Atmel AVR CAN
Don
available.
I tried asking at AVRFreaks but the message got dropped. I then tried
searching & found that there wasn't a mention of it in the mail lists.
Spooky ;-)
Erik Hermann
> Has anyone heard when the CAN version of the Mega128 will be commonly
> available.
It seems that first samples are available, the german representative
already has a working development board and wants to show it to us in
a few days.
But I think Atmel is much too late with this. They should have
integrated CAN years ago.
Anton Erasmus
They are on the EBV stock list. http://www.ebv.com which is part of
avnet. THe price was +- 6 Euro when I checked.
Regards
Anton Erasmus
Ulf Samuelsson
news:c5ores$44vsh$1...@ID-230069.news.uni-berlin.de...
> Don wrote:
>
> > Has anyone heard when the CAN version of the Mega128 will be commonly
> > available.
>
> It seems that first samples are available, the german representative
> already has a working development board and wants to show it to us in
> a few days.
> But I think Atmel is much too late with this.
That is not what I hear from my customers :-)
> They should have
> integrated CAN years ago.
--
Best Regards,
Ulf Samuelsson u...@a-t-m-e-l.com
This is a personal view which may or may not be
share by my Employer Atmel Nordic AB
Jim Granville
>>>Has anyone heard when the CAN version of the Mega128 will be commonly
>>>available.
>>
>>It seems that first samples are available, the german representative
>>already has a working development board and wants to show it to us in
>>a few days.
>>But I think Atmel is much too late with this.
>
>
> That is not what I hear from my customers :-)
You mean the customers that have not yet heard of the ST STR710FZ2 ? ;)
$/10K FLASH RAM MHz Pins Serial
-----------+------+-------+-----+------+--------+-------------
AT90CAN128 $7.50 128K 4K 16MHz 64TQFP CAN
STR710FZ2 $7.80 256K 64K 48MHz 144TQFP CAN/USB/HDLC
STR712FR1 $nya 128K 16K 48MHz 64TQFP CAN
No price seen yet on their 64pin/128K model, but if the price given
on their web site for the 256K/144 pin model is right,
that's a lot of bang to the buck.
Refs:
http://biz.yahoo.com/prnews/040330/sftu033_2.html
http://www.st.com/stonline/books/pdf/docs/10350.pdf
http://www.st.com/stonline/products/support/micro/arm/str7_10.htm
http://www.st.com/stonline/press/news/year2004/p1431p.htm
There was some talk of a monolithic FlashARM from Atmel
- is that close/canned/never really existed... ?
-jg
Erik Hermann
> You mean the customers that have not yet heard of the ST STR710FZ2 ? ;)
>
> $/10K FLASH RAM MHz Pins Serial
> -----------+------+-------+-----+------+--------+-------------
> AT90CAN128 $7.50 128K 4K 16MHz 64TQFP CAN
> STR710FZ2 $7.80 256K 64K 48MHz 144TQFP CAN/USB/HDLC
> STR712FR1 $nya 128K 16K 48MHz 64TQFP CAN
Well, I've used a lot of CAN Micros from Fujitsu years ago.
They are very good and much cheaper than PIC's with CAN.
We payed around 5 $ for a 16 Bit uC with 64K Flash and 4k RAM
(+ 10 Bit ADC, etc).
And you get the full featured C development environment for free.
I really think that Atmel is much too late.
Why use an expensive AVR if you can get a much cheaper/better HC08/HC12
Fujitsu FX or Mitsubishi M16 ?
AVR's are good for hobbyists and small volumes, not for industrial
and automotive applications.
> There was some talk of a monolithic FlashARM from Atmel
> - is that close/canned/never really existed... ?
Why not use TI or Philips?
Schwob
Hi Jim,
the ST devices sound very nice. Readin their press release though tell
me that they "plan to go into volume production in Q4 2004". In the
past ST has not been know for announcing their products conservatively
regarding schedule. So, there is already a device out there, running
60 MHz zero waitstates with 2 or 4 CAN interfaces from Philips
(LPC2129 and 2194). To me it seems that this time ST is not even a
fast follower, just a follower. There are however two things in the ST
devices that are very nice, the USB interface and the 12-bit ADC.
