PCB layout almost done.

[Reza Naima]

I'm not sure what you mean by #1.  The circuit includes an optoisolator for the limit switches.  I also don't see the need for a arduino mega and it's expensive.  I still need to finish the grounds, but I'm going to try to present only one point of shared GNDs between the motor and the signal GND - hopefully the currents will flow where they are supposed to go.  I'm done for the day -- this is where I'm leaving it off.  It's in SVN if anyone wants to play with it.  It's about 2.5x2.25" right now.  I'm big on aesthetics -- I chose some cool looking white resistors for R2 R3 R7 R8 R10 R11, and U$3 is also white. 

Any suggestions on PCB prototyping shops?  I suppose I'm in no rush and this order can be sent off to china or wherever.  I'm using 10/10/10 trace/space/drill. 

I also am using 15mil traces for the motors.  I calculated 18mils to get a 20 deg increase in temp (C) @ 2A, but I was limited by space.

I forgot to include an auto-zero interface.  And I don't think I have any more I/O pins left. 



 

Infamy wrote:

That looks nice.

A few things

1) I would bring out +5v/gnd/signal for each one of the limit switchs,
so that an optical limit switch can be used.
2) With the current placement of the terminal blocks, you cannot use a
arduino mega, i would move them to the "top" of the board or bottom,
moving the power connector off as well
3) Is only ground shared between the motor power and the Arduino? (if
so then perfect)

Alex

On Jul 18, 10:38 pm, Reza Naima <r...@reza.net> wrote:
  

I've finished picking out components and doing an initial layout.  I
found a digikey plugin that lets you sort by price which has allowed me
to optimize components by cost.  I've also added an LC filter on the
supply (1uH & 470uF) -- not sure if it's needed, but I love the look of
big coils.  I've also optoisolated the limit switch inputs to avoid
coupling noise to the MCU (which killed one of mine).  I made the
optoisolator a DIP package so that it can be replaced if it gets burnt
out somehow.  I'm not sure how the clearance will work out wrt to the
USB connector so I offset everything to the right -- the motor control
screw terminal will stick out a bit wider than the arduino board.  I
think it has a nice level of symmetry, the controllers are all lined up
so you can stick a heat sink over them, and I went with just a 2x.1"
header so you can use an IDC and ribbon cable to simplify hookup of the
limit switches.  There is also an Emergency Stop input that is by itself
(not sure if I should make it part of the same block as the limit
switches).  This is what it looks like so far..

Any thoughts?  The extra IC is a digital pot to let you control the
current per channel in software.

reza
    

[Infamy]

For the limit switchs i would bring out an extra pin, that would be
5volts. So each limit switch is gnd, signal, 5volts.

I was just more indicating that it would be a good idea to make sure
the board will work with a mega for anyone who wanted to use a mega.

In terms of prototyping shops, since the design is open source i would
try seeedstudio, they have a good deal i think its around 40$ for a
few boards. Otherwise batchpcb is a good place as well (run by
sparkfun)

Once you get that commited to the svn, I will take a try at some
simplifications, no optoisolator. and some molex headers for limit
switches and motors. But it looks like a great board!

Alex

On Jul 19, 1:02 am, Reza Naima <r...@reza.net> wrote:
> I'm not sure what you mean by #1.  The circuit includes an optoisolator
> for the limit switches.  I also don't see the need for a arduino mega
> and it's expensive.  I still need to finish the grounds, but I'm going
> to try to present only one point of shared GNDs between the motor and
> the signal GND - hopefully the currents will flow where they are
> supposed to go.  I'm done for the day -- this is where I'm leaving it
> off.  It's in SVN if anyone wants to play with it.  It's about 2.5x2.25"
> right now.  I'm big on aesthetics -- I chose some cool looking white
> resistors for R2 R3 R7 R8 R10 R11, and U$3 is also white.
>
> Any suggestions on PCB prototyping shops?  I suppose I'm in no rush and
> this order can be sent off to china or wherever.  I'm using 10/10/10
> trace/space/drill.
>
> I also am using 15mil traces for the motors.  I calculated 18mils to get
> a 20 deg increase in temp (C) @ 2A, but I was limited by space.
>
> I forgot to include an auto-zero interface.  And I don't think I have
> any more I/O pins left.
>

