Re: [CAD_CAM_EDM_DRO] Re: Interface. Gauging interest.

It is wrong to say that USB is not deterministic, it is. USB frames
operate on a strict 1 millisecond cycle. This is unlike Ethernet which
operates on a random access mechanism.

In USB each frame can carry 4 classes of data; Control, Bulk,
Interrupt, Isochronous. All of these classes of data fit within the 1023
byte frame. Each class may or may not appear in a particular frame.
It seems that for this purpose Isochronous is best suited.

Each frame can transfer 1023 bytes running at Full Speed. This is
USB 1.1 and not USB 2.0. Using Isochronous you reserve a portion
of this amount of data. You might ask for 200 bytes of the 1023.
This 200 byte per mS channel is exclusively for your use.
That means that you must supply the 200 bytes of data every 1 mS.
If you do not then the frame will leave without your data. Your
200 byte slot will be in the frame that leaves but your data won't
be there.

USB can be used in a deterministic fashion. If you do not attach
other, possibly misbehaving, devices to your USB bus it will be
completely deterministic.

Pete.

----- Original Message -----
From: "Dennis Schmitz" <denschmitz@...>
To: <CAD_CAM_EDM_DRO@yahoogroups.com>
Sent: Friday, May 19, 2006 2:17 AM
Subject: Re: [CAD_CAM_EDM_DRO] Re: Interface. Gauging interest.


Firewire would work just fine. And you're also right about the USB
interface. I did mention one company doing a synchronous version of USB, but
that's just a kludge.

There was one observation that putting brains back in the machine is full
circle from the big expensive steel boxes programmed with a keypad to
simplified electronics driven by a PC, back to brains back in the machine.
But it's not really full circle because you're not back where you started,
you're back with a WAY more powerful processor and networking interfaces.

You don't need an FPGA -- you can do the entire design in a MPC55xx series
microcontroller, including an Ethernet port. And if you have a copy of
Matlab, you can generate the control-loop code from a Simulink model. (Not
that Matlab generates especially great code, but it does generate it, and
you don't need to be a code jock.)

But the market will do what it will do. As it happens, I need such a motion
controller for another project, but I won't be able to get started on it
until after the proposal deadline next month.

If gumstix modules had a DAC <http://gumstix.com only cheaper> I'd already
be working on the software. Look at BASIX and ROBOAUDIO. Grand total of 6
16-bit PWM channels, a bunch of I/O and several differential A/Ds. No time
processing unit like in the Freescale, though, (They'd work great for
automatically reading shaft encoders with no CPU intervention.) and no DAC
outputs to drive servos, although you could drive two of them with the audio
output channels.

For that matter, with the right amplifier, you could drive servos from your
PC's audio output channels, and if you have one of the fancy-dancy ones with
surround-sound, you could drive four of them.

Maybe someone can convince Mariss to prescale his G320's to accept PC audio
output...

On 5/18/06, yracg <gcary@...> wrote:
>
>
> If PCI is not desired, has anyone considered IEEE-1394 (Firewire)? I
> don't know a lot about it except that I think it is more deterministic
> than USB. I know that it is used a lot in military avionics. Do a
> search for "ieee-1394 real-time" or "ieee-1394 avionics" and you'll
> find a number of interesting sites. Many computers have a Firewire
> port these days to interface with camcorders, and if they don't, it is
> an inexpensive card to add.
>
> My gut feel says that bit-banging a USB interface isn't going to give
> the performance that most will expect (I could be wrong). A
> higher-level protocol should be used that can stack up commands in a
> fifo. My inclination would be a design similar to the G-Rex.
> Assuming a USB interface, the design would include USB interface chip,
> an FPGA, a crystal oscillator, buffer ICs, etc. Hey! I think that is
> a G-Rex. I think part of the cost of the G-Rex is in the IP. When you
> buy one you're getting more than just hardware. If any of you want to
> make a lower cost G-Rex, I'm sure it can be done. But you're going to
> need a killer app to go with it or it will be pretty useless. You'll
> need Art and the others to support it.
>
>


[Non-text portions of this message have been removed]



Addresses:
FAQ: http://www.ktmarketing.com/faq.html
FILES: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/files/
Post Messages: CAD_CAM_EDM_DRO@yahoogroups.com

Subscribe: CAD_CAM_EDM_DRO-subscribe@yahoogroups.com
Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@yahoogroups.com
List owner: CAD_CAM_EDM_DRO-owner@yahoogroups.com, wanliker@...,
timg@...
Moderator: pentam@... indigo_red@... davemucha@...
[Moderators]
URL to this group: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO

OFF Topic POSTS: General Machining
If you wish to post on unlimited OT subjects goto:
aol://5863:126/rec.crafts.metalworking or go thru Google.com to reach it if
you have trouble.
http://www.metalworking.com/news_servers.html

http://groups.yahoo.com/group/jobshophomeshop I consider this to be a
sister site to the CCED group, as many of the same members are there, for OT
subjects, that are not allowed on the CCED list.

NOTICE: ALL POSTINGS TO THIS GROUP BECOME PUBLIC DOMAIN BY POSTING THEM.
DON'T POST IF YOU CAN NOT ACCEPT THIS.....NO EXCEPTIONS........
bill
List Mom
List Owner


Yahoo! Groups Links