Cheers, Schwob
Jim Granville
<snip>
> Hi Jim,
>
> the ST devices sound very nice. Readin their press release though tell
> me that they "plan to go into volume production in Q4 2004". In the
> past ST has not been know for announcing their products conservatively
> regarding schedule. So, there is already a device out there, running
> 60 MHz zero waitstates with 2 or 4 CAN interfaces from Philips
> (LPC2129 and 2194). To me it seems that this time ST is not even a
> fast follower, just a follower. There are however two things in the ST
> devices that are very nice, the USB interface and the 12-bit ADC.
Philips also say samples now, volume prodn Q4 04 in this
SJA2020 press release
http://www.philipssemiconductors.com/news/content/file_1015.html
LPC2294 is also tagged only as 'samples available'
I was wondering how the Philps SJA2020 & LPC2294 differ ?
[Both 256KF, 4 CAN 144 pins ]
Different die, or different marketing divisions within Philips ?
Has anyone looked closely at these two ?
Ulf Samuelsson
> Jim Granville wrote:
>
> > You mean the customers that have not yet heard of the ST STR710FZ2 ? ;)
> >
> > $/10K FLASH RAM MHz Pins Serial
> > -----------+------+-------+-----+------+--------+-------------
> > AT90CAN128 $7.50 128K 4K 16MHz 64TQFP CAN
> > STR710FZ2 $7.80 256K 64K 48MHz 144TQFP CAN/USB/HDLC
> > STR712FR1 $nya 128K 16K 48MHz 64TQFP CAN
>
> Well, I've used a lot of CAN Micros from Fujitsu years ago.
> They are very good and much cheaper than PIC's with CAN.
> We payed around 5 $ for a 16 Bit uC with 64K Flash and 4k RAM
> (+ 10 Bit ADC, etc).
> And you get the full featured C development environment for free.
>
> I really think that Atmel is much too late.
> Why use an expensive AVR if you can get a much cheaper/better HC08/HC12
> Fujitsu FX or Mitsubishi M16 ?
> AVR's are good for hobbyists and small volumes, not for industrial
> and automotive applications.
Yes, and there seem to be about 500 million hobbyists since that is how
many AVRs that has been delivered since 1987 :-)
Jim Granville
>>I really think that Atmel is much too late.
>>Why use an expensive AVR if you can get a much cheaper/better HC08/HC12
>>Fujitsu FX or Mitsubishi M16 ?
>>AVR's are good for hobbyists and small volumes, not for industrial
>>and automotive applications.
>
>
> Yes, and there seem to be about 500 million hobbyists since that is how
> many AVRs that has been delivered since 1987 :-)
Glad the smiley is there, as clearly, few of those _actually_ went to
hobbyists.
I've looked for the details behind that 500M number, and it seems to
be _total_ AVRs, and I cannot see a breakdown of ASIC / MASK / Custom /
RAM Cores / SmartCard / and Flash (merchant uC) devices anywhere.
Some simple comparisons :
Motorola claim ~6 Billion HC05/HC08 devices shipped 1990-2003.
Motorola are the biggest 8 bit uC vendor (by $), with ~24% of a ~$4.5B
market.
Microchip currently ship around 500M PICs every year.
Their average selling price remains quite low, but most
recent-release devices are FLASH.
The 80C51-core market is well over $1B, with an annual run rate
circa 1B Pcs/yr. [Even at 6Billion pcs [1990-2003], the HC05 does not
claim to be the worlds most-shipped 8 bit core - that title goes
to the 80C51]
Even Zliog, have shipped 200M Pcs of TV remote control targeted uC.
