TMC 2130 Love-Hate Relation

Dear Forum,

I am in love with TMC2130 and on the other hand I would like to kill it. Really... This controller caused me allready some failed prints because of thermal issues...

I have a Prusa I3 clone, wooden frame and well... I was using the cheap chinese stepper drivers first and changed to TMC2130. Thank you Tom... <-- Thats me beeing happy. Can you imgaine?

Details that weren't explained in the Video were stuff like accelaration/ max feed rate / current and so on... `

I am using a thicker aluminium print bed ( not the shitty 1 mm sheet the 3mm one... quite stiff) on the Y axis. Thus: More Inertia, more momentum, more Power, more current and more HEAT.
I have active cooling and heatsinks on the Chip.

Here is the result of my last print. I love it. It looks like art.

Well. Here are my Marlin Setting. Reduced the Current yesterday from 1000 mA to 900mA. Also reducd accelaration values from 200 i think to 150.... Strange side note: Before I had to change the feedrate of the X axis (double it) but NOT of the Y axis...
If I want to use stealth choop I will have to reduce the accelaration and so on more and more to have a printer that basiclly takes tripple the time to finish the prints as a normal one...

---

#define DEFAULT_MAX_FEEDRATE          { 120, 120, 8, 25 }
 *                                      X, Y, Z, E0 [, E1[, E2[, E3[, E4]]]]
#define DEFAULT_MAX_ACCELERATION      { 1000 , 800, 40, 10000 }

#define DEFAULT_ACCELERATION          800    // X, Y, Z and E acceleration for printing moves
#define DEFAULT_RETRACT_ACCELERATION  3000    // E acceleration for retracts
#define DEFAULT_TRAVEL_ACCELERATION   3000    // X, Y, Z acceleration for travel (non printing) moves

#define DEFAULT_XJERK                 10.0
#define DEFAULT_YJERK                 10.0
#define DEFAULT_ZJERK                  0.4
#define DEFAULT_EJERK                  5.0


 #define R_SENSE           0.11  // R_sense resistor for SilentStepStick2130
  #define HOLD_MULTIPLIER    0.2  // Scales down the holding current from run current
  #define INTERPOLATE          1  // Interpolate X/Y/Z_MICROSTEPS to 256

  #define X_CURRENT         900  // rms current in mA. Multiply by 1.41 for peak current.
  #define X_MICROSTEPS        16  // 0..256

  #define Y_CURRENT         900
  #define Y_MICROSTEPS        16

---

Please, can someone have a look at these values and tell me why I get hot drivers? I dont get it at all...

Cheers, Luke

When we designed the generation 2 Duets in 2016/2017 we rejected the TMC2130 because of its lower current handling capacity than even the A4988. We chose its big brother (TMC2660) instead.

The 2130 should be ok if you choose the stepper motors to match it. Prusa has his own branded stepper motors and I have no doubt that he will have chosen the right balance between rated current and inductance. It's also no surprise to me that he changed to 24V power (although that's also necessary to support the power pause facility - just as it is for the Duet) because that allows the use of lower current higher inductance motors without sacrificing top speed as much as with 12V power.

So if you find that you have to reduce the current to keep the drivers cool, and the reduced current is well below the rated current of your motors so that it doesn't give you the torque and acceleration you need, you have a couple of options:

1. Switch to lower current motors, and 24V power if you are not already using it; or

2. Change to electronics with TMC2660 drivers.

---

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].

Well... you maybe right.. I don't want to switch to a 24 V system at the moment.
Thus I change the parameters back and will try a lower Jerk value for the Y Axis.

We will see... as soon as i know whats the problem I will let you know. I have this problem especially with very small "islands" or with very high infill (small honeycombs).
So I asume that the Jerk value is to high... Maybe maybe... We will see...

Regards Luke

Well... Change the Jerk values to a lower value for Y. From 10 to 5. No overheating.

I was printing a part with honeycomb infill and 50% density.

This testing piece showed me that the X axis got hot with the old Jerk values.

Going to increase the max feedrate on Y and decrease X jerk.

further Results i will let you know.

I just binned my 2130s today because Im fed up with the drivers overheating, my motors overheating and failed prints due to the overheating issues even at slower speeds and vastly reduced accels. I just refitted the 4988s and whacked up the speed and accels to the old settings I had and now Im happy again. I now get fast, no fail prints everytime!

TMC2130s are ok on Cartesian printers if you take account of their lower maximum current rating compared to A4988 etc. This means you have to choose motors with higher resistance and inductance than usual and use 24V power. Which is what Joseph Prusa did in the i3 MK3.

I wouldn't use them on a delta printer, because the higher inductance reduces the maximum speed. TMC2660 drivers can handle around double the current of the TMC2130, so they are ok with the traditional low inductance motors.

---

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
Banana_Luke
I have active cooling and heatsinks on the Chip.

Just a comment, you may know it: the heatsink must be installed at the opposite side of the circuit board, not on the chip.

I bought TMC2130 from FYSETC, they have some disadvantes. One is that the place where the heatsink shall be installed, there is no heat conduction, because they coated lacquer over the heat conduction pins.

Here are some infos:
[discuss.toms3d.org]
I bought additional Watterott after reading this.

