Mendel prusa with 2 stepper drivers for Z axis: experiences?

Hi all,

I am finishing the build of a Mendel Prusa w/ RAMPS 1.2. Got SYSTH47-1684B steppers which are rated at a relatively high current (1.68 Amp). All goes well, but I have a couple of issues with the z axis. I'm finding the one pololu/stepstick driver per two motors solution a bit anemic, only very slow speeds in the ascending (Z+) direction are tolerated . I'm ready to try one stepper driver per motor, sharing the dir/enable/step lines, and actually have a board almost ready for this. Has anyone experience with this configuration? any known issues (sync) ? I'll report on the results. Thx for any hints!
Lanthan

I have no problems driving the two z motors from one Pololu driver. However, sharing the step/direction pins between two drivers shouldn't be an issue.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

I've got the same motors and they work fine, though I'm using Gen6 electronics from Mendel parts. Mine are wired in series (see the second post here [www.mendel-parts.com]), rather than in parallel as suggested by the wiki [www.reprap.org], as connecting in parallel can give a non-uniform distribution of current through your windings. I'm actually going to replace them with smaller NEMA17 motors as I think the 1684B motors are overkill.

Thank you all! The information is first rate.
Further testing (after reverting to a totally default configuration and gettng rid of several "improvements"):
Z+ is reasonably reliable up to F50, over this small obstacles will result in lost steps and oscillations. The motor on the left (the one that bears the load of the X motor) is indeed the most prone to losing steps, as reported in several places.
Wiring the motors serial seems a good idea.
I'm going to test the twin drivers configuration first.
Otherwise X and Y perform flawlessly. What a silence after the pirated cupcake!
L.

To close the question: couldn't test the dual driver because it appears non functional after one year gathering dust in various places..

So: I wired the motors in series as described in [www.mendel-parts.com]

that is:

pin 1 -- black -- motor1 --green -- black -- motor2 -- green -- pin2
pin3 -- red -- motor 1 -- blue -- red -- motor2 -- blue -- pin4

and Z axis performance is outstanding now. tested up to F500 ascending, no lost steps, rock solid.

In view of that: IMHO series wiring of the Z steppers should be the default wiring solution proposed in the wiki instructions, or at least should be mentioned... (going to put a note there).
Thanks again,

L

Good point, should definitely be in the wiki!

Added to the wiki Mendel Prusa page.
Noticed that information about wiring two steppers in series is quite scarce on the Internet.
Maybe we should create a specific page about that in the wiki. Would need input about theory and practice (inductance, torque, power... and also the eventual shortcomings, for now I do not see many).

The only disadvantage I can see is that the resistance will be doubled so you will need a higher voltage to push the same amount of current through the motors. With real drivers like the polou this won't be an issue as they take care of this automatically unless the voltage required to get the torque becomes more than whatever your main voltage is (12V usually) . I don't see that happening on the prusa, but it may be in some super sized design.

Do we know if the motors resistance has a negative or positive temperauture coefficient in general? If it gets less resistive with temperature then in the parallel configuration we will see the power getting more and more unevenly distributed more and more as the motors heat up.

John

droftarts Wrote:
-------------------------------------------------------

>Mine are wired in series (see the second
> post here
> [www.mendel-parts.com]
> opic.php?f=35&t=407), rather than in parallel as
> suggested by the wiki

The mendel-parts forum suggested reducing the current when connecting in series. Did you do that and how is it done with the Gen6 electronics?

(My z-axis won't run at all connected in parallel - just sits and buzzes. If I disconnect either motor, the other will run freely).

dnaman Wrote:
-------------------------------------------------------
> The mendel-parts forum suggested reducing the
> current when connecting in series. Did you do that
> and how is it done with the Gen6 electronics?

How to adjust the current: [reprap.org]
Mine is set half-way on all the stepper pots. If you are concerned, turn them right down (all the way anticlockwise), and work up from there - having a simple multimeter is a great benefit.
Though you say (in mendel-parts forum) your power supply is 12V, how many Watts does it supply? If you are using a laptop power supply, it should be supplying at least 60W, according to [reprap.org]
I use a 250W PC power supply, and haven't had any problems.

Wtih StepStick / RAMPS & SY42STH47-1684B: I ldid not decrease the driver's current. No problems so far.

Have the Z axis set in the firmware to turn off power to the motors when not in use (default config by now)

The power supply is an X-Box supply like this:

EBay PS

It's quoted at 14.2A which should be enough hopefully.

Thanks for the info on setting the current ... about the only place I didn't look, sorry.

First things first though. With the motors still in parallel, I played today with reducing the z-axis speed. It helped slightly but more importantly led me to believe that the crummy z-axis motor coupler (printed parts) are leading to considerable misalignment. I have some (metal) flexible couplings on order but I'm going to turn a pair of solid aluminum or brass couplings to try to reduce the misalignment.

After what I've read here and at MP though I think I'll probably plump for series connection ultimately.