> Infamy wrote:
> > That looks nice.
>
> > A few things
>
> > 1) I would bring out +5v/gnd/signal for each one of the limit switchs,
> > so that an optical limit switch can be used.
> > 2) With the current placement of the terminal blocks, you cannot use a
> > arduino mega, i would move them to the "top" of the board or bottom,
> > moving the power connector off as well
> > 3) Is only ground shared between the motor power and the Arduino? (if
> > so then perfect)
>
> > Alex
>
> > On Jul 18, 10:38 pm, Reza Naima<r...@reza.net>  wrote:
>
> >> I've finished picking out components and doing an initial layout.  I
> >> found a digikey plugin that lets you sort by price which has allowed me
> >> to optimize components by cost.  I've also added an LC filter on the

> >> supply (1uH&  470uF) -- not sure if it's needed, but I love the look of

[Alexandre Harvey]

Oh Also another thought, the light version of EagleCad (read free) does not do multiple page schematics, so moving the schematic all to one page maybe a good idea.

Alex

[Reza Naima]

Not sure how to move the contents from the second sheet to the first. Ill try to do that. Also the files are in svn. The opto isolator is only 1$. Why remove it?
-Reza

[Alexandre Harvey]

Sorry, I should have said digital pot, i may take out the opto as well. The main reason to loose the digital pot is I need control lines for a mosfet to run my spindle motor.

Alex

[Reza Naima]

I thought about it for a bit and I can multiplex the SPI lines with the STEP lines for the controllers -- as long as the controllers are held in a reset state while the digital pot is configured.  That will free 2 of the I/O lines.  Design a motor control circuit that can handle ac & dc and I'll integrate it.

Reza

[Garrett Mace]

i'd move the drivers much closer to the output terminals and blow the traces up to 32 mil as close to the chips as possible.

[Alexandre Harvey]

For the driver circuit, I would just put a mostect on board, if you areusing an AC load then you can drive a relay with the mosfet. SCR/Triacs are an option but ac motor speed control is always problem I find.

Alex

[Reza Naima]

I tried that but it was ugly; I'll try it again.

-r

[Alexandre Harvey]

One other thing would be to double up the connectors, so you have the terminal blocks, but also molex (2.54mm, or the 3.94mm versions) headers as well, the two footprints should be stackable with minimal space increase, so you can pick either screw headers or crimp terminals.

Alex

[Reza Naima]

can you give me the molex part number your thinking about?  and why the preference for molex connectors?

[Alexandre Harvey]

http://www.sparkfun.com/commerce/product_info.php?products_id=8097

Those are what I had in mind, here just 0.1" header connectors. Not sure on the digikey part number...

Alex

[Alexandre Harvey]

Oh and the reason for also having 5v for the limit switchs is so something like http://store.makerbot.com/electronics/electronics-kits/optical-endstop-v2-1-kit.html  can be used instead of mechanical switches for the limits/home position.

[Alexandre Harvey]

Part numbers
538-22-23-2021 Two Pin, Male, 2.54mm, polarised friction lock, (board side)
538-22-01-3027 Two Pin, Female, 2.54mm (cable side) of above

538-22-23-2031 Three pin Male, same as above
538-22-01-3037 Three pin Female, same as above

538-22-23-2041 Four Pin Male, same as above
538-22-01-3047 Four Pin Female, same as above

Alex

[Reza Naima]

The latest version is in SVN (different repository than the one for firmware; i posted a link earlier).  Feel free to make the proposed changes and post a screenshot. 

Reza

[Reza Naima]

So I had a long chat with Macegr and the issue on the table is to add an avr+ftdi to this pcb and drop the 'shield' form factor or to stick to it.  his big complaint had to do with the form factor and difficulty putting it in an enclosure/mounting it.  the pcb can be made larger to accomodate different connectors -- though i can see some appeal to making it easier to put into a case.  does anyone else have thoughts on this? 

reza

[Alexandre Harvey]

I would second this. I would also go with a Atmega64, to get a few more IO pins and allow the addition of auto-tool zero as well as spindle control.

Alex

[Reza Naima]

I would go with the 328 so the code base can remain the same. Or was that 382?  Whatever the largest one from that series is.  I'm still not convinced though. One option would be to make this a 4 layer  and add an arduino likenpcb on the backside so tat it can optionally be populated?  Suppose I could do that on the top and make it longer too.  I should figure out the costs associated.   

-Reza

On Jul 19, 2010, at 6:07 PM, Alexandre Harvey <al...@pixelfactor.ca> wrote:

I would second this. I would also go with a Atmega64, to get a few more IO pins and allow the addition of auto-tool zero as well as spindle control.

