xmega Development Board

Go To Last Post
72 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have been contemplating designing an Xmega development board. The target market is for students, hobbyists, and design engineers that want to test the Xmega without breaking the development budget.

To avoid the JTAG programmer cost, it will have a preprogrammed bootloader via USB.

All port pins will be available on 0.1" headers.

I am unsure of what other devices are desired.
Here is a list of device ideas:
1. JTAG port, External RAM, RS232, Ethernet, LCD, switches, LEDs, prototype area, beeper, and SPI memory.

2. The system will include test software for most all Xmega peripherals.

Please comment. Poll results will be used to determine possibilities.

It all starts with a mental vision.

Last Edited: Tue. Jul 7, 2009 - 10:59 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
To avoid the JTAG programmer cost,
I would not even THINK of working on a board without OCD (JTAG, PDI or DW).

You don't have a NO option. :)

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For the ethernet, what route were you thinking? A magjack + enc28j60 + supporting components is ~$14 or so. A wiznet module is ~$20. Or maybe something more exotic? You could always leave the pads for it and make it an option (easy with a module).

and I'm going to assume you've read about the elusive xplain?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes it will have a JTAG port. Sorry I forgot the NO option, I cannot seem to change that, guess I am a positive thinker.

I was thinking XPort, I see there are lower cost options.

I have experience with Xplain and STK600.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarlson wrote:

I was thinking XPort, I see there are lower cost options.

I think you will probably have to get cost to bare minimum on every single component for this crowd. I have been working on an XMega board and I'm having second thoughts about offering it as a development platform. It seems that very few Hobbyists / Engineers are going to go for something that costs more than $80 USD. You can't really do much in terms of exciting peripherals with that kind of lean profit margin unless you are producing in very large volumes. Don't know about you, but I need to pay bills.

I'm using KSZ8851SNL for Ethernet controller. It is working great. You might want to check on that one too.. much less current consumption than ENC28J60, plus it can do 10/100 link.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

One of xmega's prime advantages is the large number of serial ports it offers.

It would be nice to have a board with multiple level-converted serial ports coming out of it (via jumpers, of course), especially given the STK600 has only 1.

-- Damien.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have contemplated exactly the same thing and more or less finished something a year ago. But it's not easy to keep it low-cost and be able to make some money from it. So I gave up as it's likely not worth the trouble :)

There's more to it then just designing a sweet little board. You have to produce it, market it, sell it, offer support, keep inventory etc etc.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The STK600 TQFP100 adapter card system costs about $100 USD. While it has a socket, many not change change parts right away. Their next step may be to make a board.
The Xplain board is not being sold by Atmel. It is very limited due to the few I/O pins on headers.

The proposed board will have an improved layout resulting in significantly reduced noise levels in the ADC system compared to the STK600 w/adapter.
The board(s) may measure about 50mm x 80mm (~ 2" x 3"), small enough to integrate into hobby projects.

I appeciate all comments. I would like see more.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think for most people a simple board that just contains an XMega, a programming header and all I/O brought out to headers is sufficient. Maybe add an optional crystal or two and a power supply IC.

Whatever you design you very likely won't make serious money from it. Just like Atmel and others don't from their development tools; I would not be surprised most development boards and tools are sold below cost.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You should factor into your planning "what if Atmel suddenly decides to sell the Xplain or another Xmega eval board"?

No one knows what is going on inside Atmel and they might suddenly change their mind and take potential Xmega users more serious.

As you see from the responses, you are also not the only one thinking about providing an Xmega board. I think one freak is even already selling a larger one.

As for your board, what is your target group? Those who want an all-inclusive board, or, at the other end of the scale, those who just need some kick-start?

Wouldn't an extremely simple board, barely more than an adapter board plus a programming interface do? I am thinking along the line that there is so many possible periphery that you can't cover them all. But if you make it easy to connect all sorts of periphery many users might be happy. E.g. if you make it easy to use your board as a plug-in board. To facilitate this you might consider providing instructions (schematic, BOM, code, strip board / prototype area / PCB layout) to connect common periphery.

Your upfront costs will be higher preparing the documentation, but your per-unit cost for a bare bone adapter with a programming interface might go down.

