Arduino Due

[James Michener]

I understand the Due is being discontinued. Are there plans for a
replacement?

--Jim

[Collin Kidder]

??? This is the first I've heard of it. But, having seen your email I
just looked for other sources and it does appear that a range of
people believe this to be true. I'm not sure why they would do a thing
like that. It seems like a whole lot of people are using my due_can
library so you'd think that there must have been some interest out
there. But, I'm sure it wasn't as big a seller as the other boards.
I'm not really sure why as it is right around the price of the 8 bit
boards but many times more powerful.

Anyway, there are clone makers who are still producing their boards so
the Due won't die. But, I wonder what this means for continued support
in the IDE?

> --
> You received this message because you are subscribed to the Google Groups "Developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to developers+...@arduino.cc.

[Massimo Banzi]

Hi

When we rebuilt our manufacturing network we took the chance to redesign the product line to simplify it.

We also had to decide the sequence of release giving priority to products that are most popular.

Having said this we see an interest for the Due from the resellers and we’re working on a solution

In the meantime the SAM core is still supported and it’s getting updated even if we’re not selling the hardware

m

[Paul Stoffregen]

On 02/22/2016 06:16 PM, Collin Kidder wrote:
> But, I wonder what this means for continued support
> in the IDE?

Judging by the fact that 1.6.7's Tools > Board menu still has "Arduino
Duemilanove or Diecimila" and "Arduino NG or older", it certainly seems
Arduino isn't in the practice of removing software support for
discontinued boards!

[Massimo Banzi]

Yes we are nice people in an almost masochistic kind of way :) 

Jokes aside, there are times that ,if we don't make things ultra clear ,people draw their own conclusions and they become facts even if we never said anything like that

M

--
You received this message because you are subscribed to the Google Groups "Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to developers+...@arduino.cc.

--
Sent from my tablet

[Paul Carpenter]

Well as from your sites the Genuino Range does not list Due but Arduino range does I wonder if grey imports will occurr or more clones.

Personally UNO and Zero has too few I/O etc for most of my uses, Due has I/O timers, memory and interfaces I generally need (multiple serial, I2C/TWI, SPII yes and better speed etc.

Yes Mega has the I/O but lacks the speed and has the overheads of the AVR architecture, that has made some projects more difficult.

Having fixed speed issues in Liquid Crystal library, it now makes 4 bit interfacing more practical on all platforms

Suitable replacement with similar I/O and memory is a +1 from here

[Andrew Kroll]

Don't forget dual analog outputs!

--

You received this message because you are subscribed to the Google Groups "Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to developers+...@arduino.cc.

--

Visit my github for awesome Arduino code @ https://github.com/xxxajk

[Paul Carpenter]

With the 12 bit resolution for Analog and PWM.

Yes Dual analog as well, basically as close as possible I/O and interfaces
I would prefer with more timer pins (inputs, clks, outputs) being available
on more timers for more of the modes that are supported.

Andrew Kroll wrote:
> Don't forget dual analog outputs!
>
> On Tue, Feb 23, 2016 at 5:58 AM, Paul Carpenter
> <pa...@pcserviceselectronics.co.uk

[Massimo Banzi]

Arduino/Genuino doesn’t sell the Due full stop

There is a number of companies using the Arduino name and logo around the world who make boards that are called Due

As I said we’re evaluating an improved product for this market segment 

m

[Paul Carpenter]

See https://www.arduino.cc/en/Main/Products

Compared to https://www.arduino.cc/en/Main/GenuinoProducts

You have a website issue

Massimo Banzi wrote:
>
> Arduino/Genuino doesn’t sell the Due full stop
>
> There is a number of companies using the Arduino name and logo around
> the world who make boards that are called Due
>
> As I said we’re evaluating an improved product for this market segment
>
> m
>
>
>> On 23 Feb 2016, at 11:58, Paul Carpenter
>> <pa...@pcserviceselectronics.co.uk

[Paul Carpenter]

Oh which links through to

https://www.arduino.cc/en/Main/ArduinoBoardDue

With a Shop Now buton

That links through to

https://store.arduino.cc/index.php?main_page=product_info&cPath=11_12&products_id=243

Listing price and out of stock

Does NOT say we no longer sell this board.

--
Paul Carpenter | pa...@pcserviceselectronics.co.uk
<http://www.pcserviceselectronics.co.uk/> PC Services
<http://www.pcserviceselectronics.co.uk/pi/> Raspberry Pi Add-ons
<http://www.pcserviceselectronics.co.uk/fonts/> Timing Diagram Font
<http://www.badweb.org.uk/> For those web sites you hate

[Massimo Banzi]

Thanks for pointing out the mistake… I’ll have it fixed soon

thanks

m

[Mike]

I don't want to get off topic but this hits on the last few posts.  Now that a) we've moved technically into more powerful devices and b) people are comfortable using the Arduino ecosphere, that the "Shield" form factor that drove a good amount of popularity is actually holding connectivity back?  I believe the Arduino Zero processor has more pins that is brought out to the standard headers (one more was added - ATN). Also Arduino 101, a very powerful processing unit with only limited I/O to broken-out pins.  And I do not believe personally that wifi or bluetooth substitute significantly for pins.

Maybe we should consider Shield 2.0.  Possible backwards compatibility but allowing definition of more analog out, timers, uarts, features that modern processors provide and may be useful for beginners or engineers?  I don't want to see Beaglebone or C.H.I.P level header density though, very crowded.

[Roger Clark]

There are other ARM boards that are programmable via the Arduino IDE.

The Teensy,the PIC32 , the STM32  (I suspect there are others as well)

I believe that the Teensy is the most compatible / most well supported.

[Pieter van der Meer]

yeah, but teensy is less powerful.. far less pins.. great libs, though, as far as i can judge.. due is so great because there is such a wealth of pins _and_ the multiplexing of those pins is just very well done. SAM3X is a very nice piece of kit. 

--

Pieter van der Meer

Tasty Chips Electronics

www.tastychips.nl

[q2dg2b .]

About the commented web issue (that is to say, the shown difference between Arduino and Genuino products list) I've email months ago to sup...@arduino.cc. I even twitted to @arduino and @dcuartielles. I only asked if Genuino products would be the same than Arduino products. No response. At all. Oh, my...

[Thibaut VIARD]

Atmel SAM3X is indeed a very nice device but several others were born since.

If it could be replaced, I would see Atmel SAM4E (CM4, more RAM, 120MHz) instead if users consider ethernet and CAN associated with more power.

If people are interested in much more power, Ethernet, CAN and USB high speed, Atmel has a flash micro monster: the SAMS70/E70 (CM7, 300MHz), but its use may be a bit tricky regarding the Cortex-M7 core and the general purpose approach of Arduino framework (I mean able to serve almost all purposes).
I think this one is kind of overkill...

If Ethernet and CAN aren't part of the requirements, there are other devices in Atmel portfolio. Well, I have always missed the Ethernet on the Due but the trend is now more about Wifi and Bluetooth.
802.15.4 is also very interesting but needs bridge to communicate with computers and smartphones.

Any thoughts?

I understand the question here is kind of: what would be your perfect Due 2.0?

Thibaut

[Dennis Lee Bieber]

On Tue, 23 Feb 2016 12:24:43 +0100, Mike
<mydigi...@gmail.com> declaimed the
following:

>
>Maybe we should consider Shield 2.0. Possible backwards compatibility but
>allowing definition of more analog out, timers, uarts, features that modern
>processors provide and may be useful for beginners or engineers? I don't
>want to see Beaglebone or C.H.I.P level header density though, very crowded.
>

Less well known, I suspect, but maybe closer comparison than the
Beaglebone (slow GPIO going through Linux file-system, or fast but cryptic
PRU module) might be the TI TIVA TM4C123G Launchpad.

http://www.ti.com/tool/ek-tm4c123gxl

ARM Cortex M4F, WAY too many timers (6 64-bit and 6 32-bit, each of
which can be split into a pair of half-width timers), 80MHz, 256K flash,
32K RAM, support for CAN. They forked the older Arduino IDE into Energia,
for those that don't want the hassle of Code Composer Studio.

And cheap if you can get them direct <G> Though it needs to be pointed
out that TI doesn't bill the Launchpads as a final product -- they are
"evaluation boards" for the processor.

Though the Due is probably easier to "get into" -- TI tends to put too
many functions onto each I/O pin, requiring cryptic configuration
statements in code initialization just to activate the proper mode. And, of
course, Launchpad BoosterPacks are not Shield compatible (nor Beaglebone
Cape compatible).
--
Wulfraed Dennis Lee Bieber AF6VN
wlf...@ix.netcom.com HTTP://wlfraed.home.netcom.com/

[Paul Carpenter]

Anything requiring wifi/bluetooth is only efficiently done with it on board
or shield with CE/FCC certification as minimum. Worldwide product type
certification adds another development phase cost to base board or shield.

