Welcome! Log In Create A New Profile

Advanced

Build thread, 100 pin STM Arm Cortex board with driver chips.

Posted by grael 
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
October 01, 2009 09:34AM
ARM9 is a bit overkill for most peripherals, they are typically used in embedded display systems and some small netbooks

Any specific reason for 2 USB?

If it was one usb port you may want to look at something like an LPC21xx (ARM7-TDMI core)

Or newly released LPC17xx series with the Cortex M3 core.

No I don't work for Phillips, but they tend to be the cheaper option.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
October 19, 2009 12:59AM
Not sure why he wants 2 x USB, unless it's for USB experimental purposes ?

I used the ST chip because it has a lot of PWM, and the pins are very flexible in how they can be used.

I'm very pleased about the 5 x RC servo output I designed in, because it also give me an option of high current drive axis or tool head rotary power via brushless motor and brushless speed controller if required. Some of them will run close to 10 Kilo Watts, given the right setup. Many of the larger ones come with optical isolation to the RC receiver as well. That shunts the R&D budget over to the existing RC hobby industry developers. I'm enjoying the low prices on those already for some of my other projects.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
October 19, 2009 06:50AM
2 USB could also be implemented using a USB to TTL and one of the serial ports.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
November 05, 2009 12:32AM
The weather here is heating up, and I'm back working on my SW again for the board. [grael-minifactory.blogspot.com]
The 5 high current outputs work much better with the correct MOSFETS !

Running radio control brushless speed controllers from the servo outputs expands the system to an eyebrow raising extent, the R&D that has already been put into BESC design, and the low cost of Kong Kong sourced RC gear makes motor drive cheap, and practical. Not only that, but brushless motors are more reliable and available with much more useful output characteristics.

Hope to move my PCB and project server down to the garage soon.
Hope everybody else's projects are going well too!smileys with beer
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
November 10, 2009 04:30AM
I have two of these brushless motors and contollers with the idea of using them as direct drive milling motors as they are very light.
I have since noticed that a lot of people have a seperate belt drive motor and custom made spindles with bearing blocks some even using roller bearings.

So Im not so sure about the sutibility of the bearings on these airplane / boat motors


Bodge It [reprap.org]
=======================================

BIQ Sanguinololu SD LCD board BIQ Stepcon BIQ Opto Endstop
BIQ Heater Block PCB BIQ Extruder Peek clamp replacement BIQ Huxley Seedling
BIQ Sanguinololu mounting BIQ standalone Sanguinololu or Ramps mounting Print It Stick It Cut it


My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
November 12, 2009 05:23AM
I've been getting my ESCs and brushless motors from www.hobbyking.com in hong kong, expecting two orders to arrive soon even.

I can see why people would be doing a belt drive, to take some strain off the motor's bearings itself, and to put something heavy duty at the milling head.
I've got one of these on order:
[www.hobbyking.com]#
and and ESC:
[www.hobbyking.com]
but I'm not sure yet if it will go on my machine, or one of my RC jetboats:
[www.youtube.com]
The outrunners can give quite a large torque reaction accelerating, hence I'm trying the above inrunner.

I haven't tried it yet, but I'm pretty sure a brushless RC ESC will drive a fisher and paykel smart drive washing machine motor. NZ company, so we can get the motors from expired machines... a rotor diameter of about 600mm ! For a large machine, that would be ideal... to drive a massive screw thread or toothed belt.


My CNC/extruder creation: [grael-minifactory.blogspot.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
November 12, 2009 05:46AM
Guys

Interesting development and work.

I have to agree Kicad is good I use it quite a bit. My only winge with kicad is it really insist you start from schematic and go all the way through the design cycle.

Some times I just want to do a simple board with pads and tracks as a quick one off and jump straight in at the board end.

On the Cortex particularly if you are going the C route (IMHO the best programming language, as you can still inline assembler to get speed/size optimisations) have you considered giving your self a massive boost by running up uclinux and then having access to all the libraries that this implies.

There are images etc ready to go for this processor that could be adapted.

If you later intended adding in Ethernet the stack and protocols would be readily available for you.

I guess you would also be able to borrow at whim from the EMC source code too. To mention a few possibilities.

Thoughts for what they are worth.

aka47

PS Nophead is right do your own thing, beg forgiveness before permission.