Atmel Brochures also make this claim :
"...with three times the market share of its nearest competitor, Atmel
is also the world's leading supplier of 8 bit Flash uC"
There is also no mention of revenue, or units, so that also needs to
be clarified. viz : The 8 bit uC #1 bagging rights are currently split
between revenue and volume/units claimants.
I would like to see the substance behind that marketdroid claim - seems
a big call - but maybe the FLASH uC market is a much smaller piece of
the total 8 bit uC market than many think.
Given that Motorola, Microchip, Rensas, NEC, Philips, STm, are all
much bigger players than Atmel, and ALL these companies are ramping
their 8 bit FLASH uC, ( plus the tier 2 players, like Cygnal, Winbond,
SST, ISSI, Zilog, Natsemi, TI (etc) all have active and ramping FLASH uC
business) it does strain credibility.
Atmel can add their Flash_AVR plus Flash_89C51 numbers together, but
even with this, it's hard to imagine a 3:1 lead over #2.
Hitachi were one of the first with FLASH uC, but I think they like to
put their H8 series into the 16 bit basket, which would improve others
8 bit numbers.
-jg
CBarn24050
> I would like to see the substance behind that marketdroid claim - seems
>a big call - but maybe the FLASH uC market is a much smaller piece of
>the total 8 bit uC market than many think.
maybe, but in a few more years flash will be just about all the 8/16 bit
market.
Jim Granville
My point was not the flash growth, but to raise eyebrows at
the Atmel claim of 'leader, and 3x share of #2' in 8 bit Flash uC.
Certainly, the flash portion of 8 bit uC is growing. Nine of the
top 10 8 bit uC vendors have active FLASH family expansions - only
Infineon do not have 8 bit FLASH, but they do have
an active 16 bit Flash family.
As to 'just about all', you'd need to quantify that more.
It is interesting to note that mask is not going away, and
even a recent high end 32 bit uC from Toshiba, is comming
in FLASH and MASK versions. Surprised me, but Flash does cost
more to FAB, and also cannot be as reliable as mask, so
for high volume/stable code apps, mask still has a place.
Philips have also recently released a new OTP member of the LPC7xx
family. This seems to align with a TSSOP20 FLASH LPC9xx member, so
they could be targeting both-ends of the market. OTP for the high
volume low cost ( and even mask-flow equiv, of factory program ),
and FLASH for the less stable code, and less price critical apps.
- jg
Johnny
>Jim Granville <no....@designtools.co.nz> wrote in message news:<goGhc.2729$cY5.2...@news02.tsnz.net>...
>>
Probably it is obvious to everyone else, but I am still wondering what
applications require both CAN and USB?
Is it for cars??
regards,
Johnny.
Ulf Samuelsson
> >
> >the ST devices sound very nice. Readin their press release though tell
> >me that they "plan to go into volume production in Q4 2004". In the
> >past ST has not been know for announcing their products conservatively
> >regarding schedule. So, there is already a device out there, running
> >60 MHz zero waitstates with 2 or 4 CAN interfaces from Philips
> >(LPC2129 and 2194). To me it seems that this time ST is not even a
> >fast follower, just a follower. There are however two things in the ST
> >devices that are very nice, the USB interface and the 12-bit ADC.
>
> Probably it is obvious to everyone else, but I am still wondering what
> applications require both CAN and USB?
>
> Is it for cars??
>
> regards,
> Johnny.
>
How about an industrial controller which needs to be able to do self
programming
and to set different parameters from a PC.
* USB Device Firmware Upgrade
* USB Device Communication Class
--
Best Regards
Ulf at atmel dot com
These comments are intended to be my own opinion and they
may, or may not be shared by my employer, Atmel Sweden.
Ulf Samuelsson
> "...with three times the market share of its nearest competitor, Atmel
> is also the world's leading supplier of 8 bit Flash uC"
>
> There is also no mention of revenue, or units, so that also needs to
> be clarified. viz : The 8 bit uC #1 bagging rights are currently split
> between revenue and volume/units claimants.
>
Atmel, I believe is saying number of chips, not revenue.