Most serious users in industrial/control only put wifi/bluetooth on anything
that PHYSICALLY moves around, It is more secure and less prone to other
problems like interference or snooping to use a wire (Ethernet, serial,
modbus, CAN, etc). Networking Ethernet or Wifi or Bluetooth needs bigger RAM
and more layers of software than majority require.

[Dieter....@online.de]

Hi all,

I am a user only, but will give you an idea what some users do with your
Arduino Due (look at the image):
(This is the development environment of my garden watering system. The live
system is running well since last spring.)

On the left bottom, there is the Arduino Due with an Arduino Ethernet shield
R3:
- Ethernet: running with 14MHz
- SD-Card: running with 42MHz

On top there is one shield with:
- I2C RTC: DS1307 (100kHz)
- Watchdog and power supervisor MAX 691 CPE
- A third SPI device: 128kB NVSRAM 23LCV1024 with battery backup (16.8 MHz)
(alternatively: 128kB EEPROM 25LC1024 or 4MB EEPROM 25VF032B80-4L-S2)
- 1.2 Volt reference LT 1004 CZ-1,2 and voltage divider to measure the
12-Volt supply, 5-V, 3.3V, and CMOS Battery

The second shield with:
- I2C 128kB EEPROM 24LC1025-IP (400kHz)
- connected:
- A light sensor with analog logarithm output
- 868Mhz receiver FS20WUE (4800Baud) for up to 8 weather sensors

Two boards with 15 relays and a small controller to disable the relays on
startup.

The power control board on top left is a USV for 12-Volt:
- Precisely controls power up and power power and buffers the supply voltage
for more the 200ms.
A power down and up cycle is initiated by software at midnight.

The watering system is designed as a high available system (with pure
hardware).
The watering system is highly configurable and controllable via around 50
HTML-pages with full logging support to SD-disk cached via SDRAM.
The application uses only your basic libraries and "WebDuino".
The development environment is based on Atmel Studio with Visual Micro based
on your environment (around 15000 lines of code using 225kB flash).

The system has to run more than a month while we not at home. The live time
should be up to 30 years.

- You cannot use a computer loading the OS and application from an SD-card
(Raspberry etc.). The SD-card will fail after some thousand write operation
on the same sector.
- I need very reliable i/o's. Complex structures are sensitive to fail.
- The live time of an USV based on a battery is very limited and the failure
rate high. You cannot use it.
- All of your 8-bit boards are too slow for the highly dynamic and complex
web pages.
- An OS based system does not have the long term reliability like a pure
microcontroller system that reboots every night.
- A multiprocessor core is not suitable for a OS-less system.

The answer for me was the Arduino Due and only that one. I did not find any
other system for that price in the market.

The Arduino Due is still your flagship!

(But the board is not properly promoted, I think. How needs that board if
there is not have enough functionality.)

I would be very disappointed if you cancel then board. It would be better to
enhance the board e.g. with a full sized shield and the devices I listed.

Best Regards,

Dieter Burandt


-----Ursprüngliche Nachricht-----
From: Paul Carpenter
Sent: Tuesday, February 23, 2016 3:06 PM
To: devel...@arduino.cc
Subject: Re: [Developers] Re: Arduino Due

[q2dg2b .]

If it was based on a DIP chip (as -only- UNO is) it would be fantastic: easy fortinkering, focused on education, affordable, customizable...

[Massimo Banzi]

Thanks for the feedback this is very valuable insight!

The Due is not dead in our hearts. We just have to make decisions based on the time we have available.

One point which is very important right now and somebody touched upon is proper support.

In the “previous phase of Arduino” we were forced to launch a lot of products in a short time to satisfy the requests of manufacturing / wholesale which resulted in a lot of potentially interesting products with not enough support, care and love.

The Due never had the same level of support, examples and refinement of the previous products.

We are trying to have less products or product families but have more support, better tutorials, more compatibility across product families. This requires a different approach.

We know that market segment is there and we need to deal with it, Atmel has some nice new parts in that segment so we’ll see when time comes. (in the meantime we could just make the Due )

Let’s see but keep the feedback coming

m

[q2dg2b .]

Ok, Massimo, thanks.

So...is there some roadmap about which products will "stay" and which not?
Namely, which is the plan?

[Paul Stoffregen]

On 02/23/2016 05:36 AM, Thibaut VIARD wrote:

I understand the question here is kind of: what would be your perfect Due 2.0?

I've often ponded a similar question regarding a next generation Teensy board.  In fact, I put quite a bit of time into considering the SAM E70, but I ultimately decided against it based on the prohibitive cost and the huge amount of engineering work needed to support a whole new peripheral set.  Cortex-M7's complex bus structure that likely needs memory barrier instructions to prevent hardware race conditions was also a concern.  I suppose that's an inevitable future when (if) microcontrollers eventually move to 45 and 28 nm process nodes...

But I believe the reality today, and I hope I don't offend anyone too badly by saying this, is we lack the software infrastructure to really utilize such powerful hardware, or even the current Arduino Due, Teensy 3.x, STM32, etc.  Of course, piles of expert-oriented code like CMSIS & ASF exist, as do numerous RTOS systems.  But those are all a poor fit to Arduino's design goal to empower novices.  I have actually put quite a lot of thought into this over the last couple years, but I'm going to stop short of actually proposing anything in this message, for fear of touching off an epic bikeshedding avalanche.

My main point is a perfect Due 2.0 looks a lot more like intense software design & development than any specific hardware!

[James Michener]

I started this thread....

First a comment to Paul:  If you want RTOS, multi thread, multi user.... use Linux...the BeagleBone, Pi etc...

I thought I would state why I love about the Due.   I hoped that the Due was the replacement for the Mega.  The Mega is slow, 8 bit,  PROGMEM etc.... which is to say, I would rather have seen the Mega go away.

The Due is fast, more SRAM, more Flash, 12 bit converters (both A/D and D/A), on board switching regulator...  to a programmer you don't have to explain to people what PROGMEM is all about. 

Now I use the DUE for handheld, battery operated data gathering and presentation devices.  Where you need a graphics display (parallel)  A/D and D/A and can be very light on power consumption.   I wish it had an RF data interface... either WiFi or bluetooth.   I wish it had EEPROM... both can be supplied in the shield.

Now... a Linux based open source board would work in this application...however,  I have not had good luck keeping these things running.  The boxes / OS are too complex.   Someone always turns off the power at just the wrong point in time so that something goes wrong with the file system.  Murphy is just alive and well.   While I can keep it running,  there are just too many service calls.   The totally embedded solution... no large overhead OS ... has proven to be very robust.  It is the combination of the simple Arduino low level libraries and a fast low power processor that makes my heart race.

I see where Arduino is aiming more entry market, novice users... but I love the Due even with the current level of support.

Amore Due!
Jim Michener

[Massimo Banzi]

Paul

On 23 Feb 2016, at 18:15, Paul Stoffregen <pa...@pjrc.com> wrote:

but I'm going to stop short of actually proposing anything in this message, for fear of touching off an epic bikeshedding avalanche.

I feel you :)

Maybe the solution is to apply incremental updates to the current Arduino architecture as opposed to try to redesign something from scratch. Engineers love re-designing from scratch but that rarely brings the desired results 

m

[Paul Carpenter]

Current on bench design is Theatre/small event lighting controller
two DMX channels several triggers, buttons LEDs, LCD, Ferro Electric
RAM on I2C, so using 3 serial, I2C, SPI (24 LEDs), several timers and nearly
all the GPIO. This is custom board as very oversized 'shield', with
additional boards for Power and triggers, also 2 x DMX units.

One other mode is to act as a DMX monitor detailing timing and specification
misses or as a device on DMX. On one DMX universe, whilst controller can
deal with 2 universes. Because of this looking at tidying up Serial code
to make some classes more generic not duplicated and easier to change buffer
sizes.

[Paul Carpenter]

Well things like DIP does not worry me as gave up on new parts being DIP
nearly 15 years ago and where necessary do prototypes with Schmart boards
if making a PCB is not allowed.

All examples and libraries I currently work on are tested where possible on
Uno, Mega and Due unless Due specific.

Recently

- Supplied some minor fixes for SAM compiler warnings
- checked my PS2 keyboard libraries on all platforms
- created a simple scheduler with example on DUE
- reviewed the LiquidCrystal to get it to follow the HD44780 spec better
and give better performance for ALL platforms.
- Next stage look at good Input capture and other timer modes for Due
- Then review Serial handling first on Due to use a better Ringbuffer class
(as other similar class are in other core modules), improve error handling
and status. Then look at how best to port to AVR family)

These days AVR is 'if I have to' because it is a library to be published,
or some student or teacher thinks Arduino is Uno only.