Necessity hopefully becomes the absentee parent of successfully invented children.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
November 12, 2009 05:59AM
thanks AKA47,
I'm getting more used to C, but still caught out by a few things. right now, I have a bug that I think will be to do with a resister/s that I haven't declared as volatile.... I suspect it's an ST library routine that I made work out of context, but it will take some testing to establish.

Otherwise, it's going well again, the power of the ARM processor is making a lot of tasks quite trivial.

There is some STM starter code for ethernet I think, but this version of my board doesn't use that sub family of processor.


My CNC/extruder creation: [grael-minifactory.blogspot.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 07:19AM
After two years or more of following thee forums I have just caught up.

@ Grael..

Now I thought my 15A drive brushless motor was High current..

The size of the wire required for 70-80A would be huge.
It seems these are a little like music power v Watts RMS specifications.


Bodge It [reprap.org]
=======================================

BIQ Sanguinololu SD LCD board BIQ Stepcon BIQ Opto Endstop
BIQ Heater Block PCB BIQ Extruder Peek clamp replacement BIQ Huxley Seedling
BIQ Sanguinololu mounting BIQ standalone Sanguinololu or Ramps mounting Print It Stick It Cut it


My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 07:39AM
I don't know if you guys have seen this. It was a quick note I put up on the builders blog. It is very much pertinent to this thread though

The Aardvark IDE, it is based on Arduino and uses the same/similar code and libraries (some are extended) but is actualy targeting ARM processors. Given the compatability with the Arduino & code together with the ARM capability it shoul dmake for an easier porting route using existing code.

[antipastohw.blogspot.com]

I guess it goes without saying, that having something working quite quickly frees you up to tweak the bits that would benefit from being specifically tweaked or added on to what exists.

Phillips (trading under a different name) look to have a bunch of ARM micro controllers out which have some degree of ethernet capability in built. The guys at the link below have been using the Aardvark/Arduino IDE to program them. Linking them together in multiprocessor configs using the ethernet comms that is inbuilt to achieve this.

[www.arduino.cc]

I am in no way sugesting that we would benefit fromt he multiprocessor model at this time. Only that it is an illustration of using the Aardvark/Arduino IDE to program ARM based micro-controlers using a bunch of Arduino style code.

Interestign things and directions.

Cheers

aka47


Necessity hopefully becomes the absentee parent of successfully invented children.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 07:59AM
@ Grael..

Now I thought my 15A drive brush less motor was High current..

The size of the wire required for 70-80A would be significant.
It seems these ratings are a little like music power v Watts RMS specifications.

I have two ECS modules and two brush less motors as yet untested.

@ AK47

Kicad V Eagle is the problem ~ with most of the Reprap Boards produced in Eagle and plenty of the PCB MFG accepting Eagle files.

With all PCB cad packages I have seen so far.
The biggest pain has been finding the correct library files for the device pin outs without having to go and create your own library.

Is there a recommended set of Reprap Library required for Kicad or Eagle packages.?

I have downloaded both Kicad & Eagle My last hands on PCB design were in black tape @ 10x scale and Red and Blue tape at 2x scale a little while ago.

After that I just passed my scraps of paper to my good friends in the drawing office who made my final schematics. I then made rough drawings of where I wanted the components to go with notes of where I wanted short tracks thick tracks and Gnd planes. My last production PCB was a 2mb 128bit wide ACRTC graphics card for touch screens connected to PDP 11 in 1986-87 ish.

Since I was so spoilt for so long I've not really played much with PCB CAD packages,.. well enough to be irritated by having to create my own library components almost every time.


Bodge It [reprap.org]
=======================================

BIQ Sanguinololu SD LCD board BIQ Stepcon BIQ Opto Endstop
BIQ Heater Block PCB BIQ Extruder Peek clamp replacement BIQ Huxley Seedling
BIQ Sanguinololu mounting BIQ standalone Sanguinololu or Ramps mounting Print It Stick It Cut it


My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 08:23AM
I think any Eagle libraries can / have been converted to Kicad. I just googled around to find libraries for Kicad. Had to make quite a few parts of my own as well, but it is not hard, apart from the 3D models which are a pain in Wings 3D.

I think both produce Gerber and Excellon drill files so it shouldn't be a problem getting boards made anywhere. It is just a matter of naming the files as expected.

One thing I liked about PCB pool was that they accept the output file from GC-preview. That allows you to bundle up any Gerber files in such a way that there is no ambiguity in layers, etc. I.e. an ECAD package neutral format.


