Update from SFE / Future Dev Board Question

[N8B]

All,

I just updated the product page for the L1-64 Dev Board:
http://www.sparkfun.com/commerce/produc ... ts_id=9428
There is now a SFE.xn file available that can be used for the SFE dev board. There is also some simple example code posted that uses two threads. The code provides a simple example of how to respond to inputs and send outputs on two threads. Plus it is blinky :D. It also uses the UART module found on the XMOS website. Hope it is helpful for those that are new to XMOS.

The current PCBs in stock still have cut traces on them and the old version has yet to sell out. Here is a preview of what v15 is going to be:
http://www.sparkfun.com/datasheets/DevT ... rd-v15.pdf
The schematic is almost the same except there is now a UART over JTAG and over USB.

So here is the question:
Would people rather see the FTDI chip unpopulated on v15 and possibly bring the cost of the dev board down by five or ten bucks? Or do people like having USB/UART support on their dev boards and think the couple extra bucks is worth while?

[shawn]

Kind of a hard question to answer and or say no to.
XMOS processors enable a very tightly couple USB.
great for all that pro consumer gadgetry . SOooo keep it.
Lower prices or more affordable is really cool, so the question
balances. Thats why it a difficult question to answer. The business
person in me says that it would be negligible. If you want to
increase sales, advertise advertise advertise. Then the more
you sell the more you can offset the price. Never cheapen
your product, alway go forward. Most companies that grow
fast have a great product then they cheapen it only to never
return to their initial rate of growth or initial reputation. :|

[segher]

In my opinion, since the UART has no flow control, it is pretty much only useful as
a development console; and the UART on the XTAG gives you that already. So I
would get rid of it.

[leon_heller]

I can't get that new example to work properly. It displays the initial message, but the LED doesn't flash and the 'p', 's' and 'b' keys don't do anything. I could debug it, but I can't be bothered.

[kster59]

Before I dove into the XMOS designs I thought the Sparkfun board was a good layout.

Then I realized that the XMOS device draws around 300mA on the 1V rail and the Sparkfun board has a linear 1V regulator.

5v->1v is not very energy efficient and puts it out of spec on most USB ports.

If you use a switching regulator like the reference design then you would gain 70% efficiency and draw under 100mA.

[N8B]

In my opinion, since the UART has no flow control, it is pretty much only useful as
a development console; and the UART on the XTAG gives you that already. So I
would get rid of it.

Agreed. This thing has been nothing but trouble since the start. I'm seeing if I can get the next run of these things to not have the IC on it.

I can't get that new example to work properly. It displays the initial message, but the LED doesn't flash and the 'p', 's' and 'b' keys don't do anything.

I got it to work without any hiccups....granted the code is kinda lousy. I thought about writing a real state machine rather than using switch-case statements. I didn't put a lot of time into it.

[skoe]

I use the Sparkfun Board too. As my final board will have a UART header for debugging purposes, I also use the FTDI on the board.

Does the XTAG really has something I can use like /dev/ttyUSBx on Linux? I do not mean the XTAG stdio debugging stuff. If yes, the FTDI chip is really not necessary, Otherwise it would be nice to have, but a significant cost reduction would convince me :)

N8B: Could you thing about creating something like http://www.xcore.com/projects/xs1-l1-64-stamp-breakout? Just a minimal board with XMOS, 1V regulator, flash, reset logic (*). Nothing else. Even the 3.3 V regulator doesn't need to be on board IMHO, since most probably it's needed on the main board anyway. I guess many people would like to use it for prototypes or even low volume products if it is about 30 Euro or less.

Edit: Yes, the 1V regulator really get's quite hot. Using a switching regulator would be a good improvement on the board. If the EMC stuff isn't too difficult to solve...

(*) Edit: And clock of course :)

[Andy]

If doing a respin I'd consider the L1 design checklist more carefully (especially the section on power supply sequencing):

https://www.xmos.com/published/xs1-l-ha ... -checklist

[bsmithyman]

It looks like the SparkFun board is now available without the FTDI populated, and at a bit of a discount from the previous revision.

http://www.sparkfun.com/commerce/produc ... s_id=10113