On 25/04/13 19:13, David Bagby wrote:
> Hi Charles,
>
> On 4/25/2013 5:55 AM, Charles Steinkuehler wrote:
>> On 4/24/2013 9:27 PM, David Bagby wrote:
>>> Charles: Am I correct to assume that your LCNC stepgen PRU code is
>>> configured to use the same BBW pins as the BeBoPr (I think I remember
>>> reading that you were using that for prototyping)?
>> I am initially working on a machine configuration for the BeBoPr (which
>> I have in-hand) and Replicape pinouts.  I am trying to make the
>> low-level PRU code flexible so you can assign I/O pins at run-time, and
>> match whatever cape or home-brew hookup you're using.
> Ok, that's what I thought - I wanted to check my assumption.
>
> I'm thinking that we could get there (dynamic pin configuration) in phases - 
> see below.
>
>>> If so, I suspect that we're going to find that there are some BBB pin
>>> challenges to consider.A quick look at the BBB schematics shows that a
>>> fair number of the pins in use by the BeBoPr and Replicape PRU
>>> configurations overlap with the LCD/HDMI pin set on BBB P8. That implies
>>> at least some changes will likely be needed to move LCNC from BBW to BBB
>>> (with the current pin mappings, to run LCNC on a BBB I think it would be
>>> necessary to disable the HDMI support in the BBB).
>>>
>>> Are these questions already figured out or do these topics still need to
>>> be worked through?
>> I have spent a good chunk of the last couple days reviewing the recently
>> released docs on the BBB.  Executive summary:  I'm not worried.
> High level: "me neither";    :-)
>
>> Although as you noted, without some pin shuffling the HDMI output will
>> have to be limited (likely completely disabled with the BeBoPr).  But
>> the original 'Bone doesn't have HDMI output at all, so it's no worse
>> than before.
> In another post in this thread you also wrote:
>
> "The I/O pins used are configurable at runtime, but currently only PRU
> I/O pins are supported.  I am going to add support for the PRU to
> directly drive GPIO pins, so basically any pin on the chip could be used
> with appropriate I/O mux setup."
>
> That would be a valuable enhancement. I had not realized that the current 
> code only used PRU accessible pins.
>
> I've some additional ideas I'd like to explore & see what interest there may 
> be -
>
> My interests are around LCNC on an embedded system. I've been eagerly
> following the work you've been doing with the BBW.
> It's no surprise that  LCNC on an embedded platform can be different
> from LCNC on a PC.While LCNC on a PC is great, and I see that continuing
> to be a core supported platform, I'd also like to see PC platform
> assumptions not be constraining limitations for embeded system usage.  I
> think that some of Michael's restructuring proposals will help with this
> over time.
>
> WRT to the topic at hand, embedded platform CNC approaches tend to be a
> bit closer to the hardware. Beaglebones are an example: On this
> platform, hdw being used by LCNC is being initialized at OS boot time,
> not at LCNC start time.
>
> Here's an outline of the BB boot sequence:
>
> (The following is what I think I currently understand -- details could
> still be incorrect. I'm iteratively looking at multiple topics to figure
> this out - if someone spots an error, or a silly assumption, please
> educate me further. )
>
> 1)power up
>
> 2)OS boot process
>
> a.Hardware init
>
> i.BBB specific OS code looks for Capes (& Cape EPROMs)
>
> (this is where the kernel version seems to matter as BBB requires device
> tree support, that BBW doesn't).
>
> 1.OS Configures pins according to EPROM data on cape
>
> a.Theory: BeBoPr and Replicape implicitly assume default BBW pin setup
> (I've not verified this yet as I haven't found cape EPROM contents for
> the boards)
>
> b.Implication: these capes won't work on BBB w/o changes (pin conflicts).
>
> c.Frankly the tech details of this are a bit moot when you can't buy the
> capes...
> d. ... and we need some physical connection ability from the BBs to the
> rest of the world.
>
> 3)LCNC started
>
> a.LCNC Init - Current sequence:
>
> i.LCNC loads PRU code etc
>
>                   1.LCNC software on BBW is assuming a default BBW pin
> config (established during OS boot process).
>
> ii.Assertion: We need a different assumption for BBB...
>
>                   1.Treating BBB as BBW by "not using HDMI" does not seem
> to be an easy option
>
>               a.(disabling HDMI appears to be a sftw+hdw job -- not a
> pure software one; some hdw signals have tobe held in certain states
> during POR etc.).
>
>               b.That implies a cape is needed to convert a BBB pin out
> into a BBW pin out. It seems a bit silly to build a cape just to do that.
>
>                   2.     Issue: what assumptions to use re BBB pin out?
>
>                   3.Enabler: Some consensus on a BBB Pin config that
> could be used to jump start further BBB work and allow more folks to
> participate in LCNC on BB development.
>
>               a.We're are kind of stuck when no interface capes are
> easily available.
>
> b.Future enhancement: Hal could learn more about BBB platform and
> dynamically rearrange pin mappings and mux modes.
>
> i.More complexity -- would have to consider interactions with multiple
> capes etc.
>
> ii.     It seems preferable not to make this a LCNC on BBB prerequisite.
>
> Idea:
>
> I'd like to see a simple BBB cape that would
>
> A)Implement a simple, assumable, BBB pin mapping that could be used by
> LCNC Hal for BBB.
>
> B)Sets up the BBB pins at OS boot time (via Cape EPROM info).
>
> C)Lets us simply connect a BBB to readily available external hardware
> (e.g. Break out boards).
>
> This would enable the creation of a low cost BBB cape that could be used
> for LCNC on BB (BBW or BBB) development.I'm thinking the goal should be
> to keep it as simple and low cost as possible. This would enable
> creation of a BBB cape that enables more people to play with LCNC on a
> BBB (it would also work fine in a BBW).  I specifically think of this as
> a "LCNC proto cape", simpler than a BeBoPr or Replicape as it would NOT
> include motor drivers etc.
>
> If there is interest in this approach, I'll follow up with a minimal
> signal support proposal in the form of a straw man for people to comment
> on.As a first cut, I'd try to see what we can do using a BBB pin mapping
> that still allows use of the eMMC and HDMI features of the BBB.
>
>
> Dave
>
>
>
>
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring service
> that delivers powerful full stack analytics. Optimize and monitor your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
> _______________________________________________
> Emc-developers mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/emc-developers
>
>
following on from using the beaglebone .
i have some spare bare unpopulated replicape pcb's , if anyone wishs to 
use for lcnc use if it helps
i have not built one as yet , and the workloads increasing !

anyone interested in a board let me know ,

Dave


------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to