Watts = Volts x Amps = 170W which should be fine - good to check though!
For Prusa, don't expect metal z-axis motor couplers to last long either - they fatigue fail from the strain of misalignment and weight of the X axis. I think this is the current best design, from nophead [www.thingiverse.com]
If you're trying z-axis speed in repsnapper or similar, turn it right down and work upwards. Could well be missing steps due to too fast speed.

May I also disgress about couplers:
I first tried aluminium couplers (cheap eBay, those have a spiral cut and 2 set screws 90 degrees apart) but they are not stable on the Z direction, they just act like big fat springs when the X axis is hanging on 'em. No joy.
Then I put aquarium air hose around both motor shafts and installed "Yet another Prusa Mendel Z Coupling by Nophead" [www.thingiverse.com] I agree they are just about perfect!

About misalignment:
once you get the proper coupler, re-checked verticality, etc: check the distances between the vertical threaded rod and the smooth rod with a micrometer. They should be _the same_ at the level of the top Z motor mount, and at the level of the X axis pieces. If this not the case, you will certainly have wobble down the axis. Try adjusting the Z motor position on its mount, and/or shaving off plastic or adding tiny paper leafs in the top smooth rod bracket until the distance is right.
I grew frustrated by the lack of adjustment in the Z motor mounts, and I designed a piece with slides, but it is untested. It may do the job but it is certainly not the most satisfying solution. Files attached to the fifth message in this thread: [forums.reprap.org]
IMHO a better choice would be universal adjustment with three set screws spaced 120 degrees and captive nuts. May try that at some point in the future.

About motor power: it is not the motors that will kill your PSU, but maybe that 120W big heated bed ;-)

The SY42STH47-1684B for example is nominally rated 1.68A and 2.8V, that gives us (V x A) = 4.7 Watt theoretical per motor, and we are operating it well under that power level (cooler motor = less softening of PLA mounts).
Also well within the reach of the A3988/Pololu/StepStick who is supposed to be able to deliver up to 2 Amp at up to 35V (you'll never see that because the overheating shut-off of the driver kicks in well under that power level!). You'll need active cooling for anything over 1 Amp according to the Pololu sheet [www.pololu.com]

This motor's resistance per phase is said to be 1.65 Ohm, put two of those in series and you have 3.3 Ohm.
Supposing a power supply of 12V, without any current limiting in the driver your current in the winding would be 12V / 3.3Ohm = 3.6 Amp and you'll quickly kill your motors. Available power is not the limiting factor here.

I agree with getting these things straight, but how critical it is depends entirely on your coupling. On my repstrap I have the motors at the bottom, so the weight of the x-axis is not pulling on the couplings. They are made out of 2 sizes of vinyl tubing (1/4" and 3/8" [www.thingiverse.com]), which is stiff enough to drive the leadscrew while allowing for some deflection, which is presumably why nophead ended up using vinyl tubing. Most printing couplings don't allow this, which is why they induce problems if they are not vertical, and eventually break. I don't have problems with wavy vertical walls (any problems are usually due to x and y axis belts being too loose), despite the distance between the z smooth rod and leadscrew being 37mm at the bottom and 40mm at the top! You can see in the picture below, though it looks even more exaggerated. While this is not ideal, it's due to squeezing in the linear bearing blocks. While I could fix it, it doesn't seem to effect prints, so I'm not worried about it.

Nophead's couplers are a lot stiffer than aluminium couplers with spiral cuts indeed.

Very nice build, Droftarts, with a lot of creativity involved!
Yes with that level of deflexion of the Z axes you need very compliant couplings. It is good also to show that "this works" and that we shouldn't get too fixed on details while bootstrapping.

My point is for later evolution of the Mendel: it has been audited for low plastic/low part counts and a mighty good job has been done. As it is an iterative process, now it might be the time to audit for precision/accuracy/stiffness. Some interesting compromises might emerge.

Quote

Nophead's couplers are a lot stiffer than aluminium couplers with spiral cuts indeed.

The main point is they are solid in the vertical direction and rotationally but still allow the rod to move angularly. Since the rod is constrained at only two points it is only the angle that needs to be free to prevent it displacing the lead nut sideways. As long as the nut can rock a little, the rod can be bent and will wobble, but the axis wont be displaced. Having said that it is easy to get straight threaded rods that short in the UK.

The professional couplers are all designed to have a bearing each side, they are not designed to take axial loads. The spiral ones will stretch and also wind / unwind a little if the length changes. When constrained between bearings the overall length cannot change, so they don't unwind and cause angular errors when used as intended.

I am currently working on a simpler design, with a lower part count, but stiff in the directions it needs to be. The fundamental problem with Mendel is the triangular prism frame is strong along Y and Z, but it needs to be strong along X and Y where the dynamic loads are.

---

[www.hydraraptor.blogspot.com]

Just an update on my Prusa. I made solid couplers as planned and re-aligned everything. This reduced the z-axis runout enormously.

