Pulse duration

Anybody know what the high/low step pulse duration is on a ramps 1.4 board?

No takers on this one huh? That tells me the information is probably out there somewhere and I've just got to find it. The reason I'm asking is because I'm going to use Pololu stepper drivers on my ramps (instead of the A4988's it came with) and the documentation says the A4988's can handle step pulse durations as low as 1 micro second but the drv8825's need at least 1.9 us. I'm wondering if pulse duration relates to the frequency that the Mega is able to generate the pulses…Seems right to me. Maybe I'll start there.

The code literally toggles the pulse on/off on sequential instructions, repetier and I think later marlin builds have an optional delay between the on and off toggles.
The 4988 just requires a rising edge, as far as I can tell it doesn't need any pulse length, I ran a test over the weekend with a 100MHz ARM board and a 4988, even immediately toggling the signal back off still works even when my scope only shows it hitting about 1.9V peak.

Hmmm. Have you tried the drv8825 at all? Thanks for replying by the way.

No I was going to pick one up for testing, but I haven't had a chance to do so yet.

If you don't mind my asking, what in your opinion is the weakest link in a system running marlin firmware, ramps 1.4 with drv8825 drivers on 36 volts and these kelinginc steppers wired bipolar series?

IMO the weakest link will be the components on the RAMPS boards that are not rated to 36V.

Specifically, the PTC fuses. The 11A one is rated to 16V, and the 5A one is rated to 30V. No replacements that support higher voltages at those currents exist, so you're best replacing them with real fuses.

Also check other components, like capacitors.

Yup. My caps are 35 volt, I replaced the fuses with inline ones at the + of the power supplies and I did the FET hack successfully like you said on another thread Cefair. Thanks for your help! I'm kinda disappointed with the performance of the steppers though. I read that bipolar will give more torque with less speed and unipolar faster rapids at less torque so I was wondering what the weakest link is there. I guess I want my cake and eat it too! I have to run prints at 10-11 mm/second to keep from losing position. And yes, I've tuned the drivers to the best of my ability. I can get the motors to go as fast as 1100 or 1200 mm/minute jogging in pronterface but they skip or stall (I don't know the difference) at speeds slower than that. Usually, if I set the jog speed higher than what the motors will do, they stall. Occasionally I can tune them to where they will go as fast as they can without stalling even when I set the feed rate ridiculously high. this fast-as-they-can phenomenon is usually limited to my X axis. Rarely am I able to get the Y axis to exhibit this behavior. I think because there is a lot more mass to it than the X. Seems to me like there is a sweet spot in the tuning but that sweet spot doesn't work throughout the speed curve.
Here is a picture of my machine assembled for the most part.

And here is just my Y axis unassembled.

There are many things that limit speeds.
It looks like you are using screws to drive the axis, what's the pitch?
There is a very real possibility you are exceeding the arduinos ability to produce a step train, in practice about the fastest it can reliably process interrupts is 16K/second, repetier firmware for example processes 2 steps per interrupt for a maximum speed of 32Ksteps/second.
If you are using 16x uStepping that's only about 10 revs /sec.
Most small steppers will run reliably at speeds upto about 1000RPM or about 17revs/sec but the torque drops of DRAMATICALLY with speed, so the faster you have to run the steppers the less torque you have.
If any of the axis have significant resistance or weight, you'll also have to drop accelleration significantly, most of the firmware defaults are way too high for most printers.

Polygonhell Wrote:
-------------------------------------------------------
> There are many things that limit speeds.

Quote

It looks like you are using screws to drive the
axis, what's the pitch?

I'm using a 2.75:1 belt pulley reduction AND 3/8-10 acme screws. Yeah. The steps per mm is set at ~2564 for the Y so that's roughly 43000 steps/second if I want to go 17 mm/second. Right? What ARM board are you working with? I've been eyeballing the smoothie board myself but that would be a big learning curve for me.

Quote

There is a very real possibility you are exceeding
the arduinos ability to produce a step train

Would this manifest as a high pitched whine in the motor with no movement?

Quote

repetier firmware for
example processes 2 steps per interrupt for a
maximum speed of 32Ksteps/second.
If you are using 16x uStepping that's only about
10 revs /sec.

I'm trying to stick with the 32X ustep for the smoothness. I realize that puts a great load on 'the weakest link' of the system and that's mainly why I asked what the weakest link was. So it's the firmware and the processor? The speed at which the code can go through the loops it needs to generate the pulses?

Quote

most of the firmware defaults are
way too high for most printers.

Tell me about it. Made my mouth water when I first saw them, then reality set in.

Quote

Would this manifest as a high pitched whine in the motor with no movement?

