Hi there,
I have a pb to reach maximum update rate for a serial Quad DAC TLV 5614 with
a TMS320F240 DSP that talks to him through SPI interface.
THe SPI is configured as follow
LDP #0E0h
SPLK #0087h,SPI_CNTL1 ; Reset SPI by writing 1 to SWRST
SPLK #000Ch,SPI_CNTL2 ; Disable ints & TALK, normal clock, master
mode
SPLK #0000h,SPI_PRI ; Set SPI interrupt to high priority.
SPLK #0000h,SPI_BAUD ; Baudrate = Fatest as possible ! SYSCLK/4
SPLK #0000h,SPI_STATUS ; Clear the SPI interrupt status bits
SPLK #0006h,SPI_CNTL2 ; Enable TALK, No CLK ph, master mode +
disable SPI Int
SPLK #0052h,SPI_PORT_C1 ; Enable the SPICLK pin function. ; SPISTE
as output pin et mise � 1 ie desactive
SPLK #0022h,SPI_PORT_C2 ; Set SIMO & SOMI functions to serial I/O
SPLK #0007h,SPI_CNTL1 ; Release SWRST, clock polarity 0, 8 bits
it sends correct SCLK=(SPICLK pin), /CS=(IOPA1 pin) and /FS =(SPISTE pin)
signals to the DAC
I wrote a routine that is called every 100 us and that refresh the DAC DATA,
say I am working only with one DAC (A).
each time it will send the data and increment by 1000 the data to be sent
next time.
it sends, 0 (ie 0V), 1000 (1V), 2000 (2V), 3000 (3V), 4000 (4V) then it
comes back to 0.
the pb is that I don't see all the steps but only changes every 200 us and
not every 100 us
my code seems correct because it takes into account but not at the right
rate :-(
every 100 us, these parts of code are called
...
LACL j
ADD #1000
SACL j
SUB #1000h
BCND pasdeRAZj,LT
splk #0, j
pasdeRAZj
...
...
LACL j
SACL DAC1
...
LACL DAC1 ; Load data du DAC 1 into the ACC
SACL DataDAC
SPLK #0010h, DAC_OR
...
CS_DACon ; Active le DAC
; sort sur DAC numero DAC_OR
LDP #0E0h
LACL SPI_PORT_C1 ; SPISTE � 0 to start the transfer
AND #0BFh
SACL SPI_PORT_C1
LDP #DataDAC
LACC DataDAC,8 ; Load data du DataDAC into the Hi ACC
SACH tmp
LACL tmp
AND #000Fh ; Set the DAC control bits (4-7) ou plutot D12-D15
OR DAC_OR ; DAC voie "DAC_OR", SPD=fast mode
LDP #0E0h
SACL SPI_DAT ; Transmit HI byte
LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy�
att1 BIT SPI_STATUS,BIT6
BCND att1,NTC
LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy�
low_byte
LACL DataDAC ; Read the current value of
AND #0FFh ; data, mask off the lower
LDP #0E0h
SACL SPI_DAT ; Transmit the LOW byte
LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy�
; transfert fini ?
att2 BIT SPI_STATUS,BIT6
BCND att2,NTC
LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy�
LACL SPI_PORT_C1 ; en fait : set SPIPC1.6 � 1 => SPISTE pin � 1 ie
latch the Data !
OR #040h
SACL SPI_PORT_C1 ; latch the data into the DAC
CS_DACoff ; Desactive le DAC
where the macro ares :
----------------------
CS_DACon .macro
LDP #00E1h
LACC PADATDIR
AND #0FFFDh ; IOPA1 mise � 0 ie DAC activ�
SACL PADATDIR
.endm
CS_DACoff .macro
LDP #00E1h
LACC PADATDIR
OR #0002h ; IOPA1 mise � 1 ie DAC desactiv�
SACL PADATDIR
.endm
I tried to slower the spiclk, to make the 2nd send by spi int, same pb...
the TLV5614 doc say it has a setting time of 3 us ! yes it is true (I
verifed it) but only, if it takes it into account:-)
???
