Goldilocks = Arduino Uno + 1284p

I’m getting some Goldilocks Arduino Uno clones made, with 1284p processors.

100% Arduino Shield compatibility, with 8 times the Uno SRAM is the key justification.

IMHO, it is the Goldilocks solution; not too small (insufficient SRAM for serious fun), not too big (Arduino Mega Shield incompatibilities).

Please grab some here. http://pozible.com/goldilocks
http://feilipu.me/2013/03/08/goldilocks-1284p-arduino-uno-clone/

Constructor will be Freetronics in Australia. The result will absolutely professional.

You can get them over-clocked to 22.1184MHz too, if you want perfect serial synchronisation, or just extra 38% speed over the Uno.

Just to avoid confusion, I’m not a business. I want 10 of these for my own use, and it is cheaper and better to build 100. So this is not an advertisement, but rather a call for help.

I also do this.

Thanks, feilipu

goldilocks_133.png

Goldilocks is now 100% fully funded, so we are go!

Its been a great ride to get to this point in only 11 days. There is a lot of demand for a capable Arduino Uno clone.

http://pozible.com/goldilocks/ http://feilipu.me/2013/03/08/goldilocks-1284p-arduino-uno-clone/

We're also planning some special features to make the 16u2 Arduino serial interface device a more capable coprocessor to the 1284p. This includes breaking out its SPI bus, some analog pins, and the 16 bit PWM pins.

There are lots of applications for having two processors on the board, with one handling acquisition tasks, and the other conversion and storage tasks.

Thanks again for your support and pledges.

How do you think it would compare to a clone based on an Xmega AVR? Something like an xmega128d4 would run at 32MHz, but have less RAM than a mega1284.

Nothing against Xmega, but if it were on Xmega, then it wouldn't pass the first test of building a 100% compatible clone. ;-) Others have built similar IDE compatible clones with non ATmega MCU, such as Maple, Netduino, Arduino Due, etc, but unfortunately they're not exactly the same as Arduino Uno at the machine level.

The great thing about the 1284p is that with a simple addition of a Board description file to the standard supported Arduino IDE, everything will work as expected, right down to reading the assembly level instructions (if one wants).

There should be no surprise for new players in this upgrade option.

There should be no surprise for new players in this upgrade option.

No surprise here, CrossRoads has been supplying 1284P based PCBs for well over a year now and member maniac has supplied needed IDE files and bootloaders, so several of us are having the fun of using the 1284P 'arduino'.

Lefty

For those that are interested in Goldilocks, here is a rendered image of the prototype board.

Note, this is not the final design, as some issues are still to be resolved and are noted in the design document. But, it is pretty close.

The extra IO is brought out in the middle of the card? That's not very shield frienldy. The holes for mounting a 14-pin DIP or some Rs, Cs, or an LED or two might be handy.

I have these cards in hand right now for delivery - and if you smoke the chip, you can unplug it and drop in a new one. Onboard USB via MIKROE-483 FTDI module from Mouser.com, of offboard for programming only of embedded modules via FTDI Basic or similar. http://www.crossroadsfencing.com/BobuinoRev17/

Looks like you've got a proper switching regulator there feilipu, that will help with high VINs, 'bout Arduino designs started doing this.

What's the 2x5 pads just to the left of the 1284, JTAG header?


Rob

Hi all,

thanks for the comments.

My concept for the Goldilocks was to make the most of the 1284p and Arduino platform, but without actually adding significantly to the BOM cost of building the board or making decisions which would reduce generality for all users, and remaining 100% compatible to Arduino. The idea is that someone could use the design to make and sell it for only a few dollars more than Uno clones.

I expect that most people will just want to use the Goldilocks because they've run out of RAM or Flash on their Uno, and they've got some shield issue that prevents them from easily using a Mega. But, others will want to do more and the combination of a 32u2 and 1284p on one board will enable this.

Some of the things I've added are:

  • Power supply is specified up to 5V 3A peak with up to 28V input. Can supply enough current to run GSM shields (1.5A), and other high current shields simultaneously, without getting hot or being in danger of burning out.

  • Whilst only the standard headers will be soldered at manufacturing, the all the pins of both MCU are brought out on 0.1" pitch to enable the board to be either plugged onto a 0.1" breadboard, or have 8x1 headers applied and have the Port pins taken off board as a ribbon cable.

  • Making real use of the 32u2 (which is upgrade specified from Uno / Mega), to allow it to be a AVR-ISP, SPI slave radio controller, USB I/O, or similar. All the pins are brought out, and headers can be soldered, or jumpered across to the prototyping area.

