Welcome! Log In Create A New Profile

Advanced

TMC 2100 and ramps - any pitfalls to avoid?

Posted by DjDemonD 
Forum List Message List New Topic
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 22, 2017 03:06AM
Quote

make sure you don't have a faulty ramps with traces joining them (this is only a problem if you do the CFG 1 to GND bridge)

You will not make a shortcut, but the driver-config is still useless for 3D printing with bridged jumper pins. You'd have to cut the traces or use the LiPo cell shortcut method.
Reply Quote
GroupB
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 22, 2017 10:22AM
Next time DJ if you setting a new electronic (look like you like to test thing hehehe) try the raps128 I use them and they really good, specially in a socket, they are the only one with real top cooling and no mod needed ( socket plug and play) and on top of it they are real 128 microstep not 16 microstep X16 ( tmc suppose to have the precision of 16 microstep but the noise of 256, are rated lower in Amperage and have a know lack of torque ( can be stop with a a finger and skip if you check Thomas Sanladerer video on them)) ... that the only thing that keep me on the radds side vs duet for next electronic
Reply Quote
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 22, 2017 10:55AM
I am keen to try rapps and have heard good things about them. I will try them at some point. True 128 microstepping with 32bit boards should be doable but even with rearm it's around 100khz step pulses so at 200steps/mm @ 1/16th that's 25600 pulses/sec at 1/128 (with three axes moving you are going to start to run out). Okay on duetwifi David thinks he can get 400khz out of it.

How fast can your machine move and do you run 1/128?

Edited 1 time(s). Last edit at 02/22/2017 11:02AM by DjDemonD.

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
GroupB
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 22, 2017 02:06PM
Im using 1/64 with 1.8 stepper so far , never tried 1/128 because of the speed thing and the noise at 1/64 is very low anyway I dont see the need for going 128 right now.
The max speed im using is off course the travel and its near 230-250.

They are expensive but they are the best I tried so far I use 4988 and 8825 in the past vs my 8825 setting on same controller I got a good jump in quality but off course the 8825 have so low current zone problem that do some circle artefact on delta.

I think Angelo have a new version too
Reply Quote
GroupB
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 22, 2017 02:06PM
Double post

Edited 1 time(s). Last edit at 02/22/2017 02:07PM by GroupB.
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 23, 2017 10:53AM
Just an update on this, I had to bump up the voltage to 1.3V due to layer shifting. Took me a while to figure out. The first symptom was stripped filament in the extruder due to not being able to extrude (nozzle too close to previous layer, due to layer shifting on a delta). That had me chasing the wrong wild goose for a while.

Side note, if you try using DC42's calibration wizard, with a high acceleration, and the motors skip steps when starting a move, then you end up with some really bad calibrations!
Reply Quote
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 23, 2017 11:11AM
Yeah I'm on 1.3A as well but I am not using them for z or extruder at present. I have got two more on order.

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
GroupB
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 23, 2017 11:46AM
Probably that lack of torque thing I was talking about... I hear you can set them close to they peak rating and the controller will up and down the current when needed ( smart current control) but do some searching before I only "hear" this im not 100% sure. They probably not good enough for corexy Z because of this or direct extruder.
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 23, 2017 01:36PM
Quote
GroupB
Probably that lack of torque thing I was talking about... I hear you can set them close to they peak rating and the controller will up and down the current when needed ( smart current control) but do some searching before I only "hear" this im not 100% sure. They probably not good enough for corexy Z because of this or direct extruder.

Unfortunately the smart current config pni is also used as "enable". You'd need a tristate driver to switch in that mode ( floating= smart current ).
Also the datasheet mentioned we had to "wake up" the driver after a longer period of idle. Doesn't sound too promising for me....

My TMCs had a weightlifting session today. My CoreXY bed weights ~1.5-2kg and is moved by two ballscrews with 2mm lead. At 24V and Vref at 1.45V they were very quiet and didn't miss a step in 20mins.

Edited 1 time(s). Last edit at 02/23/2017 01:40PM by o_lampe.
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 23, 2017 08:37PM
Quote
o_lampe

Unfortunately the smart current config pni is also used as "enable". You'd need a tristate driver to switch in that mode ( floating= smart current ).
Also the datasheet mentioned we had to "wake up" the driver after a longer period of idle. Doesn't sound too promising for me....

My TMCs had a weightlifting session today. My CoreXY bed weights ~1.5-2kg and is moved by two ballscrews with 2mm lead. At 24V and Vref at 1.45V they were very quiet and didn't miss a step in 20mins.

Couldn't we just use a diode to enable smart current? High on the RAMPS pin would still get through to disable the driver, but when the RAMPS pin went low, the diode would stop the signal getting through, causing the tristate.

I'm glad you had success weightlifting today -- how hot were the drivers getting? What sort of cooling do you have set up on them?
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 24, 2017 03:04AM
The CF6_enable signal is low-active, so we'd have to fake an open state while the pin is high. Would a diode block 5V or would the TMC still see a "high" ?
If it would work, there is still the problem with waking up the driver before it receives step signals.