Getting away from the PROGMEM and similar issues is a big bonus.

This is ESPECIALLY true when mentoring novices, low grade I get who are
17-18 for their course projects. difficult enough getting them to do any
design. Let alone realise that AVR is different to their PC, Phone, Tablet
and has all sorts of restrictions, and CSV import from file over serial is
not what they think its is.

These days Uno will become the rate limiting factor for acceptance as a lot
of people not experienced in what the offerings are think Arduino is Uno
ONLY and who is this strange outfit Genuino.

Having 80MHz and above ARM with I/O like Mega shows more things can be done.
Maybe you need one with onboard SDCard interface to show how different to
Uno the range can be.

[q2dg2b .]

Sorry if I'm offtopic...only one aclaration: I told about DIP to get back to Arduino origins: a hackeable, diy, really open platform. Nowadays it isn't indistinguishable from any other free boards, as it's just a final product: I can't reproduce a Due in my classroom with discrete components. This aspect takes away magic...But surely I'm wrong or misfocused, sorry.

[Paul Carpenter]

In UK at schools soldering irons are heavily frowned upon hot glue guns
not so, which ends up with most places not wanting to do any hacking with
dip parts they cannot put on small breadboard.

I have sold something like 40 x Arduino in box kit see

https://www.facebook.com/pcservicesreading/photos/a.426416640744094.106311.326013924117700/769150216470733/?type=3&theater

I have done proof of principles on breadboard with Schmart boards for
surface mount devices, very few new devices of any large complexity come out
as DIP these days.

Picture below is fully USB 2 compliant board from about 8 years ago
as proof of principle consisting -

- USB 2 HUB
- USB FTDI device
- FPGA for multi signal generation from SPI and I2C to torque and
temperature bridge drives for custom chip tester
- Device under test

That project ended up being a long one developing Mixed signal ASIC
production testing equipment to military temperature specification.
That is all the test equipment from Wafer probe to burn in and final test.

The device under test was not a new design and that was surface mount
on its third ASIC respin.

q2dg2b . wrote:
> Sorry if I'm offtopic...only one aclaration: I told about DIP to get
> back to Arduino origins: a hackeable, diy, really open platform.
> Nowadays it isn't indistinguishable from any other free boards, as it's
> just a final product: I can't reproduce a Due in my classroom with
> discrete components. This aspect takes away magic...But surely I'm wrong
> or misfocused, sorry.
>
> 2016-02-23 21:11 GMT+01:00 Paul Carpenter
> <pa...@pcserviceselectronics.co.uk

> <mailto:pa...@pcserviceselectronics.co.uk>>:

[Tom Igoe]

The processor for the Due doesn’t come in a DIP format. Sadly, most of the current processors don’t.

t.

[Andrew Kroll]

Use conductive carbon-based non-toxic glue instead. Its water based and needs no heat to apply and dries without any special tools, etc. Also makes it possible to do circuits on plain paper.

[ODwyerPW]

Although very frustrated in the beginning with DUE as so much stuff just didn't work, I stuck with the DUEs for my own advanced project work, leaving the Unos for classes and the Minis for prototyping. However, it always appeared as if it were the red-headed step-child of the Arduino Team. I've purchased 7 DUE's (4 originals and 3 Taiguinos). I know it's not much, especially as I've purchased more than  50 Unos (5 originals and about 50 clones) and more than 50 Arduino Pro Minis (maybe 2 originals, the rest clones).

If the DUE is continued for good, I'd like to put my hat in for one who would like to see a replacement offered with the I/O. A couple of items that could have been included in a flagship type DUE would have been to utilize the built-in Ethernet of the SAM3X8E with a physical interface and jack present on the board, the built-in SDIO interface with a MicroSD slot present on the board, and a method for interacting with non-volatile memory (physical EEPROM, emulated EEPROM, or a built-in hotwire of the 16U2 chip to hijack it's EEPROM for sketch use). Many would argue using Bluetooth or Wifi is better, but Ethernet is ubiquitous... THe closest board I've used to perfection is the TI product's Launchpad.. though they didn't bring out the SD interface either.

Maybe consider an Arduino QUATTRO, based on ATSAM4E16E. Gives a nice bump in speed over the 3X8E. Utilize the interfaces so that we don't have to use up pins to give us EEPROM (you guys can figure out how to give us some EEPROM), MicroSD and Ethernet...   Mega Layout (like the DUE) would be fine. My 2 cents.

On Monday, February 22, 2016 at 6:18:44 PM UTC-7, James Michener wrote:

I understand the Due is being discontinued.   Are there plans for a
replacement?

--Jim

[Phillip Stevens]

Just for comment or discussion,

it would seem that the FT903/8 MCU could be the right size / shape to replace the Due MCU.

http://www.ftdichip.com/Products/ICs/FT90x.html

Sole supplier, but effectively so are all of the ARM core alternatives.

Has the right amount of interfaces, including two 10 bit DACs and integrated USB.

Fast, uses shadow RAM for 0 wait state operation.

GCC based compiler and environment.

On Wednesday, 24 February 2016 05:04:15 UTC+11, James Michener wrote:

I started this thread....

[Dennis Lee Bieber]

On Tue, 23 Feb 2016 15:52:23 +0100,
<Dieter....@online.de> declaimed the
following:

>- You cannot use a computer loading the OS and application from an SD-card
>(Raspberry etc.). The SD-card will fail after some thousand write operation
>on the same sector.

A good SD card should have wear leveling internally, so rewrites should
"move around" in the physical card. However, the type of card may have an
effect -- a card using LARGE write sectors may be more affected by rewrites
than one using smaller sectors. The number of "open streams" the card
supports could also be a factor -- Class 10 cards are optimized for large
streaming data and may be designed with only 2 streams (the FAT and the
data itself). If used with random file access of more than FAT and one data
stream, this could result in lots of leveling/rewrite operations as the
card closes a writable sector, then opens a new one copying partial data
from the old one into it before adding new data (and then erasing the old
location). Being able to keep lots of streams open allows for continued
writes without the copy/erase cycles.

[Roger Clark]

There are so many other processors you could use, so I don't see why you'd go for a company like FTDI as opposed to Atmel or Microchip, or Ti, or STM

Take a look on AliExpress and there are thousands of "development" boards that have 2 x 10 bit DAC's and USB etc.

e.g. Take a look at the ST32F407 e.g.

http://www.aliexpress.com/w/wholesale-stm32f407-development-board.html?spm=2114.01010208.0.331.Q3YeAh&initiative_id=AS_20160223183107&site=glo&SortType=price_asc&shipCountry=au&SearchText=stm32f407+development+board&CatId=400103

Running at 168Mhz.

(Note, I'm sure that the are PIC32 boards 
http://www.aliexpress.com/w/wholesale-PIC32--development-board.html?spm=2114.01010208.0.327.bCa0c2&site=glo&SearchText=PIC32++development+board&g=y&SortType=price_asc&groupsort=1&initiative_id=SB_20160223183243&shipCountry=au

and probably TI boards as well

But I generally find the STM32 boards to have good price / performance ratios

--

[Dennis Lee Bieber]

On Tue, 23 Feb 2016 10:04:08 -0800, James Michener
<k9...@sbcglobal.net> declaimed the following:

>I started this thread....
>
>First a comment to Paul: If you want RTOS, multi thread, multi user....
>use Linux...the BeagleBone, Pi etc...
>

There is a big gap between a hard RT OS and common Linux. RTOS (not
multi-user stuff) is optimized for interrupt responses and layered
(priority) tasks. The larger ARM systems probably beg for an RTOS to be
available <G> {Though the [apologies for mentioning it a second time] TIVA
TI-RTOS lesson book gets a bit tedious -- blink an LED: 1) without RTOS, 2)
using sleep calls in an idle task, 3) using sleep calls in a "non-Idle"
regular thread/task, 4) using a hardware interrupt on a timer, 5) using a
HWI on timer to signal a software interrupt, 6) ... message queues, event
flags, etc. to regular tasks blocked waiting for some signal...}

>I thought I would state why I love about the Due. I hoped that the Due
>was the replacement for the Mega. The Mega is slow, 8 bit, PROGMEM
>etc.... which is to say, I would rather have seen the Mega go away.
>
>The Due is fast, more SRAM, more Flash, 12 bit converters (both A/D and
>D/A), on board switching regulator... to a programmer you don't have to
>explain to people what PROGMEM is all about.
>

I just had to look it up... I can still see it used for the Due... To
my mind there is no excuse for having any compile time static data stored
in RAM. But for horrors, ever look at how a (byte-wide) look-up table is
done in the PIC? (Which really lacks the RAM to store tables -- the only
RAM is basically a large register block).

