MOTM

I have to admit that I'm a bit disappointed about the elimination of  
the kits.
However, maybe there are other ways to achieve some of the objectives.
One of the reason I like kits is so I can do mods as I build.  And it  
helps
greatly that MOTM modules are so throughly documented.

But SMT and tight layouts make hacking difficult.  So here's my  
suggestion:
as the new modules are designed, bring certain key signal input and  
output
points out to headers or jumpers.  This will give the DIY'er  
opportunities to
reroute signals, add additional controls, external circuitry, etc.

I do think the AudioEngine is a cool concept.  My suggestion on this:  
would it
be possible to create a Chameleon-like module that allows for some user
programming?  I'll have to admit I'm not sure how this should work; I  
kind of
had the idea that the Xilinx would be programmed with some kind of  
simple
DSP core, where the user could write processing code in, say, C and  
download
it to the module via serial, JTAGS, or maybe MIDI.  Non-volatile memory
would allow for code retention with the power off (and the  
possibility of
being able to store several programs and select them with a front panel
switch), so that once loaded the module can be disconnected from the
computer, for live use or just to get the computer out of the picture.

And yes, I will have to admit that I'm one of those who haven't ordered
anything in a while.  But I think I have a pretty good excuse: my  
spouse and
I are building a house; our builder abandoned the job and we are having
to finish the house ourselves and pay for everything with credit cards.
Right now, I'm so close to my credit limit that I can't even order a  
power
distribution board!

a) if it makes sense to add "hack points" on new modules, I'll certainly do so.

b) the AudioEngine has 2 programmable components: a Xilinx Spartan 3E 500K FPGA 
and an Analog Devices ARM7TDMI uP. The FPGA uses an external "Boot FLASH" device 
(also a Xilinx part) that load it's contents into the FPGA every time power is 
applied. The uP uses it's serial port for programming (like the MOTM-650 uP 
does). I have not thought about if the AudioEngine is a true 'user programmable' 
device or not. I don't have plans to make a "general purpose signal processing 
module" because, frankly, that is a *bad* business decision (everyone buys *1* 
and then that's it).

The universal way to program stuff is to use a low cost (~$15) cable that 
attaches to a parallel port (printer port) of a PC and bit-bangs a serial 
protocol called JTAG. The other end of the JTAG cable is a 6-pin 0.100 header. 
This protocol has been around 20 years. There is probably even an OSX version :)

In order to write code for the AudioEngine, you need 2 programs: one for the 
ARM7TDMI uP, the other for the Spartan 3E. I plan to use the tools from Keil 
(www.keil.com) because I have been using the 8051 version for like 10 years. The 
ARM version is very nice, you can get a free CD that will compile up to like 8K 
of code. The ARM part I will be using has 64K of code space. There are free 
tools that run under XP and Linux (Google 'ARM7 toolchain').

The Xilinx portion is a XP (possibly Linux, there is some debate if it works 
100%) program that is free, called WebPack. It's HUGE, like 400MB zipped or some 
silly thing. But it does a really impressive job. I use a variant called ISE 8.1 
that was $1200 last year but has since been "rolled into" the new WebPack 8.2 
that came out last week.

The AudioEngine will have 2 types of ADCs: 8 channels of 10-bit (maybe 12) for 
reading pots and FM CVs, and 2 channels of true 16-bit for pitch or other 
critical inputs. There are 4 channels of 12-bit accurate DAC output for 
non-audio voltage output, and a stereo 24-bit DAC for audio out. There are a 
large number (probably 40 or so) I/O pins to assign to things like LCD panels, 
rotary encoders, LEDs, whatever. There will be a USB port "hook" as well. The 
SPI buss that connects everything is brought out with additional chip selects so 
other peripherals can be easily added.

I'm sure after I build a batch I will think of an AudioEngineII that fixes 
everything I forgot :)

Paul S.