Forums

LPC213x And Ethernet

Started by dsidlauskas1 January 25, 2005

--- In , "Anton Erasmus" <antone@s...> wrote:
> On 31 Jan 2005 at 5:07, Rick Collins wrote:
>
> [Lots of stuff snipped]
> >
> > I understand that the W3100A would be simpler to use, but I believe
> > you said it did not support 100 base TX, right? The LAN91C111 is not
> > an inexpensive chip, but it includes the PHY so all you need to add is
> > the transformer and the connector and that end is ready to plug into a
> > network. Yes, it needs a lot more on the other side, but making a
> > general app board, I want to be able to support 100 base TX. Even if
> > you don't need the data rate, this can be an advantage is some cases
> > and it should still meet the selling price target of $99, qty 1.
> >
> > I will need to project a parts list in the next couple of days, now
> > that I have been promised samples of the AT91SAM7S64. This is pin
> > compatible with the SAM7S128 and SAM7S256 and will do for initial
> > prototypes.
>
> No the W3100A does support 100 base T. Although you need a seperate
> PHY and of course the magnetics + connector. They have a module which
> includes the W3100A device + the Realtek RTL8201L Phy. One only need
> to add the magnetics. If you actually use the module, then you can offer
> the board without the module at a reduced price, and people can later
> add it. Nice when one is on a tight budget.

Unless you really need some speed. You said it interfaces by I2C
which is very slow by comparison, ~400 kbps vs. >10 Mbps.

I don't plan to sell the board in versions, that gets to be a PITA.
But I will likely sell a bare board version if you want to
do-it-yourself.



An Engineer's Guide to the LPC2100 Series

> From: "Anton Erasmus" <>

> No the W3100A does support 100 base T. Although you need a
> seperate PHY and of course the magnetics + connector. They have a
> module which includes the W3100A device + the Realtek RTL8201L
> Phy. One only need to add the magnetics. If you actually use the
> module, then you can offer the board without the module at a
> reduced price, and people can later add it. Nice when one is on a
> tight budget.

Have a look at the Asix 88796 which does 10/100 and has
integrated PHY. You can get the cheap NICholas board from EDTP
Electronics for prototyping. We have one glued on the side of a
2106 running our TCP/IP stack and web server.

Stephen
--
Stephen Pelc,
MicroProcessor Engineering Ltd - More Real, Less Time
133 Hill Lane, Southampton SO15 5AF, England
tel: +44 23 80 631441, fax: +44 23 80 339691
web: http://www.mpeltd.demon.co.uk - free VFX Forth downloads




--- In , "Stephen Pelc" <stephen@m...> wrote:
> > From: "Anton Erasmus" <antone@s...>
>
> > No the W3100A does support 100 base T. Although you need a
> > seperate PHY and of course the magnetics + connector. They have a
> > module which includes the W3100A device + the Realtek RTL8201L
> > Phy. One only need to add the magnetics. If you actually use the
> > module, then you can offer the board without the module at a
> > reduced price, and people can later add it. Nice when one is on a
> > tight budget.
>
> Have a look at the Asix 88796 which does 10/100 and has
> integrated PHY. You can get the cheap NICholas board from EDTP
> Electronics for prototyping. We have one glued on the side of a
> 2106 running our TCP/IP stack and web server.

The last time I checked with ASIX, they did not have industrial temp
versions of their chips. Do you know if this has changed?



On 1 Feb 2005 at 22:52, Rick Collins wrote:

>
>
> --- In , "Anton Erasmus" <antone@s...> wrote: >
> On 31 Jan 2005 at 5:07, Rick Collins wrote: > > [Lots of stuff
> snipped] > > > > I understand that the W3100A would be simpler to use,
> but I believe > > you said it did not support 100 base TX, right? The
> LAN91C111 is not > > an inexpensive chip, but it includes the PHY so
> all you need to add is > > the transformer and the connector and that
> end is ready to plug into a > > network. Yes, it needs a lot more on
> the other side, but making a > > general app board, I want to be able
> to support 100 base TX. Even if > > you don't need the data rate,
> this can be an advantage is some cases > > and it should still meet
> the selling price target of $99, qty 1. > > > > I will need to
> project a parts list in the next couple of days, now > > that I have
> been promised samples of the AT91SAM7S64. This is pin > > compatible
> with the SAM7S128 and SAM7S256 and will do for initial > > prototypes.
> > > No the W3100A does support 100 base T. Although you need a
> seperate > PHY and of course the magnetics + connector. They have a
> module which > includes the W3100A device + the Realtek RTL8201L Phy.
> One only need > to add the magnetics. If you actually use the module,
> then you can offer > the board without the module at a reduced price,
> and people can later > add it. Nice when one is on a tight budget.
>
> Unless you really need some speed. You said it interfaces by I2C
> which is very slow by comparison, ~400 kbps vs. >10 Mbps.

That is if you use the I2C bus - I would only use this on a legacy product
that does not have enough pins for the parallel mode. The normal method
is to map it to a 8-bit 16K SRAM block. When one wants to send a packet,
one copies the data to the specific socket Tx buffer, and set a bit to transmit
the data. The hardware will handle any TCP/IP overhead.
Normal frequency for bus timing is 25MHz, but a clock up to 50MHz can be
used. the device can easily do the full 100Mbps using the normal
parallel mode.

> I don't plan to sell the board in versions, that gets to be a PITA.
> But I will likely sell a bare board version if you want to
> do-it-yourself.

I seldom buy these sort of products, hence I am obviously not the
target market, hence not in a position to say whether this is a good
or bad idea.

Regards
Anton Erasmus
--
A J Erasmus



--- In , "Anton Erasmus" <antone@s...> wrote:
> On 1 Feb 2005 at 22:52, Rick Collins wrote:
> > Unless you really need some speed. You said it interfaces by I2C
> > which is very slow by comparison, ~400 kbps vs. >10 Mbps.
>
> That is if you use the I2C bus - I would only use this on a legacy
product
> that does not have enough pins for the parallel mode. The normal method
> is to map it to a 8-bit 16K SRAM block. When one wants to send a packet,
> one copies the data to the specific socket Tx buffer, and set a bit
to transmit
> the data. The hardware will handle any TCP/IP overhead.
> Normal frequency for bus timing is 25MHz, but a clock up to 50MHz can be
> used. the device can easily do the full 100Mbps using the normal
> parallel mode.

The wiznet web site disagrees with you. I thought I would check it
out and it does seem like a nice product, but it is not fast. Here is
their speed claims.

High Performance
Processor Performance (100Mbps, PIII 500MHz,1 channel, FDX)
Atmel 89C51 300 Kbps
Atmel AVR 3 Mbps
Intel 80386 6 Mbps
Hitachi SH7709A 8 Mbps
Hyperstone
E1-16KT+DMAC control 15Mbps

I don't know what a Hyperstone... is, but that is the fastest
throughput they claim, 15 Mbps, using DMA. Still that is not a bad
speed. > > I don't plan to sell the board in versions, that gets to be a PITA.
> > But I will likely sell a bare board version if you want to
> > do-it-yourself.
>
> I seldom buy these sort of products, hence I am obviously not the
> target market, hence not in a position to say whether this is a good
> or bad idea.

If cost is of ultimate importance, I expect Olimex will have
satifactory boards.