The one counter I have to the Due (and other ARM based boards) is the
3.3V supply. I'm someone who still has a supply of 5V 74xx TTL chips lying
around -- I avoided CMOS chips, even if 74Cxx were pinout equivalents, due
to the static discharge sensitivity. {And this is were a Beaglebone is
obscene -- the A/D input is limited to 1.8V, on a 3.3V board... At least,
as I recall, the 5V Arduino's take 5V on the A/D inputs}

[Dieter Burandt]

Hi Dennis,

I agree with you in reagrd to the SD-card.

In my system I use the SRAM to store configuration data and in addition use the SRAM to write the logs into a memory. Normally at midnight I write the logs from the SRAM to the SD-card. Normally, I don't delete the log data at all. So the SD-card is reliable.

Writing 256 byte to the SRAM cost around 150 micro seconds. Much faster than writing to an EEPROM or to the SD-card.

I didn't found any information about Wear Level Handling of SD-cards. So your information is very interesting.


Using a multitasking OS is an option, but is getting very expensive if you need a high available system.
With today’s Linux variants a very rough estimation of the required CPU power you need around 10 times the power.
Let's say for a 84MHz Arduino Due you need a 800MHz CPU. This is surprising, but comes mainly from the expectation of the user, who expect tools as known from Windows or so.

A very, very long time ago, I worked with a 4MHz system with a real time OS. That did work very well, but only using Teletype Terminals.

A general question for the Arduino group is, do they like to spend a lot of power into a new board with a new processor, or do the like to enhance it with minimal effort and getting a (board and) a shield that fit the requirements of the users.

If you have a combination of a board and a shield that contains an RTC, a battery buffered SRAM with proper size, an EEPROM, a well working watchdog and power fail logic, and ..., then you are on the right way even though, not all need all of the additional devices. Surly no big problem to add jumpers to the board to enable specific devices or not.

But that is a decision of the Arduino Group.


Best Regards,

Dieter Burandt

"Dennis Lee Bieber" schrieb im Newsbeitrag news:<3t4qcbllj3u5bpfhf...@4ax.com>...

> On Tue, 23 Feb 2016 10:04:08 -0800, James Michener

> declaimed the following:
>
> >I started this thread....
> >
> >First a comment to Paul: If you want RTOS, multi thread, multi user....
> >use Linux...the BeagleBone, Pi etc...
> >
> There is a big gap between a hard RT OS and common Linux. RTOS (not
> multi-user stuff) is optimized for interrupt responses and layered
> (priority) tasks. The larger ARM systems probably beg for an RTOS to be

> available {Though the [apologies for mentioning it a second time] TIVA

[Thibaut VIARD]

Dieter,

The inner wear leveling of sdcard is totally hidden to user: an sdcard is composed of a nand flash (single or multi layer) and a controller.
This controller handles wear leveling, spare blocks to replace bad blocks and all ugly features coming with nand flash techno.

[Dieter....@online.de]

Hi Viard,

 

What you are writing is also my knowledge.

 

If that is correct, a Linux system on an SD-card is a good toy, but not more. You can use the SD-card as long as you are aware of it and may be a good, cheap variant for learning. But not more.

 

Dieter Burandt

 

 

From: Thibaut VIARD

Sent: Wednesday, February 24, 2016 2:03 PM

To: devel...@arduino.cc

Subject: Re: [Developers] Re: Arduino Due

 

Dieter,

The inner wear leveling of sdcard is totally hidden to user: an sdcard is composed of a nand flash (single or multi layer) and a controller.
This controller handles wear leveling, spare blocks to replace bad blocks and all ugly features coming with nand flash techno.

On 24 Feb 2016 12:58, "Dieter Burandt" <Dieter....@online.de> wrote:

Hi Dennis,

I agree with you in reagrd to the SD-card.

In my system I use the SRAM to store configuration data and in addition use the SRAM to write the logs into a memory. Normally at midnight I write the logs from the SRAM to the SD-card. Normally, I don't delete the log data at all. So the SD-card is reliable.

Writing 256 byte to the SRAM cost around 150 micro seconds. Much faster than writing to an EEPROM or to the SD-card.

I didn't found any information about Wear Level Handling of SD-cards. So your information is very interesting.


Using a multitasking OS is an option, but is getting very expensive if you need a high available system.
With today’s Linux variants a very rough estimation of the required CPU power you need around 10 times the power.
Let's say for a 84MHz Arduino Due you need a 800MHz CPU. This is surprising, but comes mainly from the expectation of the user, who expect tools as known from Windows or so.

A very, very long time ago, I worked with a 4MHz system with a real time OS. That did work very well, but only using Teletype Terminals.

A general question for the Arduino group is, do they like to spend a lot of power into a new board with a new processor, or do the like to enhance it with minimal effort and getting a (board and) a shield that fit the requirements of the users.

If you have a combination of a board and a shield that contains an RTC, a battery buffered SRAM with proper size, an EEPROM, a well working watchdog and power fail logic, and ..., then you are on the right way even though, not all need all of the additional devices. Surly no big problem to add jumpers to the board to enable specific devices or not.

But that is a decision of the Arduino Group.


Best Regards,

Dieter Burandt

"Dennis Lee Bieber" schrieb im Newsbeitrag news:mailto:3t4qcbllj3u5bpfhf...@4ax.com...

[Dennis Lee Bieber]

On Wed, 24 Feb 2016 12:58:51 +0100, Dieter Burandt
<Dieter....@online.de> declaimed the
following:

>
>I didn't found any information about Wear Level Handling of SD-cards. So
>your information is very interesting.
>

Unfortunately, I can't find the URL for the site that I got that
information on -- it came up only in the last two months, probably on
either the Raspberry PI or Beaglebone forums... Best I could find currently
is:
https://web.archive.org/web/20150326122100/http://ugweb.cs.ualberta.ca/~c274/resources/hardware/SDcards/WPaperWearLevelv1.0.pdf

This one doesn't talk wear, but timings, and might still be of interest
https://github.com/openenergymonitor/documentation/blob/master/BuildingBlocks/TimeSeries/writeloadinvestigation.md


<eek! 9:15 and I still need to take a shower to get to work by 10:00>

[Dieter Burandt]

Hi Dennis,

that is not only the first document I got about wear level of SD-cards, it is also a good explanation.

Thank you so much!


The document relies on the I-Grad SD-cards that are also defined as professional cards. But on the web pages of SanDisk I could not find any information to these devices.

What I read in the document is, that the endurance cycle specification of those SD-cards is 2,000,000 cycles. That is really a lot.

When I look to EEPROM 100,000 is already very good and I heard often of values from some thousand to some tenthausand.

I expect So the live time of to 149828 years to 79.3 years will be significantly lower for the low cost SD-cards, but sounds good.


Best Regards,


Dieter Burandt

"Dennis Lee Bieber" schrieb im Newsbeitrag news:<59ercb55vgbgq8jc6...@4ax.com>...

> On Wed, 24 Feb 2016 12:58:51 +0100, Dieter Burandt

> declaimed the
> following:
>
> >
> >I didn't found any information about Wear Level Handling of SD-cards. So
> >your information is very interesting.
> >
> Unfortunately, I can't find the URL for the site that I got that
> information on -- it came up only in the last two months, probably on
> either the Raspberry PI or Beaglebone forums... Best I could find currently
> is:
> https://web.archive.org/web/20150326122100/http://ugweb.cs.ualberta.ca/~c274/resources/hardware/SDcards/WPaperWearLevelv1.0.pdf
>
> This one doesn't talk wear, but timings, and might still be of interest
> https://github.com/openenergymonitor/documentation/blob/master/BuildingBlocks/TimeSeries/writeloadinvestigation.md
>
>
>

> --
> Wulfraed Dennis Lee Bieber AF6VN
> wlf...@ix.netcom.com HTTP://wlfraed.home.netcom.com/
>

[Dennis Lee Bieber]

On Wed, 24 Feb 2016 12:58:51 +0100, Dieter Burandt

<Dieter....@online.de> declaimed the
following:

>
>
>Using a multitasking OS is an option, but is getting very expensive if
>you need a high available system.
>With today’s Linux variants a very rough estimation of the required CPU
>power you need around 10 times the power.