It might, basically if the interrupt rate is too high, you start missing interrupts, which reduces the speed and introduces a rather large amount of jitter into the step output, which could stall the motor. But you get the same thing if you run out of torque, run the motor too fast or your acceleration or jerk are set too high.

Quote

The steps per mm is set at ~2564
I'm trying to stick with the 32X ustep for the smoothness.

I wouldn't bother with 32x uStepping with a steps per that high, to give you some point of comparison one of my printers here is set to 53.3 steps/mm @ 8x uStepping, and there are a lot of other errors that are significant long before the steps/mm become an issue.

Quote

So it's the firmware and the processor?

Depends what you're asking it to do, with the 10TPI screws and the 2.75:1 reduction, about the fastest you can practically expect to move with steppers even with the best stepper drivers and controller is 1000rpm * 0.1 /2.75 inches/minute ~40 ipm which is 17mm/s assuming I understand what you mean by reduction.
The arduino is a 16MHz microcontroller, with an interrupt frequency of 16KHz, you get 1000 cycles to process the moves, it's pushing it. Repetier has a mode that will output 4 steps per interrupt, which would get you in the ballpark.

Quote

What ARM board are you working with?

I'm using an LPC1768 based board, with my own software and some breadboarded components at the moment, so it's probably not much use to you.
Smoothie is is based on the LPC1769, so it's pretty close to what I've run tests with, I haven't looked at the smoothie software stack.

mmm i did measure the step once on a scope, i cant rem exactly but i recall it has a weird duty cycle of over 90% ...

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

I suspect that part of your problem is that you're missing several microsteps due to the high voltage you're running and the rather long blanking time on the DRV8825 drivers. Something similar can happen with a stock A4988 as explained by Nophead here:
http://hydraraptor.blogspot.com/2012/04/stepstuck.html

The DRV8825 has a minimum on-time of 3.75μs with an effective off-time of around 30μs (It's fixed frequency, not fixed off-time). Running through Nophead's motor maths, with an on-time that long at 35V and running at full 1.7A peak, it looks like the minimum current the DRV8825 will output is 55% of peak. Which means if you're doing 1/32 micro stepping, the motor is jumping from step 1 to around step 12 and then sitting there until the commanded current catches up to the minimum it can output. The A4988 lets you pick a full-time mixed decay mode that makes it possible to hit lower currents, but that's not an option on the DRV8825. The best you could do is pull the DECAY pin high to permanently set it to fast decay, which would give you a huge current ripple, but would atleast get you closer to the correct current level.

Of what consequence is a 'huge current ripple'? Isn't there some kind of sacrifice when you give up slow decay? I could look it up but I want to reply quickly to keep you talking to me. :-)

HA ! ... i love the speak / talk about power ripples ...

i did some experiments (see here --> [3roomlabs.blogspot.sg]) ... remedy --> add more condensers.

there were some comments that the power rail noise does not affect the Rsense ability to feedback correctly. some say the A49xx have ability to float Vsense over the noise, i am really not so sure that A49xx are really that robust against a noisy supply. any ideas out there on this?

you can see here (http://3roomlabs.blogspot.sg/2013/04/044-stepsticks-concentration-camp.html), i directly attached the caps to the Vmot pins.

coincidently, i wanted to order the DRV8825 stepstick variant to try this measurement, then again, anybody wanna post additional pics of power rail ripples?

cheers

Edited 1 time(s). Last edit at 05/27/2013 06:03PM by redreprap.

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

wow your rig is solid metal ... i bet they are heavy ?


cncjoe Wrote:
-------------------------------------------------------
> Yup. My caps are 35 volt, I replaced the fuses
> with inline ones at the + of the power supplies
> and I did the FET hack successfully like you said
> on another thread Cefair. Thanks for your help!
> I'm kinda disappointed with the performance of the
> steppers though. I read that bipolar will give
> more torque with less speed and unipolar faster
> rapids at less torque so I was wondering what the
> weakest link is there. I guess I want my cake and
> eat it too! I have to run prints at 10-11
> mm/second to keep from losing position. And yes,
> I've tuned the drivers to the best of my ability.
> I can get the motors to go as fast as 1100 or 1200
> mm/minute jogging in pronterface but they skip or
> stall (I don't know the difference) at speeds
> slower than that. Usually, if I set the jog speed
> higher than what the motors will do, they stall.
> Occasionally I can tune them to where they will go
> as fast as they can without stalling even when I
> set the feed rate ridiculously high. this
> fast-as-they-can phenomenon is usually limited to
> my X axis. Rarely am I able to get the Y axis to
> exhibit this behavior. I think because there is a
> lot more mass to it than the X. Seems to me like
> there is a sweet spot in the tuning but that sweet
> spot doesn't work throughout the speed curve.
> Here is a picture of my machine assembled for the
> most part.
> [i1135.photobucket.com]
> /photobucket-39740-1353389544484.jpg
> And here is just my Y axis unassembled.
> [i1135.photobucket.com]
> /Repstrap/DSC02182.jpg

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

Quote

for newbies who do not know the exact science, this will present many hours of tinkering and pondering about the source of strange problems (like steppers going up in smoke? having a nice burn on the fingers?)

