Typical speed settings for a MM 1.5 / Marlin?

[T i m]

Hi all,

The MM 1.5 I recently built for a mate it is generally working ok ... just that we get the odd joggle, typically on the X axis (the last was about a single 1mm joggle in a tall test box after about 3 hours printing).

https://dl.dropboxusercontent.com/u/5772409/Spooooni.jpg

We have checked all the mechanicals and all seems fine so it's possibly a speed related setting that we have left (inappropriately) at default (possibly for the more common 1.8 Deg steppers?) or not set properly etc. The print measurements seem to be pretty close so I think we have that bit ok. It is having to push 2 x direct drive extruders about and I'm sure that's not as easy as just one (half the mass to start and stop)?

Now, the kit came with 0.9 Deg steppers and that straight away seemed to cause some issues as few of the defaults in Marlin seemed to work and those we did spot needed reducing quite a bit.

So, we haven't seen exactly when these joggles typically appear but I suspect it's during one of the accellerated moves (or stops from such).

FWIW the motors are all Wanti 42GYGHM809'S (0.9, 1.68 A, 4.2 kg⋅cm) driven via GT2-20's from a 30A 12v PSU though a Mega / RAMPS and the red A4988 boards (both the motors and the drivers run cool / warm).

When we first tried the motors with the defailt settings it seemed that they were being overdrivewn step-rate-wise so I set as many of the speed_related settings I could slower (and that seemed to sort it out at the time), however, I feel we have missed some and would like some feedback from the panel as what sort of values might be reliable / conservative please? For us it's reliability of speed any day, however, I'm aware that sometimes the printer needs to move fast, when bridging etc and I assume thefore it would actually be detrimental to limit all the speeds arbitarily?

Cheers, T i m

p.s. Please note the printer isn't mine or in front of me so it isn't easy for me to 'just try stuff' etc.

[Michael Anton]

On Monday, June 22, 2015 at 9:03:02 AM UTC-6, T i m wrote:

Hi all,

The MM 1.5 I recently built for a mate it is generally working ok ... just that we get the odd joggle, typically on the X axis (the last was about a single 1mm joggle in a tall test box after about 3 hours printing).

https://dl.dropboxusercontent.com/u/5772409/Spooooni.jpg

We have checked all the mechanicals and all seems fine so it's possibly a speed related setting that we have left (inappropriately) at default (possibly for the more common 1.8 Deg steppers?) or not set properly etc. The print measurements seem to be pretty close so I think we have that bit ok. It is having to push 2 x direct drive extruders about and I'm sure that's not as easy as just one (half the mass to start and stop)?

That is caused by the stepper missing steps.  Lower your acceleration settings.  I run 1500 on mine.  I usually print at 20mm/s for outside shells, 60mm/s for infill, non-printing moves at 180mm/s.  I've been able to run as high as 300mm/s reliably.

 

Now, the kit came with 0.9 Deg steppers and that straight away seemed to cause some issues as few of the defaults in Marlin seemed to work and those we did spot needed reducing quite a bit.

You may need to drop microstepping down to 8X from 16X to get your 0.9degree steppers to be reliable, at high step rates.

 

So, we haven't seen exactly when these joggles typically appear but I suspect it's during one of the accellerated moves (or stops from such).

FWIW the motors are all Wanti 42GYGHM809'S (0.9, 1.68 A, 4.2 kg⋅cm) driven via GT2-20's from a 30A 12v PSU though a Mega / RAMPS and the red A4988 boards (both the motors and the drivers run cool / warm).

What do you have the motor current set to?

 

When we first tried the motors with the defailt settings it seemed that they were being overdrivewn step-rate-wise so I set as many of the speed_related settings I could slower (and that seemed to sort it out at the time), however, I feel we have missed some and would like some feedback from the panel as what sort of values might be reliable / conservative please? For us it's reliability of speed any day, however, I'm aware that sometimes the printer needs to move fast, when bridging etc and I assume thefore it would actually be detrimental to limit all the speeds arbitarily?

I find that I get better bridging by slowing down, as it gives more time for the fan to cool the filament.  I find that there is a lower limit to the speed, and that below that the quality is affected.  For me that is around 20mm/s.

 

Cheers, T i m

p.s. Please note the printer isn't mine or in front of me so it isn't easy for me to 'just try stuff' etc.