Don't confuse a general purpose multi-processing OS with an embedded
RTOS. The former typically has a file system from which it can load and run
arbitrary applications. The latter tends to have all tasks predefined into
the single application, and only really is responsible for making it easier
to handle interrupts and the defined tasks. In that aspect, it is closer to
multiple threads within a single program under Linux/Windows/what-have-you.

Since all memory allocations tend to be made during initialization
(especially in hard real-time stuff -- where dynamic memory allocation
during run-time could change task timing), pretty much everything is fixed
in place. Time-slicing may not even be part of an RTOS; each task must
execute some blocking operation to permit the scheduler to move to another
task (Shades of the original cooperative multitasking of the Macintosh).

[Dennis Lee Bieber]

On Wed, 24 Feb 2016 16:40:25 +0100, Dieter Burandt
<Dieter....@online.de> declaimed the
following:

>

>I expect So the live time of to 149828 years to 79.3 years will be
>significantly lower for the low cost SD-cards, but sounds good.
>

It also depends upon the file-system. I really wish I could find the
article that discussed the number of "streams" cards supported.

Remember that most cards (and cameras, et al.) run with a version of
the FAT file system. FAT is not a journaling file system. A card with only
two 'streams" can use one stream to buffer changes to the FAT, and the
other stream to buffer the data for a single output file. If you use such a
card with a journaling file system (NTFS, ext3, etc.) which uses three to
four streams simultaneously, a 2-stream card will be performing a lot of
activity -- each stream change (inode, data block, journal data, say) will
trigger a full block allocation and erase cycle, wearing out the blocks
much faster. On a card with, say, 4 stream capability, each of those
(inode, data, journal) can be buffered in the controller chip
simultaneously, and only written to the actual memory when the buffer fills
or the file is closed. Furthermore, formatting a card might result in
"mis-aligning" the format's idea of sectors with the card's idea of blocks,
resulting in more wear. See
http://www.pcper.com/reviews/Editorial/SD-Card-Speed-Classes-Grades-Modes-and-File-Systems-Explained
which has a link to an SD specific formatter (but again, that would be a
FAT file system, so not common for those systems running Linux)

http://www.jeffgeerling.com/blogs/jeff-geerling/raspberry-pi-microsd-card
shows speeds for large stream and 4K random I/O for a selection of cards.
Doesn't say anything about wear leveling but the odds are good that the
cards with better 4K I/O have multiple streams so aren't doing as many
allocate/erase of blocks.

http://haydenjames.io/increase-performance-lifespan-ssds-sd-cards/
has hints for helping to extend life -- like turning off the Linux "access
time" update when reading files (since updating the access time means
writing to the card... I should see about doing that for my Beaglebone)

[Dieter Burandt]

Hi Dennis,

RTOS is not a typical multitasking OS, with full protected tasks. What I wrote with regarding a
very, very old system with a 4Mhz CPU was such multiuser, full protected realtime OS and
it did work very well. It does not exist anymore. The things got changed. But a lot of the basics
are still correct. But that all depends on the requirement for real time and must be analyzed precisely.

But it is not that, what I was writing about, except that a system based on  e.g. Linux OS is a complete
different system like an embedded microcontroller system without an OS and the user must
use different methods of programming his application. (On Linux and so on you have a task
with things you can do in parallel, on a microcontroller and you have an "Init" and big "Loop"
and you have do wait an everything to get finished.)

The microcontroller can be not only much more faster then a Linux system, as long as it is build
up properly, but also needs significant lower resources. And that is, what it makes interesting.

If you should use Linux system or a typical microcontroller without an OS depends on what you
expect form the system.

RTOS is something between of both in various versions.

(For my big project for example need: Extreme reliable. Long live system. Proper recovery from
a power fail. Support of all required interfaces. Acceptable responds time for HTML-pages.
Low Power. In a critical situation full controll over all of the software. => I selected the right board.)


Best Regards,

Dieter Burandt

[Jack Rickard]

We have fallen completely in love with the Due.  But my sense is it really hasn't been adopted by the Arduino community.  Perhaps the 3.3v and shield differences?

I did have a couple of things we just couldn't live with.

1. MicroUSB.  They just break completely off the board at the slightest twist.  We like the more robust printer port Mini-B.

2. EEPROM.  We don't need much.  But it is RARE to have an application that doesn't have some configuration items that need to be saved from session to session.  I know this is really an omission in the SAM3X, but it is a glaring one.

3. CAN.  It's onboard. But you can't have it.  We have done a shield with a couple of transcievers and an EEPROM.

So yes, I'd like to see Ethernet and CAN ports on the next version.  And EEPROM somewhere.

We'e done our own version of the Due addressing a couple of these items.  We don't miss the second USB port actually.

I'd like to see it continue to evolve to more powerful processors and faster speeds.  But I don't want to give up the easy C++ programming interface.  And yes, libraries and tutorials are really what binds Arduino together.

Jack Rickard

http://store.evtv.me/proddetail.php?prod=EVTVDue2

--

You received this message because you are subscribed to the Google Groups "Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to developers+...@arduino.cc.

--

--------------------------------

http://EVTV.me

---------------------------------

[Andrew Kroll]

MicroUSB is fine with a robust footprint and proper choice of connector. 
My boards will break before the connector comes off. The problem is the foot print does not have enough area to be solid enough.

[Massimo Banzi]

Jake

Interesting take on the Due

where did you share the HW files?

m

<EVTVDue5-3.jpg>

[andre]

Hello all

My five cents.

I have been looking at buying a Due now for a very long time. Yes one can do magic with that amount of processing power and RAM.

The 3.3v is a pain, all my own shields self designed, mostly work with 5V, now I have to change the PCB and have new ones made to have a zener and resistor added to make it not blow the Due.

Over and above that, all my analog electronics, which tries to work in the range of 0V to 5V to get max out of 10bit ADC has the same problem.

I will in future, try to design things for 3.3 or 2.7 V as well..l But I cannot take my pump house and irrigation controller running on a Mega and plug a Due in there. with the processing power I could have done all sorts of fancy calculations on pump motor current, voltage, phase shift analysis and powerfactor measurement, etc.

Obviously if I had not build and designed the solution in 2011 I would have used a Due, and yes I want a Due and I am going to buy one before they vanish off the shelves.

The 3.3V I think delays the uptake, and eventually it should sell.

Then maybe, people looking for that sort of processing power, buy's a Raspberry, or PC Duino, OR?

i personally like the Arduino more, although I have a PI, as I don't have to deal with an OS, etc. I work on linux every day, develop even my Arduino code on Linux, but my pumphouse controller boots reliably in 2 seconds!

Regards
A

[Brian MacLeod]

I totally agree the 3.3v is the reason it wasn’t adopted, it might as well have a different pin out, because it’s not compatible with any shield, because of the 3.3v

 

BMac

 

From: andre [mailto:andre...@andjo.co.za]

Sent: Tuesday, March 01, 2016 12:13 AM
To: devel...@arduino.cc

[Paul Carpenter]

That is what the IOREF pin is for and yes it is possible to use many devices
even I2C devices with 3V3 or 5V. Even ancient NXP PCF8574 and PCF8591 work
down to 2.7V as meant for battery operation.

Needs more careful design, but many newer devices are 3V3 or lower, and the
more complex devices have a core supply voltage and usually selectable
I/O voltage supply.

Right choice of even 74 series in ACT or similar can make devices work on
either. Lots of AHC or AHCT devices are capable of running of either.
Digital level translation is better done with TI TXB010x/TXS010x series for
bidirectional level translation. An 8 way level translator uses less board
space.

Analog needs more work but not impossible.

Even LCDs will interface with 3V3 dirrectly just got to watch your
thresholds.

3V3 or 2V7 or 1V8 power levels people have been dealing with for over a
decade. Solutions already exist.

Brian MacLeod wrote:
>
>
> I totally agree the 3.3v is the reason it wasn’t adopted, it might as
> well have a different pin out, because it’s not compatible with any
> shield, because of the 3.3v
>
> BMac
>

> *From:* andre [mailto:andre...@andjo.co.za]
> *Sent:* Tuesday, March 01, 2016 12:13 AM
> *To:* devel...@arduino.cc
> *Subject:* Re: [Developers] Arduino Due

--
Paul Carpenter | pa...@pcserviceselectronics.co.uk
<http://www.pcserviceselectronics.co.uk/> PC Services
<http://www.pcserviceselectronics.co.uk/pi/> Raspberry Pi Add-ons
<http://www.pcserviceselectronics.co.uk/fonts/> Timing Diagram Font
<http://www.badweb.org.uk/> For those web sites you hate

[Paul Stoffregen]

Anyone know if the Ethernet Shield has met a similar fate as Arduino Due?