Alex

On Mon, Jul 19, 2010 at 5:58 PM, Reza Naima <re...@reza.net> wrote:

So I had a long chat with Macegr and the issue on the table is to add an avr+ftdi to this pcb and drop the 'shield' form factor or to stick to it.  his big complaint had to do with the form factor and difficulty putting it in an enclosure/mounting it.  the pcb can be made larger to accomodate different connectors -- though i can see some appeal to making it easier to put into a case.  does anyone else have thoughts on this? 

reza

Reza Naima wrote:

The latest version is in SVN (different repository than the one for firmware; i posted a link earlier).  Feel free to make the proposed changes and post a screenshot

[Alexandre Harvey]

the 328 is the part you are thinking of, the 1280 atmega work with Arduino IDE.

Also attached is image of what i would do to add molex connectors.

Alex

[Reza Naima]

Looks good. Can you check in changes or do I have to give you permission to do so? I was also wrong, there are currently 2 free io pins, and I can free another 2 ifneeded

"Alexandre Harvey" <al...@pixelfactor.ca> wrote:

>the 328 is the part you are thinking of, the 1280 atmega work with Arduino
>IDE.
>
>Also attached is image of what i would do to add molex connectors.
>
>Alex
>
>On Mon, Jul 19, 2010 at 6:48 PM, Reza Naima <re...@reza.net> wrote:
>
>> I would go with the 328 so the code base can remain the same. Or was that
>> 382? Whatever the largest one from that series is. I'm still not convinced
>> though. One option would be to make this a 4 layer and add an arduino
>> likenpcb on the backside so tat it can optionally be populated? Suppose I
>> could do that on the top and make it longer too. I should figure out the
>> costs associated.
>>
>> -Reza
>>
>> On Jul 19, 2010, at 6:07 PM, Alexandre Harvey <al...@pixelfactor.ca> wrote:
>>
>> I would second this. I would also go with a Atmega64, to get a few more IO
>> pins and allow the addition of auto-tool zero as well as spindle control.
>>
>> Alex
>>

>> On Mon, Jul 19, 2010 at 5:58 PM, Reza Naima < <re...@reza.net>re...@reza.net

>>>> al...@pixelfactor.ca> wrote:
>>>>
>>>>> <http://www.sparkfun.com/commerce/product_info.php?products_id=8097>

>>>>> http://www.sparkfun.com/commerce/product_info.php?products_id=8097
>>>>>
>>>>> Those are what I had in mind, here just 0.1" header connectors. Not sure
>>>>> on the digikey part number...
>>>>>
>>>>> Alex
>>>>>
>>>>>
>>>>> On Mon, Jul 19, 2010 at 12:46 PM, Reza Naima < <re...@reza.net>
>>>>> re...@reza.net> wrote:
>>>>>
>>>>>> can you give me the molex part number your thinking about? and why the
>>>>>> preference for molex connectors?
>>>>>>
>>>>>>
>>>>>> Alexandre Harvey wrote:
>>>>>>
>>>>>> One other thing would be to double up the connectors, so you have the
>>>>>> terminal blocks, but also molex (2.54mm, or the 3.94mm versions) headers as
>>>>>> well, the two footprints should be stackable with minimal space increase, so
>>>>>> you can pick either screw headers or crimp terminals.
>>>>>>
>>>>>> Alex
>>>>>>
>>>>>> On Mon, Jul 19, 2010 at 12:32 PM, Reza Naima < <re...@reza.net>
>>>>>> re...@reza.net> wrote:
>>>>>>
>>>>>>> I tried that but it was ugly; I'll try it again.
>>>>>>>
>>>>>>> -r
>>>>>>>
>>>>>>>
>>>>>>> Garrett Mace wrote:
>>>>>>>
>>>>>>> i'd move the drivers much closer to the output terminals and blow the
>>>>>>> traces up to 32 mil as close to the chips as possible.
>>>>>>>
>>>>>>>
>>>>>
>>>>
>>

-Reza

[Alexandre Harvey]

You would need to give me permission, I also need to remove the silk since you don't want silkscreen going over holes.

Alex

[Reza Naima]

Ok, try again.  Also, most board houses will remove the silk over the soldermask holes automatically.  I've also found that when they don't, the silk doesn't affect the solder (it floats on top, kinda neat).

Reza

[Alexandre Harvey]

Hmm, tried to commit it won't let me. I have the schematic and the pcb updated. (I rechecked out and tried and it still give me a permission error)

Alex

[Reza Naima]

What email address do log into google with?

-Reza

[Alexandre Harvey]

The one you added to the project as a contributor, that seems correct, it just wont let me commit for some reason. I will dig into more tomorrow.

Alex