Mike 

[T i m]

On Monday, 22 June 2015 22:19:43 UTC+1, Michael Anton wrote:

On Monday, June 22, 2015 at 9:03:02 AM UTC-6, T i m wrote:

Hi all,

The MM 1.5 I recently built for a mate it is generally working ok ... just that we get the odd joggle, typically on the X axis (the last was about a single 1mm joggle in a tall test box after about 3 hours printing).

https://dl.dropboxusercontent.com/u/5772409/Spooooni.jpg

We have checked all the mechanicals and all seems fine so it's possibly a speed related setting that we have left (inappropriately) at default (possibly for the more common 1.8 Deg steppers?) or not set properly etc. The print measurements seem to be pretty close so I think we have that bit ok. It is having to push 2 x direct drive extruders about and I'm sure that's not as easy as just one (half the mass to start and stop)?

That is caused by the stepper missing steps.  Lower your acceleration settings.  I run 1500 on mine.  I usually print at 20mm/s for outside shells, 60mm/s for infill, non-printing moves at 180mm/s.  I've been able to run as high as 300mm/s reliably.

Thanks very much for those Mike. I'll check what we currently have set and adjust accordingly.

 

Now, the kit came with 0.9 Deg steppers and that straight away seemed to cause some issues as few of the defaults in Marlin seemed to work and those we did spot needed reducing quite a bit.

You may need to drop microstepping down to 8X from 16X to get your 0.9degree steppers to be reliable, at high step rates.

Ok, something else to try. Is that because with that fine a natural step angle and trying to sub step to 16X it's all a bit much? If it cures the skipped steps,  should doing that affect the print quality much (could it make no difference or even make it better)?

OOI, these were two prints we started before I turned up the current a bit to the X axis and printed the Spooooni (that only had one joggle after a few hours printing).

 

So, we haven't seen exactly when these joggles typically appear but I suspect it's during one of the accellerated moves (or stops from such).

FWIW the motors are all Wanti 42GYGHM809'S (0.9, 1.68 A, 4.2 kg⋅cm) driven via GT2-20's from a 30A 12v PSU though a Mega / RAMPS and the red A4988 boards (both the motors and the drivers run cool / warm).

What do you have the motor current set to?

Erm, I'd have to answer 'I don't know' Mike. What I do know is that the original instructions suggested setting the voltage on the pot on the driver to .55V when just powered by the Mega? I think we may have raised all the drivers from that and now tend to go on how reliably it's working and how hot the drivers and motors get after some prolonged use.

With the X driver turned down a bit we got joggles on nearly every layer on one print. ;-(

 

When we first tried the motors with the defailt settings it seemed that they were being overdrivewn step-rate-wise so I set as many of the speed_related settings I could slower (and that seemed to sort it out at the time), however, I feel we have missed some and would like some feedback from the panel as what sort of values might be reliable / conservative please? For us it's reliability of speed any day, however, I'm aware that sometimes the printer needs to move fast, when bridging etc and I assume thefore it would actually be detrimental to limit all the speeds arbitarily?

I find that I get better bridging by slowing down, as it gives more time for the fan to cool the filament.

Ah, the reverse of my thoughts then. ;-)
 

 I find that there is a lower limit to the speed, and that below that the quality is affected.  For me that is around 20mm/s.

Check.

Cheers, and thanks again for all your advice, T i m

[Michael Anton]

On Monday, June 22, 2015 at 5:15:49 PM UTC-6, T i m wrote:

On Monday, 22 June 2015 22:19:43 UTC+1, Michael Anton wrote:

On Monday, June 22, 2015 at 9:03:02 AM UTC-6, T i m wrote:

Hi all,

The MM 1.5 I recently built for a mate it is generally working ok ... just that we get the odd joggle, typically on the X axis (the last was about a single 1mm joggle in a tall test box after about 3 hours printing).

https://dl.dropboxusercontent.com/u/5772409/Spooooni.jpg

We have checked all the mechanicals and all seems fine so it's possibly a speed related setting that we have left (inappropriately) at default (possibly for the more common 1.8 Deg steppers?) or not set properly etc. The print measurements seem to be pretty close so I think we have that bit ok. It is having to push 2 x direct drive extruders about and I'm sure that's not as easy as just one (half the mass to start and stop)?

That is caused by the stepper missing steps.  Lower your acceleration settings.  I run 1500 on mine.  I usually print at 20mm/s for outside shells, 60mm/s for infill, non-printing moves at 180mm/s.  I've been able to run as high as 300mm/s reliably.

Thanks very much for those Mike. I'll check what we currently have set and adjust accordingly.

 

Now, the kit came with 0.9 Deg steppers and that straight away seemed to cause some issues as few of the defaults in Marlin seemed to work and those we did spot needed reducing quite a bit.

You may need to drop microstepping down to 8X from 16X to get your 0.9degree steppers to be reliable, at high step rates.