thanks for your kind help
DAC with a TMS320F240 via SPI interface
Started by ●April 21, 2004
Reply by ●April 22, 20042004-04-22
I solved by myself the pb was with clock polarity (wrong one) this conf works better :-) SPLK #00C7h,SPI_CNTL1 ; Reset SPI by writing 1 to SWRST !!!clk pol1 SPLK #000Ch,SPI_CNTL2 ; Disable ints & TALK, normal clock, master mode SPLK #0000h,SPI_PRI ; Set SPI interrupt to high priority. SPLK #0000h,SPI_BAUD ; Baudrate = Fatest as possible ! SYSCLK/4 et �a marche ! SPLK #0000h,SPI_STATUS ; Clear the SPI interrupt status bits SPLK #0006h,SPI_CNTL2 ; Enable TALK, No CLK ph, master mode + disable SPI Int SPLK #0052h,SPI_PORT_C1 ; Enable the SPICLK pin function. ; SPISTE as output pin et mise � 1 ie desactive SPLK #0022h,SPI_PORT_C2 ; Set SIMO & SOMI functions to serial I/O SPLK #0047h,SPI_CNTL1 ; Release SWRST, clock polarity 1, 8 bits "Lotfi" <llbaghli(nospam)@ifrance.com> a �crit dans le message news: c65hoh$h3d$1@arcturus.ciril.fr...> Hi there, > > I have a pb to reach maximum update rate for a serial Quad DAC TLV 5614with> a TMS320F240 DSP that talks to him through SPI interface. > THe SPI is configured as follow > > LDP #0E0h > SPLK #0087h,SPI_CNTL1 ; Reset SPI by writing 1 to SWRST > SPLK #000Ch,SPI_CNTL2 ; Disable ints & TALK, normal clock,master> mode > SPLK #0000h,SPI_PRI ; Set SPI interrupt to high priority. > SPLK #0000h,SPI_BAUD ; Baudrate = Fatest as possible ! SYSCLK/4 > SPLK #0000h,SPI_STATUS ; Clear the SPI interrupt status bits > SPLK #0006h,SPI_CNTL2 ; Enable TALK, No CLK ph, master mode + > disable SPI Int > SPLK #0052h,SPI_PORT_C1 ; Enable the SPICLK pin function. ; SPISTE > as output pin et mise � 1 ie desactive > SPLK #0022h,SPI_PORT_C2 ; Set SIMO & SOMI functions to serial I/O > SPLK #0007h,SPI_CNTL1 ; Release SWRST, clock polarity 0, 8 bits > > > it sends correct SCLK=(SPICLK pin), /CS=(IOPA1 pin) and /FS =(SPISTE pin) > signals to the DAC > I wrote a routine that is called every 100 us and that refresh the DACDATA,> say I am working only with one DAC (A). > each time it will send the data and increment by 1000 the data to be sent > next time. > it sends, 0 (ie 0V), 1000 (1V), 2000 (2V), 3000 (3V), 4000 (4V) then it > comes back to 0. > the pb is that I don't see all the steps but only changes every 200 us and > not every 100 us > > my code seems correct because it takes into account but not at the right > rate :-( > > every 100 us, these parts of code are called > ... > LACL j > ADD #1000 > SACL j > SUB #1000h > BCND pasdeRAZj,LT > splk #0, j > pasdeRAZj > ... > ... > > LACL j > SACL DAC1 > ... > LACL DAC1 ; Load data du DAC 1 into the ACC > SACL DataDAC > SPLK #0010h, DAC_OR > ... > > CS_DACon ; Active le DAC > ; sort sur DAC numero DAC_OR > LDP #0E0h > LACL SPI_PORT_C1 ; SPISTE � 0 to start the transfer > AND #0BFh > SACL SPI_PORT_C1 > > LDP #DataDAC > LACC DataDAC,8 ; Load data du DataDAC into the Hi ACC > SACH tmp > LACL tmp > AND #000Fh ; Set the DAC control bits (4-7) ou plutot D12-D15 > OR DAC_OR ; DAC voie "DAC_OR", SPD=fast mode > LDP #0E0h > SACL SPI_DAT ; Transmit HI byte > LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy� > att1 BIT SPI_STATUS,BIT6 > BCND att1,NTC > LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy� > > low_byte > LACL DataDAC ; Read the current value of > AND #0FFh ; data, mask off the lower > LDP #0E0h > SACL SPI_DAT ; Transmit the LOW byte > LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy� > > ; transfert fini ? > att2 BIT SPI_STATUS,BIT6 > BCND att2,NTC > LACL SPI_BUF ; clear le SPI_STATUS,BIT6 : donn�e_envoy� > > LACL SPI_PORT_C1 ; en fait : set SPIPC1.6 � 1 => SPISTE pin � 1 ie > latch the Data ! > OR #040h > SACL SPI_PORT_C1 ; latch the data into the DAC > CS_DACoff ; Desactive le DAC > > > where the macro ares : > ---------------------- > CS_DACon .macro > LDP #00E1h > LACC PADATDIR > AND #0FFFDh ; IOPA1 mise � 0 ie DAC activ� > SACL PADATDIR > .endm > CS_DACoff .macro > LDP #00E1h > LACC PADATDIR > OR #0002h ; IOPA1 mise � 1 ie DAC desactiv� > SACL PADATDIR > .endm > > > > I tried to slower the spiclk, to make the 2nd send by spi int, same pb... > the TLV5614 doc say it has a setting time of 3 us ! yes it is true (I > verifed it) but only, if it takes it into account:-) > ??? > > thanks for your kind help > > >
