@ThantiK I was kinda expecting this. Try configuring them for spreadCycle instead of stealthChop!
@Thomas_Sanladerer , that’s what I plan on doing next. I’ve got them bumped up to 1.0v on the vref - seems to be a 1:1 correlation between voltage and amperage I think. They came pre-set at 1.2v but overheat at that value for me.
@ThantiK
it is also possible that the firmware of your printer has double step option active, this can cause also such problems.
Take a look at http://www.repetier.com/documentation/repetier-firmware/rf-installation/ --> Stepper timings
@Watterott_electronic @ThantiK I think it is the stealthChop option. The double steps generated by the firmware should not cause missing steps. (It can generate more noise because the 16ms => 256 microstep interpolation is not working)
These drivers probably are an improvement but more testing is needed. The current settings could also be different due the changed current control algo.
I look forward to the TMC2130, It can be even more interesting then the TC2660. (In combination with an ARM / Edison)
In my opinion the trinamic drivers are much better the the allegro and ti drivers.
@Watterott_electronic , we use Marlin firmware - No double step enabled as far as I can tell. It seems to happen on very fast travel moves, and quick Z-lifts. We do Z-lifts at 150mm/s, with 20 tooth pulleys on belt drive. It’s very quick going up and then rapidly back down on all motors. I’m going to extensively attempt getting these drivers optimally situated - if that requires lowering the acceleration or whathaveyou, that’s fine too. The drivers are worth changing our settings for if we can manage to get them to be reliable.
@ThantiK Marlin also does the double steps. It is needed to get the higher speeds.
@erik_vdzalm , just looked through our configuration - is this not an exposed option in Marlin?
It is not an option because it does nog give problems with any of the known drivers. And it is needed to get acceptable speeds.
If the speed is low enough it is not active. For higher speeds it is needed.
We will make some test (video) after Christmas, with the TMC2100.
We build a large 3D printer (1000x800x650mm building volume) in the last month and we got some trouble with the double steps.
And we lost steps if this option was active, because you can not double the steps without driving a ramp.
A other way for fast moves is to change the microstep level to 1/4
P.S.
The TMC2130 will be also available soon.
Maybe the double steps makes the interpolator stepping to fast to the next step. (But the effect is small, Then the system was probably already on the edge of skipping. Disabling the double steps also reduces the max speed, This could also fixed the problem)
On the TMC260 i do not use the interpolator. I use them on an ARM board and use 64 microsteps. (I did not hear any sound improvement above 64 microsteps)
I am interested in the TMC2130 because it has some nice new features compared to the TMC2660. But I hope they did not increase the spi register length.
For these drivers an ARM/Edison/DSP board is better. Then you can use the drivers features together with better path planning. The speed of the atmega is just to limited to get the most out of these drivers.
Silence means less vibration means less wear and tear on the moving parts of our printers. Aside from the polarity of motion, were these drop in replacements for the pololu drivers? Do they automatically provide 256 usteps/step or does that have to be handled elsewhere in the setup? I’d love to get these on my Deltamaker if the replacement is straightforward and they are reliable.
@Jason_Doege , there were some small issues when I tested these. “StealthChop” is worthless on them. However, the spreadCycle mode I managed to get printing well. It didn’t, however, have enough power for the extruder - but that may be fixed in the newer revision that @Watterott_electronic made to the drivers. They’ve basically flipped the driver around, thermally sinked through the board, allowing you to put a LARGE heatsink on the back.
spreadCycle isn’t silent, but it’s certainly WAAAAAY quieter than 1/32 that we use now.
You would have to change your steps per mm settings in firmware to get the right speeds/dimensions. Specifically you’d need to halve them.
Hopefully whenever we get the updated versions from Watterott, they’ll prove to be a perfect drop in. These were their first run, and I’ve gotta say they were great to play with!
By default, they do 16 to 256 step interpolation. It’s a little more intricate than that, but that’s the gist.
@ThantiK Thanks for that. I watched @Thomas_Sanladerer 's video review of these and realized they weren’t what I thought at first. Really, the best way to go is with proper commutation of the drives, but that is a bit more expensive to implement.