I'm trying to help someone resolve an ethernet problem. Turns out he's
got a cheap clone shield, which appears to be based on an older Arduino
Ethernet Shield, which is no longer on the Arduino website (so no
schematic... of course, the clone company doesn't have a website or any
documentation). He wants to buy the genuine shield, so we can both test
with the same known-good hardware. Adafruit's out of stock, Sparkfun
says it's retired, and even though it's still on Arduino.cc, it's not
for sale on the Arduino store. :(

Worse yet, Musto's company now seems to be selling an "Arduino Ethernet
Shield 2" based on the W5500 chip, and an "Arduino Ethernet R3" which
appears to be the W5100 and '328 processor, so not really a shield at
all. Those guys sure are good at "confusingly similar"!

I just want to get to the bottom of a mysterious ethernet compatibility
issue....

[Massimo Banzi]

Paul

In general you know you can email me directly if you have questions.

In any case we are rebuilding our product line evaluating what kind of products make sense for Arduino to make and what makes sense to leave to partners. Ethernet shield is on the radar but I want to make something better or avoid it completely


the current discussion is useful for me to understand that there is a need for something in the Arduino Due range (again I rather build something more exciting maybe incorporating some of the suggestions we got here)

Too many product in the product line are hell to understand (I was reading an interesting article the other day about the complexity of the Intel product line for PC/Laptop processors which is hard to understand even for professionals)

3.3v is one of the many issues which make the Due and Zero more difficult for the market but not the only one.
I believe the main issue is that we didn’t invest the same amount of energy documenting and supporting the Due as we did with the previous product families..

I would like to contain the number of product we make and support them better, with better code, more compatibility and more documentation.

m

[Mike]

Paul - you are 100% correct, with good designs one can use the IOREF pin or make a 3V3 + 5V compatible shield.  So for custom designs, interoperability is very possible.  But for custom designs it is less expensive to just select the operating voltage of the design you have in mind.  Only in an environment where one might mix and match components (schools, hobbyists, etc.) would a dual voltage solution be most beneficial.

Commercial companies have not filled the educational/hobbiest need for dual voltage shields, usually picking 5 volts for Uno compatibility.  3V3 designs are more common in custom "shields" for specific clone lines.  I looked around for some IOREF dual voltage boards a couple months back, I may have found one.  I doubt there are more than 5 on the market for general use (again discounting specific user applications like a company or enthusiast crowd).  I'd like to be shown wrong though.

If this is indeed the state of shields, I'd like to hear cogent ways to get from today to tomorrow on a logical path.  No sense throwing stones, better to work together.

[Paul Carpenter]

Well last design for Due was custom and for education and IOREF driven
to be Due or Mega compatible where possible.

A lot of the problems are historic (always done this way), or beyond scope
of even a lot of CS/ICT teachers - "what's this voltage thing I thought this
was like plugging in USB, plug and play -. The biggest problem is most
educators and others think Arduino/Genuino is Uno ONLY, even shield
developers. Saw an evaluation board for some analog circuitry from a chip
manufacturer the other week on a basic Uno shield and only expected to work
with Uno.

Maybe next 3V3 board should have level translators on board to cope with
users expectation levels.

Yes I have some understanding of the education market, other half is head of
computing and ICT at local Secondary school in UK, so I often get asked to
'help'. Been involved with UK exam boards as well.

In education for most where they want a class of up to 30 to use devices
Uno is only possibility as price point difference in bulk of even 50 pence
per unit can make any device a non runner. Non-existant budgets.

Mike wrote:
> Paul - you are 100% correct, with good designs one can use the IOREF pin
> or make a 3V3 + 5V compatible shield. So for custom designs,
> interoperability is very possible. But for custom designs it is less
> expensive to just select the operating voltage of the design you have in
> mind. Only in an environment where one might mix and match components
> (schools, hobbyists, etc.) would a dual voltage solution be most beneficial.
>
> Commercial companies have not filled the educational/hobbiest need for
> dual voltage shields, usually picking 5 volts for Uno compatibility.
> 3V3 designs are more common in custom "shields" for specific clone
> lines. I looked around for some IOREF dual voltage boards a couple
> months back, I may have found one. I doubt there are more than 5 on the
> market for general use (again discounting specific user applications
> like a company or enthusiast crowd). I'd like to be shown wrong though.
>
> If this is indeed the state of shields, I'd like to hear cogent ways to
> get from today to tomorrow on a logical path. No sense throwing stones,
> better to work together.
>
> On Tue, Mar 1, 2016 at 1:46 PM, Paul Carpenter
> <pa...@pcserviceselectronics.co.uk

[Dennis Lee Bieber]

On Tue, 1 Mar 2016 07:13:10 +0200, andre
<andre...@andjo.co.za> declaimed the
following:

>The 3.3v is a pain, all my own shields self designed, mostly work with
>5V, now I have to change the PCB and have new ones made to have a zener
>and resistor added to make it not blow the Due.
>

Unfortunately, it seems anything using an ARM core is going to be 3.3V
(and the A/D inputs on Beaglebones are nearly half of that, at 1.8V).

[tzi...@gmail.com]

Related to that, is there a "shield design guidelines" documentation, similar to the library design guide? I haven't seen one, and I would venture a guess that most of the shields that only support UNO are from people who wanted to get the job done, and wouldn't even have a reference doc that highlights the things to look out for when designing a shield.

A sample CAD file as a reference that shield designers should build on would be even better, and/or an EAGLE part for the R3 pinout that specifies what pins should be connected and how

Vasilis

[Phillip Stevens]

Would like to see a SAMC21 solution.
That is the 5V ARM C0+ option.

http://www.digikey.com/catalog/en/partgroup/sam-c21-xplained-pro-evaluation-kit/53507

[q2dg2b .]

There are many good ideas here (and/or many good requests). I know I'm a little dumb but I insist: are these ideas/suggestions taken into consideration in a formal way? There is a clear roadmap to new/deprecated products? Why don't open to community the design of news boards? I'm very anxious with the (subjective, surely) sensation of Arduino is going nowhere...

Thanks

[Andrew Kroll]

Every MCU platform has its own positives negatives and pitfalls.

The way to approach things is to choose the correct components for the job. If that means working with 5 volts, or 3.3 volts, or 1.8 volts it really doesn't matter. That's the whole idea behind hacking... If something is not fitting your design, for example a shield that only runs at 5 volts, hack it!
Make it work to your needs, and then share your modifications with everybody else. Complaining does nobody any good.

[Thibaut VIARD]

How can you say Arduino is going nowhere? They just introduced the MKR1000 and co-organized the Hackster contest!

Do you think bringing a new board while setting a full production line for the previous ones is done in a week?
You can add to that all the work done by software people on maintenance and support every day.

Please take a look at github repositories and come back with more positive suggestions.

T.

[q2dg2b .]

Ei, you can see https://github.com/arduino/Arduino/search?utf8=%E2%9C%93&q=q2dg if you want to know my contributions to Arduino Github, it's easy. And yes, they are only issues (not pull requests) and they are related only to documentation, but that's because I'm a teacher, not an engineer.

When I say "nowhere" I'm talking about the (apparent) lack of vision. Will Due (or similar) stay? Will Yún (or similar) stay? Will Ethernet Shield (or similar) stay? (in fact, will the other shields stay?) When will be an official guide/pcb files to show hot to design shields in the "right" way to expand the ecosystem? What happens to the issues/pull requests that are hanging weeks and weeks without response?

You say MKR1000...well, it's a Zero+Wifi without the compatible shield form-factor...ok, but, why it doesn't have Bluetooth/BLE? Ah, for that there's the 101, but, wait, this board (based on a third architecture!!??) hasn't wifi, and wifi shield is not compatible. Oh, my! You can't have the two things at the same time...and meanwhile there is the RPi with all this stuff,

Many third-party compatible boards filling the gaps official Arduino products don't offer, which is good but at the same time it's sad. How many suggestions are put here or in the forum which are drowned in the sea of Internet? Why not have a community debate about how should be the new board?

[William Westfield]

> That is what the IOREF pin is for…
:
> even 74 series in ACT or similar can make devices work on either…

Yes, “In theory.” But IOREF does nothing to make the existing fields compatible and shield vendors are not interested in making their shields twice as expensive by adding buffers (and frequently there is not much room.) (I don’t think I’ve seen either a Design Guide on how to use IOREF, OR a shield that uses it :-( ) (although there are some shields with 3V chips that use buffering to work with the 5V arduinos…)

> 3V3 or 2V7 or 1V8 power levels people have been dealing with for over a decade. Solutions already exist.

But NOT in the arduino world. (For the most part.)

