TMC2130 vs TMC2224?

Which driver is better, TMC2130 or TMC2224? It appears the biggest difference between them is UART and the other over SPI which is 10 times faster and better supported.

EDIT: soon after I posted, I found there's an issue with TMC2224 turning off when driven fast: https://github.com/KevinOConnor/klipper/issues/196

Edited 1 time(s). Last edit at 03/21/2018 04:11PM by newbob.

The 2224 is more advanced in some ways and costs less than the 2130, but doesn't have stall detection. However stall detection for 3D printers is oversold IMO. I agree that the UART interface in the 2224 is a backward step. Nevertheless, we have a build of RepRapFirmware that supports the 2224 and its sibling the 2208.

Edit: there is no problem with 2224s turning off unexpectedly when RRF drives them.

Edited 1 time(s). Last edit at 03/21/2018 04:14PM by dc42.

---

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

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

Ok, I found another difference: steahlthChop in TMC2130 and steahlthChop2 in TMC2224:

quote from TRINAMIC:
"steahlthChop™ is using a pure current regulation without any pre-regulation. Depending on the steepness of the acceleration the target current might not be reached during the accereleration phase. We improved this in the version, stealthChop2™. Here we are using a pre-regulation which allows more torque during accerleration phases."

Above quote was from a thread about stealthChop loosing steps in stealthChop mode.

"However stall detection for 3D printers is oversold IMO"

I'm new to reprap ( a month or so). I have 2130 boards on my RAMPs and also flyback boards (all wetterott). I've never used the "regular" drivers so I can't compare the sound, but I will say that my motors seem awfully quiet with StealthChop. I have no X or Y endstops - I'm using stall detection and it seems to work just fine to me. As for it being oversold, if it's true that a lot are using it, I haven't come across that in my searches. But, I'd agree that it's probably not that useful. I could have just done switches and who knows ... maybe I will revert some day.

From my persepective, I enjoy experimenting and learning, so I was immediately drawn to the 2130. If I build a second printer I'll end up experimenting with 32-bit and will evaluate other driver offerings then.

The TMC2100, TMC2130, TMC2660 and TMC2224 drivers are indeed very quiet, mostly because of the interpolation from x16 to x256 microstepping they perform. The TMC2224 can interpolate to x256 from other microstep settings as well as from x16.

Using stall detection to replace X and Y endstops isn't difficult, but it is less accurate than microswitches or other regular endstops because the stall status is only updated every 1 or 4 full steps of the motor. This means that the homing position could shift significantly with repeated homing, if the stall position is on the edge between two full steps. There are mechanical design steps that can be taken to try to avoid this situation.

So I am tempted to say that stall detection in place of endstops is best used on low-cost printers where homing accuracy is less important than minimising the parts count - except that drivers with stall detection cost more than those without, so stall-detecting drivers are not really suited to minimum-cost printers at present.

---

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].

The 2130 is also 256 steps.

They sure do cost more! My first batch of Chinese boards weren't configured correctly for SPI and only two from the batch of five survived my attempts at soldering SMT jumpers. The second batch were real ones, costing $14 each, plus another $14 each for flybacks. I killed one of those, too. Meanwhile, I already own the microswitches that they are replacing.

So ... educate me ... why does it matter if stall detection is late? Say it's homing and it tries to step 3 times more than necessary because the 2130 board has triggered the end stop yet ... why does that matter? The moment it does trigger we know that it's at 0 and the firmware then zeroes. Am I missing something?

Quote
tbeaulieu
The 2130 is also 256 steps.

That's what I said.

Quote
tbeaulieu
So ... educate me ... why does it matter if stall detection is late? Say it's homing and it tries to step 3 times more than necessary because the 2130 board has triggered the end stop yet ... why does that matter? The moment it does trigger we know that it's at 0 and the firmware then zeroes. Am I missing something?

It matters if you are doing any of the following:

- using the facility provided by some firmwares to resume a print after a power failure
- also using stall detection to detect missed steps and to re-home the printer
- homing a delta printer to max endstops, and not always running auto calibration immediately after homing

In all of these cases you need the homing position to be highly reproducible.

---

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].

For me stall detection is only a shortcut in homing procedure. What really matters is the possibility to reduce stepper current.
When the carriage bumps into the homing position, it doesn't hurt so much. ( my ears and the printer ). Then stall detection cuts off the fruitless attempt of the steppers to move further.

My procedure to increase stall detection precision:
Loosen the grubscrews of the (delta-) drive pulleys
Configure the steppers to run in full step mode
Manually move the carrier to it's home position
Tighten the pulley grub screws

From now on, the steppers will be at a full step position when the carrier bumps into the home position. ( theoretically, for peace of mind )

Edited 1 time(s). Last edit at 03/24/2018 04:16AM by o_lampe.

Hmmm. I'm not entirely sure that the full step thing matters much, but considering that a typical belt drive seems to be about 90 steps/mm (Give or take) that's an awful lot of precision that you'd need in terms of the position for a hard stop. 0.011mm isn't an awful lot of distance to be away from the hard stop/stall position before it's 1 microstep out. In fact, the 0.178mm or so that the printer will travel at full step would probably be pretty difficult to have complete repeatability after re-tightening screws, though maybe possible.

---

MBot3D Printer
MakerBot clone Kit from Amazon
Added heated bed.

Leadscrew self-built printer (in progress)
Duet Wifi, Precision Piezo parts

Quote
SupraGuy
Hmmm. I'm not entirely sure that the full step thing matters much, but considering that a typical belt drive seems to be about 90 steps/mm (Give or take) that's an awful lot of precision that you'd need in terms of the position for a hard stop. 0.011mm isn't an awful lot of distance to be away from the hard stop/stall position before it's 1 microstep out. In fact, the 0.178mm or so that the printer will travel at full step would probably be pretty difficult to have complete repeatability after re-tightening screws, though maybe possible.

Being 1 full step out (typically 0.2mm) in X or Y after pausing the print, rehoming and resuming would cause a noticeable layer shift. I doubt that being 1 microstep out (typically 0.0125mm if using x16 microstepping) would be noticed.

---

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].

TMC2224 has INDEX output that sends a pulse every 4 steps. I wonder if that could that be used for stall detection:

if 0 INDEX pulses arrive after sending 5*(micro steps per full step) step pulses then it means movement have stalled.

Edited 1 time(s). Last edit at 03/27/2018 11:36AM by newbob.