Reply by ●April 22, 20042004-04-22
hi again unfortunately, what works when I no longer MUX (switch between DACs) ie I set only one DAC, does not work if I try to through a signal each 100 us on each of the 4 DACs... it seems complicated... does any one know about MUX time ??? is it more than some micro-secs ?? thanks
Reply by ●April 22, 20042004-04-22
Lotfi wrote:> hi again > > unfortunately, > what works when I no longer MUX (switch between DACs) ie I set only one DAC, > does not work if I try to through a signal each 100 us on each of the 4 > DACs... > > it seems complicated... does any one know about MUX time ??? is it more than > some micro-secs ?? > > thanksDo you have a spec sheet? There is not only the MUX switching time to consider, but also the time for the DAC inputs to settle. Depending on the logic speed, the total could be longer than you like. Can you slow the clock for a test? By the way, in "I have a pb to reach ...", what is a pb? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●April 26, 20042004-04-26
"Jerry Avins" <jya@ieee.org> a �crit dans le message news: 4087f381$0$16472$61fed72c@news.rcn.com...> Lotfi wrote: > > > hi again > > > > unfortunately, > > what works when I no longer MUX (switch between DACs) ie I set only oneDAC,> > does not work if I try to through a signal each 100 us on each of the 4 > > DACs... > > > > it seems complicated... does any one know about MUX time ??? is it morethan> > some micro-secs ?? > > > > thanks > > Do you have a spec sheet? There is not only the MUX switching time to > consider, but also the time for the DAC inputs to settle. Depending on > the logic speed, the total could be longer than you like. Can you slow > the clock for a test? >thank you for your answer> By the way, in "I have a pb to reach ...", what is a pb? >the pb is that it works but not at full speed... But I re-looked at my routine friday and... I discovered a software error :-((( I had a pb on a suspicious DataPage initialisation prior to send my 2nd byte... so it workq fine now... by the way, settling time is 3 us and I had verified it it was correct, I also suspected CLK phase and polarity... in fact it was only sending 8 high bytes... so while my variables change to reach a value that changes over 2^8... times is past... so it it vicious... since it seems to work but not completely... Thanks for your support, it helps to see that people are here and may read and give you a kind of help :-) Ciao
Reply by ●April 26, 20042004-04-26
"Lotfi" <llbaghli(nospam)@ifrance.com> wrote in message news:c6jc4s$bh$1@arcturus.ciril.fr...> > "Jerry Avins" <jya@ieee.org> a �crit dans le message news: > 4087f381$0$16472$61fed72c@news.rcn.com... > > Lotfi wrote: > > > > > hi again > > > > > > unfortunately, > > > what works when I no longer MUX (switch between DACs) ie I set onlyone> DAC, > > > does not work if I try to through a signal each 100 us on each of the4> > > DACs... > > > > > > it seems complicated... does any one know about MUX time ??? is itmore> than > > > some micro-secs ?? > > > > > > thanks > > > > Do you have a spec sheet? There is not only the MUX switching time to > > consider, but also the time