Ok, something else to try. Is that because with that fine a natural step angle and trying to sub step to 16X it's all a bit much? If it cures the skipped steps,  should doing that affect the print quality much (could it make no difference or even make it better)?

OOI, these were two prints we started before I turned up the current a bit to the X axis and printed the Spooooni (that only had one joggle after a few hours printing).

The issue is that there is a maximum step rate that is possible due to processor speed (I'm assuming that you are using an Arduino, or equivalent).  Also, as the steppers warm up, the magnetic field will drop some, which reduces your torque.  Running 0.9 degree stepper requires double the step rate.  It shouldn't affect the print quality much, as microstepping is not very accurate past a certain point.  At 8x microstepping, with 0.9 degree motors, you are still probably better off compared to someone running 16x, and 1.8 degree steps, as you are less reliant on microstepping.

That said, I have my Y running on a 0.9 degree stepper, with 16X microstepping and I haven't really had problems, but I did need to drop my acceleration further than before I changed the motor.  Likely the 1.8 degree motor had a higher torque output, which didn't help the situation.

 

 

So, we haven't seen exactly when these joggles typically appear but I suspect it's during one of the accellerated moves (or stops from such).

FWIW the motors are all Wanti 42GYGHM809'S (0.9, 1.68 A, 4.2 kg⋅cm) driven via GT2-20's from a 30A 12v PSU though a Mega / RAMPS and the red A4988 boards (both the motors and the drivers run cool / warm).

What do you have the motor current set to?

Erm, I'd have to answer 'I don't know' Mike. What I do know is that the original instructions suggested setting the voltage on the pot on the driver to .55V when just powered by the Mega? I think we may have raised all the drivers from that and now tend to go on how reliably it's working and how hot the drivers and motors get after some prolonged use.

With the X driver turned down a bit we got joggles on nearly every layer on one print. ;-(

 

When we first tried the motors with the defailt settings it seemed that they were being overdrivewn step-rate-wise so I set as many of the speed_related settings I could slower (and that seemed to sort it out at the time), however, I feel we have missed some and would like some feedback from the panel as what sort of values might be reliable / conservative please? For us it's reliability of speed any day, however, I'm aware that sometimes the printer needs to move fast, when bridging etc and I assume thefore it would actually be detrimental to limit all the speeds arbitarily?

I find that I get better bridging by slowing down, as it gives more time for the fan to cool the filament.

Ah, the reverse of my thoughts then. ;-)
 

 I find that there is a lower limit to the speed, and that below that the quality is affected.  For me that is around 20mm/s.

Check.

Cheers, and thanks again for all your advice, T i m

Mike 

[T i m]

<snip>

OOI, these were two prints we started before I turned up the current a bit to the X axis and printed the Spooooni (that only had one joggle after a few hours printing).

The issue is that there is a maximum step rate that is possible due to processor speed

Understood.
 

(I'm assuming that you are using an Arduino, or equivalent).

I am.
 

  Also, as the steppers warm up, the magnetic field will drop some, which reduces your torque.

Understood (and hence why it might print ok at the beginning but not later on).

 

  Running 0.9 degree stepper requires double the step rate.

Ok, and 'for the same speed' I'm assuming? ie, If you stepped a .9 at the same rate as a 1.8 you would be at the exact same rate but just have half the speed / distance?
 

 It shouldn't affect the print quality much, as microstepping is not very accurate past a certain point.

This ios something I'm not sure I fully understand (as in see how well it all works in the real world). Ok, I understand how stepper motors work (attracting and / or repelling(?) a permenant magnet with an electro magnet in specific magnetic 'steps') and I understand how microstepping of that base principal would work but I'm unsure at what point you still have full (rotational) control of a motor under such conditions. Like, if the motor is told to come to a halt at the 12 of 16 microsteps, can it? Or is it just that the motor can be rotated using the 16 microsteps (where inertia would also play a part) but it would only be able to stop on a whole / natural step?

 

 At 8x microstepping, with 0.9 degree motors, you are still probably better off compared to someone running 16x, and 1.8 degree steps, as you are less reliant on microstepping.

Ok, that sorta goes along with my previous thought process. (Ignoring all other factors ...) The inherent accuracy is going to be down to the base resolution of the motors. However, ITRW, there are other considerations, like the accuracy of the rest of the system so there will come a point where the accuracy of the motor is lost in the errors in the rest of the hardware?
 