Cannot judge real numbers but many mobile phones has AVR chips inside.
5 million here, 5 million there quickly adds up over time.
With 400+ M phones delivered each year, this is OK.
AVR is also very popular in whitegoods.
Jim Granville
>> Atmel Brochures also make this claim :
>>"...with three times the market share of its nearest competitor, Atmel
>>is also the world's leading supplier of 8 bit Flash uC"
>>
>> There is also no mention of revenue, or units, so that also needs to
>>be clarified. viz : The 8 bit uC #1 bagging rights are currently split
>>between revenue and volume/units claimants.
>>
>
>
> Atmel, I believe is saying number of chips, not revenue.
> Cannot judge real numbers but many mobile phones has AVR chips inside.
> 5 million here, 5 million there quickly adds up over time.
> With 400+ M phones delivered each year, this is OK.
> AVR is also very popular in whitegoods.
Volumes, not revenue would help the case, and Atmel were probably
the first with sub $1 Flash uC. However, sub $1 flash is now the norm.
I still remain less than convinced, and it would take a ship load of
~30c ATtinyXX's to ramp the volumes.
(Atmel would, of course, exclude the OTP/Mask AVRs from this data ? :)
We may have to wait for releases from Microchip and Motorola on how
their 8 pin / sub $1 flash devices are ramping, for a 'reality check'.
Hitachi claim to have a 2:1 lead, and appx 25% of the market in
16 bit uC, but that's by revenue. They do tag some of their H8's
as 8 bit, but mostly that revenue goes into the 16 bit basket.
Note that the text above does NOT state AVRs only ( Tho it often
downstream-morphs into that ).
Atmel shipped 89C51s well before they shipped AVRs.
That combine effect also helps lift the numbers.
-jg
Johnny
Johnny
<u...@NOSPAMatmel.com> wrote:
>> >Hi Jim,
>> >
>> >the ST devices sound very nice. Readin their press release though tell
>> >me that they "plan to go into volume production in Q4 2004". In the
>> >past ST has not been know for announcing their products conservatively
>> >regarding schedule. So, there is already a device out there, running
>> >60 MHz zero waitstates with 2 or 4 CAN interfaces from Philips
>> >(LPC2129 and 2194). To me it seems that this time ST is not even a
>> >fast follower, just a follower. There are however two things in the ST
>> >devices that are very nice, the USB interface and the 12-bit ADC.
>>
>> Probably it is obvious to everyone else, but I am still wondering what
>> applications require both CAN and USB?
>>
>> Is it for cars??
>>
>> regards,
>> Johnny.
>>
>
>
>How about an industrial controller which needs to be able to do self
>programming
>and to set different parameters from a PC.
>
>* USB Device Firmware Upgrade
>* USB Device Communication Class
Yes. thats true. I suppose there are CAN applications like Devicenet
as well.
Anyway I suppose like a lot of people, I am currently faced with the
same problem, ie. death of the serial port on PCs. Maybe you can
tell me when Atmel will produce an ARM7 processor with USB, a decent
amount of Flash, and RAM on board?
I would like about 1MB Flash and 256kB RAM.
regards,
Johnny.
Ulf Samuelsson
> >programming
> >and to set different parameters from a PC.
> >
> >* USB Device Firmware Upgrade
> >* USB Device Communication Class
>
>
> Yes. thats true. I suppose there are CAN applications like Devicenet
> as well.
>
> Anyway I suppose like a lot of people, I am currently faced with the
> same problem, ie. death of the serial port on PCs. Maybe you can
> tell me when Atmel will produce an ARM7 processor with USB, a decent
> amount of Flash, and RAM on board?
>
> I would like about 1MB Flash and 256kB RAM.
Unfortunately, that is "indecent" with current 0,18u Embedded Flash process.
I think that you will see devices with USB + CAN and 256 kB Flash quite
soon.
(AT91SAM7A3 will arrive this year).
Atmel has ARM7 + USB for large volume ASSP customers.
