I write this post with some caution. Let me say up front that there is probably between 0% and 1% chance of this actually happening - it is very much a personal pipedream. Let me also request that you don't email or PM me with questions about what features might or might not be included in such a hypothetical thing - I'll probably just ignore you even if you do have a red or blue name on this board

With that out of the way, let me explain what is on my mind. I am considering commissioning a design for an OMAP3 (ARM) single-board computer for a personal project unrelated to EMPEG but which just so happens to require a decent audio output stage, a dirty 8v-18v power input and the ability to sense or switch a number of 12v lines. Furthermore, it would have Ethernet and USB OTG, an LCD interface, be low power and likely to be similar physical dimensions to the EMPEG main board.

This is probably a good point to remind you of the content of the first paragraph. I should also add that in this increasingly hypothetical scenario, it would only be a main PCB - it would be up to the community to get software running on it and design any daughter boards / interface boards (such as front panel).

Now to the questions.

1. If I could offer a PCB with identical dimensions and fixing points to the EMPEG MK2a PCB and a target price tag of $200-$300, would that be of interest?

2. If you answered yes to the previous question, which specific subset of rear connectors on the MK2a do you feel it is important to retain the exact mechanical position/location of?

3. Following on from 1 and 2, how would your answers differ if a completely different car dock connector was used?

Many thanks for reading this far. Please now read paragraphs 1 and 3 one more time

--
mdavey

I don't have any input regarding the inputs (ha!) or outputs, but even if the hardware was completely FREE, by far the most daunting task would still be the software.

To put this into perspective, we all currently have empeg boards that are paid for. With hardware that's largely understood. With a working and well hacked kernel. No one has stepped forward to completely re-write the player software.

I think the idea would be that the player software wouldn't need to be rewritten. With a suitably patched up kernel the player software could be convinced that it was still running on the old hardware.

If that's the case, then I don't see a point really. I still have two empegs that function well.

Yes, that's the way it would go, for sure. Assuming the OMAP chip in question here is 100% percent instruction-set compatible with the current empeg CPU.

I like my empegs, but do wish that they (1) had a fast comms/expansion port, and (2) had a quicker CPU at times.

Cheers

Ok, I'll give you this. It would be nice to get a much higher transfer rate moving stuff onto or off of the empeg drives. But apart from that, for music playback, the limitations are now, and would continue to be, in the software.

If you built something brand new you could potentially slap some other existing software onto it. But at that point it's not really any part empeg.

So for me, with the existing software, I'd like to be able to move music to the device at a much faster speed. I wouldn't want to give up the existing display unless it could be replaced with something with greater clarity.

If we're talking about new software. Then the sky's the limit and I'm sure it can be discussed to death forever. With fast enough hardware it might even make sense to just get Slim Center and SqueezePlay running on it. At least that's something achievable, unlike writing brand new software.

When this has been talked about before we've talked about keeping the existing display.

Maybe.

The only connector that needs to not move is the ethernet port, so that Mark's docks still work.


The connector is separate from the main board isn't it ?

Wait, what?
I've got power, serial, sound, and ethernet hooked through a ML dock. Any new arrangement on the back of an empeg would require all those connections to remain constant for the dock to work.

Am I lost?

Yes

Those are all wired via the docking connector.

Wow, $200-$300 in low volume (I assume low volume as you mention it's a personal project)? Your distributor must be giving you a hell of a price on the OMAP3 processor.

We were looking into using the OMAP3 for a project last year but went down another route.

For me, I'd have to say no. Don't get me wrong I love my Empeg but it rarely gets used any more basically because of it's form factor. It just won't fit in the dash of my car anymore, which makes it a little redundant to me.

Now if it were in a different format which gave me lots of mounting, display and control options whilst running the current software then I would be very interested indeed

In an ideal world I'd like something that is similar to the current Parrot units, a small readable display along with steering wheel controls that blends in with my current factory headunit. If it did hands free phone too it would be really perfect.

Cheers

Cris.

I could be interested if the new mainboard was a direct drop-in for the old mainboard. Even better would be if a new frontpanel with the switches in the same place could be developed, but with a display which is more readable in daylight. Ethernet should be in the same place since, as what have been said earlier, the dock connector isn't affixed to the mainboard.

Stig

I would be interested if the board had at least 100Mbit and preferably a Gigabit ethernet connector, and a SATA disk interface to go with it (or the faster ethernet connector would be of no use). Everything else is perfect on my Empeg as-is, on the motherboard that is. There's of course always room for a better screen, a more 'modern' looking front etc... but we're only talking about the mobo here, right?

I think, unfortunately, that this is probably not worth doing. I'd love it if it was, but I feel as has been said by others that the software would be the problem.

I'm working on a couple of designs using the OMAP3 processor at the moment, and I and a couple of others did discuss exactly this project some months ago. The hardware is fairly easy, although the processor itself is a bit of a sod to lay out a PCB for, being 515 pins on a 0.4mm grid, as I'm sure you're aware

We have been quoted some very nice prices, which would make the cost you mention plausible. But I suspect that either rewriting the software without access to the source code, or writing an empeg emulator, would be somewhat tedious.

It's a pity, as a 600MHz processor with 256MB of ram would be a nice performance bump from the existing system. That said, the existing one has kept the thing going far, far longer than anyone would have expected 11 years ago.

Also, as the processor doesn't have that sort of hardware built in, adding 100/1000 Mbit ethernet, Sata, etc, would certainly bump up the cost and complexity significantly. The thing is aimed more at graphics and audio. You do get a hell of a lot of processing power for the price, though.

pca

IMO, if anyone is going to make custom hardware, they'd be better off building something with a decent SoC and keeping it TINY - smaller still than some of the others out there, with all the required bits to make auto connection simple (12v trigger relay input for instance). Then with integrated WiFi or Bluetooth, pair it up with an iPod Touch or iPhone. On the little computer you'd run Linux, Slim Center and the new player, and on the iPhone/iPod you'd run one of the controller programs such as iPeng. Hopefully with landscape support.

That said, you can probably do this now with existing micro PCs, though I'm not sure which ones would make for a nice glove-box install.

Ya think?

tanstaafl.

The Squeezecenter developers have squeezecenter running on a sheevaplug, all you need to add is a USB wifi, hardrive, and soundcard. The schematics are available for the sheevaplug to wire in a 12V-5V switching supply directly to the board instead of using a AC inverter. I'm not sure if the sheevaplug has enough processing power to run squeezecenter and squeezeplay; the sheevaplug is not high on priority.

That post was written a long time ago. More of the pieces are available off-the-shelf as open-source than were then; and some of the code that was so valuable then is effectively obsolescent now. Nobody needs to bother rewriting the disk cache code, for instance: just use SSDs and keep them permanently "spun" up.

Peter