Still very little action when the motors were connected in parallel - intermittent spurts of activity but mostly it just sat and buzzed. So I wired them in series. Still no drive just buzzing.

While looking through the Printer Definition options in Repsnapper, I chanced on the one called "Enable Acceleration" which was off by default. So I turned it on. Suddenly the z-axis worked just fine. Smooth as all-get-out. The weird thing is though that I later turned it off again and it still worked just as well. I've searched high and low for a bad connection somewhere and can't find a thing. But for now it's working.

One thing is confusing me. When I test print a part (no extruder - just going through the motions) it first homes the table. The Z-axis is fine but both the X and Y axes move away from the limit switches instead of towards them. I've double checked that I placed them where the Gary Hodgson "Visual Instructions" specified. However, I have a 3D model set for the assembly which shows them in alternative locations (which would make more sense of the way mine is acting).

I don't doubt that the X and Y directions could be reversed in the firmware or by re-wiring the motors. Or I could just move my limit switches. I don't suppose it matters which home position is used but I'd be interested in how other peoples' machines behave.

So assuming the front of the machine where the Y-axis motor is, is towards you ....at the home position:

- is the table all the way towards you with the extruder all the way to the right? or

- is the table at it's farthest position with the extruder all the way left?

Hi dnaman,

Some combinations of firmware / host software do not play together very well.

BTW: which firmware are you using?

Give a try to printrun (proterface or pronsole): for both of my machines it works very reliably. Haven't looked back at repsnapper excepted with the rare plotting gcode generated by some Inkscape plugin, that manages to get printrun and pronsole to barf.

The way I configured the mendel : Y homes all the way near me, and x at my left. Note I have set the idler Y end towards me. You play with "invert axis" in the firmware to get what you want, no need to touch the wiring.
So the printing/plotting [0,0,0] position is up (far from me) and at the left.

The home origin needs to be front left or back right (that's position ON the bed, not position OF the bed), or your prints will be mirrored. Most Mendels/Prusas default front left, I didn't know this when I built my wooden 1X2 repstrap (picture above, Y is homed, X is away from the microswith on the right) and initially went back left. I realised after the first couple of prints they were flipped, and inverted the X axis rather than the Y. The advantage of this way around is, when I home the X and Y axes after a print, the bed is out the front, ready to take the part off. The advantage of having front left origin is that the X axis limit switch will be on the same side as the motor, making the wiring easier.

droftarts,

I wired mine front-left (on the bed) but the software/firmware homes back-right and therefore doesn't find the limit switches. It did occur to me that either of the other two corners would flip the part. OK, one way or another I need to get the limit switch positions to agree with the software. The advantage of the switch being near the motor and simplifying the wiring doesn't really apply the way I have mounted my electronics.

Lanthan,

Not sure of the firmware revision .... whatever Mendel-Parts shipped with the latest Gen6 run. I'll dig around for printrun (proterface or pronsole). It might help me out actually. When I do a dummy run using some of the published Mendel stl files it seems to go through the motions OK. When I try a file that I've generated myself (Pro/Engineer) it homes and then sits there echoing something back and forth.

Thanks for the help guys - I truly appreciate it.

Like Lanthan said, you can invert the axes in the firmware. It's in 'configuration.h'.
Just in case you missed it (I'm sure you haven't), here's the Gen6 electronics page, specifically the firmware bit that talks about inverting axes: [reprap.org]

Yes, thanks, I did have that droftarts and I will change the firmware for this and also the z-axis scale.

Right now though I'm getting an avrdude error when I try to back up the existing firmware ... see separate thread.

Lanthan Wrote:
-------------------------------------------------------
>
> pin 1 -- black -- motor1 --green -- black -- motor2 -- green -- pin2
> pin3 -- red -- motor 1 -- blue -- red -- motor2 -- blue -- pin4


I'm having the hardest time parsing this chart (and since the Z motors are all that's missing from my Prusa, I'm pretty keen to find out!) Is it the wiring in my attachment correct for RAMPS?

My Z axis won't turn at all. Using prusa mendel with sanguinolulu 1.3a on Pronterface.
I get a loud hum and the motors vibrate but won't turn. When i disconnect either one of the motors, the other starts working as expected.
I'm really new to this and and don't know what to do. Do i need to set the driver current twice as high for 2 motors?

Here a movie of someone who has got the same issue. Aalthough his option of oil did not make a change.


Prusa mendel Z axis Not turning

Yes you need twice the current if the two motors are in parallel as they get half the current each. You can wire them in series if that is a problem but then the top speed will be less.

---

[www.hydraraptor.blogspot.com]

i am kinda confused re the color of the wiring, but electrically when you guys say series, you mean as in --> motor A coil A start --> end --> motor B coil A start --> end .... and so on for coil B ... am i right ?

Yes.

---

[www.hydraraptor.blogspot.com]

i only have it sometimes when it tries to move to fast so i set it to lower max speeds. works but maybe i have to check if placing in series can be a solution