The weightlifting drivers were cooled by a 40mm fan mounted 10mm above the heatsinks. ( hurray for qtips and toothpicks winking smiley )
They were cool to the touch.
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 09:34AM
There's good news and bad news about the white V1.2 TMC's:

Bad news:
The jumper settings are not recognized by the chip. It will run in 16/256 stealthmode, no matter if you do the CF1_to_GND wire hack.
You have to bridge the outermost solderpads on the backside. ( The ones I thought were CF0-5 ).
I found out, because I wanted to configure the Z-stepper in quadstep/256 interpolating mode to reduce z-steps from 1600 to 400/mm. Therefor I also had to bridge the next two pads ( CF2 )

Good news:
The Vref can be reduced a lot afterwards. I've set it to 0.9 instead of 1.5 on my X and Y axis.
You will definitely hear a difference in spreadcylcle mode; the steppers chirp like canaries on extasy.
Reply Quote
Attachments:
open | download - TMC_bridged.jpg (291.5 KB)
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 10:05AM
So just to be clear for 1/16 to 1/256 spreadcycle just bridge the outermost pad, not the two pads shown in your example unless you want 1/4 to 1/256?

Good work BTW where did you find out about this?

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 10:15AM
Like you wrote, only bridge CFG1 if you want to use spreadcycle. Then alternatively do the wire hack or make sure, your controller has a pull down resistor ( like RAMPS, Rumba )
The 1/4 step interpolation mode is annoyingly loud, I switched back to 16/256 when I found out, it wasn't the cause for the Rumba board to cancel my vase_mode_burn_in_test file.

I found out by beeping the chip pins. There was no continuity between the jumper pins and the chip.

Edited 1 time(s). Last edit at 02/28/2017 10:17AM by o_lampe.
Reply Quote
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 10:26AM
Okay so I did solder both, lol, desoldered the middle set, just the outer set bridged. Removed the cf1 bridge. Now printing and definitely at 1/16th as its moving the right distance, and very quiet so must be interpolating. My extruder motor sounds so loud now.

Reduced my vref to 1.0v now printing one of the black cubes from my post above that layer shifted to see if it prints it okay.

Edited 1 time(s). Last edit at 02/28/2017 01:28PM by DjDemonD.

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 12:07PM
Okay so don't be too quick to lower your motor current if you have a fairly hefty corexy like mine:

Left one is at 1.0v the right one at 1.2v. (yes I know its underextruding a bit).
Motors running at about 40 deg C so not really a problem. I run my motors on my kossel XL at 1.1A using TMC 2660's in the same mode and no issues there.

Edited 2 time(s). Last edit at 02/28/2017 12:19PM by DjDemonD.

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 02:28PM
Maybe I could reduce my current so much because it was to high before. I'm just dialing in my new CoreXY and started with max. possible current for the TMCs.
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
February 28, 2017 10:06PM
Quote
o_lampe
There's good news and bad news about the white V1.2 TMC's:

Bad news:
The jumper settings are not recognized by the chip. It will run in 16/256 stealthmode, no matter if you do the CF1_to_GND wire hack.
You have to bridge the outermost solderpads on the backside. ( The ones I thought were CF0-5 ).
I found out, because I wanted to configure the Z-stepper in quadstep/256 interpolating mode to reduce z-steps from 1600 to 400/mm. Therefor I also had to bridge the next two pads ( CF2 )

Good news:
The Vref can be reduced a lot afterwards. I've set it to 0.9 instead of 1.5 on my X and Y axis.
You will definitely hear a difference in spreadcylcle mode; the steppers chirp like canaries on extasy.

Wow, thanks so much for doing this troubleshooting. I suspected that the drivers were in stealthchop, not spreadcycle. Hard to tell really, without having any experience of them beforehand. I got the feeling that the torque was way too low for the set current Vref. I did try asking the manufacturer for info on those pads, but no info was forthcoming.

I did notice that at 1.3V, the drivers did make some noise, not nearly as much as the A4988 or DRV8825, but definitely not silent. I guess the heated bed is coming off _again_. It just. never. ends.
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 04, 2017 07:23AM
Has anyone noticed resonance issues?

I switched from 8-bit Rumba board to 32bit RADDS, but kept the same TMC drivers and 0.9° steppers. I ran the steppers with belts off, because I wasn't sure about my CoreXY config.
The steppers made resonant noises at the same speed as before with the Rumba board. Initially I blamed it on the CoreXY frame, but they do the same without any load?!
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 04, 2017 07:32AM
Yep, mine definitely resonate at a certain speed. I've noticed it when going from the home position to the bed at the start of a print, when the speed multiplier is at 200%. Perhaps we can all note the speed and try to correlate this to steps/second?

Be interested to hear if the speed is the same for all.
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 04, 2017 09:36AM
At 200 steps/mm and 50mm/s I had the highest vibrations. Equates to 10k-steps/s. That was the 8bit CPU.
With 32bit I only tested it for a few minutes and vibes started from 50mm/s but they were "all over the place". Now using RAPS128 for X/Y and TMC for Z.