That said, I have my Y running on a 0.9 degree stepper, with 16X microstepping and I haven't really had problems,

Nor have we, now anyway. Apart from the potentiall difference in mass between the X and Y, the Y is running just the shory (30cm?) lead that comes with the motor, whereas the X is using an additional extension lead. Now, considering the (low) resistance / inductance) of these motors and the restricted diameter of wire one can use with those std 'header' type connectors, I'm guessing it is possible for our X to be running under quite different circumstances to our Y because of this?

 

but I did need to drop my acceleration further than before I changed the motor.

Because I am (we are) new to all this and it's onlt really being done for fun / exploration, I will confess we focussed on getting it built and running rather than being fascidious about every detail. Initially very litte worked (the motors just made noises like they were being over driven frequencywise) and then we got it to do more using Repetior but not directly from the LCD and now we can generally print the same from both. However I see there are still values that we haven't explored and haven't really touched the storing ov values in EPROMS and typically (I) make changes to the Marlin config file and upload the adjustments to the Arduino / RAMPS that way.

 

 Likely the 1.8 degree motor had a higher torque output, which didn't help the situation.

Sorry Mike, didn't help because it confused matters?

So, I don't know what to go for now re our stepper motors ... do I go for .9's as they are inherently more accurate and we can still step them /2 and have the same load on the CPU as a 1.8 stepped at x16? Or do I (we) go for 1.8's as they seem to be the norm?

Someone suggested that a .9 was likely to be more accurate than the rest of the hardware so a bit pointless (in our usage)?

I would really like to make the best decision re the motors, not only because my mate will probably also get what I get but because unlike say the use of threaded rod V leadscrew, it 'will' cost quite a bit if we make the wrong decision and have to replace all the motors. ;-(

<Devils advocate mode on> Like, does a cartesian printer really put a substantial load on it's CPU, expecially compared with a Delta? If an Arduino Mega *can* run a Delta, allbeit, not with a graphics display etc, is it possible it can easily cope with something less demanding, possibly including a .9 degree stepper? Your own experiences seem to suggest it can (and I'm not really questioning anything you have said Mike, just thinking out loud and in general).

I say this will full respet to all those who do know what's going on under the hood of all this malarky but I'm betting there is also quite a bit of 'follow_the_leader' with this game, just as there is in many other 'hobby' subjects.

I used to race RC electric cars and a couple of mates got into it recently and invited me along to a club meet. I dug out all my old gear and turned up to learn just how much things had moved on (brushless motors, Lithium batteries etc). Other racers were interested in my stuff because they hadn't seen anything like that (that old) before! Who runs NiMH these days! ;-)

Anyway ... in spite that I hadn't actually raced for ages and my gear was out of the dark ages, I somehow managed to get into the 'A' final and didn't come last in that either <g>.  The bottom line, the evolution of the technology wasn't an advantage on that track and to (many of) those drivers ... and for some it was 'all the gear and no idea ..' <weg>

Now 'of course', anyone with real world experience of all the aspects of 3d printing *IS* likely to be in possesion of those useful facts about things that really make a difference, but like you said, people were doing all this using solid bearings and drill rod and probably doing ok. ;-)

Just me thinking out loud and trying to get the bigger picture. ;-)

Cheers, T i m

[Michael Anton]

On Tuesday, June 23, 2015 at 1:24:24 AM UTC-6, T i m wrote:

<snip>

OOI, these were two prints we started before I turned up the current a bit to the X axis and printed the Spooooni (that only had one joggle after a few hours printing).

The issue is that there is a maximum step rate that is possible due to processor speed

Understood.
 

(I'm assuming that you are using an Arduino, or equivalent).

I am.
 

  Also, as the steppers warm up, the magnetic field will drop some, which reduces your torque.

Understood (and hence why it might print ok at the beginning but not later on).

 

  Running 0.9 degree stepper requires double the step rate.

Ok, and 'for the same speed' I'm assuming? ie, If you stepped a .9 at the same rate as a 1.8 you would be at the exact same rate but just have half the speed / distance?

Yes, if you stepped each at the same frequency, the 0.9 deg. steppers would move at half the RPMs. 

 

 It shouldn't affect the print quality much, as microstepping is not very accurate past a certain point.

This ios something I'm not sure I fully understand (as in see how well it all works in the real world). Ok, I understand how stepper motors work (attracting and / or repelling(?) a permenant magnet with an electro magnet in specific magnetic 'steps') and I understand how microstepping of that base principal would work but I'm unsure at what point you still have full (rotational) control of a motor under such conditions. Like, if the motor is told to come to a halt at the 12 of 16 microsteps, can it? Or is it just that the motor can be rotated using the 16 microsteps (where inertia would also play a part) but it would only be able to stop on a whole / natural step?