Stealing Proteus doesn't make you an engineer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

isn't the market for a 3rd party xmega dev board pretty tiny?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There are a lot of third party boards for other processors, so why not for an xmega?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think the xmega development is slow due to many factors:
1. I think stocking by distributors started in June 2009.
2. It does not have ISP programming, very few JTAG programmers work with it. I have found only the STK600 and JTAG ICE mkII USB. There might be others.
3. The cost. By the time you buy a STK600 and the adapter card you have ~ $300 USD. If you need debug capability of the JTAG ICE mkII USB that is more $.
4. Some may be scared off by reading the xmega_a and xmega128a1 manuals, about 550 pages. When you go to develop it seems like there should be more information. It is possibly a fear factor.

Why xmega?
The xmega peripherals are new and improved. The I/O, clocks, timers, SPI, USART, DAC, DMA and ADC are able to do many things not possible on other AVR devices.

It is a paradigm shift. It may take time for some to adjust to the power and capability. We must now scale up our thoughts and creativity to go further on project with the xmega.

I hope this simple and lowcost development board will help others get started.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
We must now scale up our thoughts and creativity to go further on project with the xmega.
Which, of course, is of not much use untill people can reliably BUY the chips. :?
BUT you are an optimist instead of a realist. :wink:

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My biggest hurdle is not purchasing them, but soldering them!

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I was considering this as well and thought it would be most profitable to just sell the bare boards. I have some XMEGA samples that are looking for a home, but nothing to power it up.

This would make inventory easy and shipping even easier. Just put it in an envelope.

I know you can get a good quality 4 layer board made fairly inexpensively when you buy 100. I'd pay $20 for a bare board with places for regulators, bypass caps, pin headers, and a proto area. (Like the PICPROTO from melabs.com)

official AVR Consultant
www.veruslogic.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jayjay1974 wrote:
There are a lot of third party boards for other processors, so why not for an xmega?

lack of design-wins (so far)?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

stevech wrote:

lack of design-wins (so far)?

Huh?

KitCarlson,
The market for an Xmega dev board does seem to be small and not very profitable when looking at comments like this and others in this forum, but you have to remember that this forum is primarily composed of hobbists who are nervous about trying something new or just like to complain about things in general. Most Engineers will probably just design their own boards. I can say after 3 weeks or so of playing with the XMega, I'm quite certain I will not be using previous AVRs for new designs. This new MCU is sooo much nicer. I've decided not to offer an Xmega dev board, so I'm not just trying to push it here. If Atmel doesn't do something horrible, I predict the XMega will take over as THE MCU to use within a year or two.

OK... flame away :?:

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
THE MCU to use within a year or two.
..well we have been waiting for it's availability for about 3 years now, so what's a year or two more. :lol:

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you think the Xmega will take over the world then put your money where your mouth is. Buy Atmel stock.

Stealing Proteus doesn't make you an engineer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

js wrote:
Quote:
THE MCU to use within a year or two.
..well we have been waiting for it's availability for about 3 years now, so what's a year or two more. :lol:

There is nothing stopping you from ordering thousands of them right now John. From most of your posts regarding XMega, I'm pretty sure you will still be complaining about not being able to get them for 3 more years :)

What I meant by 1-2 years, is the adoption time for a new product. Most people just aren't willing to be out front on these things for fear of the unknown. Most XMega posts here result in reply comments about lack of availability, mistakes in documentation, errata, nothing-new-here, ARM is better choice, etc. Once people actually start trying them out and realizing what they are capable of, there will be a big move to it. Until then, I think there will much complaining.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If it is anything besides the 128A1, then I can't comment but I am looking at a 128A1 I just bought from mouser, and I take it to be the real deal.

(kind of interesting it wasn't cut-tape, it was in a nifty little cardboard padded box the size of my palm)

There are other people thinking of 3rd party dev boards as well.

http://forum.sparkfun.com/viewto...

and http://www.mikrocontroller.net/t... (warning, HUGE PAGE)

but a low-cost kit I can see appealing to many

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ArnoldB wrote:
If you think the Xmega will take over the world then put your money where your mouth is. Buy Atmel stock.

Why? Are you a stock holder? ;)

I don't think it will take over the world, I just think it will take the lead in Atmel's 8-bit line-up. It is cheaper because of the new fabs, the peripherals are much more consistent and powerful, it uses very little power. It is very easy to write code for.

Time will tell...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
There is nothing stopping you from ordering thousands of them right now John
And when will I get them? :) Say some of the 44 TQFP types, 16K flash? Maybe I could stretch to 64 pin type 64K flash but it would be just a playaround board of not much practical use to me.

