[sdiy] Odp: Re: ARM Dev Boards Possibly Lunatic Idea

Sat May 13 18:05:20 CEST 2017

mars at pingdynasty.com wrote:
>We've made a similar mini-probe to fit a 5pin 1.25mm pitch picoblade 
>footprint. For debug builds I can fit a connector and have the 
>programmer permanently attached so I can do on chip debug over SWD. For 
>production I leave the connector off and use the pogo-pin probe to flash 
>them. I'll take a pic when I'm in the workshop later.
>
>I have no experience with the SAMs but they probably have a built in DFU 
>or serial bootloader which is activated by pulling some pins high, 
>allowing the device to be flashed over USB or UART. It's a real 
>limitation though if the debug interface is not broken out: no on-chip 
>debug!
>
>Scott, you say:
>>> the only reason I'm interested in that particular chip is that the 
>>> speed
>>> is almost 40% higher than the STmicro ones
>
>You mean 300MHz as opposed to 216MHz on the STM32F7? The SAM can only do 
>300MHz with a 1.2V supply, and the tsunami from what I can tell runs at 
>3.3v. There's probably a bunch of peripherals that won't work at the 
>lower voltage, the datasheet says USB and DAC requires > 2.5V.

yeah, the Tsunami doesn't use USB for data transfer AFAIK, there's a micro B port, but
I think it's just for 5V.

>The STM32H7, which is supposedly in production now but doesn't seem to 
>be widely available yet, runs at 400MHz. But there will always be newer, 
>faster parts - what do you need the speed for? To take full advantage of 
>an MCU like this will take years of development time (at least it has 
>for me!). You can start with a lower powered device and work your way 
>up. If you can do 2 voice polyphony on e.g. an STM32F4 then the STM32F7 
>is a drop-in replacement.
>
>Martin

I'm certainly not going to try to build an instrument of 128 voices and multitimbral
with mulitple effects.  This started because of a Karplus-Strong project I did on a
dsPIC.  It currently supports 10 voices and includes a chorus effect.  However, as so
many have said, the dsPIC internal DACs suck (audible noise mainly).  And even at
that, I had people ask me if I was selling it.  It supports pitch bend to -12/+12 and
has some other interesting features that are kind of "secret" to me at the moment :) 
including a way to eliminate the usual KS tuning inaccuracy esp. at higher pitches.

As for the need for speed, I wasn't aware of the 300 MHz limitations (voltage), so
that's a bummer.  The idea of replacing the dsPIC with ARM is to get away from the
crappy DACs and to be able to use float (in C) instead of 1.15 fixed (in assembly) and
hopefully get more voices (16 would be nice).  I licked my finger and stuck it in the
air to decide that a Flash based ARM of the Tsunami class should be able to do that
(in bare metal) and now I need to try it.  The F7 Discovery board is being shipped as
I write this and it is well priced for my purposes.  

I also find interesting the ARM 12 bit DACs which can run up to 1 MHz sample rate
though I'm not sure I'll have pin access to those yet.  That's a lesser consideration
anyway.

I've got a C version of the KS synth working on a Rpi3 at 32 voices, but the pi runs
at 1200 MHz and I'd not want to hide an Rpi3 in a box and since there's an OS
involved, crashes are always possible.  So this has now become an experiment in "how
much better is a Flash based ARM than a dsPIC".

