Interesting stepper driver + controller

[Jason von Nieda]

I ran across a really interesting stepper driver today completely by accident:

http://www.ebay.com/itm/ST-L6472-Based-SPI-Stepper-Driver-7-0A-Peak-Current-NEMA-17-Mounting-for-Arduino-/230891366885?pt=LH_DefaultDomain_0&hash=item35c232a9e5

And here is the chip on it:

http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/DM00047136.pdf

It's a complete stepper driver and motion controller all in one IC package. Supports up to 7 amps at 45 volts. The interesting thing about this is that it handles acceleration profiles and everything right on board. No Grbl or friends needed. I really like the form factor the seller has, where the board just bolts right into the back of the stepper. This is interesting for OpenPnP because while this chip is not great for coordinated motion (like in 3D printers and mills), we aren't worried about coordinated motion. All we care about is that the thing gets to the right place - doesn't really matter how it gets there. So this could be a really interesting option for a low cost motion controller + stepper driver combo board for an OpenPnP kit.

Jason

[Daniel Dumitru]

Hello JAson,

Indeed ST it's sending out to market interesting products.

Regarding this specific one :
- that 7A peak it's something like dust in eyes. In that setup (PCB), probably you can count on a 2A

- as you said needs SPI commands to move. Right now are no SW ready available to use. (it may be integrated pretty easy but right now doesn't exist)

Daniel

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To post to this group, send email to ope...@googlegroups.com.
To unsubscribe from this group, send email to openpnp+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Jason von Nieda]

Hi Daniel,

I agree, on that board 7A would be a dream. But I am more interested in the chip than the board.

While SPI is needed, it simplifies the amount of control circuitry you need, and greatly simplifies the software, i.e. Grbl. I still have not come across an open source motion controller that meets my needs for OpenPnP, although TinyG comes close and Smoothie (when released) will likely do the trick. These chips would eliminate the need for a complex, fast motion controller firmware and be simply a program to translate g-code (or simpler) to the SPI commands required for the chip.

In any case, this is just another option to add to the arsenal. For the time being I'll be using TinyG.

Jason

[Daniel]

It's fantastic that we have almost the same opinions.
I have started to wrote a module for Smoothie for rotating nozzle.

Daniel

[Danh Trinh]

I evaluated the voltage mode control version called the L6470 for a mobile robot application. The chips had great specs on paper but lab tests showed that it over heated at 3.0 amps peak (which is less current than the 3.0 amps RMS in the specs). The ST applications Engineer confirmed that the 3.0 amps RMS rating is an "absolute" spec but the chips can not do 3.0 amps RMS continuously due to thermals. 

I spent a few hundred bucks on 2 ST eval boards and frankly felt duped by these "paper" specs.

Here's the correspondence to the app Engineer:

https://my.st.com/public/STe2ecommunities/mcu/Lists/motor_control/Flat.aspx?RootFolder=%2Fpublic%2FSTe2ecommunities%2Fmcu%2FLists%2Fmotor_control%2FEVAL6470H%20Application&FolderCTID=0x01200200770978C69A1141439FE559EB459D7580005A810128153A7244A21980BD49736843&currentviews=142

Scroll down to the last reply where the Engineer basically says that at 3.0 amps RMS the chip package would theoretically heat up to 279 Celsius but the chip is rated for only 160 C

[Miroslav Pejic]

Hi Jason, Daniel, Danh,
I'm constantly monitoring this group an I realised there is permanent need for more current for stepper motors.
Since I'm an old fashion electronics engineer, maybe my idea will be strange. Anyway, here it is:
Can we use one  stepper motor controller (mentioned here) and simply add external H-bridge module with more amps and better possibility for cooling? Does anyone tested this combination?
Please, any opinion, or idea, or module recommendation.
Regards,
Miroslav

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To post to this group, send email to ope...@googlegroups.com.
To unsubscribe from this group, send email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msg/openpnp/-/Fr-qIWLZbcsJ.

[Richard Spelling]

Actually, if I understand it correctly, in stepper motors you never
reverse the current, so just a power fet for each of the coils should
work theoretically.

The problem is that most of the chips use current sensing to be able to
run the steppers at higher voltages than they are speced for, which
would break unless you use some kind of current mirror/amp circuit.

I like the Pololu stepper modules myself. With enough heat sinking you
can run them at near 2A, and they are relatively inexpensive.

Zippy Mark II will use them on a RAMPS board on top of an arduino Mega.

Laser cutter is using them without problems.

Other end of the scale is the Gecko drivers. And I'm sure there are others.

I find though that if I need more than 2A to run the motors, I have done
something wrong. I can see on a (steel) milling machine, but a pick and
place machine should never have any resistance to motion.

Unless you are trying to go super fast.

Anyway.

> http://www.ebay.com/itm/ST-__L6472-Based-SPI-Stepper-__Driver-7-0A-Peak-Current-NEMA-__17-Mounting-for-Arduino-/__230891366885?pt=LH___DefaultDomain_0&hash=__item35c232a9e5

> <http://www.ebay.com/itm/ST-L6472-Based-SPI-Stepper-Driver-7-0A-Peak-Current-NEMA-17-Mounting-for-Arduino-/230891366885?pt=LH_DefaultDomain_0&hash=item35c232a9e5>
> And here is the chip on it:
>

> http://www.st.com/internet/__com/TECHNICAL_RESOURCES/__TECHNICAL_LITERATURE/__DATASHEET/DM00047136.pdf

> <http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/DM00047136.pdf>
>
> It's a complete stepper driver and motion controller all in one
> IC package. Supports up to 7 amps at 45 volts. The interesting
> thing about this is that it handles acceleration profiles and
> everything right on board. No Grbl or friends needed. I really
> like the form factor the seller has, where the board just bolts
> right into the back of the stepper. This is interesting for
> OpenPnP because while this chip is not great for coordinated
> motion (like in 3D printers and mills), we aren't worried about
> coordinated motion. All we care about is that the thing gets to
> the right place - doesn't really matter how it gets there. So
> this could be a really interesting option for a low cost motion
> controller + stepper driver combo board for an OpenPnP kit.
>
> Jason
>
> --
> You received this message because you are subscribed to the Google
> Groups "OpenPnP" group.
> To post to this group, send email to ope...@googlegroups.com

> <mailto:ope...@googlegroups.com>.

> To unsubscribe from this group, send email to
> openpnp+u...@googlegroups.com

> <mailto:openpnp%2Bunsu...@googlegroups.com>.

> To view this discussion on the web visit
> https://groups.google.com/d/msg/openpnp/-/Fr-qIWLZbcsJ.
>
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "OpenPnP" group.
> To post to this group, send email to ope...@googlegroups.com.
> To unsubscribe from this group, send email to
> openpnp+u...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
Visit my online store for solar electronics: http://www.richardsfoundry.com
------------------------------------------------------------------------------
NOT sent from a Blackberry. Sent from a free hand-me-down PC running
free open source Ubuntu Linux... :-P

[Karl Lew]

If you bolt it to the stepper, wont the board get hot?

[Jason von Nieda]

Thanks for posting that Danh, very interesting! I suppose if you intend to use it at it's maximum ratings you have to add water cooling :)

The L6472 is supposed to be rated for 7A, although with such a small package I imagine it's limits will be much the same as you discovered.

With that said, I've found so far that 3A is about the upper limit of what I've needed. Even on my metal milling machine I only use 3.2A motors. For the PnP I think we'll need much less. This is something I want to do some real testing on soon.

As Richard says later in the thread, if you want real power, go with Gecko. They can't be beat. I am more interested in these chips for their built in motion controller.

Jason

[Danh Trinh]

"Can we use one  stepper motor controller (mentioned here) and simply add external H-bridge module with more amps and better possibility for cooling? Does anyone tested this combination?"

If I understand you right I think that is what most folks have tried here - using an Arduino as a controller to generate steps/dir signals for stepper motor drivers. Some solutions integrate the processor and driver chips onto one board.

[Danh Trinh]

" I am more interested in these chips for their built in motion controller."

There was talk in the 3D printing forums about using these SPI stepper drivers. The motion controller does work well since it takes care of accelerations/decelerations. Biggest challenge is it's not clear how to synchronize all the driver chips through SPI since they each have a separate internal clock that is generating the stepper pulses. With the conventional solution, the main controller is responsible for generating the driver pulses so synchronization is much easier. 

[Jason von Nieda]

Yep, but luckily we don't need synchronization with PnP. Just need to know when the command has completed.

Jason

To view this discussion on the web visit https://groups.google.com/d/msg/openpnp/-/2QIOYqVZOGQJ.

[utopcell]

This is a very interesting controller Jason, but I see that the maximum speed it can support is 15610 steps / sec.

Isn't this a bit on the low side ?

-- Stergios

[Jason von Nieda]

Ah, shucks. I completely missed that. Nice catch! Back to the drawing board :)

Jason

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To post to this group, send email to ope...@googlegroups.com.
To unsubscribe from this group, send email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msg/openpnp/-/06l73sV9EUIJ.

[Danh Trinh]

"Yep, but luckily we don't need synchronization with PnP. Just need to know when the command has completed."

Good point.

[Riley Porter]

Yah I too saw this board and was like NO WAY!

Our TinyG driver does up to 2.5AMPs and we have quite a bit of copper!

I plan on playing more with open pnp just not have enough time yet!

ril3y

On Thu, Jan 10, 2013 at 5:27 PM, Danh Trinh <dttw...@gmail.com> wrote:

"Yep, but luckily we don't need synchronization with PnP. Just need to know when the command has completed."

Good point.

On Thursday, January 10, 2013 5:02:08 PM UTC-5, Jason von Nieda wrote:

Yep, but luckily we don't need synchronization with PnP. Just need to know when the command has completed.

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To post to this group, send email to ope...@googlegroups.com.
To unsubscribe from this group, send email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msg/openpnp/-/cwN9xSQ3GiUJ.

[varun s]

Guys I personally feel this stepper mounted driver cum control board is quite good but wat if if u want to debug or rectify a driver wat if it fails due to some problem then you have to turn the whole machine to reach the driver controller. I would prefer a main control board with onboard drivers so that I can rectify it or debug or hack when I need if in one place it is good but what Jason found is quite interesting a SPI based is good but not the stepper mounted idea and not for my pnp...  

Just my thought guys.. :-)

[Vinay Dand]

Hello,

Have a look at Stepper Controllers / drivers at

http://www.trinamic.com/index.php?option=com_content&view=article&id=330&Itemid=285

as an alternative to ST ICs. Trinamic ICs perform well in our field operations.

Vinay

[Karl Lew]

I am going with a rather dull choice of $14.95 :
http://www.amazon.com/SparkFun-EasyDriver-Stepper-Motor-Driver/dp/B004G4XR60/ref=pd_sim_sbs_misc_1
The EasyDriver design proven and is OSHW. It could also be combined with an Arduino clone on a custom board to drive one axis.
Will prototype with $16 Teensy Arduino to do accel profile, limit test, flash pretty lights, etc.. on the axis.
Computer would only have to tell each axis to "wait for D millis, then be at position X at time T". Then, "all together now, make it so!"

Not an exciting design, but it is cheap, extensible and easy.

[Jason von Nieda]

Karl,

You may want to look at Grbl before you get into writing your own software. We actually have a modified Grbl that gets us most of the way there.

Doing the work of calculating acceleration and accurate step timing for 4 axes on an Arduino is actually a fairly tricky problem and the people working on Grbl have put a lot of time into making it work.

You can find the OpenPnP specific version at https://github.com/openpnp/grbl

Jason

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To post to this group, send email to ope...@googlegroups.com.
To unsubscribe from this group, send email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msg/openpnp/-/RHQ34mRTSP4J.

[Karl Lew]

Cool! Thanks Jason. Will do.

My last machine drove three axes with one Arduino. Calculating the trajectories was the fun part and I did it on Windows using parametric quintic curves as described by Prof Rida Farouki in his book on Pythagorean Hodographic Curves. The computer would send the Arduino a single 3d trajectory of up to 100 points and the Arduino would drive all three steppers in sync through that trajectory. But that is another story.

I am fascinated by your earlier comment about PnP not requiring coordinated motion. That is very insightful. To me it implies that the axes can be driven separately as long as they arrive at feeder and pcb at the same time. In other words, the accel prrofiles dont need to be coordinated, which simplifies the math a lot.

That said, I didnt know about Grbl and will read up on it.

Thanks,
Karl

[Karl Lew]

Jason,

Grbl is really interesting. It takes g code subset and the arduino orchestrates the steppers with individual accel profiles. This is good cool stuff!

I might need multiple grblduinos. My PnP will have 5-9 steppers. On my last CNC machine I did not use gcode at all but my Arduino coding was analogous to grbl. In particular, I noticed that my Arduino Mega had its hands full orchestrating three steppers. Based on that experience I am thinking that perhaps a single grblduino might at best only support at most 4 steppers well?

If that is indeed the case, then I believe I will need to use 3 or 4 grblduinos more or less like this:

1 grblduino for each yz axis fixed gantry with 3steppers. I.e. no x-axis gcodes
1 grblduino for component tray feeders with 1-2steppers. i.e. only x axis gcodes
1 grblduino for pcb tray feeder with 1stepper. i.e., only x axis gcodes.

Will I be able to help make openpnp support the above configuration?
I can happily write software with your guidance.

Karl

[Riley Porter]

We wanted a 4axis grbl board too.  This is what we came up with.

http://www.flickr.com/photos/rileyporter/7993079751/in/photostream

4 axis grbl + drivers + arduino in one.  It uses the DRV8818's will do 2.5AMPS.  MUCH MUCH copper.  But tinyg supports up to 6 motors.. 4 onboard and 2 off.

Riley

Karl

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To post to this group, send email to ope...@googlegroups.com.
To unsubscribe from this group, send email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msg/openpnp/-/bgNSzXgWgQ0J.

[varun s]

Hi Riley,

Nice board look's awesome but is it open source or your custom development...

[Riley Porter]

That board is not release yet.  In fact the only one we have we gave to Edward Ford at the shapeoko project.  He tested it for us.  He says it runs 24x7!  So he likes it a lot.  Anyhow,  all of our boards are CC-NC.  Once released..

Riley

[Jason von Nieda]

Hi Riley,

Is your grblBoard intended to replace TinyG, or will you still continue developing TinyG?

Jason