for the DAC inputs to settle. Depending on > > the logic speed, the total could be longer than you like. Can you slow > > the clock for a test? > > > > thank you for your answer > > > By the way, in "I have a pb to reach ...", what is a pb? > >I suspect a 'pb' here refers to a 'problem' :-)...if you haven't figured that out already. Cheers Bhaskar> the pb is that it works but not at full speed... > But I re-looked at my routine friday and... I discovered a software error > :-((( > I had a pb on a suspicious DataPage initialisation prior to send my 2nd > byte... > so it workq fine now... > by the way, settling time is 3 us and I had verified it it was correct, I > also suspected CLK phase and polarity... > > in fact it was only sending 8 high bytes... so while my variables changeto> reach a value that changes over 2^8... times is past... so it itvicious...> since it seems to work but not completely... > > Thanks for your support, > it helps to see that people are here and may read and give you a kind of > help :-) > Ciao > > > >
Reply by ●April 26, 20042004-04-26
Bhaskar Thiagarajan wrote: ...> I suspect a 'pb' here refers to a 'problem' :-)...if you haven't figured > that out already. > > Cheers > BhaskarI arrived at that conclusion too. There are plenty of acronyms that people assume everyone knows (but it turns out that I'm not everyone), but it pretty well turns me off when someone who wants my help isn't willing to push the extra keys needed to spell out ordinary words. I'm inclined not to bother with people who are inclined not to bother. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●April 26, 20042004-04-26
"Jerry Avins" <jya@ieee.org> wrote in message news:408d7cf0$0$28899$61fed72c@news.rcn.com...> Bhaskar Thiagarajan wrote: > > ... > > > I suspect a 'pb' here refers to a 'problem' :-)...if you haven't figured > > that out already. > > > > Cheers > > Bhaskar > > I arrived at that conclusion too. There are plenty of acronyms that > people assume everyone knows (but it turns out that I'm not everyone), > but it pretty well turns me off when someone who wants my help isn't > willing to push the extra keys needed to spell out ordinary words. I'm > inclined not to bother with people who are inclined not to bother.But you do bother to ask :-)...which kinda defeats your intention of ignoring the OP.> > Jerry > -- > Engineering is the art of making what you want from things you can get. > ����������������������������������������������������������������������� >
Reply by ●April 26, 20042004-04-26
On Mon, 26 Apr 2004 17:19:27 -0400, Jerry Avins <jya@ieee.org> wrote:>Bhaskar Thiagarajan wrote: > > ... > >> I suspect a 'pb' here refers to a 'problem' :-)...if you haven't figured >> that out already. >> >> Cheers >> Bhaskar > >I arrived at that conclusion too. There are plenty of acronyms that >people assume everyone knows (but it turns out that I'm not everyone), >but it pretty well turns me off when someone who wants my help isn't >willing to push the extra keys needed to spell out ordinary words. I'm >inclined not to bother with people who are inclined not to bother. > >JerryI agree completely. [-Rick-]
Reply by ●April 26, 20042004-04-26
Bhaskar Thiagarajan wrote: ...> But you do bother to ask :-)...which kinda defeats your intention of > ignoring the OP.I didn't promise to ignore him. I asked because it might just have been a bit of jargon I'm not up on. I find his obtuseness when I did ask amazing. Very little could be a better indication of how far his world is from mine. Lofti: read the part in http://users.erols.com/jyavins/procfaq.htm about cryptic abbreviations. Read the rest, too. Several regulars wrote it. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