It all depends on the load placed on the motor.  It doesn't really have full torque when it is in between steps, so it may not come to the expected position.  Microstepping makes the motion smoother, but not necessarily more accurate, or with a higher degree of precision.  It does to some degree, but not to the degree that we try to push them.  It will stop between steps, but the actual position may not be exactly what was commanded.

 

 

 At 8x microstepping, with 0.9 degree motors, you are still probably better off compared to someone running 16x, and 1.8 degree steps, as you are less reliant on microstepping.

Ok, that sorta goes along with my previous thought process. (Ignoring all other factors ...) The inherent accuracy is going to be down to the base resolution of the motors. However, ITRW, there are other considerations, like the accuracy of the rest of the system so there will come a point where the accuracy of the motor is lost in the errors in the rest of the hardware?

Right.

 

 

That said, I have my Y running on a 0.9 degree stepper, with 16X microstepping and I haven't really had problems,

Nor have we, now anyway. Apart from the potentiall difference in mass between the X and Y, the Y is running just the shory (30cm?) lead that comes with the motor, whereas the X is using an additional extension lead. Now, considering the (low) resistance / inductance) of these motors and the restricted diameter of wire one can use with those std 'header' type connectors, I'm guessing it is possible for our X to be running under quite different circumstances to our Y because of this?

No, it should be running exactly the same current, since the drivers act as a constant current source.  There will be a greater voltage on the motor as seen by the driver, but that won't matter too much, and in truth, the wire resistance is still probably pretty insignificant compared to the motor winding resistance.  Remember you are driving your motors with at least 12V, and the motors are probably only rated at 4V or so.

 

 

but I did need to drop my acceleration further than before I changed the motor.

Because I am (we are) new to all this and it's onlt really being done for fun / exploration, I will confess we focussed on getting it built and running rather than being fascidious about every detail. Initially very litte worked (the motors just made noises like they were being over driven frequencywise) and then we got it to do more using Repetior but not directly from the LCD and now we can generally print the same from both. However I see there are still values that we haven't explored and haven't really touched the storing ov values in EPROMS and typically (I) make changes to the Marlin config file and upload the adjustments to the Arduino / RAMPS that way.

If you have an LCD, you should be able to change a lot of the motion constants from there for testing, and then tell it to write the new config. when you are sure of the values.

 

 

 Likely the 1.8 degree motor had a higher torque output, which didn't help the situation.

Sorry Mike, didn't help because it confused matters?

So, I don't know what to go for now re our stepper motors ... do I go for .9's as they are inherently more accurate and we can still step them /2 and have the same load on the CPU as a 1.8 stepped at x16? Or do I (we) go for 1.8's as they seem to be the norm?

Someone suggested that a .9 was likely to be more accurate than the rest of the hardware so a bit pointless (in our usage)? 

I would really like to make the best decision re the motors, not only because my mate will probably also get what I get but because unlike say the use of threaded rod V leadscrew, it 'will' cost quite a bit if we make the wrong decision and have to replace all the motors. ;-(

It probably won't matter which motors you use, as unless you end up with are really high voltage, or really high inductance (bad), you will be able to get them to work.

 

<Devils advocate mode on> Like, does a cartesian printer really put a substantial load on it's CPU, expecially compared with a Delta? If an Arduino Mega *can* run a Delta, allbeit, not with a graphics display etc, is it possible it can easily cope with something less demanding, possibly including a .9 degree stepper? Your own experiences seem to suggest it can (and I'm not really questioning anything you have said Mike, just thinking out loud and in general).

Yeah, it is a pretty significant load.  If you are running Marlin, it has a double, and quad step modes that it uses when it has to move fast.  This in effect reduces the microstepping.  I don't know at what rates these kick in though, but it is documented somewhere.  Most people attempting to use Arduinos on deltas run into a lot of limitations, which is why you see everyone moving to ARM, as it really is significantly better.  I've been tempted to do the same myself, even on a cartesian.  There are times when the printer can get pretty bogged down with data.  Printing from the SD card does help with this though, as at least you aren't bound by serial port speeds.

 

I say this will full respet to all those who do know what's going on under the hood of all this malarky but I'm betting there is also quite a bit of 'follow_the_leader' with this game, just as there is in many other 'hobby' subjects.

