SPI interword gap

View: New views
6 Messages — Rating Filter:   Alert me  

SPI interword gap

by David Baur :: Rate this Message:

| View Threaded | Show Only this Message

I'm attempting to send 4K 16 bit words over the SPI bus at 24Mhz.  Theoretically it possible, but I've noticed that there is a clock gap between words.  Does anyone have any experience on how to reduce or eliminate this gap?  I've discovered a 'turbo' mode, but haven't really noticed a reduction in gap size.

Thanks

--
David G. Baur
505 286-0293 h
505 284-8031 w
505 362-4443 c

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: SPI interword gap

by jumpnowdev :: Rate this Message:

| View Threaded | Show Only this Message

The Linux SPI subsystem is not designed for things like this.

It's based on workqueues.

http://www.linuxjournal.com/article/6916

Use the OMAP3 McBSP controller instead.

McBSP3 can be mux'd out pins 27-30 on most of the expansion boards (gpio
144-147)

I've used it in a number of projects and you can get very good,
deterministic performance from it.

It can do 'SPI' protocol, but with only one 'slave' device.


--
View this message in context: http://gumstix.8.n6.nabble.com/SPI-interword-gap-tp4964698p4964699.html
Sent from the Gumstix mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: SPI interword gap

by William Hue-2 :: Rate this Message:

| View Threaded | Show Only this Message

David,

If you make use of the hardware FIFO, there should be no inter-word gaps.  If you still get the occasional gap with the FIFO, then using DMA should get rid of that.  Are you using the standard driver or a custom one that you wrote?

        William


On 2012-06-22 12:46, David Baur wrote:

> I'm attempting to send 4K 16 bit words over the SPI bus at 24Mhz.  Theoretically it possible, but I've noticed that there is a clock gap between words.  Does anyone have any experience on how to reduce or eliminate this gap?  I've discovered a 'turbo' mode, but haven't really noticed a reduction in gap size.
>
> Thanks
>
> --
> David G. Baur
> 505 286-0293 h
> 505 284-8031 w
> 505 362-4443 c
>
>
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>
>
>
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: SPI interword gap

by AKS-4 :: Rate this Message:

| View Threaded | Show Only this Message

Hi

How about bit bashing, switching on and off the GPIO ports instead
of using SPI hardware registers? Then you can control everything.

On 6/22/12, David Baur <dgbaur@...> wrote:

> I'm attempting to send 4K 16 bit words over the SPI bus at 24Mhz.
> Theoretically it possible, but I've noticed that there is a clock gap
> between words.  Does anyone have any experience on how to reduce or
> eliminate this gap?  I've discovered a 'turbo' mode, but haven't really
> noticed a reduction in gap size.
>
> Thanks
>
> --
> David G. Baur
> 505 286-0293 h
> 505 284-8031 w
> 505 362-4443 c
>

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: SPI interword gap

by Akram Hameed :: Rate this Message:

| View Threaded | Show Only this Message

Pretty sure he'd never manage 24MHz bitbashing. If he did, I'd be
surprised!

-----Original Message-----
From: AKS [mailto:aungkyas@...]
Sent: Saturday, 23 June 2012 10:01 AM
To: General mailing list for gumstix users.
Subject: Re: [Gumstix-users] SPI interword gap

Hi

How about bit bashing, switching on and off the GPIO ports instead of
using SPI hardware registers? Then you can control everything.

On 6/22/12, David Baur <dgbaur@...> wrote:

> I'm attempting to send 4K 16 bit words over the SPI bus at 24Mhz.
> Theoretically it possible, but I've noticed that there is a clock gap
> between words.  Does anyone have any experience on how to reduce or
> eliminate this gap?  I've discovered a 'turbo' mode, but haven't
> really noticed a reduction in gap size.
>
> Thanks
>
> --
> David G. Baur
> 505 286-0293 h
> 505 284-8031 w
> 505 362-4443 c
>

--------------------------------------------------------------------------
----
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and threat
landscape has changed and how IT managers can respond. Discussions will
include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: SPI interword gap

by David Baur :: Rate this Message:

| View Threaded | Show Only this Message

I'm using the mcspi low level driver and code based on Scott Ellis' spike protocol layer.  I turned on 'Turbo' mode which did reach the speed requirement. 

But subsequently, my mmc ssd failed and I can no longer create a kernel that talks over the ethernet 'Invalid chip ending'. 

And after much consternation, now I can't even build a kernel that starts.

I'm losing my faith in the Gumstix boards.

dave

On Fri, Jun 22, 2012 at 5:27 PM, William Hue <william@...> wrote:
David,

If you make use of the hardware FIFO, there should be no inter-word gaps.  If you still get the occasional gap with the FIFO, then using DMA should get rid of that.  Are you using the standard driver or a custom one that you wrote?

        William


On 2012-06-22 12:46, David Baur wrote:
> I'm attempting to send 4K 16 bit words over the SPI bus at 24Mhz.  Theoretically it possible, but I've noticed that there is a clock gap between words.  Does anyone have any experience on how to reduce or eliminate this gap?  I've discovered a 'turbo' mode, but haven't really noticed a reduction in gap size.
>
> Thanks
>
> --
> David G. Baur
> 505 286-0293 h
> 505 284-8031 w
> 505 362-4443 c
>
>
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>
>
>
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users



--
David G. Baur
505 286-0293 h
505 284-8031 w
505 362-4443 c

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
gumstix-users mailing list
gumstix-users@...
https://lists.sourceforge.net/lists/listinfo/gumstix-users