Re: USB interface

On Mar 21, 4:57 am, <castlebravo...@xxxxxxx> wrote:
<don> wrote in message

news:lIednd0pKPGWEVnUnZ2dnUVZ_jaWnZ2d@xxxxxxxxxxxxxxxxxx



castlebravo...@xxxxxxx wrote:

By altering the way in which data is stored, using a counter rather than
a memory location to indicate time, I can reduce my memory requirements
from 2 Mbyte to 128 Kbytes. I only store when there is data to be stored,
not on every clock cycle. I need to do my homework to figure out the usb
protocol end of the deal.

http://geocities.com/castlebravo...@xxxxxxx/walking.pdf

Bob
Hi Bob,

I have followed this thread from the beginning.

I think you have not thought this all the way through.

Your RAM interface may give you an easy front end, but does nothing for
the USB backend.

One of your posts mentioned an heart rate monitor and a few switches.

That kind of data is not that fast.

Most higher level processors can do that.

I would guess that you are not a embedded programmer.

I see this entire project can be done with a processor and some A/D
converters.

Once that processor is selected and the interfaces are designed, the rest
is software.

Good luck on getting this done.

don

PS: Let us know when you get something working. It will be interesting
what direction you go.

A/D converters? I do not need to measure analog levels, only edge
transitions. I am interested in time measurements. The time of each
heartbeat, the time of each foot fall and the time of each button push. I
have created some test files with software and played with processing it. I
intend to do like wise with the new proposed format.

Bob

So, just like our project, It doesn't make sense to have USB build-
in. We have an external USB to SPI converter to map the uC EEPROM to
a mass storage device. It can also share a micro-SD card between the
uC and PC.

If you don't need A2D, your entire project can be done with a 50 cents
uC and a micro-SD card.
.

Quantcast