You could try the ARM9 (AT91RM9200) which of course needs external
flash/SRAM.
> regards,
> Johnny.
Jim Granville
<snip>
> Anyway I suppose like a lot of people, I am currently faced with the
> same problem, ie. death of the serial port on PCs. Maybe you can
> tell me when Atmel will produce an ARM7 processor with USB, a decent
> amount of Flash, and RAM on board?
>
> I would like about 1MB Flash and 256kB RAM.
That's quite a bit - currently off most single chip road-maps.
Intel, & Atmel, do have stacked-die ARM/FLASH/RAM devices where the
process and yields can be optimised for each die in the stack.
Intel have some numbers in that series that make 1MB/256K look small :)
-jg
Johnny
<u...@NOSPAMatmel.com> wrote:
>> >How about an industrial controller which needs to be able to do self
>> >programming
>> >and to set different parameters from a PC.
>> >
>> >* USB Device Firmware Upgrade
>> >* USB Device Communication Class
>>
>>
>> Yes. thats true. I suppose there are CAN applications like Devicenet
>> as well.
>>
>> Anyway I suppose like a lot of people, I am currently faced with the
>> same problem, ie. death of the serial port on PCs. Maybe you can
>> tell me when Atmel will produce an ARM7 processor with USB, a decent
>> amount of Flash, and RAM on board?
>>
>> I would like about 1MB Flash and 256kB RAM.
>
>Unfortunately, that is "indecent" with current 0,18u Embedded Flash process.
>I think that you will see devices with USB + CAN and 256 kB Flash quite
>soon.
>(AT91SAM7A3 will arrive this year).
>
>
>Atmel has ARM7 + USB for large volume ASSP customers.
>
>You could try the ARM9 (AT91RM9200) which of course needs external
>flash/SRAM.
Thanks for the info. But how come there are other AT91 chips with 2M
Flash and 256KB RAM?
Basically I am looking for a highly integrated and low-cost solution.
I think a SOC approach is a great from a reliability and manufacturing
point of view. External SRAM also tends to be expensive especially
towards the end of its production life a few years from now, which
could become a problem for us.
I need a lot of flash for storing about 400KB of configuration data.
So an external Flash might be a decent solution.
regards,
Johnny.
Jim Granville
> On Thu, 29 Apr 2004 17:15:04 +0200, "Ulf Samuelsson"
>>>I would like about 1MB Flash and 256kB RAM.
>>
>>Unfortunately, that is "indecent" with current 0,18u Embedded Flash process.
>>I think that you will see devices with USB + CAN and 256 kB Flash quite
>>soon.
>>(AT91SAM7A3 will arrive this year).
>>
>>
>>Atmel has ARM7 + USB for large volume ASSP customers.
>>
>>You could try the ARM9 (AT91RM9200) which of course needs external
>>flash/SRAM.
>
>
> Thanks for the info. But how come there are other AT91 chips with 2M
> Flash and 256KB RAM?
These are dual die solutions in BGA packages. See my other post about
Intel also making stacked-die XSCALE (ARM) pocessors for Cell Phones.
>
> Basically I am looking for a highly integrated and low-cost solution.
> I think a SOC approach is a great from a reliability and manufacturing
> point of view. External SRAM also tends to be expensive especially
> towards the end of its production life a few years from now, which
> could become a problem for us.
>
> I need a lot of flash for storing about 400KB of configuration data.
> So an external Flash might be a decent solution.
Serial FLASH is getting quite cheap, and there are ARM uC from Philips
and STm with 256K bytes FLASH and 64K Bytes of SRAM, which I would have
called 'quite a lot' in a single chip device.
Of course, one designers' "Quite a lot" is anothers "not nearly enough" :)
Why do you need 256K Bytes of RAM ?