And then there is the 3.6V limitation but I guess that will become more standard in the future instead of 5V.

Who knows next seminar in Oz I may even see a real chip.

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Say some of the 44 TQFP types, 16K flash?

I am with you on this one. My Atmel FAE suggested a dual footprint for the time being - That was in January with a promise of parts in March.

That was two projects ago!

I'm now considering the 44 pin AVR32 parts that are on the shelves.

official AVR Consultant
www.veruslogic.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

telliott wrote:
Why? Are you a stock holder? ;)
I am not that mad.

Stealing Proteus doesn't make you an engineer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

syntroniks wrote:
If it is anything besides the 128A1, then I can't comment but I am looking at a 128A1 I just bought from mouser, and I take it to be the real deal.

(kind of interesting it wasn't cut-tape, it was in a nifty little cardboard padded box the size of my palm)

There are other people thinking of 3rd party dev boards as well.

http://forum.sparkfun.com/viewto...

and http://www.mikrocontroller.net/t... (warning, HUGE PAGE)

but a low-cost kit I can see appealing to many

I would guess that the German link would be the sort of board that would appeal to most people most of the time.

Does anyone know what they charge for it ?

Of course you can always bundle other support chips but then the price goes up. I would appreciate one MAX3232 which would also enable a RS232 bootloader.

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Telliott said:

Quote:
is primarily composed of hobbists who are nervous about trying something new or just like to complain about things in general.

If you are going to slam me, and others, you might at least have the courtesy to provide some data for your claims!

I am primarily one of those "hobbyists", and I happen to be excited about a bunch of features on the XMega, (but not worth listing in this thread).

I happen to have already read the general and chip specific data sheets, several times, and already have a testbed board made up to tinker with. As I eluded to earlier in this post, my biggest hurdle right now is soldering the chip, all 100 pins, without getting any bridges or cold joints. That part is still a work in progress.

JC

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think we should got something around 40$!
why all of you are thinking in higher prices.
these days we can get a 500k gate fpga having the speed of about 700MHz and the price is about 20$.do not you know that Xmega's full speed is 32MHz.i think they are a bit over priced(inspite that they have a bunch of good peripherals)!!!

I love Digital
and you who involved in it!

Last Edited: Fri. Jul 10, 2009 - 05:26 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC wrote:

If you are going to slam me, and others, you might at least have the courtesy to provide some data for your claims!
JC

Hey JC,

Sorry.. I'm sure I should have used some different words there. I did use the words 'primarily' and 'or', but still no excuse to offend. My response was primarily to the OP who is probably feeling a bit let down by all the negativity in these forums about the idea of an XMega dev board and XMega in general. I know it certainly changed my plans of offering an XMega dev board.

Glad to hear you are excited about this new silicon too. Again, sorry for the generalization.. I'll try to avoid that in the future.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi Telliott,

Relax!

It has been a while since there was a good Which language is best, or Pic v AVR, or 8-bit vs ARM, similar such "discussion".

You ended with Flame On, I just took you up on it :)

Perhaps start a Hobbyist vs Professional Developer "discussion"...

Nah, not worth it; and anyways, I don't want to overly Off Topic the OP's thread.

Take Care,

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I appreciate all comments. My idea of what to do for the boards is starting to become clear. I hope this project helps many get started using the xmega. That is my reason for doing this work. There needs to be a critical mass in the cosmic mind :)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Here is a draft of the board.
http://216.77.188.54/coDataImage...

It measures 2" x 3". It is bootladable via USB or RS232. It has JTAG port and all pins on 0.1" headers.
Users can make their own expansion boards.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Looks good.

I've been thinking that you might be able to get rid of the max3232. The XMega allows you to invert pins. For something that needs to be really cheap, I think you could get away with just sticking a current-limiting resistor on RX (maybe a tvs diode.. just for a little extra protection)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

True RS232 swings negative. Seems like $1.50 is money well spent to have correct RS232 levels with esd protection. Users will be able to interface with a wide variety of RS232 equipment with confidence.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarlson1 wrote:
True RS232 swings negative.

Agreed. I've never seen what I described cause problems with any serial communications, but I probably wouldn't sell that to customers either (unless that is what they wanted). I think it would be OK for a dev board that people want to get dirt cheap and will likely only be connected to a PC with a cable less than 3 meters. $1.50 is just the cost of the IC. There is also board size, part procurement, part placement, profit (I assume). Probably really adds $7.50 for small runs.