I used to race RC electric cars and a couple of mates got into it recently and invited me along to a club meet. I dug out all my old gear and turned up to learn just how much things had moved on (brushless motors, Lithium batteries etc). Other racers were interested in my stuff because they hadn't seen anything like that (that old) before! Who runs NiMH these days! ;-)

Anyway ... in spite that I hadn't actually raced for ages and my gear was out of the dark ages, I somehow managed to get into the 'A' final and didn't come last in that either <g>.  The bottom line, the evolution of the technology wasn't an advantage on that track and to (many of) those drivers ... and for some it was 'all the gear and no idea ..' <weg>

Now 'of course', anyone with real world experience of all the aspects of 3d printing *IS* likely to be in possesion of those useful facts about things that really make a difference, but like you said, people were doing all this using solid bearings and drill rod and probably doing ok. ;-)

Just me thinking out loud and trying to get the bigger picture. ;-)

Cheers, T i m

Mike

 

[T i m]

On Tuesday, 23 June 2015 10:31:09 UTC+1, Michael Anton wrote:

<snip>

 

Ok, and 'for the same speed' I'm assuming? ie, If you stepped a .9 at the same rate as a 1.8 you would be at the exact same rate but just have half the speed / distance?

Yes, if you stepped each at the same frequency, the 0.9 deg. steppers would move at half the RPMs. 

Ok.
 

 

 It shouldn't affect the print quality much, as microstepping is not very accurate past a certain point.

This ios something I'm not sure I fully understand (as in see how well it all works in the real world). Ok, I understand how stepper motors work (attracting and / or repelling(?) a permenant magnet with an electro magnet in specific magnetic 'steps') and I understand how microstepping of that base principal would work but I'm unsure at what point you still have full (rotational) control of a motor under such conditions. Like, if the motor is told to come to a halt at the 12 of 16 microsteps, can it? Or is it just that the motor can be rotated using the 16 microsteps (where inertia would also play a part) but it would only be able to stop on a whole / natural step?

It all depends on the load placed on the motor.

Understood.
 

 It doesn't really have full torque when it is in between steps, so it may not come to the expected position.

That was what I was imagining migh be the case.
 

 Microstepping makes the motion smoother, but not necessarily more accurate, or with a higher degree of precision.

Ok.
 

 It does to some degree, but not to the degree that we try to push them.  It will stop between steps, but the actual position may not be exactly what was commanded.

I think you can sort of feel how that might work when you slowly rotate a stepper by hand.

 

 

 At 8x microstepping, with 0.9 degree motors, you are still probably better off compared to someone running 16x, and 1.8 degree steps, as you are less reliant on microstepping.

Ok, that sorta goes along with my previous thought process. (Ignoring all other factors ...) The inherent accuracy is going to be down to the base resolution of the motors. However, ITRW, there are other considerations, like the accuracy of the rest of the system so there will come a point where the accuracy of the motor is lost in the errors in the rest of the hardware?

Right.

Cool. I might be getting there! ;-)
 

 

 

That said, I have my Y running on a 0.9 degree stepper, with 16X microstepping and I haven't really had problems,

Nor have we, now anyway. Apart from the potentiall difference in mass between the X and Y, the Y is running just the shory (30cm?) lead that comes with the motor, whereas the X is using an additional extension lead. Now, considering the (low) resistance / inductance) of these motors and the restricted diameter of wire one can use with those std 'header' type connectors, I'm guessing it is possible for our X to be running under quite different circumstances to our Y because of this?

No, it should be running exactly the same current, since the drivers act as a constant current source.

Again, after posting I (thought I) had read that somewhere ... doh!
 

 There will be a greater voltage on the motor as seen by the driver, but that won't matter too much, and in truth, the wire resistance is still probably pretty insignificant compared to the motor winding resistance.  Remember you are driving your motors with at least 12V, and the motors are probably only rated at 4V or so.

I think I was thinking on from that and wondering also about the stopping power when done over a greater resistance cable. The whole back-EMF and braking control thing is a bit out of my understanding of such things.

 

 

but I did need to drop my acceleration further than before I changed the motor.

Because I am (we are) new to all this and it's onlt really being done for fun / exploration, I will confess we focussed on getting it built and running rather than being fascidious about every detail. Initially very litte worked (the motors just made noises like they were being over driven frequencywise) and then we got it to do more using Repetior but not directly from the LCD and now we can generally print the same from both. However I see there are still values that we haven't explored and haven't really touched the storing ov values in EPROMS and typically (I) make changes to the Marlin config file and upload the adjustments to the Arduino / RAMPS that way.

If you have an LCD, you should be able to change a lot of the motion constants from there for testing,

Yers, we have done a little of that.
 

and then tell it to write the new config. when you are sure of the values.

Not done that or many of the Gcode / command stuff. It took us quite a while to comprehend that you have several places to set things ... RAMPS jumpers (microstepping), driver current, Marlin settings, Host software settings, Slicer settings etc etc.

