Idea for future revision

Mainly USB2 and DDR, but as well GPMC , LCD and Camera in case everyone begin making their own routing for all kind of different boards with no really knowledge about how/what to do. J This being said I have even once experienced problems with I2C due to having too fast rise times compared to the length of the I2C lines…

Secondly it would potentially end up in a huge mess with different pin-muxing configurations for all the different boards :slight_smile:

The above concerns taking into consideration I though in general like the idea being an ideal concept J

Best regards

Søren

Beagle is not intended to be the ultimate development board. At $149 there is only so much you can add.

GPMC will not be added.
DDR is as short as it can get and it will never be brought out to anything on Beagle.
Camera is a slight maybe, but no plans to do it as it would force a re-layout of the board and add cost.
I2C can be looked at assuming of course you are more specific as to which one it was. This could be addressed in layout or changing the resistor values.

Gerald

Hi Gerald,

Please don’t get me wrong – I didn’t suggest the changes mentioned. I just tried to comment on the fact, that what might sound easy and convenient from a SW point of view (bringing out all connections to 0.1” pin headers) might not be that easy from a HW point of view…

I don’t and won’t expect SDRC and GPMC to be brought to pin headers. I can perfectly live with the Beagle as it is – It’s currently already a very great deal fitting many people’s need. J

In case I should suggest anything I would like to see the keypad matrix from TPS65950 brought out to pin headers for easy interfacing to some kind of home build keypad and/or having more accessible GPIO. That being said this can be done with external I2C chips as well, so I don’t consider it a bit deal, but something more convenient in case the routing change is easy?

Best regards

Søren

Keypad idea is a good one. I will take a look and see what we can do.

Gerald

Not likely. It is a cost adder. If it costs money, it most likeley will not happen.

Gerald

Maybe SATA interface is a good choice

2009/5/28, Gerald Coley <gerald@beagleboard.org>:

What would people think about make the existing “Expansion Header” the same pitch/spacing as the LCD headers. That would save quite a bit of space which could be used to bring out more pins like ADCIN, GPIO, Keypad pins…?

Not likely. It would have to go on the USB port using a USB to SATA converter. To get these chips, you have to buy at least 25K to ge the cost down. But then, we are stuck with all those parts. I know this because I have already designed a SATA interface for the OMAP3 on another board.

Plus, we would have to add a hub to connect it to the one USB port that we have and still have an external USB port availoble. Then we would need to increase the board size and pay the NRE for a new layout.

Plus, we would have to ad the 12V supply to the board, unless we only supported 5V versions. History says if we did that, someone would then want to support their older larger drives, and want the 12V.

I estimate these changes would add $15 to the cost of the board. I just priced an external USB to SAT at $15. This would mean only those people that wanted USB to SATA would have to pay the $15. Putting it on the board makes everyone have to pay the $15. This is what is behnd the “Bring your own” perpheral philosphy of the Beagle Board. We will not be abandoning this philosophy.

Gerald

I have gotten nothing but grief for the pitch of J4 and J5. If I were to change the pitch of the expansion connector and add the signals, then we would need to go to 8 layers and have to do another layout. It would increase the cost of the board by about $3 and would render any expansion card that anyone has made useless on this revision of the board.

Let me ask this. Would everyone be fine with the price of Beagle going to $199? In affect, that is what all of you are saying as each of you want your feature added to the board and there has been no mention of what these changes are worth.

Gerald

Sorry, just suggesting ideas and wondering what people’s thoughts were… Since you are probably too far away to give you a coffee and a man-hug, lets start with a huge “Thank you”. The board absolutely rocks… Its great and suggesting ideas is in no way a criticism. I’m sure there are people out there who think they know-it-all and moan, but on the whole most of us are trying to come up with helpful suggestions.

I don’t know if it would seriously inconvenience anyone to have the pitch on the expansion header reduced. That’s why I asked…

Sadly, I started this conversation - at the suggestion of one the core team. Perhaps to stop these email flurries in the future we should just point people to a Wiki page which is checked periodically for realistic ideas…? You can then state the various “rules” guiding the design once…

Assuming these ideas don’t cause major problems for anyone, here are some cent saving ideas: (a) remove the S-Video connector (b) remove the JTAG header © remove the DC power connector.

I wouldn’t mind paying the extra $50 for additional features. The BeagleBoard is an incredible value.