It was just a thought...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

nice and compact

Only thing I can think of is to add silkscreen info on the headers (a la STK500).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarson,

Nice board.
I think I missed it above, are you planning at the moment to sell the bare board, or a populated board?

If you do not have a silk screen, how about some Labels for PortA, PortB, ... in copper on the top?

JC

Edited.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@KitCarson:

Nice board and nice project. I'm also interested in this silicon, but the main problem for me is the debugging feature, since I'm used to JTAG (MkI), and the JTAGMkII is too expensive to justify it to my wife, at least, for now. I also know of some commercial products that will be using 44 pin Xmega's.

@DocJC:

Hey Doc, this is a damn nice board, and these really tiny vias around Xmega are driving me nuts! Which size are them? Too sexy for my designs ;) The smallest I could do right now are full through board and 0.2mm hole diameter plus 0.15mm corona, total 0.5mm. Too big for some things (I'm targeting BGA's).

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes the board have will have silk screen.

Not sure how to sell the board. Some may not be able to solder the TQFP100 or 0402 caps. Buying the parts in quantity helps reduce the cost. Not sure of the quantity. That is the problem.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Not having to solder a TQFP100 is a major (maybe the only) reason people would want to buy your board.

I think you need to make at least several hundred, maybe thousands of boards if you want to be able to sell it for a low price. With the risk you cannot sell all of them.

Marketing is always difficult and is a major activity of most companies I think. Actual product development is only a small part.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarlson wrote:

Do you desire a low cost xmega development board?
1. Yes, but it has to cost $100 or less USD
2. Yes, but it has to cost $80 or less USD
...
The proposed board will have an improved layout resulting in significantly reduced noise levels in the ADC system compared to the STK600 w/adapter.
The board(s) may measure about 50mm x 80mm (~ 2" x 3"), small enough to integrate into hobby projects.

I appreciate all comments. I would like see more.

The design is the fun part. Are you prepared to support the on slot of orders at that price point? See my blog entry on the subject:

"Is there a rule of thumb for estimating the cost of getting circuit boards assembled?"

http://blog.designer-iii.com/contract_manufacturer/20090616-211514-Is-there-a-rule-of-thumb-for-estimating-the-cost-of-getting-circuit-boards-assembled

"A reoccurring theme I see on message boards is why does it cost so much to have some electronic widgets manufactured. Here is some background for you that might help explain that, from a thread I wrote on an other AVR list..."

Rather long to post the full thing here.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In a previous life, I used to calculate manufacturing costs. Gross estimates usually equal the BOM itself, so the cost tends to double the bare products bill, although that depends too much on the batch size and manufacturing processes. A good design can reduce them a lot, being a bare board with only one side populated with SMD components or with little THT that can be pick'd and place'd the cheapest. Component arrays can also help.

Design for Manufacturing should be another subject at EE universities.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In quantity things can be become very cheap. One design I developed a tiny bit of software for was a miniPCI DVB-T card. Eight layers, only half a mm thick or so, filled microvias, 4 mil tracks etc etc. Cost: $0.80

Cost of the FPGA that was on it... just a couple of $, through regular distribution channels it's tens of dollars, even in quantity.

Of course this was a big multinational with all the resources to pull off this kind of things.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you want to get a few PCBs out there, as well as for yourself, might be worth checking this out:
http://www.seeedstudio.com/depot...

You could ask them to produce extras and if they would be willing to give you a certain percent of profits. I couldn't seem to find anywhere that would let you sell the design, without making it "free" though.

As for quantities, here is what it would cost me (single qty from digi-key):
max3232: $1.50
xmega128a1: $8.00
FTDI232: $4.50
11x headers: ~$5
mini USB: $1.20
passives, switches, etc. ~$5
+ ~$10 for the PCB (just made this one up)
= $35

I would think it would be reasonable if you could sell a kit for about this, or an assembled board for a bit more (~$45). I suspect there is not enough interest to allow such large quantities to make this worth your while though.

As a (cruel) comparison, you can see this board which doesn't have USB, but does use a micro that would cost $15 in single qty: http://futurlec.com/ET-STM32_Sta...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi Guillem,
Thanks for the kind words.

I'm away from home, but can let you know the details when I get home, in a week or so.

I rushed the board to get it done to take on a trip, and in the process took some short cuts, and made two non-critical errors in the lay out.