>On 2017-05-13 07:16, Roman wrote:
>> This is exactly how I program all my PICs. My "programming pen" isn't
>> so nice, it doesn't even look like a pen. Actually it's so ugly that
>> I'm embarassed to show any picture. Done it probably 10 years ago with
>> the intention to make proper one in spare time next week. Well, that
>> never happened.
>> 
>> By the way that website is full of good stuff! Go and see it all if
>> there's anybody except me who haven't seen it already. That guy, Mike
>> Harrison, is a genius!
>> 
>> Roman
>> 
>> Dnia 12 maja 2017 23:37 Richie Burnett 
>> napisaÅ‚(a):
>>
>>> I've had success with using this type of improvised "programming
>>> pen" for
>>>
>>> in-circuit microcontroller programming before.
>>>
>>> http://www.electricstuff.co.uk/penprobe.html [1]
>>>
>>> -Richie,
>>>
>>> -----Original Message-----
>>>
>>> From: Scott Gravenhorst
>>>
>>> Sent: Friday, May 12, 2017 10:22 PM
>>>
>>> To: synth-diy at synth-diy.org
>>>
>>> Subject: Re: [sdiy] ARM Dev Boards Possibly Lunatic Idea
>>>
>>> No idea, other than the FTDI port, I don't see a connection that
>>> hooks to
>>>
>>> the appropriate pins on
>>>
>>> the ARM to do (for example) parallel programming of the Flash -
>>> there are no
>>>
>>> connections to the
>>>
>>> required pins for that. JTAG pins aren't connected. From what I see
>>> (and I
>>>
>>> could be "blind"), the
>>>
>>> only way they could have done it is via FTDI or possibly using a
>>> socket that
>>>
>>> accepts the SMD part,
>>>
>>> program it first and then solder it, but that seems a bit labor
>>> intensive to
>>>
>>> me. The datasheet
>>>
>>> lays out (I believe) 4 methods of programming the Flash, and the
>>> schematic
>>>
>>> didn't seem to support
>>>
>>> any except for FTDI. Note though that the FTDI port was just .1"
>>> spaces
>>>
>>> holes into which pins
>>>
>>> could be soldered, but the board comes with nothing soldered there.
>>> I
>>>
>>> suppose it would be easy
>>>
>>> enough to deal with that using some clip-thingies.
>>>
>>> And that's why I'm asking this question here.
>>>
>>> MTG wrote:
>>>
>>> How did the developer do it? There must be some kind of ISP
>>> connection
>>>
>>> on the board.
>>>
>>> On 5/12/2017 12:29 PM, Scott Gravenhorst wrote:
>>>
>>> Here's a possibly lunatic idea I've been mulling for awhile.
>>>
>>> I was thinking that the Tsunami Super WAV Trigger Board has
>>> essentially
>>>
>>> everything I'd
>>>
>>> want on a dev board, including the Microchip (Atmel) ATSAMS70N20 ARM
>>> (and
>>>
>>> again, the
>>>
>>> only reason I'm interested in that particular chip is that the speed
>>> is
>>>
>>> almost 40%
>>>
>>> higher than the STmicro ones).
>>>
>>> What I first wondered is whether it would be possible to completely
>>>
>>> replace the program
>>>
>>> in it. There is a firmware update function, but I doubt if that
>>>
>>> _completely_ erases and
>>>
>>> replaces the code. If I'm not mistaken (from looking at the schemo
>>> and
>>>
>>> reading the
>>>
>>> datasheet) I might be able to program the board through the FTDI
>>> port on
>>>
>>> the board.
>>>
>>> However, I'm not sure, so maybe someone here has advice. Anyway, my
>>>
>>> whole idea was to
>>>
>>> buy a second WAV trigger and blow out the WAV trigger program and
>>> replace
>>>
>>> it with my own
>>>
>>> designs - and use it as a dev board. It's a bit more expensive than
>>> the
>>>
>>> other
>>>
>>> selections we've discussed, but failing finding a dev board for
>>>
>>> ATSAMS70N20, if that
>>>
>>> could actually work, I wouldn't mind spending the money. The
>>> development
>>>
>>> platform for
>>>
>>> the Microchip/Atmel parts is Atmel Studio, which itself is free, but
>>> I
>>>
>>> don't know if as
>>>
>>> you use it you find that it needs bits and pieces that cost money -
>>> can
>>>
>>> anyone say if
>>>
>>> that is true or not?
>>>
>>> So is this total lunacy?
>>>
>>> -- ScottG
>>>
>>> 

-- ScottG
________________________________________________________________________
-- Scott Gravenhorst
-- http://scott.joviansynth.com/
-- When the going gets tough, the tough use the command line.
-- Matt 21:22