[www.hydraraptor.blogspot.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 09:31AM
As nophead suggested there are loads of library files kicking around on t'internet for kicad. I can definitely confirm that a quantity of these are indeed eagle ports. Some of them still have my name in the top of them from stuff I contributed to eagle a few years ago.

The other suggestions are bang on too.

IMHO Being able to create/modify your own library files is definitely and advantage it is quite rare that you will find a library file for many pinned devices that has the pins where you want them to get a good schematic drawing. (a good one being one that instantly conveys the structure and meaning of the circuitry together with it's functional blocking at a glance)

Board houses once up on a time used to only accept large scale (pref colour coded red,blue for double sided) photo reducible physical artwork or Gerber/Excellon Format Files. They would often photo-plot the Gerber files to produce the artwork and use the Gerber to drive the drilling machine. Best bet with most board houses these days if your not sure is ask them, what they can really handle will sometimes be most surprising.

Re rolling your own. I personally don't bother with the 3d stuff, the chances of their being 3d stuff that exactly matches the mechanical specs of your chosen component well enough are slim.

It is a nice to have thing to be able to see how your spacing of components might work out but my experience has been that manufacturers change their mechanical specs. ALternatively when buying in you will take the lowest price from a competing manufacturer with different mechanical specs. Rendering the potential benefits questionable.

For mechanical spacing etc I usually position the foot prints, printout the board so far on paper at 1:1 size and then place my components on/through the paper and tweak to suit. When happy I put the tracks down.

It is also good practice to place the holes etc for mechanical fixing and dimensions of the board first so that you wont inadvertently put a track somewhere you need to drill later. When laying up board that mount via slides (Eurocards or Extruded aluminum cases) I place a bounding box or some such to make sure I don't lay a track that will be rubbing up and down the slides (on plastic slides it can wear on the tracks, on metal slides the mask or conformal coating wears off and makes a short to the slide)

Doing a quick paper mock board makes it really easy to see if the mechanical dimensions for the board fit you chosen housing and fixing method as well.

The great thing about electronic CAD versus tapes and light box is that you can very easily rip up parts/tracks and re-lay them until you get something you can live with (Whilst the end result might be, the process is not really an exact engineering practice, I guess this is why it is called artwork)

On BLDC currents. Yes you are pretty much looking at peak rather than average or steady state currents. Using chunky wires though is still a good move. Peak current for a given motor is pretty much always stalled current. IE the motor draws most current when still and least when running at max RPM with no load.

Do also bear in mind when looking at FET switching, the maximum quoted currents are pretty much never attainable. The dissipation, packaging, leads and bond wires will impose lower limits. Particularly if you consider steady state currents as opposed to peak for the switching device. (on resistance versus current versus dissipation versus switching time period)

Hope this helps

Cheers

aka47


Necessity hopefully becomes the absentee parent of successfully invented children.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 06:19PM
[www.mcmanis.com] <-- an EXCELLENT discussion on MOSFET current ratings, both on how to figure out how much heatsinking you need, and how to calculate the advertised maximum current for a mosfet.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 06:50PM
Yup, thats where I think I read the full story too.

Well worth the reading, its a little tough going in places but worth sticking with it.

thumbs up


Necessity hopefully becomes the absentee parent of successfully invented children.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 07:35PM
Thank you both.. KiCad it is..

@Nophead
Thank you I will start to build up some KiCad Library's, if there is a way to edit or import one Library component to another, maybe we should look at maintain an RepRap Library for both CAD packages to try to cut down on repeated re-inventing the wheel.


@ AK47
I had not thought of having a layout so tight that the exterior dimensions would be a problem. I Like the idea of the physical paper test though would have hoped it to not be necessary. In view of the cost of producing boards its a great safe guard / sanity check.

BLDC and ECS devices.. without measuring the wires or calculating the cross-sectional area they just look far too small for the quoted current.

Particularly looking at the photos from the links Grael gave for the two devices. Also from the Two ECS modules and BLDC motors I have here

Edited 1 time(s). Last edit at 12/03/2009 07:37PM by BodgeIt.


