Help with Portable Radio/Media Player coding

rsavas · Joined: 10 Aug 2010 Posts: 197 Location: Baldwin, Ontario

Hello everyone reading this post.

I have designed a product that I call my Portable Radio/Media Player. I have written much of the code, enough to have tested out all the HW(hardware) and make the design very functional to this point in time. There is a lot of coding left to do, as I have lots of ideas for additional features.
I will include a document describing the functionality, pictures of the HW etc. Sorry it is in MS word, not to good at HTML!!
I do not mind sharing this design with anyone that is interested in contributing. The parts costs are fairly expensive, around the $250 USD level, so I won't be covering those costs for folks that want to contribute. There is not much sense in contributing to the coding, if you do not fabricate the design. Since there is a lot of surface mount (SM) parts, you may require my help to solder one up for you. I can help in that regard, if you are not confident in your SM soldering skills. Depending on the amount of people wanting to be involved, I do have limited amount of time that I can offer for assembly. We can work things out on a per person basis. The BOM is loaded into Digi-Key, as that supplier was the one who had all the parts available. The bare PCB is setup at a US based PCB fabricator. I only fab'ed 2 prototype PCB's to start debugging.
This was my first design using Atmel Xmega and BASCOM-AVR. I found it to be a great tool for learning both the HW and SW aspects of product development. My intention was to make something of use to myself and possibly make this into a product for sale. I do not believe that I will try to market this product, so this is why I am proposing to open up this design into the public domain, for others to use as both a useful product and a learning tool.
Let me know if anyone is interested. We can either take our discussions off line or we can use this forum to expand our discussions for everyone to learn from. The whole idea here is to share our knowledge and make this design availabe to the BASCOM community. If Mark Alberts is interested, I can release the design to MCS Electronics, to sell if they are interested. Mark & myself would have to discuss this in more detail, if he is interested.
At this point, I will retain design control of the HW.
I actually presented this design to Elektor this week, for use in a magazine article. Their response to me was "thank you for proposing your project to Elektor. I’ve had a look at the material supplied and am deeply impressed by your work. Sadly projects of this size and complexity do not fit the current publication strategy for Elektor magazine." Oh well, I tried
I will try the MCS approach, to see if anyone from the BASCOM-AVR forum, is interested in contributing to the SW development effort.

Kind regards
Rick Savas

Paulvk · Joined: 28 Jul 2006 Posts: 1257 Location: SYDNEY

Very nice!!

Would be good to know how much the boards cost you. Some things that would interest me and others like me is to have narrow band width abilities plus single side band. I am like many others, a HAM radio operator this is where the people are that would build and have the skills to build this. Also have a look at http://members.home.nl/bzijlstra/ Ben has a transceiver project there that I am certain will give you some ideas. I am busy at the moment with the web server but would like to build one as soon as I have time and having the ability to add/change the code is great. Some of the things to add would be to have software on a PC that puts a radio on the screen that you can control and also TCPIP control with wiznet this should not be hard. So if any bugs can be ironed out hardware and software wise. I get a journal that I think would publish this as a project if it can have the narrow band an single side band abilities.

Regards Paul

albertsm · Joined: 09 Apr 2004 Posts: 5921 Location: Holland

hi Rick,

Very impressive. Most people start a simple project when using a new chip, but you did not mind. It is super that you got this working. The PCB looks well designed. You implemented a lot of features. I wonder how much time you spend at his project. you must have learned a lot. Great that you want to share your project. The SMD and high part cost will be a problem for some, but it will be also an excellent project for those having the time and interest. Part of the fun is the soldering, but for some, the fun is in the programming. I think it is best to organize some kind of interest poll.
- pcb only
- pcb with smd parts
- ready pcb

I will send this info to some people i know. We could also publish this project as an AN. You can then combine this with a poll. I will discuss this off line with you.