Such is life. That is why there are Revision Numbers :)

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@DocJC: AFAIK, a product released to the market with everything Ok (regulations, FCC, EMI, CE, UL, etc) with a board revision 3 or less is a real success. The PCB and/or product designers of such product can be considered good engineers.

@thmjpr: I have one of this nice board from futurlec, and I agree that they are damn cheap. But I will notice you that they have few advantages hard to beat by any occidental production: low cost for labour hours, and high voume production facilities and infrastructure that also lowers costs. Another important advantage is that buying max3232 directly to Maxim by Internet by full reels or boxes (that have few trays) will reduce your cost for that IC when compared to digikey (that are cheap also, compare them to RS or Farnell), and futurlec probably buys more than one reel at once, since they re-use the same components on many different designs.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Attached is an Seeeduino Mega board, Basically a minimalist dev board with 100 pin ATMEGA1280, built with the arduino bootlader. The board is pretty close to what I'd like to see for an XMEGA128A1-based board (without the arduino bootloader ofcourse). The price is $49.50 which i consider to be good.....

http://www.seeedstudio.com/depot...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For those ones who do not have around any wires, tools, solder station (home), would be nice to provide a RS232 connector and few push buttons and LEDs.
George.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guillem noted:

Quote:
these really tiny vias around Xmega are driving me nuts! Which size are they?

I use ExpressPCB, (I've not learned Eagle, yet, ? ever), and the small vias around the Xmega are 0.031" pad size and 0.014" hole, through plated, with the known possibility that at this size the via may end up being filled, and not an open hole.

Sorry for being Off Topic, but it kind of relates, as it is about an Xmega layout / board...

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have placed the schematic and PCB layout in the projects section under "low cost xmega".

I am moving on. Thank you for all the comments.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I was wondering if there would be any interest in some XMEGA dev boards I have made for a school project. It has an AT90USB162 on board as well to handle USB. All programming of the XMEGA will be over USB using the PDI (no external programmer required). The XMEGA can also use the USB as a general communication interface via a USB-SPI bridge. There is a serial SRAM, SD card slot, audio amp, temp. sensor, 5-way switch, and 4 LED's. All of the XMEGA port pins are routed to headers and the external peripherals are connected via solder jumpers so that all pins can be available. It is about the size of an AVR Dragon. A photo is available at http://xmega.mattair.net/JM-XD20...

https://www.mattairtech.com/
ARM Cortex M and XMEGA development boards / Gentoo Linux

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I was wondering if there would be any interest in some XMEGA dev boards

What determines your path to millionaire-dom here is exactly what price you think you can sell these for. ;-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Pricing is the difficult part for me. I was thinking around $100. I have to hand assemble them, which is a pain, so that may be too low. I can't mass produce them (yet).

https://www.mattairtech.com/
ARM Cortex M and XMEGA development boards / Gentoo Linux

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It's tricky isn't it? - I think the person with the first sub-$50 board is going to sell them like the proverbial hot-cakes but it takes quite a gamble to buy the components in the kind of quantity where you could benefit from economies of scale to make a profit on something at that price.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

To be honest, I think you will loose money. You probably have to sell them below cost price to be able to sell them, perhaps your board becomes a hot item and maybe then you can consider having them produced in bigger quantities (with the risk you won't be able to sell them). Maybe then you can make a bit of profit. But big sales also means a lot of cost and time to handle incoming orders, administration, storage etc etc. Don't forget warranty, returns and support. What about marketing?

I think that if you can sell a few hundred you have done very well.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

But I bet you sure will be tired of soldering after 100 or so of em..... thats what board houses do. Take orders and deposit in advance contingent on getting about 80 or 90 on order?

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think if you plan on making a profit off those the minimum order quantity you should aim for is one thousand per order. Price breaks for parts starts there for any serious discount, and 1000 qty drastically reduces board cost. For such a small and simple board, you can have whatever fab house assemble them for you, and still come out with a price less per board than if you ordered in 100 qty and assembled yourself.

For prototypes, never EVER order boards in less than 10 qty. Price of 10 boards often == price of one board.

I for one would be very interested in a smallish board all populated with pre-programmed bootloader for about 50$.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If someone would just make an Xmega based Arduino. The 1280 based Arduino Mega is $65: http://www.nkcelectronics.com/ar... surely it must be possible to simply replace the 1280 with a 128A1 for similar money?

Actually I'd be pretty happy if the mega128 on this $32 board: http://www.saelig.com/product/IC... were replaced with a 128A1

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@physicist
I sent you an PM.

