uCore

[Braintu]

Hi everybody,

first I have to say, that I am totally exited by the MIDIbox NG concept!

While I was reading the tutorial, I came up with an idea that I want to present you and ask for your oppinion.

Lately I was working with the BeagleBone and some time ago I worked (had to) with the Arduino environment - and the realy nice thing about these boards is the form factor. I thought that it would be great to have something like this for the MIDIbox environment. So let me explain this a bit more:

In my oppinion there are three main issues that make, e.g. the BeagleBone and the Arduino so handy:
1. The physical appearance
- size (fits in your hand and in small housings)
- data transfer and power are managed via usb
2. The programming environment incl. the library which are easy to use.
3. The documentation and the community support.

So here is how I see these points with the MIDIbox envoironment - let's start with the last point:

3. I realy like the way the user manual for NG is written, with all these nice headlines 'Remote me!', 'Toggle me!' and so on. Also lot of pictures witch is really good for newbies. And everybody here knows how good the community support is.

2. With the MIOS-Studio and the new 'midi scripting language' I think this point is checked.

1. I realy like the idea of having a core module and putting all the peripherals around it as you need it is a great concept. Also the idea to have it all in through hole parts makes it very good for beginners.

But:
From time to time I build some midi-devices - sometimes just for test reasons, sometimes to have a small controller. So it is a bit of an effort to build a new core everytime or even disassemble old projects to get a core module free - so I was thinking, why not build a core module with some basic in- and output possibilities. Than I started to play around with the core module schematic and try to figure out what is needed in small midi projects. At the same time I did some basic pcb test, just to figure out, what could be put on a small BeagleBone-sized pcb. I came up with the following:

- 8 digital inputs
- 8 digital outputs
- 8 analog inputs

First I tried to free some pins but it didn't seem to work and than the compatibility would mess up. So I put one 74HC595 and a 74HC165 one the bord. But the pcb was much bigger than the BeagleBone or Arduino - with the LPCXpresso and all the huge through hole parts. Therefore I had to make some radical
decissions:

- everything smd
- no Xpresso, instead the bare LPC1769
- no Ethernet connection
- no midi connectors on the board

At the first glance this seems to be a bit to radical, but I think it could be worth a thought.

Very often for small midi projects - or just to get something to work USB is fine. The midi connector is only used for hardware equipment - but could be connected to an extension header. Ethernet is mainly used for OSC, but nearly all the devices that speek OSC can also speak MIDI (not true for the mobile devices). With protocolls like Copperlan, OSC can be tunnel through USB, and as TK showed in a thread (sorry - can't find it right now) MIDI via USB is superior to OSC. Also I decided to put a microSD header directly on the board. 

 

I put a serial ADC and the DIO connector on the board so it would still be possible to add external boards like AINSER, DIN, DOUT.

So I was thinking about layouting a pcb for this. Here is a first rough sketch of the idea:

So please tell me:
- what you think about this concept?
- what could be changed/added/...?
- is there a general interest/need for this board?

I am curious for your comments.

Best,
braintu

Edited January 30, 2013 by nILS

[Duggle]

My take on this kind of idea I call "minicore".

It's a STM32F103RET6 ($7) on an SMD adaptor PCB ($0.90), from Futurlec. 

It runs MIOS32 great, as it should.

The nice thing about the adaptor is that you only put the interfaces on the underlying veroboard that you need for the project.

I have a development baseboard (pictured) with connectors such as JTAG (for flashing a bootloader or debugging) that wouldn't normally be needed on a "finished" item.

So the finished adaptor and baseboard would not need to be much larger than the SMD adaptor itself in many cases.

Very economical in both cost and size!

 

[nILS]

Braintu: Please don't manually add line breaks to your post. This may be fine for your screen but it's most annoying on bigger and smaller ones.

[ilmenator]

So please tell me:

- what you think about this concept?

- what could be changed/added/...?

- is there a general interest/need for this board?

