're a different mapping than the pololu drivers.

[Jamie Osborne]

Looks like our jumpers are selecting the 1/32 mode that isn't on the Pololu boards. Is there any advantage to this, can the fw make use of the increased resolution?

Jamie

On 22 Dec 2014, at 10:47, Damian Axford <dam...@axford.me.uk> wrote:

the other option to get the purple ones to work is to adjust the micro-stepping jumpers - I believe they're a different mapping than the pololu drivers.

[Damian Axford]

wouldn't bother with 1/32 stepping - the mechanicals aren't good enough to see any benefit.

another fairly easy/cheap improvement is to move to T2.5 belt (i think it's currently using T5 belt) - I have some continuous belt spare and it's a quick job to print off some diff drive pulleys

[Jamie Osborne]

Ah that's a shame.

A belt upgrade with half the pitch sounds like a good plan - but I think Steve has already printed up the bigger belt pulleys you suggested on Wednesday!

Jamie

On 22 Dec 2014, at 14:38, Damian Axford <dam...@axford.me.uk> wrote:

wouldn't bother with 1/32 stepping - the mechanicals aren't good enough to see any benefit.

another fairly easy/cheap improvement is to move to T2.5 belt (i think it's currently using T5 belt) - I have some continuous belt spare and it's a quick job to print off some diff drive pulleys

[Damian Axford]

yeah - I was a bit slow off the mark.... sorry Steve :)  at least they're small prints...

[Gyrobot]

 

And they're beautiful too, such a waste ;-)!!

[Jess Robinson]

Aye from me.

Jess

Jamie Osborne

Jamie Osborne

Jamie Osborne

To unsubscribe from this group and stop receiving emails from it, send an email to swindon-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "swindon-hackspace" group.

To unsubscribe from this group and stop receiving emails from it, send an email to swindon-hacksp...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Using Opera's mail client: http://www.opera.com/mail/

[Robert Longbottom]

Right, I've ordered them.  Delivery estimate from Amazon is 31st - 3rd Jan, so hopefully we'll have them for the first new year hackspace.

Rob.

[Jamie Osborne]

Excellent, thanks Rob!

Jamie