That's me! Thanks for the reply Red. Yes. All metal. I suppose if I were a carpenter it would be all wood! I've abandoned the Ramps board and taken on the headaches that come with the Azteeg X3 now. It seems like I can only just barely get these reprap electronics to work on this design so I'm printing mendel and rostock parts. It really sucks starting a long print only to find later that it shifted some time in the middle.

i find this entire 3d printing to be very enticing. coupled with my limited knowledge, i decided to go the long way of trying to ascertain the exact science and try to DIY the entire thing. so lets see if my patience will win, or my curiosity will force me to buy a kit (its kinda 50/50 now LOL !!!)

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

Limited knowledge? Your name is ALL OVER this forum...

cncjoe Wrote:
-------------------------------------------------------
> Limited knowledge? Your name is ALL OVER this
> forum...

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

So what do you make of this if you don't mind. I have THIS CONTROLLER with THESE DRIVERS trying to run THESE STEPPERS. I can't get them to run throughout the speed range I'm shooting for. Using the same controller, THESE DRIVERS, and THESE MOTORS I've had much better results.

mmm interesting, some questions

1) does the azteeg run the same "sketch" as arduino?
2) whats ur Vmot?
3) can you try running those in 1/1 step mode?
4) what are the Vref or drive current settings?

cheers

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

i just had another thought about testing the limits of your setup. if you can program a test "sketch", make the motors step at roughly 0.5s intervals (say 20 steps clockwise, and then 20 anti clockwise). try to measure the step intervals and see which is the most consistent setup by changing the following

1) stepping from 1/1 to 1/32
2) current adjustments (or Vref?)
3) if you think you can try mod the Rsense, not quite sure what is a good optimum for DRV8825. i changed my A4988 Rsense to 0.18ohm so it can drive about 1.2A with (theoretically) maxed Vref.
4) if you have a sensitive DMM, try measure the power supply to Vmot resistance. ramps does about 0.07ohms, not sure how well is Azteeg design.
5) i find that there is a slight difference in the torgue if you have a large reservoir cap sitting right on top of the stepstick Vmots. you could try this, but i did this mod only after noticing the power drain ripples on a scope, not sure if you could try and observe this.

i think by doing item 1, you could determine at which point the drive would start to go wrong. for the A4988 i even mod the R4 the drive frequency till it was very whiney LOL. so i guess at best, this will help give you an idea at what limit you can use the stepper -driver combo?

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

The drivers should be more or less identical, aside from the extra heatsinking on the SureStepr drivers. The difference would be with your steppers. I'd offer an opinion, but I don't know if you're running the keling steppers in series or parallel. Also, the specs listed for the QU-BD steppers don't make any sense. They say either 1.2A or 1.05A at 10.5V, with a resistance of 3 (ohms?). 3 ohms definitely wouldn't be right for either of those currents at 10.5V. If I knew what the actual phase resistance was for the QU-BD stepper, I'd have a better idea of what the difference between the two steppers was.

mmm i think u r right. 3ohm looks more like a 2A+ coil

i have a 0.38A stepper, i could run it at 30v. and a new 1A NEMA23 (20kgf), is also rec to run at 30V. somehow i love steppers with high coil resistances

KSevcik Wrote:
-------------------------------------------------------
> The drivers should be more or less identical,
> aside from the extra heatsinking on the SureStepr
> drivers. The difference would be with your
> steppers. I'd offer an opinion, but I don't know
> if you're running the keling steppers in series or
> parallel. Also, the specs listed for the QU-BD
> steppers don't make any sense. They say either
> 1.2A or 1.05A at 10.5V, with a resistance of 3
> (ohms?). 3 ohms definitely wouldn't be right for
> either of those currents at 10.5V. If I knew what
> the actual phase resistance was for the QU-BD
> stepper, I'd have a better idea of what the
> difference between the two steppers was.