A number of 5V ARM chips have been showing up “recently”, and I notice that someone has released an Uno-like board with a 5V ARM. It’ll be interesting to see what happens. (http://forum.arduino.cc/index.php?topic=383029.0 )

> MicroUSB is fine with a robust footprint and proper choice of connector.

Another situation where theory and reality diverge. TI Launchpads as used in Utexas EDX class: “Be really careful because the connectors are fragile” (along with “oh - I ripped by connector off.”) Atmel JTAG-ICE3: “the micro-usb connector is especially fragile.” Atmel Xplained mini eval boards: “In addition to using micro-usb, let’s put it off on a long narrow piece of pcb.”

BillW/WestfW

[Andrew Kroll]

Two short answers for you.

Hack it.

Make it so.

--
You received this message because you are subscribed to the Google Groups "Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to developers+...@arduino.cc.

--

Visit my github for awesome Arduino code @ https://github.com/xxxajk

[Stefan Dzisiewski-Smith]

I wouldn't normally chip in to this (as it's very tangential) but I think it might help some.

The whole "micro USB is fragile / fine" debate is one where I have a lot of personal experience. I have tens of thousands of products that I have designed out there using micro USB and have NEVER had a customer complain about a connector coming off the PCB (and believe me, I've had stranger support requests). I put this down to using a high quality SMD connector with pin-in-pad mechanical retention - the model I normally spec is the Hirose ZX62D-B-5P8. 

The only issues I tend to see are "why can't I upload" (answer: you're using a charging cable without data conductors) and in 1 case a customer who somehow managed to force a bizarre camera cable which I had never seen before into the receptacle. In fact, the main thing that might put me off micro USB in the future is the existence of these cables with no data conductors due to the proliferation of micro USB as a charging standard - going with mini USB would reduce the number of support requests I end up fielding.

I mention this not to wind people up or to try and show off - merely to give a suggestion to those who have bad experiences of something to try. I also apologise for the fact that this has nothing to do with the thread title.

S

--
You received this message because you are subscribed to the Google Groups "Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to developers+...@arduino.cc.

--

—————————————————————

Stefan Dzisiewski-Smith

Head of Technology

—————————————————————

T   + 44 207 650 7977    |   Web      bareconductive.com

M  + 44 779 302 9265    |   Twitter  @BareConductive

                                                                            

       TOUCH BOARD  |   Pin Function Map      

                                                                            

This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager.

[Paul Carpenter]

William Westfield wrote:
>> That is what the IOREF pin is for…
> :
>> even 74 series in ACT or similar can make devices work on either…
>
> Yes, “In theory.” But IOREF does nothing to make the existing fields compatible and shield vendors are not interested in making their shields twice as expensive by adding buffers (and frequently there is not much room.) (I don’t think I’ve seen either a Design Guide on how to use IOREF, OR a shield that uses it :-( ) (although there are some shields with 3V chips that use buffering to work with the 5V arduinos…)

Look to make something being compatible means careful thought, in a lot of
cases it is new PCB, and DIFFERENT VERSION of chips being used that can work
in both domains. It is NOT always adding extra chips or buffers, if designed
correctly in MOST cases.

The issues about how to deal with analogue is different and requires
extra work. Not insurmountable in all cases depends on a shield by shield
basis, and how many are actually being sold to see if investment is
worthwhile.

>> 3V3 or 2V7 or 1V8 power levels people have been dealing with for over a decade. Solutions already exist.
>
> But NOT in the arduino world. (For the most part.)

Because most only see UNO and sometimes Mega and don't even think anything
else exists, that is users and vendors of shields.

>> MicroUSB is fine with a robust footprint and proper choice of connector.
>
> Another situation where theory and reality diverge. TI Launchpads as used in Utexas EDX class: “Be really careful because the connectors are fragile” (along with “oh - I ripped by connector off.”) Atmel JTAG-ICE3: “the micro-usb connector is especially fragile.” Atmel Xplained mini eval boards: “In addition to using micro-usb, let’s put it off on a long narrow piece of pcb.”

I have seen all sorts of products with broken USB or power connectors of
ALL types, some break easier than others. The standard USB B is not
safe either having seen a collection of HP printers where the THROUGH
hole types of USB B connectors were sheared off and rattling around inside.

I personally would prefer mini B as different from phone and charger leads
so in education less 'pocketable' for kids phone use.

[William Westfield]

> Why don't open to community the design of news boards?

Lots of “community members” have designed new boards. PRJC (teensy/etc), Adafruit (feather, metro), Sparkfun (redboard, samd21 breakout), Seeed (86duino, LinkIt Duo), Crossroads with his 1284 boards. And then there are the designers that I’d normally consider outside of the community (Simblee, esp8266, xmegaduino, energia, chipKit, etc) There are lots of players in the sort of “small to moderate numbers of users” space that Arduino used to occupy (Where selling 500 boards/year is “major success.”) But it’s a lot more difficult when you get to the size of Arduino NOW, where you’d really like a new board to “justifiy itself” by taking over a fair percentage of (just guessing) 50,000 boards/year.


> I'm very anxious with the sensation of Arduino is going nowhere...

In a real sense, it doesn’t need to go anywhere. The Uno is near-perfectly placed (cost, speed, program size, robustness, simplicity) for maybe 90% of the applications where people want to use an Arduino-like board. The fancier boards make us feel “warm and fuzzy” about being able to get “more”, but we’re not really ready to start buying them instead (especially with “really exciting” technology in the same price range - Raspberry Pi, BeagleBone, and etc.) (I mean, are you really going to pay $50 for a Due, plus $45 for an ethernet shield, to make an IoT gadget that would be easier to develop on $50 worth of RPi gear?)

Personally, I wish Arduino would focus on the $30 price point, instead of the form factor(s). But that might be just me, being cheap :-) (incremental improvements: xmegaE - twice as fast, twice the RAM… Or 328pb - extra I2C/Serial…) OTOH, I figured out a long time ago that implementing every “interesting chip” in arduino form-factor is not necessarily a good thing. It would be better to narrow down the HW selections and concentrate on improving the software (for instance, it’s a bit sad that Due and Zero, which are both ARMs, are implemented quite differently.)

BillW/WestfW

[Dennis Lee Bieber]

On Tue, 1 Mar 2016 15:55:58 -0800, William Westfield
<wes...@mac.com> declaimed the following:

>Another situation where theory and reality diverge. TI Launchpads as used in Utexas EDX class: “Be really careful because the connectors are fragile” (along with “oh - I ripped by connector off.”) Atmel JTAG-ICE3:

Yeesh -- this comment made me look at my TIVA 123... I'd expected to
see a through-hole mount for the USB shell, but it's just solder-tacked to
the top surface.

[Massimo Banzi]

Dear q2dg2b (it would be nice to know your name at some point :)

I resent your statement that Arduino is lacking vision. We are moving forward innovating in different areas despite the difficult situation we sometimes have to face.

Off the top of my mind I can think of these:

* Launching a range of compact wireless products with battery management etc. The upcoming MKR1000   (with ATMEL ARM) was the subject of the most successful maker oriented competition with over 4900 submissions (organised with Microsoft who sponsored us in giving away 1000 of them to makers)

* Reimagined (With Intel) the classic Arduino UNO format with a product which provides makers and learners with much more power, connectivity and sensors at more or less the classic price of the UNO

* Created CTC, a very effective programme to teach Programming, Electronics and Robotics in school. We delivered it in several countries in the last 2 years and now working with Intel to deploy it worldwide.

* Delivered a whole new experience with our Create online platform where the IDE is only 1 of the modules we’re releasing. A cloud platform helping makers to discover, learn, code, share and document using the same tool. More innovation is coming. (1000 people are already beta testing the platform and we have more partners getting on board)

* Made our IDE modular and enabled anybody to add support for their board to the IDE. 

* Removed the pain of dealing with Libraries and Cores with the Library Manager and Board Manager.

* Ported the Arduino API on several different architectures . We’re now running on AVR, ARM, X86 and ARC plus running on top of RTOS and Linux ( see the ARTIK platform with Samsung.) You can now run Arduino from an 8 pin AVR micro controller with 512byes of RAM to an 8 core Linux machine.

* Adding Debugging to the IDE, that’s been worked on with members of the Arduino community.

* re-designed the Arduino builder so that it’s easier to embed the official arduino compiled in makefiles, external IDE, cloud platforms etc. 

* Designed the ESLOV modular sensors platform which we delivered to dozens of partners in Europe as part of an EU research project. This project enables people to create connected devices with zero code and zero knowledge of sensors. We also developed a visual IDE to program the modules.

we have more work being released as open source in IOT, cloud platforms and education. We’re also working on a lot more content to help people make sense of the new devices that are being made available on the market.