I just suggested he contact my old employer and ask for a quote on assembling his/her board.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:
It's tricky isn't it? - I think the person with the first sub-$50 board is going to sell them like the proverbial hot-cakes
My concern would not be so much to reach the sub $50 range, but that Atmel might have managed to miss the window of opportunity for the xmegas with their policy. So there might not be much demand.

There are boards coming close to the sub $50 range, e.g. this one, but there doesn't seem to be a rush towards the board.

http://www.alvidi.de/shop/produc...

39,90 Euro = $56.8 single qty incl. tax. Excluding tax you are below $50, but your customs will than make sure it won't stay this way.

If you take 10 the price is still $52.6 incl. tax.

Stealing Proteus doesn't make you an engineer.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

physicist wrote:
I was wondering if there would be any interest in some XMEGA dev boards I have made for a school project. It has an AT90USB162 on board as well to handle USB. All programming of the XMEGA will be over USB using the PDI (no external programmer required). The XMEGA can also use the USB as a general communication interface via a USB-SPI bridge. There is a serial SRAM, SD card slot, audio amp, temp. sensor, 5-way switch, and 4 LED's. All of the XMEGA port pins are routed to headers and the external peripherals are connected via solder jumpers so that all pins can be available. It is about the size of an AVR Dragon. A photo is available at http://xmega.mattair.net/JM-XD20...

nice!

If you have a ton of PCBs some people may be interested in just the bare board.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks for the feedback. I have worked at an assembly house and have experience with most processes (stencil, solder paste, pick and place, reflow, wave solder, rework, etc) so I was considering doing the manufacturing myself (I just need a stencil and paste). Overhead costs shouldn't be that much based on past experience. Marketing might be tricky though. Bear in mind that this is not my main job, and the boards would be "brown box". I want to gain experience with the XMEGA now, because I think it will eventually do better, especially after the smaller ones become available. I have considered selling the same board without the peripherals for around $65. It would have just the essentials (including USB). Maybe I should look into porting Arduino?

https://www.mattairtech.com/
ARM Cortex M and XMEGA development boards / Gentoo Linux

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I got tired of waiting for someone to release a simple XMega board to try out the new chip. So I made one myself; basically just landing for the atxmega64a3 TQFP-64, 3.3V regulator, spot for crystal (which you don't really need since it has a nice internal PLL which can run up to 32MHz), standard 5V jack and bring out all the pins including the new PDI ISP connection and standard JTAG 2x5 connector.

The boards are out for Fab and should be ready in a couple of weeks. And yes, I had a few extra made to sell to anyone interested. :-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am interested, could you post a picture of the design?

thanks

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Sure. Here's a screen shot of the Eagle Layout. I'm thinking of spinning a second smaller board with one of the TQFP-44 parts. Leaning towards the Xmega64a4.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why not make a board in Arduino format (or Arduino Mega format) with an Xmega on it? You'll have to come up with an Arduino compatible boot loader, but that's probably already available.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I just recently finished a development board for my university as a theses project.

It features an XMEGA A3, ENC28J60 ethernet controller, power over USB and USB controller, JTAG, 4 switches and 4 LEDs, 10 pin connectors to PORT A, B and C, and is fully compatible with the Tuxgraphics TCP/IP stack. (implemented and tested)

The "do it yourself" price for 1 unit was about 60 USD.

Let me know if this sounds interesting and i can post a separate thread with more details.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

https://www.mattairtech.com/
ARM Cortex M and XMEGA development boards / Gentoo Linux

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok, posted in this thread:

https://www.avrfreaks.net/index.p...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When I ran out of timers I started looking for options and read about Xmega.
Reading this chain of conversation was interesting, as many opposed developing an xmega dev board.
it is 6 years since this conversation, I do not see xmega being popular till now.

What do you think, what is going to happen with xmega.

Thanks
K

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

it is 6 years since this conversation, I do not see xmega being popular till now.

What do you think, what is going to happen with xmega.


What makes you think the Xmega is not popular. Atmel have now released 10's of different chip designs and the thousands of posts in the Xmega forum shows that they are being widely used. Atmel wouldn't continue to release new chip designs (like the very interesting, recent E5) if the design were a lame duck and you wouldn't get all the chatter about them if no one was using them.

(having said that it's true that with Cortex M0/M0+ at 8bit prices quite a few people might consider a tiny/mega->ARM move rather than a tiny/mega->Xmega move these days).