rsavas · Joined: 10 Aug 2010 Posts: 197 Location: Baldwin, Ontario

Hello Mark, Paul & everyone else

Thanks for your kind reviews and possibly opening the door to make this available to all interested. Yes a poll is a good idea to judge the interest level. This is certainly not a beginners project for SMT assembly, but once you have working HW, then it is dream for someone to learn some fairly advanced programming techniques and first & foremost a great learning tool. What is most important, it is something of everyday use and will offer years of enjoyment. One of my objectives is to have something that has a very long life cycle. This not cheap through away electronics from offshore. I was at a bit of a loss with Elektor response but I look at it as their loss and BASCOM gain. To bad Elektor has a low opinion of the customer base capabilities.
I started the project using the XMega128a1BOB from Sparkfun, just to get my feet wet, about this time last year. So, I did learn to swim before jumping into the deep end Wink

I designed the board in the summer of 2011 and have been working on it, in my spare time, ever since. (The winter project of 2011/2012)
The raw PCB cost me $50 each, as there was a new customer promotion at the fab house. It is a 4-layer PCB. Price goes down with QTY in the build. In QTY I think it does down to ~$20 USD = reasonable for the complexity.
The DSP radio portion, is based on the Elektor design using the Silicon Labs Si4735 part. One feature that the part does not have any SSB support. This has been one criticizm by HAM operators but SILabs designed this part for broadcast use only, so I guess they consider SSB not of broadcast quality. I really only use it for FM. There are however selectable IF filters and other noise enhancement features. Will have to read the documentation, I can supply this or you can get a login at SILab's to get the large documentation set for yourself. I have a huge folder with all the datasheets,app notes etc.
Parts costs are high, because of the quantity of parts involved. This can be scaled down to use only what you require and stuff addition features as you wish. There is one problem with this approach because soldering the SMT parts is the first step in assembly. Trying to solder some fine pitch SMT parts lastly, just makes the task that much more difficult. The MAX9729 DFN was a problem for me, good thing the part is only ~3$ as I destroyed one in my first attempt. Obviously this design is geared for automated assembly and I have designed it with this in mind.
If you review the Elektor DSP radio articles, Mr Kainka has implemented a virtual UART using the FTDI part. I obviously did the same, so you can control the board using either a terminal or there was also a few VB6 applications, that he wrote, as another method to control the board. I have tested this out and it works as originally implemented. Very simple ASCII control characters are used to send control codes and to transfer data. This opens up the door for lots of additional control and a very nice GUI on a PC. This area requires a lot of work, which I have almost no experience at all!! = Call for some HELP!! I realize that this is out of the scope of BASCOM in its real sense, as it is a separate effort on its own.
I did not implement any TCP/IP control in HW on the prototype, but that is a logical next step in feature set if you think that it buys you features that you can not do, with what is currently available. I guess the design can be made to work as a web page, a whole different approach.
As with any prototype, it is pretty hard to get everything right in one go. Mechanical design is another weak point. I should have known from my professional career, in that I should have done some ME design first or at least in parallel with the HW tasks. It was not until I got the basic SW functions going that I realized that I would use the bottom row of the LCD, to display the encoder and switch menus, as I switched through the operation modes!! Oh well, a one man show has it limitations!!

So lets see what becomes of my offerings to the BASCOM community. I welcome everyone's involvement, at any level. The BASCOM community has offered me lots, to get to this point, so this is my offering in sharing of the fruits of my labour.

Best regards
Rick Savas

Paulvk · Joined: 28 Jul 2006 Posts: 1257 Location: SYDNEY

Hello Rick

With the board will you have to pay tooling costs again to get more made and does that fab house have the ability to source and mount the surface mount devices. Also is the board at its final without errors. The lack of SSB is not a show stopper but puts some limits to the market is there a drop in substitute to get it. VB is not a problem, for many here like me started with it and it can even go to Linux with Gambas.