We can’t always work completely in the open for a number of reasons that I won’t discuss here because there are legal implications. Sorry about that. Any major change to the platform has always been discussed here, sometimes at length so we always worked with the community.

I’m happy to discuss our product strategy and respond to individual questions , sometimes I won’t be able to get into details until a certain product is out in the open.

If this seems to you like we’re going nowhere….

PS: can we  break this thread? it’s not about the DUE anymore.

M

[andre]

Hello

I have seen nothing that compares to Arduino and it has made my life fun since I have bought the first one in 2011.

I am working the the commercial software world at banks with software from IBM, Oracle etc, and Arduino at least comparable in quality and even better when it comes to usability, especially since I can access all the source, designs, etc.

Keep up the good work.

Regards
A

[Dennis Lee Bieber]

On Wed, 24 Feb 2016 09:18:00 -0500, Dennis Lee Bieber
<wlf...@ix.netcom.com> declaimed the following:

>On Wed, 24 Feb 2016 12:58:51 +0100, Dieter Burandt
><Dieter....@online.de> declaimed the
>following:
>
>>

>>I didn't found any information about Wear Level Handling of SD-cards. So
>>your information is very interesting.
>>
> Unfortunately, I can't find the URL for the site that I got that
>information on -- it came up only in the last two months, probably on
>either the Raspberry PI or Beaglebone forums... Best I could find currently
>is:
>https://web.archive.org/web/20150326122100/http://ugweb.cs.ualberta.ca/~c274/resources/hardware/SDcards/WPaperWearLevelv1.0.pdf
>
>This one doesn't talk wear, but timings, and might still be of interest
>https://github.com/openenergymonitor/documentation/blob/master/BuildingBlocks/TimeSeries/writeloadinvestigation.md
>
>
><eek! 9:15 and I still need to take a shower to get to work by 10:00>

Resurrecting a dead thread...

Stumbled onto the link that described the "streams" (# of open
allocation units supported).

https://wiki.linaro.org/WorkingGroups/KernelArchived/Projects/FlashCardSurvey?action=show&redirect=WorkingGroups%2FKernel%2FProjects%2FFlashCardSurvey
https://lwn.net/Articles/428584/


And something related: http://wiki.laptop.org/go/NAND_Testing

[LM]

Only problem with Due I have found is the SPI bus. Those pins are in the programming interface, not where the normal Arduino IO pins are.
So how to fix it in Due2. Find a Cpu where the ISP interface has dedicated pins or at least not in SPI1, Uart1, Can1 or whatever1 pin.

On Tuesday, February 23, 2016 at 3:18:44 AM UTC+2, James Michener wrote:

I understand the Due is being discontinued.   Are there plans for a
replacement?

--Jim

[Daniel De Kock]

It looks like Atmel/Microchip plans to discontinue teh SAM3X series -it might take a while
but a suitable replacement for the Due could be the SAM-C series 5V ARM with the same peripherals as on the 3X series :
http://www.atmel.com/products/microcontrollers/arm/sam-c.aspx#samc21
or if one were to stay with 3v3 the SAMG55 : http://www.atmel.com/devices/ATSAMG55.aspx

[Phillip Stevens]

+1 SAMC with 5v IO.

Identical to SAMD from software side.

[Jim Leonard]

In the places where I'm using a Due it's because I need both the speed and the RAM on top of having a range of I/Os. I would not look forward to a downgrade.

--jim

[Paul Carpenter]

Seconded, slower speed with less RAM and Flash, and quite a few of us can
cope with 3V3 I/O, 5V I/O is not a show stopper.

5V I/O is less of a show stopper than speed and memory and choice of I/O
as quite a lot of devices connected to outside world should be buffered
or level converted anyway, whilst majority of decent I2C or SPI devices are
either 3V3 or can run on multiple voltage rails.

[James Michener]

Speed RAM and Flash size are important to me.

[Pierce Nichols]

Thirded. 3V3 is much more common in new peripherals, and interfacing
5V and 3V3 is easy.

Why not replace it with an Cortex-M4 device? There are plenty of
excellent candidates in Atmel's product line.

-p

> --
> You received this message because you are subscribed to the Google Groups
> "Developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to developers+...@arduino.cc.

--
Pierce Nichols
Principal Engineer
Logos Electromechanical, LLC

[Victor Aprea]

Just to be clear, I'm not dissenting, but, suffice it to say, the consensus we are likely to reach on a Developer's List may not be the same consensus that would be reached on a User's List or Educator's List. 

Kind Regards,

Vic

Victor Aprea // Wicked Device

victor...@wickeddevice.com

T: @vicatcu

[Paul Carpenter]

LOTs of OLD devices work on 3V3

74AHCxxx and others
I2C devices even ANCIENT PCF8574 and PCF8591

Last year on Education basis I had a 6th form student do a project with
5V DIP parts (see http://www.pcserviceselectronics.co.uk/arduino/logic.php )

25% of potential 74xx series parts had to be dropped as OBSOLETE
15% had to be dropped as difficult to source

It is a matter of TIME before very few DIP and 5V parts will be
easily available.

I am regularly involved in education (like daily) and computing
or ICT type subjects involving all sorts of things including
Arduino and Pi.

Victor Aprea wrote:
> Just to be clear, I'm not dissenting, but, suffice it to say, the
> consensus we are likely to reach on a Developer's List may not be the
> same consensus that would be reached on a User's List or Educator's List.
>
> Kind Regards,
> Vic
>
> Victor Aprea // Wicked Device

> victor...@wickeddevice.com <mailto:victor...@wickeddevice.com>
> T: @vicatcu <http://twitter.com/vicatcu/>

>
> On Tue, Apr 19, 2016 at 2:34 PM, Pierce Nichols
> <pie...@logos-electro.com <mailto:pie...@logos-electro.com>> wrote:
>
> Thirded. 3V3 is much more common in new peripherals, and interfacing
> 5V and 3V3 is easy.
>
> Why not replace it with an Cortex-M4 device? There are plenty of
> excellent candidates in Atmel's product line.
>
> -p
>
> On Tue, Apr 19, 2016 at 11:20 AM, Paul Carpenter
> <pa...@pcserviceselectronics.co.uk

[William Westfield]

On Apr 19, 2016, at 12:27 AM, Daniel De Kock <dekock...@gmail.com> wrote:

> It looks like Atmel/Microchip plans to discontinue teh SAM3X series

What leads you to conclude this?

BillW

[Paul Carpenter]

Just had a conversation with an educator, their basic comment

"Rather have more memory than worry about 5V I/O"

Paul Carpenter | pa...@pcserviceselectronics.co.uk
<http://www.pcserviceselectronics.co.uk/> PC Services
<http://www.pcserviceselectronics.co.uk/pi/> Raspberry Pi Add-ons
<http://www.pcserviceselectronics.co.uk/fonts/> Timing Diagram Font
<http://www.badweb.org.uk/> For those web sites you hate

[Andrew Kroll]

I would like to put a final nail in the coffin of this discussion...

Times change things change.
Microcontrollers and variants come and go.
It is usually best not to depend on something to be forever because it never will be.

***Change is a good thing embrace it!***

[Dennis Lee Bieber]

On Tue, 19 Apr 2016 20:35:13 +0100, Paul Carpenter
<pa...@pcserviceselectronics.co.uk>
declaimed the following:

>
>It is a matter of TIME before very few DIP and 5V parts will be
>easily available.
>

Except in my parts bin -- where I still have a naked Intel 8080, 8228,
a few 21l02s etc. 74xx TTL stuff...

[q2dg2b]

Well, in https://www.arduino.cc/en/Products/Compare Due board appears as retired. And in Products page has disappeared. So the future is clear.

On the other, I've seen all shields except Wifi and Proto have misteriously disappeared from Products page, too...I suppose soon there will be some news about all these movements

PD: On the other hand, there's no information about "porting" Lilypad family to Genuino brand...will be disappear too?

[Jim Leonard]

At least we still have all of the 3rd party manufacturers of these. It's sad that nothing comparable was created to replace it.

[Adrian Godwin]

Yes, I was planning to migrate to that as a standard device instead of the AVR until I saw the support failing. I'm tending to use Teensy instead now.

[Massimo Banzi]

As I mentioned a while ago we’re cleaning up the product line removing the products we no longer manufacture.

For a number of those products there are replacements coming at different times in the future. 

They won’t be identical to the old ones because in the meantime the world has moved forward :)

BTW We adopted a new policy of announcing only when something is actually available.

In the past we had multiple instances when something seemed to ready to go next week and then some big issues was found and we had to delay the actual delivery by months…

m

--

Massimo Banzi <m.b...@arduino.cc>

Arduino LLC

[q2dg2b .]

Thanks a lot, Massimo