I don't think we need another core module. Those with a broader knowledge in electronics will easily come up with solutions such as Duggle's. For beginners, adding more choices in terms of board variety is bad because it will confuse them even more. And supporting them will be more difficult.

 

I came up with the following:

- 8 digital inputs

- 8 digital outputs

- 8 analog inputs

We do have a lot of different projects running on the MIDIbox platform. The choice of components sitting on the core board right now is actually a very good compromise. Your minimal config of 8 I/Os and 8 analog inputs is already there!

 

- everything smd

- no Xpresso, instead the bare LPC1769

Going SMD is fine for advanced users. It scares most beginners. That would call for pre-assembled boards - very problematic, we had this with the Core32 boards that had the processor pre-soldered. Using an LPC "shield" is much more comfortable and makes for potentially more successful projects IMHO.

 

The midi connector is only used for hardware equipment - but could be connected to an extension header.

Having the MIDI connectors on the board is crucial for me. You probably don't remember the old 8bit PIC core board which used breakout cables - I am really glad we have left that era behind...

 

So it is a bit of an effort to build a new core everytime or even disassemble old projects to get a core module free...

That is something even your project won't solve :smile: .

[TK.]

Hi Braintu,

 

first of all: thanks for the compliments! :smile:

 

A general statement to an alternative core module: just two days ago Fairlightiii introduced his plans for an alternative PCB to me. It's smaller (10x10 cm), and has an off-board MIDI IN/OUT circuit (provided as a second PCB for those who really need this - two PCBs can be combined for 4 MIDI IN/OUTs). It has an SD Card on board as well.
I like it (sorry Ilmentator!) - not at least because I know that it will be perfectly documented in the Wiki, and supported by Fairlightiii.

 

So, partly it matches with your idea, partly you are trying to reduce the part count even more by creating some kind of LPCXPRESSO replacement dedicated for MIOS32. But I fear that such a solution isn't attractive enough for most people.

 

It isn't newbie friendly. Not only due to the SMD soldering, but also because somebody would need an external JTAG programmer to flash the bootloader into the chip, which increases the costs (and complexity) for people who don't own such an equipment yet.

 

Best Regards, Thorsten.

[yogi]

The current core carrier/LPCxpresso combo seems like a good design to me. For a special case design, a smaller footprint base could  be used. The LPCexresso (less JTAG side) is really just a breakout board, so having a different mother board, scaled to an app's needs, could be prototyped easily. The STM32 core was far less flexible.

The only fear I have is that at some point, the cost of the LPC board will jump. I saw this happen with Analog's ADSP2181 dev brd; went from $75 to $250, (when they saw plps using them for dedicated projects, I guess). As long as NXP keeps the price point for the LPCxpresso in the $20 ballpark, not sure how a 'new' board would be better without a production run of at least 500. At this point in time, you can afford to have a few LPCs, They'r just as cheap as a Arduino UNO.

Nice idea but not sure how needed it is.

[Braintu]

Hi everybody,

sorry for the delayed answer...

 

My take on this kind of idea I call "minicore".

This seems to be a very effective way but you still have to solder and build all the parts around it, depending on your project.
 

Braintu: Please don't manually add line breaks to your post.

Sorry, I was in the 80 char-style.

 

 Your minimal config of 8 I/Os and 8 analog inputs is already there!

As far I understand the schematics, that's not 100% true. Yes, there are GPIOs available, but they do not have a e.g. 220 Ohms resistor for LEDs or a pull up for buttons. So I would not totally agree with this.

 

That would call for pre-assembled boards - very problematic,

Why so? What would be so problematic with pre-assembled boards? I mean from my point of view, building a midi controller is not so much about soldering - for me it is about designing the UI and build up great ideas in terms of building something that exactly fits to your needs.

 

Having the MIDI connectors on the board is crucial for me.

Yeah, maybe because you have hardware equipment, like synth you want to control. In my opinion, today most of the sound generating equipment is software so a USB would perhaps be enough for lot of people. As mention in a later answer something like a shield could be the answer.