<snip>
 

 

I would really like to make the best decision re the motors, not only because my mate will probably also get what I get but because unlike say the use of threaded rod V leadscrew, it 'will' cost quite a bit if we make the wrong decision and have to replace all the motors. ;-(

It probably won't matter which motors you use, as unless you end up with are really high voltage, or really high inductance (bad), you will be able to get them to work.

Ok, so if *you* and for a MM 1.5, and given the same spec, price and even if we have to play with microstepping jumpers, would you go for a .9 degree as it might be considered more versatile than a 1.8?

 

<Devils advocate mode on> Like, does a cartesian printer really put a substantial load on it's CPU, expecially compared with a Delta? If an Arduino Mega *can* run a Delta, allbeit, not with a graphics display etc, is it possible it can easily cope with something less demanding, possibly including a .9 degree stepper? Your own experiences seem to suggest it can (and I'm not really questioning anything you have said Mike, just thinking out loud and in general).

Yeah, it is a pretty significant load.  If you are running Marlin, it has a double, and quad step modes that it uses when it has to move fast.

Ah, again, more things to confuse the innocent! ;-(
 

 This in effect reduces the microstepping.  I don't know at what rates these kick in though, but it is documented somewhere.

The trouble is with me that at 58 and trying to do a million other things (family, duty and hobby), I can't seem to remember the detail like I used. Or I can but it takes a lot longer to dully sink in than it did.
 

 Most people attempting to use Arduinos on deltas run into a lot of limitations, which is why you see everyone moving to ARM, as it really is significantly better.

Yes, I understand that for a Delta as I believe the 'live maths' is *much* more complicated. Like, it has to calculate the Z axis all the time, whereas I would imagine a cartesian only has to deal with it in one direction and only at the end of each layer? That could be one 1/3rd less data? (but probably isn't that easy or logical). ;-)

 

 I've been tempted to do the same myself, even on a cartesian.  There are times when the printer can get pretty bogged down with data.

What are the symptoms of that please Mike? I'm guessing you would see the printer hesitate or slow? I don't know that we have seen that yet but may have without realising it (and that would be with .9 motors and 16x microstepping on all axis (inc the extruder)).
 

 Printing from the SD card does help with this though, as at least you aren't bound by serial port speeds.

Understood and we have done some of that and in general with greater sucess. We did generally see the printer just lock up more often owhen printing directly from the PC and it seemed that the serial port just dropped out somewhere.

Anyway, I popped round to my mate with the printer earlier to drop of some stainless steel bits I had worked on for him and a replacement batch of filament I'd been sent by the same supplier to see if it was a batch / colour thing as to why the filament would be broken between the spool and the extruder when you came to the printer each morning.

I tried to print an 80 tooth GT2 pulley (for the belted extruder I'm trying to make) and it joggled pretty well straight away. Then we removed the extension lead to the X stepper and adjusted the voltage on the X driver to exactly match that of the Y (that doesn't joggle now). Then I went into the Slic3r setting via Repetier and noted the 'Speed for non print moves' was set to 130mm/s. So, because it seemed it might be that sort of action that was causing issues I lowered it to 60, re-sliced and the pulley printed fine.

https://dl.dropboxusercontent.com/u/5772409/80T%20Pulley.jpg
(not sure if it should be considered a 'good' print or not but it completed!). ;-)

I then re-fitted the X motor extension lead and left it printing another / bigger pulley (also printing fine as I left).

This is how it is set now:

https://dl.dropboxusercontent.com/u/5772409/Slic3r1.jpg

When I'm there next we need to go though the settings more carefully and compare them with yours etc.

Thanks again, T i m

[Michael Anton]

You can't go wrong with any motor recommended by the Reprap group.  Somewhere on the wiki there is a list of motors.  Ideally you want to find higher torque ones, that have the lowest inductance, and the lowest resistance.  Kysan makes a motor that everyone uses, and I can attest to it working very well.  The only reason I ended up putting a 0.9 degree motor on my Y was I broke a motor shaft from running my belt too tight, so I figured I'd try it.  I'd say it works ok, but doesn't seem to have the same amount of torque as the original Kysan motor did.  The part number of the Kysan motors I bought is 42BYGH4803.  I have also tried an even higher torque motor from RobotDigg, with a part number of 17HS6002.  If you have room for these, I really recommend them, even over the Kysan motors.  They run much quieter, which is very nice, and seem to have plenty of torque.  The inductance on them is a little on the high side though, so that my limit the top end stepping rates if you are only running a 12V supply (I run mine at 24V).