If the layout/connectors are going to change, then I would ask for a single 100-pin .05" pitch header. You can still solder wires into the holes if desired, or use a header.

Alternatively a higher-density SMT connector could be used as long as it is readily available from stock at Digi-Key in low quantity and reasonable price. That would then create a need for an expansion board for prototyping purposes.

I guess I would prefer a through-hole connector, but the 0.05" pitch doesn’t bother me if it would buy us access to more features. I would put more GNDs on the connector also.

-Howard

Removing the DC connector:

  1. makes the EHCI host port basically useless as the USB port will not provide enough power to the EHCI port.
  2. We plan to double the memory, which will take even more power further taxing the OTG port power
  3. The remote possibility of adding a feature, which will require more power.
  4. The need for the ability to add power for the expansion boards and LCD boards that are coming which certainly cannot be handled by the OTG port.

JTAG header adds no cost as it is just holes in the PCB. If we were to add a simple “free” feature, then that could be an advantage. Of course, the DSP people would hunt me down and shoot me, but I suppose I could change my name if needed.

S-Video, could go. But, we just improved it on rev C3, so I would like to get some feedback as to how affective they were before I saved the $1 by removing the S-Video.

Just because each of you may not need a feature, it deos not mean that it may not be needed elsewhere. We spent three months arguing about the features that ended up on Beagle. A lot of thought went intot it. Having designed over 20 OMAP EVMS, I have had some experience on the various tradeoffs. I have seen features added that 1% of the people even used and watched money sitting on people’s desks from HW they never even used.

We added more memory and the EHCI port and never raised the price. We just ate it. I had rather spend money on more MHZ and memory, than adding a feature that only a few people need. It is a tough balancing act we have.

Gerald

The RS232 chip already has several power control pins and iirc,
autoshutdown modes. Its just not wired as such. That'd be a change I'd
vote for, as well as wiring up some of the hardware flow control
lines. Perhaps they could be used to wake the processor from sleep.
(eg, processor goes to sleep, CTS goes low. Host needs to send data,
raises RTS, BB wakes, raises CTS, host sends data). The extra control
lines may require a larger RS232 chip which would then need more power
so it'd be a project to work out.

My vote would be for bringing out more of the TWL4030 pins. Such as
the power button and the battery pins. It would still work as it does
when shipped, but could be modified to support power from an external
battery.

I will add it to the “list” of things to consider if we do a respin of the board somewhere in the future.

The battery pin is already brought out on Rev C3.

Gerald

I don’t know it Russ referred to the Backup Battery Pin or to the pins needed for charging control? Anyway, some pins for charging control might be another useful set for people intending to do real handheld battery driven equipment with their Beagles…

Best regards

Søren

While some of these items that have been suggested may find their way into a future Beagle design, I can assure that support for a battery and it’s associated charging is NOT one of these.

Gerald

I understand the problems associated with bringing these pins out,
such as the good probability of someone catching a battery on fire and
the possible associated liability. Even a threat of a lawsuit would
likely be enough to shut the project down. I can't say that I'd do any
different in Gerald's position. It'd be nice if providing necessary
warnings with the product and making the modifications required to
support charging reasonably difficult were enough, but I understand
Gerald's position on this issue.

You have the right perspective. If this were a full blown EVM, say one that costs $1500, and it would mainly be bought by those with lots of experience, things would be different. But then of course, that would not be what the Beagle is all about. As I said before, we spent months looking at the feature list on Beagle. We started at $1500 and ended up at $149. In the process, a lot of things were taken out.

As it relates to the battery circuitry on the TWL4030, I will just say that I spent two years supporting the battery interface on the TWL4030. I will leave it at that!:^)

Gerald

I like the idea of more memory. I held out about 3-6 months before
buying the RevC2 as my first BB for this reason.

For those wanting a power switch... splice one into your power cable.

For the most part, I really want to see improvements in the software
side (ie: everything "just working" with any of the various linux
distros). I lean a little more towards the hacker/user side of the
spectrum (not so much "developer"), so I feel just a small twinge of
buyer-remorse because newer/better versions of the hardware will come
out and eventually be usable out of the box, while I currently have to
just sit and wait several months while the experts get things working.

The BB is still a very cool powerful and low-power device and I intend
to work it into my (robotics/wearable) projects as soon as I'm able.
:slight_smile: