Batch buy of L6470 Stepper Driver Breakouts

[Charles Yarnold]

Hello all,

One of my upcoming projects is making my own pick and place and on my list of things to build for it are stepper motor drivers.

After asking around I have settled on the L6470 stepper driver, it seems an awesome chip. It works over spi and all you need to do is tell it where to go, it handles driving the stepper for the correct number of steps, acceleration, deceleration, correcting for missed steps, stall detection, homing switch control and a whole lot more.

You can see more here: https://www.sparkfun.com/products/10859

The snag is I need 10 of them, and at ~£25 a pop its not cheap, so I'm going to make my own breakouts.

Its looking like I can get the price down to ~£11 each driver, with boards containing 3 and 4 drivers on each (3 drivers on a board being ~£33, 4 being ~£44).

Would anyone else be interested in buying some of these if I was to make a surplice, how many would you be interested in? If we get interest I will get a fixed cost and pop up a wiki page for orders.

Hugs,

Sol

[Matte]

I would definitely be interested in this :D


www.matthearn.co.uk

[Simon Howes]

Count me in  for  2 :)

[Adrian Godwin]

Has anyone here used the chinese stepper motor controllers based on
the Allegro 6560 ?

eg

http://www.ebay.co.uk/itm/280953028748

I appreciate they're nothing like the ones Sol's looking at, I'm just
curious what the quality is like.

-adrian

[Simon Howes]

Talking of which - anyone know a cheap source of ~1amp steppers? I'm building an slr stop motion camera dolly and want to move/pan the camera (and move on wheels) but dont have the battery capacity for bigger nema 23 steppers... (but still need the holding torque to keep a heavy slr with lens from tilting back down)

[Toby Catlin]

I would like to raise my hand. I would like 1 with 3 drivers.

toby

[IrradiatedHaggis]

Hey,

These are awesome. I would be interested in 2 four-driver boards... Or 8 single driver boards. These are great drivers for a 3d printer, CNC mill, etc... I have a couple of projects in mind... If you had surplus I'd consider buying more in the future even.

Cheers,

Troy 




Sent from Samsung tablet

[James C]

Hey,

I would be interested in one 3 and one 4 driver board.

Cheers

[Andy "Bob" Brockhurst]

On 26/09/2012 00:37, Charles Yarnold wrote:
<SNIP>

>
> Its looking like I can get the price down to ~�11 each driver, with
> boards containing 3 and 4 drivers on each (3 drivers on a board being
> ~�33, 4 being ~�44).

<SNIP>
> Hugs,
>
> Sol
Yes I would be interested.

Probably one of each 3 and 4 driver versions.

Bob
--
a.k.a b3cft/Andy

[Nick Johnson]

I'm interested, though I'd find 2 axis boards more useful. Is there substantial shared overhead that makes it simpler or cheaper to make multi-axis boards rather than smaller single-axis ones?

[Charles Yarnold]

Awesome, lots of interest! Will finalise the design and costing this week and post when the sign up sheet is on the wiki.

Nick: There is a small saving being able to use the same board for differing amounts of drivers. (leaving non used slots unpopulated and jumpered) But its mostly personal preference to have them all on one board. 

[Simon Howes]

Hello Charles!

Question: what form would this come in?  Would soldering be required or would these  be lovely chairman Charles greatleapforward prefab pcbs? :)

How would you foresee current  limit being set? Jumpers? Resistor? Turnpot? Or is it all (including microstepping) controlled via software?

(Also, i want 2x4 channel pls)

[Charles Yarnold]

Hia,

So I'm planning to ship fully soldered and tested boards.

Everything is set via the spi link "All commands and data registers, including those used to set analogue values (i.e. current control value, current protection trip point, deadtime, PWM frequency, etc.) are sent through a standard 5-Mbit/s SPI." Its quite a nice chip.

http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00255075.pdf

I also forgot to mention, the biggest reason to put multiple drivers on one board for me is that they can be chained for up to 8 steppers, meaning that all of this is integrated in one board rather than jumping control wires between them.

Sol

[Toby Catlin]

Any idea on time scales? I don't want to be pushy or anything but i am thinking of ordering one of these: http://makerslideeurope.com/eshapeoko-mechanical-kit.html and it would arrive until mid october. Your driver board would be perfect, but I couldn't sit around too long with a stationary mill.

slight tangent but anyone have any opinions on shapeoko? It seems like a good kit to me but i have never made anything like this.

thanks

t

[Robert]

I'd love to get a 4 axis board.

I tried to make up a 3 axis board but I got distracted so I only came up with the eagle files up to a point.

https://github.com/pinski1/dSMD

By the way I tried making up a 3 axis carrier, 

[Dean Forbes]

Hi Sol I would be interested in seeing where this goes please keep me in the loop Dean Forbes

 
On Wednesday, 26 September 2012 00:37:35 UTC+1, Solexious wrote:

[Charles Yarnold]

Revising my design and looking at costing, I'm now thinking of making it individual boards with chaining connectors, any thoughts, objections?

Sol

On 26 September 2012 17:59, Charles Yarnold <charles...@gmail.com> wrote:

[Nick Johnson]

Very much in favor of this.

-Nick

[Kurtnelle]

How are you going to solder the thermal vias?

[Charles Yarnold]

So, my initial design is up, it needs some tweaking and silkscreen fixing. I would welcome any comments: https://github.com/solexious/L6470Breakout

As chaining individual boards with a connector shouldn't effect anyone who wants multiple boards I will be going in this direction unless I get any strong objections. 

Nick Johnson: Brill

Kurtnelle: ah, I recognise your photo from a certain comment section of sparkfun, hello! As its broken out already by being dual with the busy pin I was going to leave it up to people to chain that if wanted as its not needed to be chained. What were you thinking? These boards will be solder pasted and popped in my reflow oven so should have no problems.

Sol

[Adrian Godwin]

ST don't seem to give much indication of heating in this chip and it
has quite a high thermal resistance (22 deg C/W to their sample
board). If you need to heatsink it, that layout will make it quite
difficult, because there are components all around the device.
Apparently some users have found that the evaluation board isn't much
good above 1A for that reason.

Maybe you could move the capacitor and connectors so that, if
necessary, a bar could be bolted horizontally across the chip (and
continue across several boards) ? Mounting holes in line with the chip
would help with that, too.

-adrian

On Fri, Sep 28, 2012 at 6:44 PM, Charles Yarnold

[Robert]

I really think you'd be better served putting 3 on a board rather than making them separate.

Will you be using 1oz or 2 oz copper? Also will you be getting the chips as Farnell don't sell them?

Have you taken a look at the ST demonstration boards, they seem to do what you're currently proposing so it may be work "borrowing" a few ideas off of them.

On Wednesday, 26 September 2012 00:37:35 UTC+1, Solexious wrote:

[Charles Yarnold]

On 29 September 2012 11:27, Robert <pin...@gmail.com> wrote:

I really think you'd be better served putting 3 on a board rather than making them separate.

Why? 

Will you be using 1oz or 2 oz copper? Also will you be getting the chips as Farnell don't sell them?

Yes I will be getting the chips. 

Have you taken a look at the ST demonstration boards, they seem to do what you're currently proposing so it may be work "borrowing" a few ideas off of them.

Yes, Ive seen those boards and their schematic. As Adrian has pointed out there are big flaws in how they are laid out.

Best,

Sol

[Robert]

Because most people use steppers on 3 axes machines, so having 3 on a single board (daisy chained SPI or no) simplifies some of the connections for power as well as the mounting of the board rather than boards. Of course you do have the problem of if you need 4 you need to make do with 3 or 6, but possibly a single chip and a triple chip boards? Either way it'll all be pretty good.

[Charles Yarnold]

I agree if your looking for a fixed number of boards then having them on one pcb is marginally more convenient than plugging in connectors between separate ones, but would it be a show stopper for you? :)

Sol

[Daveb]

I have been confused by the 'daisy chain' feature in this chip. The datasheet doesn't say much about it,  just a few references to how you can save controller pins.

I found this datasheet from maxim about daisy chaining in their chips - http://www.maximintegrated.com/app-notes/index.mvp/id/3947

This scheme has the command executed by the device that holds it in the shift register on the rising edge of the CS line.

I guess that the chip selected gets to drive the flag lines - busy, sync 

am I on track for this?

[Adrian Godwin]

A common way to handle multiple SPI chips is to provide each with a
chip-select and send / receive data to each one, enabling the chip
selects in turn. But if the chip actions the SPI command only at the
rising edge, it's possible to connect the chips in a daisy chain and
pass all the data with a single long SPI transfer that leaves the
appropriate contents in each register when CS rises.

This is inefficient if you need to send data at odd times to different
chips, but if you're updating them all together anyway (as you usually
would in a multiple axis controller) it saves a lot of chip select
lines, and is particularly convenient when all chips are of the same
type.

Flag outputs are generally always active and would need to go
separately to the controller (I haven't checked this on the 6470) but
any data on the SPI bus would also need to be shifted back through the
daisychain to be read. Ideally, each new transfer shifts back the
results of the previous one, but that's not always appropriate.

-adrian

[Robert]

Definitely not a show stopper!

I'm not too sure about daisy chaining either as there are some 1 byte commands but also some 4 byte commands. How do I send a 4 byte commands to one chip but a one byte to another? MSB padding or LSB padding? I think I was gonna make my own daisy chained board and experiment.

[Charles Yarnold]

Hi Toby,

Sorry I seemed to have missed your email in amongst the others.

I'm not too sure on time scales as of yet, with China in its holiday period I would put a rough estimate on 3-4 weeks.

Best,

Sol

[Charles Yarnold]

Very good points, I had just thought of using heat-sinks like these https://tindie.com/ellisgl/raspberry-pi-copper-heat-sink-kit/ but will have a look at a bit of a shuffle of the cap to accommodate your suggestion too.

Sol

[Charles Yarnold]

The honest answer is I'm not sure, I was planning on figuring out the software once they arrived.

On the note of chaining boards I will be adding option jumpers to the board so you can choose to have the BUSY lines etc either linked or taken from each chip individually.

With this and heat changes I'm getting close to a final prototype.

Sol

[Charles Yarnold]

So, the final prototype design is ready and off with seeedstudio for PCB manufacture. You can see it here: https://github.com/solexious/L6470Breakout

The main change from my original plan for this board is its customasable-ness. I have added jumpers so that the board can be used stand alone, or as a chained board. With chaining you can select what gets shared between boards, either with solder jumpers, or with pins and jumpers for a easier to change configuration.

Once I have a working unit I will have a page up for orders. I'm still confident I will meet my target cost, but due to the Chinese holidays I'm awaiting on final conformations.

Thanks to everyone for your input and support!

Sol 

[asc]

Hello Sol,

I think I'd be interested in either 2-3.

I'm still a bit confused (being electronically incompetent and
struggling to follow the thread fully) about how many stepper motors
that you plan to run off each board, and whether the (roughly) £11
cost is per pcb or based upon the number of drivers per pcb - can you
confirm this?

-asc

[Charles Yarnold]

Hia,

The plan has changed somewhat from my first email, to clarify:

Each board has one driver that can work one stepper motor, this will cost roughly £11.

These boards have link cables so you can easily chain upto 8 of the boards together to control upto 8 steppers from minimal IO pins. (they can also be used as stand alone units)

Hope that clears things up!

Sol

[TonyD]

Hi

I'm a bit of a late coming to this group buy.

Is there still time to get an order in, or have I missed the boat?

Also any plans just to sell the blank PCB just by itself?

thanks
Tony

On Oct 4, 8:25 pm, Charles Yarnold <charlesyarn...@gmail.com> wrote:
> Hia,
>
> The plan has changed somewhat from my first email, to clarify:
>
> Each board has one driver that can work one stepper motor, this will cost
> roughly £11.
>
> These boards have link cables so you can easily chain upto 8 of the boards
> together to control upto 8 steppers from minimal IO pins. (they can also be
> used as stand alone units)
>
> Hope that clears things up!
>
> Sol
>

[Charles Yarnold]

Hi Tony,

I'm currently waiting on the post to deliver my prototype before I put in the large order. Will post to the list when that happens.

No plans on just selling the PCB at the moment.

Sol

[TonyD]

OK, thanks for the update.

Tony

On 26 Oct, 23:07, Charles Yarnold <charlesyarn...@gmail.com> wrote:
> Hi Tony,
>
> I'm currently waiting on the post to deliver my prototype before I put in
> the large order. Will post to the list when that happens.
>
> No plans on just selling the PCB at the moment.
>
> Sol
>

[Charles Yarnold]

Hello all,

So I have the first 2 prototypes made up and working, you can check out the design here: http://www.youtube.com/watch?v=aAK0OnfMe1g

I'm working on test code for multi-driver support at the moment to do the final tests of my design before sending off the initial production run of 100.

I welcome suggestions or comments on the design!

Sol

[Jim Hayes]

Why the plastic grass in the final test?

[Charles Yarnold]

Its my automated creeper from crazy golf :)

[Tim Storey]

Well done, that's great.

\t
On 12 Nov 2012, at 21:11, Charles Yarnold <charles...@gmail.com> wrote:

> http://www.youtube.com/watch?v=aAK0OnfMe1g

[Charles Yarnold]

Hi all,

So, I'm finally all done and ready with the final version of the board. The last tests were to check the chaining design was fully functional: http://www.youtube.com/watch?v=KOVeZNFIaik

Taking onboard advice from Adrian, the final version is 10mm longer in one direction, this gives me room to move around components and allow for a bar to be connected along multiple driver boards. This also gives me enough room for an extra screw terminal for power chaining.

Thanks to everyone for their help, suggestion and encouragement both on the list and at the hackspace. Now I need your help to get these boards made. I will be posting a new thread on the list with purchase details for those new to this topic (or have muted it due to noise ;)

Thanks again,

Sol