 

<Devils advocate mode on> Like, does a cartesian printer really put a substantial load on it's CPU, expecially compared with a Delta? If an Arduino Mega *can* run a Delta, allbeit, not with a graphics display etc, is it possible it can easily cope with something less demanding, possibly including a .9 degree stepper? Your own experiences seem to suggest it can (and I'm not really questioning anything you have said Mike, just thinking out loud and in general).

Yeah, it is a pretty significant load.  If you are running Marlin, it has a double, and quad step modes that it uses when it has to move fast.

Ah, again, more things to confuse the innocent! ;-(
 

 This in effect reduces the microstepping.  I don't know at what rates these kick in though, but it is documented somewhere.

The trouble is with me that at 58 and trying to do a million other things (family, duty and hobby), I can't seem to remember the detail like I used. Or I can but it takes a lot longer to dully sink in than it did.

I feel your pain.  I turn 50 this year, and my brain does not absorb info like it did when I was younger.  I design instrumentation for a living, so there is a constant need to keep learning about new things...

 

 

 Most people attempting to use Arduinos on deltas run into a lot of limitations, which is why you see everyone moving to ARM, as it really is significantly better.

Yes, I understand that for a Delta as I believe the 'live maths' is *much* more complicated. Like, it has to calculate the Z axis all the time, whereas I would imagine a cartesian only has to deal with it in one direction and only at the end of each layer? That could be one 1/3rd less data? (but probably isn't that easy or logical). ;-)

The problem comes in calculating between cartesian space, which is what the g-code is in, to delta space.  It's not just the Z that is calculated, but the X and Y as well.  The mathematical transforms to do this are complex, and given how much processing power is used just to generate the pulses for the motors, the poor Arduino is overwhelmed. 

 

 

 I've been tempted to do the same myself, even on a cartesian.  There are times when the printer can get pretty bogged down with data.

What are the symptoms of that please Mike? I'm guessing you would see the printer hesitate or slow? I don't know that we have seen that yet but may have without realising it (and that would be with .9 motors and 16x microstepping on all axis (inc the extruder)).

Right, it will pause periodically, and the motion is not as smooth.  You will see it more often when sending data from the computer, rather than printing from the SD card.

 

 Printing from the SD card does help with this though, as at least you aren't bound by serial port speeds.

Understood and we have done some of that and in general with greater sucess. We did generally see the printer just lock up more often owhen printing directly from the PC and it seemed that the serial port just dropped out somewhere.

I found that if printing from Repetier host, that periodically the printer would just stop.  Press the enter key, and it will continue.  I think if the USB interface drops the return "ok" that is sent from the printer, that the host stops sending data.  I found it to be very annoying, and unreliable, so I've just always used the SD card.  I did find that a Raspberry Pi running Octoprint seemed to work pretty well though.  I think I read somewhere that this is caused by ground loops due to having the USB cable connected, which actually makes some sense.

 

Anyway, I popped round to my mate with the printer earlier to drop of some stainless steel bits I had worked on for him and a replacement batch of filament I'd been sent by the same supplier to see if it was a batch / colour thing as to why the filament would be broken between the spool and the extruder when you came to the printer each morning.

I tried to print an 80 tooth GT2 pulley (for the belted extruder I'm trying to make) and it joggled pretty well straight away. Then we removed the extension lead to the X stepper and adjusted the voltage on the X driver to exactly match that of the Y (that doesn't joggle now). Then I went into the Slic3r setting via Repetier and noted the 'Speed for non print moves' was set to 130mm/s. So, because it seemed it might be that sort of action that was causing issues I lowered it to 60, re-sliced and the pulley printed fine.

I designed an extruder that uses one of the RobotDigg geared extruders, and it works really well.  I'd highly recommend doing this.  It sure makes for a nice compact extruder.

Try putting the extension lead back in.  Unless the wire gauge is really fine, it should make no difference.

 

https://dl.dropboxusercontent.com/u/5772409/80T%20Pulley.jpg
(not sure if it should be considered a 'good' print or not but it completed!). ;-)

I then re-fitted the X motor extension lead and left it printing another / bigger pulley (also printing fine as I left).

This is how it is set now:

https://dl.dropboxusercontent.com/u/5772409/Slic3r1.jpg

When I'm there next we need to go though the settings more carefully and compare them with yours etc.

Like I said, take a look at the acceleration settings on the printer.

The bridge speed that I use is 20mm/s, but I can't claim to have the best bridging performance around.  I've never spent much time working out the optimal values for this.

 

Thanks again, T i m

No problem,

Mike