Regards Paul

rsavas · Joined: 10 Aug 2010 Posts: 197 Location: Baldwin, Ontario

Hi Paul,

Yes tooling costs are just part of getting a board fab'ed. Part of tooling costs are DFM analysis, so that has been done already, at that fab PCB supplier. Bare board testing is also an additional charge. For my initial build, I did not pay for test, but they did it anyways for free.
Assembly costs are kind of the same cost model. The setup charge is significant, but if amortized over a large build quantity, setup cost become less on a per board cost.
I looked at assembly costs from that fab house, out of the question for me at that time, so that is why I did it by hand. It takes a fair bit of time/skill but in a hobby sense your time is not worth a lot of money, so I used hand assembly, as it is the only economical solution for a few. If we are talking a large build, then that is a whole other story. Since this is a mixed technology board, I can see doing the SMT parts by machine and do what we call a back load operation by hand. There are a few build models to follow, all depend on quantities/costs/parts involved.
There is no SSB variant available at this time, so that is one feature that will not be on the list.
As with any prototype, of this complexity, a clean design at rev 1.0 is a challenge. I did pretty good in this regard. What errors I have, amount to a few component substitutions. There are really are no layout errors as far as HW functionality. The AM band has some noise issues, that require the Si4735 circuitry to be shielded from the many oscillators on the board and external sources. I did think about this issue and have made the layout with this in mind. So the design is not 100% clean, in what I would call a design ready for mass production or what we used to call at "HP" Manufacturing Release status. I have a list of enhancements, which I have documented. For a hobby product, it is fine at this layout revision.
In the Mechanical sense, to make this a production ready design for sale to the masses, I would tweak the layout to match a proper mechanical design. Like I said earlier, I figured out that I'd use the 4th row of the LCD as menus for the encoders/switches after I did the SW. This means re-centering the encoders/switches in the middle of the LCD module. I'd move the UART/USB "B" connector & circuitry to the left of the first encoder/switch to make it symmetrical. The vertical USB "A" conn. for the USB memory stick, I'd move over a bit, on the right, to leave a little more finger room between Enc/SW6 and this connector.
I also found that for the TI PCM2902B that you can loopback SPDIF. I was unsure of how it would deal with copyright management. I added one barnacle (wire), for this function, so that I could record audio in digital format from an internet audio stream. Oop's, did I say I found a way to defeated the copyrights management scheme built into the PC? If I did a revision I'd put a 2:1 MUX so that I could select SPDIF sources for the PCM2902B transmit side.
Nice to hear that you are experienced in VB6 or equivalent SW to run a app/GUI on the PC using MS or Linux for that matter. This is the sort of help that I am asking for!!
So, I think I answered your questions sufficiently. Maybe more detail than you require, but sometime more is better than less.
There are lots of details, I am sure that you and others may want to ask. Ask away!!

Regards
Rick

Paulvk · Joined: 28 Jul 2006 Posts: 1257 Location: SYDNEY

So if I read what you are saying correctly you will not have to pay another set of setup costs to get more boards made I know some do charge every time. I think for a number of people having the SMT already on the board will be the diffrence between building it or not. Good job getting the board ok first off I know how hard that is I have made many boards and its hard not to have some small errors. I gather your using a character LCD the cost of 128 x 64 graphic ones is about the same so I have a number of these as well, it is much easyer to put prompts on these. Is there an rs232 port avaliable to control with this way add ons could be built to control it eg a colour touch screen you can get these for $18 mounted on a board with sd card holder and touch chip to your door, see Bens hobby corner for the style, about Iphone size or an infrared/wireless remote.
In the past I have used a china pcb fab house (very good customer service) so I did a quick quote on a board around the size/layers of yours the cost was about the same it would be interesting though to know what their cost to mount the SMDs would be but from my experience around 100 is a starting point to make things cost effective so now need 99 others Very Happy

.

Regards paul.

rsavas · Joined: 10 Aug 2010 Posts: 197 Location: Baldwin, Ontario

Hi Paul,

If you want a board or more, I suggest you use the same PCB shop in the US and go with the $50 per board one time deal, as I did. I can give you the mfg data. We can change the file names of the gerber files, so that they will not recognize the same board name. Parts I have setup on Digi-key, so I can share the BOM with others.
LCD is a standard 4x40 backlit, two controllers. I thought of graphics but biggest issue with them is that they do not come in the same width as the 4x40, at least not the cheap ones. The 4x40 LCD is around $25.00, a good deal for what you get, runs off 3.3V as well!!
To be honest, not really interested in changing the display at this point. As it is, I did not design the I/F to be 8-bit bi-direction as required for graphics displays. I used the 4-bit I/F and Luciano's 2-controller control code. If this was to be a real product I'd consider a re-design with a graphics screen. When you open up the design for feature input then you are going back to the product definition stage. Well this product is at the prototype stage, so can't go back on HW feature input because the board is done already. It is fine the way it is, for me at least, the only customer at this point in time. The 6 encoder/switch combos work quite well. Once you get the basic functions coded up, then there is million and one ways to make the UI. I struggled with this at first, since I have switches and encoders to work with. I am not a big fan of touch screen interfaces as it is., good for some things, others not, just a grimy screen. Gotta have a rotary controls for volume & freq controls.
The USB control port is using the FT232RL part, a type B connector and the FDTI PC driver.