-jg
Heinz-Jürgen Oertel
>
> Refs:
> http://biz.yahoo.com/prnews/040330/sftu033_2.html
> http://www.st.com/stonline/books/pdf/docs/10350.pdf
> http://www.st.com/stonline/products/support/micro/arm/str7_10.htm
> http://www.st.com/stonline/press/news/year2004/p1431p.htm
>
Jim,
You mentioned the ST devices STR710FZ2 and STR712FR1. Following your links
and searching the ST site gives me no answer about the CAN module
integrated. Can you help. Or may be one of the ST listeners?
The data sheet only says "CAN 2.0B" compatible. That says nothing about what
I can expect from the CAN controller. Besides this, CAN2.0B is the old
Bosch Spec. Now a CAN controller should be compatible with ISO DIS 11898.
This standard is available since years. No idea why manufacturers follow
the old Bosch Spec.
Regards
Heinz
--
with best regards / mit freundlichen Grüßen
Heinz-Jürgen Oertel
+===================================================================
| Heinz-Jürgen Oertel port GmbH http://www.port.de
| mailto:o...@port.de
| phone +49 345 77755-0 fax +49 345 77755-20
| Regensburger Str. 7b, D-06132 Halle/Saale, Germany
| CAN Wiki http://www.CAN-Wiki.info
| Newsletter: http://www.port.de/engl/company/content/abo_form.html
+===================================================================
Johnny
I think stacked dies would be fine for my application.
> Why do you need 256K Bytes of RAM ?
the product has two functions, one is a network bridge/gateway and
also as something similar to a PLC controller.
I am guessing that I need around 120k for cacheing of status for up to
1000 network nodes. I also need at about 40kB for buffering of
network data. At the early stages of the project is is difficult to
tell exactly how much we need, so more is generally better.
regards,
Johnny.
Johnny
<no....@designtools.co.nz> wrote:
>Johnny wrote:
>> On Thu, 29 Apr 2004 17:15:04 +0200, "Ulf Samuelsson"
><snip>
>>>>I would like about 1MB Flash and 256kB RAM.
>>>
>>>Unfortunately, that is "indecent" with current 0,18u Embedded Flash process.
>>>I think that you will see devices with USB + CAN and 256 kB Flash quite
>>>soon.
>>>(AT91SAM7A3 will arrive this year).
>>>
>>>
>>>Atmel has ARM7 + USB for large volume ASSP customers.
>>>
>>>You could try the ARM9 (AT91RM9200) which of course needs external
>>>flash/SRAM.
>>
>>
>> Thanks for the info. But how come there are other AT91 chips with 2M
>> Flash and 256KB RAM?
>
>These are dual die solutions in BGA packages. See my other post about
>Intel also making stacked-die XSCALE (ARM) pocessors for Cell Phones.
I some negative feeling about Xscale because it seems to me that they
are used for consumer products such as PDA's and phones. This is a
technology that is still evolving. They might be good value for money
at the moment while the volumes are high, but I don't want to have to
revise the hardware design when the processor becomes obsolete 2 or 3
years from now. I would hope there is a chance that whatever silicon
I choose will still be available in 5 years from now. I know people
who had bad experiences with StrongARM designs from Intel being made
obsolete.
regards,
Johnny.
Jim Granville
<snip>
> I some negative feeling about Xscale because it seems to me that they
> are used for consumer products such as PDA's and phones. This is a
> technology that is still evolving. They might be good value for money
> at the moment while the volumes are high, but I don't want to have to
> revise the hardware design when the processor becomes obsolete 2 or 3
> years from now. I would hope there is a chance that whatever silicon
> I choose will still be available in 5 years from now. I know people
> who had bad experiences with StrongARM designs from Intel being made
> obsolete.
A good point. I'd talk with Intel, as they are still a big player in
Embedded MCU/MPU/Comms controllers - it's just mormally buried under
the Pentium revenues - under that even $1B looks small :)
Intel do have embedded product lines, where they extend the design
lifetimes.
I think some of the StrongARM issues were fab closure related, in the
move from Digital to Intel ? XScale is all Intel, and it is being
used in some longer-life semi-consumer products.
-jg