This is a long overdue answer... I spent a lot of time fidling around with the TMC2130s... douh... never stop a running system... and... I couldn't get them running... partially yes...
Now I switched to the DRV8825 (1/32 microstepping).
Whats the main difference? Well... Its cheaper (5 of them for about 15 €) and they are also "very" silent... Of course not to compare with the TMCs silent mode but coming from an A4988 it is definitly a difference between night and day, and I really would recommend everyone that builds a 3D printer to buy those, because of the good cost-value ratio...

Anywas... I am happy with that one...

Regarding TMC2130s cooling. I have the original ones from Waterott, I Placed the heat sink on the holes you can see in the link below. That should be right?
I am using a 12V system... people say with 24V it should be fine/better...

[www.filastruder.com]

regards Luke

I fell into the same "trap" with the 2130's. Purchased them to make the machine quieter..
Installed on the X and Y axis, configured the firmware and fired it up....holy sh@t I thought the axis were coming apart...jerking and rattling to no end.
Re-checked the firmware settings and connections...all ok.

Lowered the current settings (from 1000 to 800) and fired it back up. Same result when moving the axis...tried one more time by reloading "fresh" firmware (1.1.8 vers)
Then nada from either driver...
Ok rechecked it all..everything correct...

Loaded the simple sketch provided with the 2031 Arduino file, attached another motor...no dice. Installed another 2130 driver and voila motor works.
So seemingly the x-y drivers had blown shortly after install and after the jerking and rattling of the motors.

Put the 8825 back in the x and kept the 2130 in the Y. Printed a test cube and wow layer shifts from the Y side! Unreal!
Was about to give up, but decided to change acceleration on the x/y from 3000(default) to 1000 and jerk from 10 to 2.

Printed a test cube, looked ok but corners were blah...Increased jerk and layer shift came back!
Where I ended up is with acceleration at 500 and jerk at 20 for the x/y....

Going to try this for a while...see what happens...

Final thoughts:
It makes a very quiet machine, but settings have to be tweaked quite drastically (in my case). I don't think these drivers are capable of their max current rating without faulting.
Looked at the 2660 as well, but they are not available in the same board form as the 2130's...

OP's title is wright....a love/hate relation

Like I said in this post [reprap.org], TMC2130s are good if you have low-current motors to suit them and preferably 24V power too. So you may need to change your motors and up the power supply voltage to get good results from them. What is the specification of your motors?

When we designed the Duet WiFi, we rejected the TMC2130 in favour of the TMC2660 because we realised that too many users would have motors that needed more current than the TMC2130 can handle comfortably.

---

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].

This is currently what I have on each axis. Sticking with 12V supply since that's what I got...

84oz Bi_polar, 1.8Deg, 2A, 1.4Ohm Phase resistance...

I'm thinking of going to these for my X and Y

62oz Bi_polar, 1.8Deg, 0.85A , 6.2 Ohm Phase resistance...(42x42x48)

Or these..

85oz Bi_polar, 1.8Deg, 0.64A , 15 Ohm Phase resistance... (42x42x60)

What are your thoughts?

Edited 1 time(s). Last edit at 06/23/2018 06:01PM by Bartje64.

The 0.85A motors are a possible choice, but with only 12V power I suggest you use the motor EMF calculator at reprapfirmware.org to check the XY movement speed you will get before the torque starts to drop.

The 0.64A motors have too much resistance and inductance for 12V operation.

---

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].

Well I plugged in the motor data (EMF calculator) for what I have currently and what I was looking at.
It has me confused....


Current: 84oz Bi_polar, 1.8Deg, 2A, 1.4Ohm Phase resistance...
If I'm reading this correctly......and reduce the motor current to 250ma..my torque slips low at 243, high at 260..
Increase it to 850 and it gets worse....low 175, high 223
Both have peak back emf of 8.2v and 0.8 v.


With the "low current" motor.: 62oz Bi_polar, 1.8Deg, 0.85A , 6.2 Ohm Phase resistance
the motor current to 250ma..my torque slips low at 116, high at 137..
Increase it to 850 and it gets worse....low 58, high 76
Both have peak back emf of 14.4v and 0.9 v.

If I'm interpreting this correctly, the motors I have are OK, but I need to reduce the current settings to get the speed?

I must be mis-interpreting this....

Bart

That makes sense to me, since the inductance stores energy. The higher the current, the more energy. The driver has to switch the coils against their natural current flow, so there is extra time needed when there's more energy stored than necessary.

Thus, cranking up the stepper current beyond the 'sweet spot' makes them slower.

I wish, I could explain it better...

250mA motor current would be too low. I suggest you base the calculations on the actual motor current being about 75% of the rated current.

---

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].

Correct, at 250ma it just did nada....(but vibrate)..
So I've been playing with this now for some time running the 2130, then back to the 8825,
Playing with the current on the 2130..up..down..playing with different speeds but have not yet gotten the 2130 where I wanted it...
Ordered a 12-24V dc-dc boost and am going to try that with one axis using the 2130....

For the heck of it, I checked the trimpot voltage on the 8825's, right where they should be at .75V (Rated motor current is 2A so at .75 Max current is at 1.5A).
Dropped it down to .42 which dropped the max current to ~850ma. Boy that made a difference! It got much quieter.
So far no layer shifts...
So I guess the lesson here is that there is no need to run at higher current levels than needed, all it seems to do is create noise (in my case)

Still going to play with the 2130 when the dc-dc boost comes in...