New group XK1 User Group

Archived group discussions. Please post all group discussions under their relevant XCore group page.
dougxmos
Junior Member
Posts: 6
Joined: Thu May 27, 2010 10:13 am

Post by dougxmos »

DanB wrote:
jonathan wrote:They are now obsolete.

http://www.xmos.com/products/xk1a
But still suitable for people that wish to use libflash; The XK-1A uses a revised flash device that requires the user to write their own configuration file if they wish to use libflash. Taken from XK-1 Development Kit Update document, available here:

https://www.xmos.com/download/public/XK ... -10%29.pdf
1.2 Revised flash device
The Atmel AT25DF041A-FLASH device on the XK-1 has been replaced with a Winbond
W25X10BVSNIG chip on the XK-1A. Both devices are supported in XFLASH, but
only the Atmel AT25DF041A-FLASH device is supported in libflash. To support the
Winbond device in libflash you need to write your own configuration file—see the
Tools User Guide for further instructions.
So, make your choice... Would you rather spend your time wishing hopelessly to use to the XScope [XK-1] or writing configuration files [XK-1A]? :) Heh.
Just wanted to let you all know that we've produced an SPI spec file for the Winbond part for those wishing to use libflash to perform field-upgrades. We'll be making this file available within the next few days on the XMOS website, and it will be included natively in the next tools release. I'll follow up with a link here in due course.


User avatar
jonathan
Respected Member
Posts: 377
Joined: Thu Dec 10, 2009 6:07 pm
Contact:

Post by jonathan »

So yes, they are now totally obsolete.

Given that the respin was clearly done to fix this XScope incompatibility, what would be a really awesome way of recognising the mistake - and demonstrating how seriously customers are taken - would be for XMOS to volunteer a replacement XK-1A for anyone with an XK-1 on request (rather than expect them to buy another!).

It might even generate some great positive coverage for the company and the XK-1A - which will no doubt help XMOS and probably kit sales too!

Hey, let's make it even better. Why not - if you give away your old XK-1 to your local school/university/hackspace/neighbour's hacker daughter, XMOS will ship you a new XK-1A. Now that would be truly awesome.
Image
User avatar
leon_heller
XCore Expert
Posts: 546
Joined: Thu Dec 10, 2009 10:41 pm
Location: St. Leonards-on-Sea, E. Sussex, UK.
Contact:

Post by leon_heller »

I was hoping that they'd dispensed with the XTAG and knocked $50 off the price, but no such luck.
m_y
Experienced Member
Posts: 69
Joined: Mon May 17, 2010 10:19 am

Post by m_y »

dougxmos wrote:
XK-1-Development-Kit-Update wrote: 1.2 Revised flash device
The Atmel AT25DF041A-FLASH device on the XK-1 has been replaced with a Winbond
W25X10BVSNIG chip on the XK-1A. Both devices are supported in XFLASH, but
only the Atmel AT25DF041A-FLASH device is supported in libflash. To support the
Winbond device in libflash you need to write your own configuration file—see the
Tools User Guide for further instructions.
Just wanted to let you all know that we've produced an SPI spec file for the Winbond part for those wishing to use libflash to perform field-upgrades. We'll be making this file available within the next few days on the XMOS website, and it will be included natively in the next tools release. I'll follow up with a link here in due course.
Both the 10.4.2 and 11.2.0 tools include a SPI spec for the Winbond W25X10.

A longer explanation:

The tools come with a large set of SPI specs, each defined as a macro. From the 11.2.0 tools installation directory they're in "target/include/SpecMacros.h" (this file is included by "flash.h" and should not be included separately as it has no multiple-inclusion guards). This set of SPI specs includes FL_DEVICE_WINBOND_W25X10 which corrosponds to the part used on the XK-1A.

Xflash automatically supports all the devices for which there are macros in "SpecMacros.h" (but adding to this file won't expand xflash's range; it's compiled in). Thus the XK-1A is supported by xflash to exactly the same extent as is the XK-1.

For libflash, things are more complex. It has two functions which connect to a SPI flash device: fl_connectToDevice() which accepts a list of SPI spec structures as one of its parameters, and fl_connect() which relies on an internal table of SPI specs. This internal table includes specs for all the flashes used on XMOS dev boards but not the XK-1A. It is my impression that this will not change; we will not continue to augment the internal table because the associated bloat will likely cause some applications to fail to link due to memory contraints.

Going forward it recommended that applications use fl_connectToDevice(). In this respect, the XK-1A is supported to exactly the same extent as is the XK-1.
User avatar
russf
XCore Addict
Posts: 146
Joined: Thu Dec 10, 2009 10:17 pm
Contact:

Post by russf »

On Friday, I received a new XK-1 from Digikey. Tuesday the announcement was made about the XK-1A.

I just contacted Digikey for an RMA & replacement. (Glad I did not hack that board, yet)

Apart from whether that gambit flies, I like Jonathan's suggestions for sending out free replacements.

It's going to be inconvenient for early adopters if they obtain newer devices to augment older development tools, and have to juggle with the incompatibility. If you build a reputation for punishing early adopters, guess what!? No adopters!

We already have the G4/Lx link incompatibility, which is quite a pain. A lot of the appeal of these devices and the code modules, is 'composability' - the ability to grow your product big enough for your needs, then when the design is stable, to shrink it down again as part of cost reduction. I'm following this strategy with a 3xL1 board, that will probably become a 2xL1 board later. (Or the Ethernet that is running on an external device will be grafted into one of these devices.

For my needs the XScope looks fantastic. I see it being the ideal way to debug inside the deliverable network, but also to be able to make external testing rigs that monitor I/O for board bringup. This is what I'm planning for my own development, and the XK-1(a) is the basis of that test-rig.

Can we try to get rid of this extra impediment?

Thanks XMOS!

Love means being able to talk about things!

--r
User avatar
jonathan
Respected Member
Posts: 377
Joined: Thu Dec 10, 2009 6:07 pm
Contact:

Post by jonathan »

By the way, you can now use your XK-1s with XScope, courtesy of pytey.

https://www.xcore.com/forum/viewtopic.php?f=8&t=1214
Image
Locked