 

You probably don't remember the old 8bit PIC core board which used breakout cables

Indeed I do - build my first sequencer with it and also the Clockbox...

 

 

It isn't newbie friendly. Not only due to the SMD soldering, 

Do you also think that pre-soldered boards are not a way to go? - I mean what would be the problem?

 

but also because somebody would need an external JTAG programmer to flash the bootloader into the chip, which increases the costs (and complexity) for people who don't own such an equipment yet.

Not so sure. If a pre-soldered board would be an option this could be already done, if not - on the programming part of the LPC-board there is a 10 pin SWO/JTAG connector so a JTAG programmer would cost 20$.

 

 

Even if the feedback is not that positive, I am still confident of the idea of a uCore. - Maybe because it fits so good to my needs.

 

I think I will draw the schematic and try some layouting and then post it here to get some feedback again.

 

Best,

braintu

[TK.]

Hi Braintu,

 

 

 

or a pull up for buttons

 

note that the LPC1769 has internal pull-up devices, which are normally activated by MIOS32 when GPIOs are used as inputs.

Therefore you won't find such pull-ups on typical IO ports on the MBHP_CORE_LPC17 module, such as J5A/B or J10

 

 

 

Do you also think that pre-soldered boards are not a way to go? - I mean what would be the problem?

 

The biggest problem is, that we are living in Germany... ;-)

This not only means that higher production and investment costs have to be considered (somebody told me once, that assembly companies typically expect at least 100 boards for a run in the fab).

But this also means, that you've to consider all this strange legal stuff such as CE, RoHS, EAR, WEEE, etc... here a good entry point for the never ending discussions about such topics: http://www.mikrocontroller.net/topic/176076

 

The most simple way to bypass such restrictions would be to sell the pre-assembled boards outside Europe. This is what we did for the MBHP_CORE_STM32 module via SmashTV - but it was a different situation, since it was the only 32bit module at this time; the demand was high enough, and Smash fortunately managed to phase out the module before the next, more powerful MBHP_CORE_LPC17 came up, so that his pre-investions didn't get lost.

 

In your case I fear, that the demand is not high enough - and that you will have to organize and handle the whole topic by yourself.

Of course, I could be wrong... let's see how many people evince their interest in this thread.

 

Best Regards, Thorsten.

[Braintu]

Hi Thorsten,

 

note that the LPC1769 has internal pull-up devices, which are normally activated by MIOS32 when GPIOs are used as inputs.
Therefore you won't find such pull-ups on typical IO ports on the MBHP_CORE_LPC17 module, such as J5A/B or J10

 

- ja, you're right with the buttons/pullups. But for the LEDs there still has to be a e.g. 220 ohms resistor. I was thinking about having these soldered already to the board to avoid extra soldering. Maybe a combination of 220 ohm and a pullup is possible - but I am not quite sure at the moment. I think I will start drawing a schematic the next days so there it is possible to discuss about concrete problems...

 

But this also means, that you've to consider all this strange legal stuff such as CE, RoHS, EAR, WEEE, etc... here a good entry point for the never ending discussions about such topics: http://www.mikrocont...et/topic/176076

 

I must admit that I haven't thought about this! But there are several things coming to my mind:

1) In the past there was a shop from a guy out of Germany who sold only pcbs. (mikes shop or something like this) Not sure if there are different laws regarding not soldered pcbs.

2) I thought maybe for a forum community the legal issues could be avoided - but that's quite naive.

3) I've seen this discussion in the mikrocontroller forum - but in my opinion these people are always taking about what is not possible instead of figuring out a way.

4) There are a lot of shops who sell pre-soldered pcbs, like sparkfun, adafruit, dangerous-prototypes, ... but, yes, they are all outside of Germany. I once ordered some pcbs from smash tv and the shipping it self was already 50 $ - So maybe it is time to find a shop inside Germany or at least EU?

 

If it is alright with you, I would start a wiki-page about the uCore and the idea behind it. I had a look at Fairlightiii's wiki page and it's really great documented. If you agree to a wiki-page I would try to make it a bit like this page.

