IDEAS Trinamic TMC2100 - quieter (and better) stepper drivers (?)

yes go ahead, be our guinea pig )))

I see no use for the Astrosyn since the do little when compares to simple nema mounts in nylon to reduce vibration noiceon the acrylic. And fans are the biggest source of noise (so I am changing them all including a huge 80 con electronics), but reducing the pitching sound of the drivers would be nice. Only issue i see is that these Sticks seems not best suited in a Rumba

 

I got mine yesterday and have to solder them but thats not an issue. Need to figure out the microstepping and voltage. It says these drivers are sensitive to cutting power in motion. They actually have a protector on their site for this. The same applied to voltage when powering on.

Is this an issue with other drivers as well?

 

This is a deal breaker since have decoupled the raspberry pi form the rumba and he USB will always be on.
OH well, I will look at the Drv8825 then

 

More from watterots FAQ:

 

@Stian Indal Haugseth thanks. It is important to know about thesis constraints

 

Oh my... There is a 5V regulated out on the TCA2100 and the data sheet states you can feed the 5V to VCC_IO which is VIO in the StepStick. This way it would power its own logic correctly. Not very impressive engineering as the SilentStepStic doesn't use this. I need to locate the 5V on the driver and fix this my selv

 

This should work. This way I can feed the TCA2100 5V to VIO (VCC_IO). Just removing the pin from the driver to the RUMBA. This will leave the driver powerless when the RUMBA has no motor power even if USB is connected.

The red trace is from pin 25 (5VOUT) on the TCA2100. I will have to be careful and add a small wire (blue line) wherever I am able to solder. Will of course test with just one first

 

Arch.. look complicated to me, so it is not just unplug and plug

 

As the only useful mode in this case (with the Rumba) is 1/16 spreadcycle, it's best to omit pin CFG1 to CFG3 and solder a bridge from CFG1 to GND on the driver board, like that you don't have to worry about jumper settings or pulldown to ground on CFG1 on your board. Note: CFG2 and CFG3 have to be open.

 

Where do you have this from?

I did some research now and I found a forum not recommending skipping the pins and doing your own pulldown. This thread on Ultimaker forum https://ultimaker.com/en/community/11571-step-by-step-installation-of-silentstepstick-drivers-on-umo. I don't think he got the current limit correct as the SSS adjusts the RMS current.

This is what I will try:
Keep all the CFG pins. Open all the jumpers on the RUMBA as this will give me CFG1 -> GND, CFG2 -> open and CFG3 -> open. This would give me 16 microsteps and 256 interpolated steps.

So far all is simple. Then I want to remove VIO pin and use the 5V from the TMC2100 to feed itself VCC_IO. This way I do not have to be careful with USB power being present when there is no motor power.

PLEASE CORRECT if this is wrong! Take a look at the quote below.

Update (minor braindysfunction):

As far as I know deviding something by 1.41 and then multiplying with same should end up with the same number

Vref = Imax
1.5V = 1.5A

From Watterotts SSS FAQ:

 

After first trial on Y I'd call it a failure.

1. It was reversed so I had to reverse the motor cable on the RUMBA.
2. I just noticed Y has 1/8 stepping and SSS does not support this. So I doubled the steps with M92. Don't know why only this stepper driver was running in 1/8 and not 1/16 like the others. Probably a good reason.
3. It seemed the Y axis was staring to become disabled sporadically. It was turning itself on and off every 4 - 8 second. First I thought it was overheating and disabling itself but I did notice some errors in the log on OctoPrint that it was not responding. Maybe 1/16 is just too fast?

TMC2100 only supports full, 1/2, 1/4 and 1/16 microstepping.

Next I'll try 1/4. Would I loose a lot of resolution doing this?

The funny or not so funny thing was it was just as noisy as the A4988 when it wasn't cutting out. Just on that back plate resonance frequency.

 

I suggest to set Vref to 1.4V =~ 1A, thats how I run them. There is written somewhere not to exceed 1.2A. On some boards you can't trust that leaving the jumpers sets CFG2 and CFG3 to open, so it's better to omit the pins. For sure 1/16 is not to fast for the board.

 

Please note that this will increase the thermal load on the driver, as it will now be using its own voltage regulator.
This means you'll have less headroom for experimentation (they already heat up more than other drivers, especially with interpolation or stealthchop on).

I strongly suggest mounting some heatsinks on the driver packages.

 

Thanks, will try to lower the current some more.

Heat sink is already mounted. But I think I will get some bigger ones as there is more room for a big one.

 

So my Geetech drivers turned up. Put one in tuned the motor to around .66 but it's still skipping steps, only when trying to start homing it's fine when printing?! Do have any advice, tune it higher?

 

@Syko_Symatic : My voltage is on 0.62 at the moment, and no problems.

 

Hmm... @Rob Heinzonly whats the acceleration on yours set to?

 

@Syko_Symatic : It's at 1000 mm/s²

 

Hmm wonder what the issue is then!?

 

@Rob Heinzonly didn't need to move a jumper or anything? Maybe it just needs to be plugged in again? It's odd it fails every time I start a print or home, eventually it moves but is way out of position!?