Rick

albertsm · Joined: 09 Apr 2004 Posts: 5921 Location: Holland

Like with all projects, there are always more idea, wishes and features. but you need to take care that it will not take for never and you need to accept that it can not toast your bread in the morning.
So the option you have ; use it as is, so you save setup costs.
Or try to have some users involved and try to add the things you need in one go. I think that soldering is out of the question for most. But there is an alternative too: you can also have some of the parts assembled. this will save costs too. and a user can then assemble the other more simple parts.
The mechanical part (enclosure) is very important. Some kind of standard enclosure with a front panel would be needed.
I think it is time for a poll. Which questions need to be asked???
I think something like this :
"introduction..."
- Would you prefer a ready made and assembled board?
- Would you prefer a PCB only with parts (KIT)
- Would you prefer a PCB only
- Would you prefer the design only and have it made yourself

Do i miss something?

Paulvk · Joined: 28 Jul 2006 Posts: 1257 Location: SYDNEY

Hello Mark

I think the only thing would be "Would you prefer a ready made programmed and assembled board?" as some may not be able to program the xmega.

Now I have your problem Mark not enough time for one project and now three I want to do Very Happy

Regards Paul

rsavas · Joined: 10 Aug 2010 Posts: 197 Location: Baldwin, Ontario

Yes Mark, you pretty well summed it all up!!

The problems with doing partial soldering (SMT only) for sale
1) no guarantee that it works, after all is done and if it doesn't = problem? Upset customer!!
2) no way to test, must complete assembly in order to test
3) test only consists of functional, no ICT, mixed signal design
Full assembly & tested or forget it, if not = problems, that customer can not resolve
Two choices, bare PCB & parts and go for it (like I did) OR Full assembly & tested
Yup mechanical design/enclosure is a sore point at this time. Any ME's out there?
Like I said, I did it ass backwards, should have done ME first or in parallel with layout!!
Do not sweat it Paul, I'll be at this for a while, you will probably catch up later!!
Parts are supplied by Digi-Key, one stop shop. Makes no sense to have to deal with supplying parts.
I do not want to be a parts distributor, nor should anyone else. Leave that to Digi-Key & Mouser.
I got the BOM all setup & second sources (database) as well.
Because there are so many part types, you have be careful with out of stock items. Some items have second sources!!
There are over 100 different parts in the BOM. A box full of parts!!

Rick

Paulvk · Joined: 28 Jul 2006 Posts: 1257 Location: SYDNEY

Trouble with Digi-Key & Mouser is they are very expensive for their shipping to Australia $50 US or more Element14 /Farnell are in Australia so if they have parts I want its often cheaper even though the parts may be more expensive even had manufacturers send me parts as samples when I have asked if they have another agent that may not want $50 shipping for a $2 part. But with a list I can just look at my normal suppliers. Yes customer satisfacion is a problem with ICs only, its case of your dammed if you do & dammed if you do not. Well lets see how many others show interest.

Regards Paul

rsavas · Joined: 10 Aug 2010 Posts: 197 Location: Baldwin, Ontario

Hi Paul,

For both Digi-Key & Mouser, shipping is free for purchases over $200. I wait until I have enough parts to make my order over the $200 threshold. I agree, to order a few parts shipping can be more than the value of the parts themselves, so you have to be careful. I have found that Mouser is generally cheaper than Digi-Key, on most items. Splitting an order between the two can be a problem in that it can bring you below the $200 threshold, so your saving are erased, by the shipping costs.

Regards
Rick

albertsm · Joined: 09 Apr 2004 Posts: 5921 Location: Holland

albertsm · Joined: 09 Apr 2004 Posts: 5921 Location: Holland