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

I think you guys are awake when I am sleeping. Some awesome suggestions!
I'm using the Marlin firmware if that's what you meant Red. As far as Vmot, I've tried 12V, 24V, and 36V (36V only on the Ramps). I've found that I get the best results with the 24V though. I've tried full step, half step, eighth step, sixteenth step, and thirty-second step. I'll have to do them again and document the results. All I can remember is they work best at eighth or sixteenth stepping. When I did 32nd stepping I was limited to around 900 mm/min. I think I've got that one figured out though. Marlin has a max step frequency " #define MAX_STEP_FREQUENCY 40000 // Max step frequency for Ultimaker (5000 pps / half step)" so with pronterface, I was not able to get the machine to go faster. I might be wrong on that but that's my best guess.

As far as the Vref settings, I not sure I know exactly what I'm doing on that. The 22Kgcm stepper (they say they were custom wound by the way) supposedly are rated at 1.05A per phase so I set the Vref to .5 ohm. That was very wrong for some reason. Either I'm doing it wrong or there is some other consideration I'm not taking into account. I (usually) set the Vref by sound.
I turn the pot all the way counter clock wise (anti clock wise) and jog the axis 10mm starting at 100mm/min. If it stalls I wait for it to stop growling then turn the pot clockwise a very small amount and repeat the process until it runs quietly. I progress up in feedrate until it won't tune. Strangely, if I tune the motor for higher feedrates, it might stall at the lower ones later??? Anyway, I ended up at around .25 ohms on the Vref for those motors. I'm not sure I understand the equation exactly. Current Limit = VREF × 2
I'm running the Kelings bipolar-serial. Odd. The resistance on the qu-bd motors measures 9.2ohms???
The component mods you do Red are a little over my head and I don't have a sensitive DMM or an oscilloscope. I'll try the changes in micro stepping again and take better notes. I can post videos too if you guys would find it helpful? My channel

i have a permanent insomnia problem (its 3am here hahaha)

i have no idea how to equate stepping frquency to the parameters of a stepper/driver (perhaps nophead found a way to calculate this? hee hee hee)

when you say 0.5 ohm you mean 0.5volts yes?

i have a NEMA23 china cheapo stepper, it is rated for 1A 20kgf. it is a 6 wire, so when configured for 4 wire i series the adjacent coils and it gives 16ohms. unofficial spec wise it could go up to 30v++.

i think DRV8825 has its own Vref equation. if your stepper is 1.05a rated, i will try a limit of 1.4A, see its heat dissipation before turning it down. also every make of stepper have a torque/speed performance curve, the slower it spins the more torque you can get, the faster = reduction in torque.

what are the Rsense like on those DRV8825? 0.1ohm?

so @ 1.4A, ... 1.4A = Vref x 2 = 0.7v

so i think if you are going to get maximum speed, you either sacrifice number of u-steps or torque. or ... NEMA34 ! hahaha which is a very large stepper compared to stepsticks, i dont think it will fit the driver current somehow. or like what viktor (VDX) uses, 3 phase motor. i still dont understand how does a 3 phase stepper work with a 325V drive o.0"

btw does the DRV8825 whine like this ? [www.youtube.com]

Edited 1 time(s). Last edit at 05/28/2013 03:39PM by redreprap.

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro

Oops. Yes. .5V not .5ohms. My eyes don't even work at 3AM! You obviously have a better understanding of electronics than I do. There is a diagram here for the pololu drivers I use. It looks like there is a 10k resistor on the Vref??

Quote
Redreprap
see its heat dissipation before turning it down

It won't even turn at higher Vref. It just hums. Makes me think there is something else wrong but what? I'll video it to show. Get some sleep bud!
Thanks!

Quote

perhaps nophead found a way to calculate this?

[hydraraptor.blogspot.co.uk]

The stepper in the video is chopping in the audio range. Normally it is ultrasonic but the beat between the two coils can be audible when it is stationary. That is why it is so much louder than normal.

---

[www.hydraraptor.blogspot.com]

based on [www.pololu.com]

the Vref is suppose to be able to go up to 1.25v, so u r saying yours wont do that?

---

______________________________________
__my mixed bag blog || aka --> [http] || ___ so 3D printing is everywhere ... dont worry, hospitals can now 3Dprint body parts, they will charge you $1million excluding surgical fees ... you will die paying your debts. thats their aim ___ if every patent expires tomorrow, everybody will surely get a 3dprinter and make EVERYTHING ! ____ there is a "DIY-DTG" t shirt printing forum, you can mod an EPSON printer to PRINT like a pro. ___ CNCzone? overly commercialized it seems ___ my country? they will be taxing you for every cm of road you use and track you to your grave using GPS and its government authorized, now they will fire all the traffic wardens instead.___ EEVBLOG? there is only 1 way to do things --> take it apart like a pro