Bodge It [reprap.org]
=======================================

BIQ Sanguinololu SD LCD board BIQ Stepcon BIQ Opto Endstop
BIQ Heater Block PCB BIQ Extruder Peek clamp replacement BIQ Huxley Seedling
BIQ Sanguinololu mounting BIQ standalone Sanguinololu or Ramps mounting Print It Stick It Cut it


My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 03, 2009 11:59PM
BodgeIt Wrote:
-------------------------------------------------------
> @ Grael..
>
> Now I thought my 15A drive brush less motor was
> High current..
>
> The size of the wire required for 70-80A would be
> significant.
> It seems these ratings are a little like music
> power v Watts RMS specifications.
>
> I have two ECS modules and two brush less motors
> as yet untested.
>
I've since picked up two 40 amp ESC, a 50, and a 70 that will run to 22.2 volts.
there is one at hobby king that goes up to 200 amp and almost 50 volts. However, I'm using mine so far for radio controlled boats.
The 2200mAHr 11.1 volt 35C batteries I have, and the 5000mAHr 7.4 V 20C LiPos have 12 guage wire, but there are batteries like this:
[www.hobbyking.com]
with double that discharge rate, around 200 amps. Only good for about a minute and a half at that rate though...

The only brushless speed controllers practical for CNC seem to be the car versions though, none of the others do reversing, although it would seem to be a simple rewrite of the software. After all, each winding is already bidirectionally powered by a brushless speed controller.


>without
> having to go and create your own library.

I got my boards made by PCBCART, and they found a few flaws that came from bad source components, things like solder resist over solder pads.

>
> Is there a recommended set of Reprap Library
> required for Kicad or Eagle packages.?
If you get hold of a Kicad or Eagle file, I think it will contain those components already, so you could save them individually.

Graham.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 04, 2009 07:07AM
Thanks for your reply Graham,

I'm not sure why I would need to have reversing ability thou.

Currently I'm using an Aldi ~ Dremel look alike, I'm hoping that using a BLDC and ECS with drive belt and home made bearing block will be significantly quieter than a real or look alike Dremel.

Thank you I will look at PCBCart as well as PCB pool.

I'm playing with KiCad now to create a simple PCB first so I can try milling a PCB using Bertha I will also give Home etching another chance again though I never really liked the photo etching process in the early 70's


Bodge It [reprap.org]
=======================================

BIQ Sanguinololu SD LCD board BIQ Stepcon BIQ Opto Endstop
BIQ Heater Block PCB BIQ Extruder Peek clamp replacement BIQ Huxley Seedling
BIQ Sanguinololu mounting BIQ standalone Sanguinololu or Ramps mounting Print It Stick It Cut it


My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
December 17, 2009 06:44PM
I think creating project specific libraries for kicad and referencing/keeping them with the rest of the project files is probably a useful idea.

More so if the whole of the project folder/directory is stored in the RepRap repository.

I guess it is then available to all along with the rest of the source for the design.

I don't know about you guys but I usually need to tweak each library on a per project basis anyway so as to get a schematic layout I can live with. At least the Major IC's anyway.


Necessity hopefully becomes the absentee parent of successfully invented children.
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
March 07, 2010 01:26PM
grael Wrote:
-------------------------------------------------------
> Thanks for approving my choice of generic chip
> type for my own board design ! >grinning smiley<

Hello!

I am very interested on cheap ARM Cortex! And now we have Arduino based on ARM: [www.xduino.com]

grael, how is the status of this project? is there any reprap wiki page with more info?

I would like to help if the boards suits my needs for RepRap :-)


---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]
TC
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
March 25, 2010 09:36PM
Graham,

RE: JTAG... don't overlook it. While you don't need it to load code (you can load via a UART as you pointed out) you will likely want JTAG for development (debug).

I would recommend that you at least put the footprint for the standard JTAG header on the board even if you don't insert the connector. That way, if you get into trouble you can always add the connector.

[infocenter.arm.com]

[www.keil.com]

TC
TC
Re: Build thread, 100 pin STM Arm Cortex board with driver chips.
March 25, 2010 09:53PM
Graham,

RE: SD connectors... this Wiki page is pretty good...
[en.wikipedia.org]

But it doesn't tell the whole story. If you still are interested in having SD capability then the simple thing is to look at the implementation from an evaluation board schematic.

I'm not that familiar with development boards for ST Microelectronics devices so I can't make a specific suggestion. However, I did look at a number of boards for the NXP LPC17xx family and found many good examples.

The most common thing seems to be micro-SD on these boards which requires a single SPI interface (sclk, cs, sdout, sdin).

TC
Sorry, only registered users may post in this forum.

Click here to login