  • Fully utilising the 1284p. Supporting JTAG, by adding pads. Allowing a RTC functionality, by pinning the 32kHz crystal and caps. Proper ground plane and AVcc decoupling for lower noise in analogue circuits. Real crystal oscillator to allow timing choices (eg 16MHz, 20MHz as built, 22.1184MHz).

  • Building an SPI interconnect between the MCU, and SS options to allow either one to take Master role.

  • Fixing the broken Arduino R3 specification by allowing the SCL/SDA pins to be optionally connected to A4/A5 for older (non-R3) shields.

  • Adding a uSD card cage, and using PB0 and PB1 to drive it (keeping 1284p SS free for shield use).

  • Adding a stitched prototyping area, to allow easy jumpering and attaching radio / RTC / sensor modules.

  • Adding cut/join solder pads for every pin, to allow I/O to be moved around the board as desired.

Lots of things that have been done, with the idea that there is a lot of capability locked up in the two MCU (1284p and 32u2), so let's bring it all to the surface, for those that need this stuff. But also, keep it really easy for those that just want some more space for their sketches.

Really looking forward to getting my hands on one next week. Very excited.

Just following up on progress. We've got the prototype Goldilocks boards on the way now. While there are still some things we're planning to add, and testing to be done on the existing new features. It is getting closer.

Looking good.

5V 3A peak with up to 28V input.

How do you get 3A with such a small inductor?


Rob

The current (prototype) power supply implementation is identical to the Freetronics EtherMega R3. The power supply component selection is noted in the schematic https://github.com/freetronics/EtherMega/raw/master/EtherMega.pdf CDRH6D38 6R8 http://www.sumida.com/products/pdf/CDRH6D38.pdf EUP3476 http://www.eutechmicro.com/PDF/EUP3476.pdf

My design specification for the Goldilocks was that it should happily drive (knockoff) GSM/GPRS shields, needing 1A to 1.5A, as well as being driven at a wide varieties of voltages up to 28V, without burning fingers or blowing up.

So whilst the SMPS regulator is rated to 3A continuous, the inductor is only rated to support 2.4A at 20 C. I still see that being adequate for my purposes (and also way better than the Uno or its peers).

I'd also note that the SMPS is one area that we're not entirely happy with. Because of the lack of reverse current protection in the regulator chip, we need to use a jumper to select the power source. This is under development, and I'm hoping that this will be solved for the finalised Goldilocks boards.

Does anyone see a requirement to have a 5V 3A continuous supply on an Arduino platform? If so, then we can work out how it can be done.

Phillip

Because of the lack of reverse current protection in the regulator chip, we need to use a jumper to select the power source.

Can you bump the regulator output a bit and pass it through a schottky diode?

Does anyone see a requirement to have a 5V 3A continuous supply on an Arduino platform?

3A might be nice but if you currently have 2.4A I would think that's enough. It's heaps better that any standard Arduino.


Rob

Really 2.4A continuous though? Without saturating? If so, I can't imagine that not being enough. If you're driving high-current loads than that you should probably be powering them separately.

Agree on the diode isolation and accounting for slight voltage drop in the feedback divider. Next question. :-)

This is outstanding. :) As soon as they're available, I'll be buying two. 1 because I know it'll replace a Mega I have dedicated to a task. The other to have for next time. You might like my comments about the eleven here and here.

I like your improvement on the UNO the most so far. :D

Just another update on progress, which is ticking along.

The production PCB for the Pozible pledge rewards have been completed, and components are now being “placed” and tested.
We should have the finished boards in house around 2nd August, so we can get them ready for the supporters.


Also, Jon has decided that the Goldilocks will become a Freetronics product too,
which should ensure longevity (code updates, forum, bug fixes, etc) even if something were to happen to me.

It's looking good, the 1284 is "just right" IMO to, a great combination of features.


Rob

feilipu, thank you SO MUCH, this really is the perfect Arduino for me! My projects tend to vary somewhere near the RAM limit of the UNO but do not require all the pins of the Mega. I'm buying 2 or 3 as soon as they are selling and shipping from freetronics!

Thanks, Chris

Another Update note on the Goldilocks.

Finally, the production boards have been tested and shipped to Pozible supporters.

I've written a user manual, that is available here: https://tinyurl.com/Goldilocks1284p

Soon, they'll be available from Freetronics too. http://freetronics.com/goldilocks.

When we recover from a marathon testing, packing, shipping session, the design files will go up via Freetronics web site and github.

I just received my Goldilocks 20MHz and 22.1184MHz(overclocked) boards. They are working great. I have one driving four 16x32 RGB Matrix panels in a 16x128 pattern. The extra overclock is just enough speed to eliminate the flicker when driving four panels together. Now to try it with four 32x32 RGB panels.