PS: I don't have RADDS schematics, so I'm unsure what the microstepping switches really do, when set to OFF. Don't want to blame TMCs until I dig deeper into this subject.

Edited 2 time(s). Last edit at 03/04/2017 09:41AM by o_lampe.
Reply Quote
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 04, 2017 09:39AM
Ive noticed this but am struggling to replicate it. I tried from 3000-6000 mm/min travel moves in 100mm/min increments and couldn't make it do it. Maybe its happening dynamically with x+y moves?

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 04, 2017 09:45AM
I'm wondering, how the interpolation actually works? Does it spread the sub-steps equally between incoming step signals or is it a fixed frequency, which more or less fits between the real steps?
Don't own a 'scope, but would like to see if above a certain step rate the interpolation lags behind.
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 04, 2017 10:02AM
Mine is still in stealthchop mode, and resonates at around 150 mm/s at 160 steps/mm. It resonates between around 140 and 160 mm/sec.

That's 9000 mm/min for those who are that way inclined spinning smiley sticking its tongue out

This was on a delta, moving the carriage in Z direction for 100 mm, using Repetier and RAMPS.

Just another data point.
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 05, 2017 02:29PM
The reason for the excessive vibrations with RADDS is the missing pull down resistor on CFG1. After wiring it to GND, the vibes were gone.
Reply Quote
DjDemonD
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 05, 2017 05:09PM
Interesting, would be why I cant replicate it my re-arm/ramps combo has the pulldown resistor.

The only odd resonance I am seeing is on first startup there's a definite crunch then they work smoothly.

Also I am back on 1.3v vref after I just got this object set at 1.2v.


Reprinting it now.

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 07, 2017 05:37AM
RADDS unleashed! eye popping smiley

After finding the reason for the vibrating steppers, I replaced the RAPS128 drivers on X & Y axis of my CoreXY with the CF1-modded TMC's
Not only can I print small circles much faster, but I could also crank up the acceleration from 500 to 2000mm/s²!
The printer almost jumps from the table now grinning smiley
And it is still silent in spreadcycle mode.

Replacing drivers: save ~10$ each ( TMC2100 vs RAPS128 )
Fat grin on my face: priceless
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 07, 2017 09:03AM
Quote
o_lampe
RADDS unleashed! eye popping smiley

After finding the reason for the vibrating steppers, I replaced the RAPS128 drivers on X & Y axis of my CoreXY with the CF1-modded TMC's
Not only can I print small circles much faster, but I could also crank up the acceleration from 500 to 2000mm/s²!
The printer almost jumps from the table now grinning smiley
And it is still silent in spreadcycle mode.

Replacing drivers: save ~10$ each ( TMC2100 vs RAPS128 )
Fat grin on my face: priceless

What is wrong with the RAPS128 drivers? I have a few of these, but haven't tried them yet:
[www.panucatt.com]
which I think uses the same driver chip (THB6128). Would you say that the TMC2100 is better? Why?

Can the RADDS not drive those chips fast enough? Have you considered 1/64 microstepping instead?

Also, can you share what the reason for the vibrating steppers was?

Sorry for the firehose of questions spinning smiley sticking its tongue out

Thanks in advance.
Reply Quote
o_lampe
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 07, 2017 11:45AM
The Panucatt driver uses the same chip but has different features. (i.e. idle current reduction vs. decay potmeter )

RADDS and RAPS128 work fluently together, but with 0.9° steppers and a ( theoretically ) fast CoreXY, 1600steps/mm is borderline.
The carrier started to stutter in curves above 70mm/s. I could have used 1/64 or 1.8° steppers, but with TMC I have best of both words: fast and quiet.

The reason for the vibrations was the CF1_to_GND bridge, that is necessary with RADDS ( no pulldown resistor like on Ramps or Rumba )

Edited 1 time(s). Last edit at 03/07/2017 11:51AM by o_lampe.
Reply Quote
nebbian
Re: TMC 2100 and ramps - any pitfalls to avoid?
March 08, 2017 10:13AM
After this happened:



I went ahead and soldered that bridge that olampe posted. It definitely changed something, and I'm pretty sure that this changes the drivers from stealthchop to spreadcycle. I kept the current at the same value (1.3V).
I know what you mean about canaries on ecstasy! It sounds like a very high pitched warbling, maybe 15 khz. Definitely not silent.
The behaviour was similar, I couldn't tell the difference in prints between the two. But it's louder in this mode.

I checked my board, and the pin labelled CFG1 is certainly grounded on my RAMPS board. So it really looks like these white TMC2100 drivers are mislabelled.

The problem with that print turned out to be a lump of plastic sitting on the motor drive gear teeth. No wonder it skipped.
Reply Quote
Page 2 of 4    Pages: Previous1234Next
Newer Topic Older Topic
Sorry, only registered users may post in this forum.

Click here to login