So I will start to do some schematics design and also investigate on the legal issue and then see how far I will get...

 

Best,

braintu

[smashtv]

I once ordered some pcbs from smash tv and the shipping it self was already 50 $

Most PCB orders do not cost this much to ship.

Shipping is done at my cost - I use "First-Class Mail International Large Envelope" service for the Low Value Orders, and either "First-Class Package International Service" or "Express Mail International" for the main shop depending on total value.

The big difference between Express and First-Class is tracking, and my parcel insurer sets a 'requires live tracking for coverage' value threshold.

The USPS rates are at https://www.usps.com/calculateretailpostage/welcome.htm for comparison.

Best regards

Tim

[Braintu]

Hi Tim,

 

sorry, I did not meant to make bad advertisement. 

But since you are here, could you imagine to have a reseller in e.g. Germany?

 

I do not mean myself, but maybe there are some people willing to.

Lot of these arduino/breakout-board-shops are popping up these days and maybe there is somebody how would resell your pcbs.

 

I am just thinking about making the live/ordering easier, cheaper and faster inside Germany.

 

Please let me and us know what you think.

 

Best,

Daniel

[John E. Finster]

reseller in e.g. Germany

 

I don´t think reselling the SmashTV boards in Germany would be much easier or cheaper compared to SmashTV.

 

If one would want to do that, one would have to order a huge amount of boards from Smash to keep the shipping costs per board to a minimum. SmashTV already puts the minimum price on the board itself, so there is just no way to get cheaper boards even if one would order a huge amount of them. Then one would have to pay german sales tax and, due to the amount one would have to order, also custom fees. So per board one would have to raise the price to add the sales tax, the custom fees and the shipping costs from the states (thats board + 19% tax + ~ 5% custom fee + shipping + 19% tax for shipping + 5% custom fee for shipping). Then the reseller probably wants to add a small expense allowance (for the time and the effort) and BOOM!

 

i guess the price per board would have been doubled by now.

 

One could save some money when one has a small business here in Germany to avoid paying extra sales tax. But then other costs would arise like income and business taxes plus the annual fee to the IHK (i hate those guys!!!). And the sales tax would have to be paid by the customer anyway. If a commercial seller would take care of the reselling then the boards would be even more expensive since he wants to make a reasonable profit out of it.

 

Plus think about the effort and time SmashTV puts into this community. He is always around to help if you have problems with the boards or with the kits he provides. He also has an excellent customer communication policy. If you order something you can check the status of your order on his site. If anything doesn´t go as planned you get the info straight away. One would have to be ready to do the same. 

 

As for fast: When i order some boards from SmashTV they normally get here in a week. That´s fast enough for my taste. I have to wait longer for the new boots my wife orders from OTTO here in Germany.

 

I know this sounds kind of bathetic, but really don´t think changing the staus quo would make anything faster/easier/cheaper for anybody. It would just complicate everything.

[technobreath]

Agree with all of the above. and:

 

I also think we should consider the scenario, that many sellers of the same products in a "small" community, would probably cause all of the sellers bleed money. I am pretty sure that putting together kits, get PCBs produced, managing the shop, costumer support - and all the rest which I don't even have the knowledge to list here :smile:, takes a lot of time.

 

A seller / shop also needs to be solid, steady and not on / off when it comes to all of these points.

 

Running a shop like this is basically a full time job, and I for sure wouldn't like to do all that as a sparetime activity. I simply wouldn't have time for anything else. And most of us have families, many have kids, including me, and finding time to even build the kits can be challenging.

 

So hands in the air for you SmashTV for making this available for us! :smile:

Edited February 7, 2013 by technobreath

[yogi]

Running a shop like this is basically a full time job, and I for sure wouldn't like to do all that as a sparetime activity. I simply wouldn't have time for anything else. And most of us have families, many have kids, including me, and finding time to even build the kits can be challenging.

 

So hands in the air for you SmashTV for making this available for us! :smile:

Couldn't have said it better, nothing but RESPECT to Tim. \O/