The dsPIC does _not_ support DMA.

Bob Ammerman
RAm Systems
(contract development of high performance, high function, low-level
software)


----- Original Message -----
From: "Ed Browne, Precision Electronic Solutions" <Ed_B_PES@SWBELL.NET>
To: <PICLIST@MITVMA.MIT.EDU>
Sent: Monday, August 20, 2001 2:31 PM
Subject: Re: [EE]: SPI to Serial


> Thanks.
>
> I looked at this part early on, but it has no FIFO so the DSP would be
> interrupted for every 8 bit transmit.  Currently a DMA buffer is used on
the
> serial port so the DSP is not interrupted at all.  Haven't looked at the
new
> PIC DSP enough to notice if there are DMA's - almost mandatory.  Maybe I
> should look at it ....
>
> -----Original Message-----
> From: pic microcontroller discussion list
> [mailto:PICLIST@MITVMA.MIT.EDU]On Behalf Of Matt Pobursky
> Sent: Monday, August 20, 2001 1:15 PM
> To: PICLIST@MITVMA.MIT.EDU
> Subject: Re: [EE]: SPI to Serial
>
>
> You also might look at Maxim's MAX3100 series SPI based UART chip. It
sounds
> ideal for your application. It has an external interrupt pin as well. I'm
> not
> sure about non-standard baud rates, but I'm guessing that since it has an
> external clock input you should be able to find a clock (crystal)
frequency
> that will work for you.
>
> http://dbserv.maxim-ic.com/quick_view2.cfm?qv_pk=1731
>
> Matt Pobursky
> Maximum Performance Systems
>
> On Mon, 20 Aug 2001 10:39:20 -0500, Ed Browne, Precision Electronic
> Solutions
> wrote:
> >Hi, all.
> >
> >I am building an interface between an ADSP2181 SPORT (serial port) and an
> >RS232.  I have been able to make the DSP talk SPI before, so I thought a
> >fast PIC may be able to do the conversion.  Selected PIC for this project
> is
> >an 'F877.  I had hoped to spurt SPI data out as a very fast stream,
> >concurrently building the packet frame, then resend it as a more or less
> >continuous asynchronous stream.  The problem gets further complicated
> >because the RS232 is at 220.5kbaud - fast and non standard.  That problem
> is
> >solved by judicious selection of the crystal - either 14.318MHz or 18MHz.
> >
> >The DSP is a battery powered data acquisition system interrupt driven by
an
> >ADC.  Data is continuous and not all the SPORT handshake lines are
> >available.  The DSP does some pretty serious filtering before passing the
> >data so tasks are already pretty tight and optimized: get data, filter,
> send
> >data.  It already works and works well - 1 bit in 20 noise after a gain
of
> >103 (like I said, lots of filtering):-).
> >
> >My problem is this:  Using PICC, I have written simple minded code to
> >generate data so I can see if the PIC can keep up.  RS232 seems to be
able
> >to, but [[[I have seen very long latencies between SPI words (operating
at
> >CLK/4)]]].  Is this normal?  If these can be shortened, how and how much?
> I
> >have browsed archives and am starting the process of understanding the
> >compiler listing to see if there is something obvious.
> >
>
> --
> http://www.piclist.com hint: PICList Posts must start with ONE topic:
> [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
>
> --
> http://www.piclist.com hint: PICList Posts must start with ONE topic:
> [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
>
>

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads