TheBackShed.com

Forum Index : Microcontroller and PC projects : MMII - SPI Bulk Send Problem

Author

Message

srnet
Senior Member

Joined: 08/08/2014
Location: United Kingdom
Posts: 164

Posted: 04:53am 01 Feb 2015

Copy link to clipboard

Print this post

I have been porting my RFM98 LoRa routines from PICAXE to MMII.

I wrote a routine to print out the register contents of the RFM98 with SPI commands. This works and the output display is correct for a RFM98 after reset;

RFM98Print:
Reg 0 1 2 3 4 5 6 7 8 9 A B C D E F
00 00 09 1A 0B 00 52 6C 80 00 4F 09 2B 20 08 02 0A

I next wanted to set the frequency registers by writing values to registers $06, $07, $08, the default values, after reset, are as above $6C, $80, $00

For 434.400Mhz the values are, $6C, $99, $99

So this is the basic routine;

GOSUB RFM98Print 'print values at reset

PIN(NSS) = 0
SPI WRITE 4, &H86, &H6C, &H99, &H99
PIN(NSS) = 1

GOSUB RFM98Print 'check frequency is programmed

However at the second RFM98Print all the registers report as $00, the RFM98 appears to be blank;

RFM98Print:
Reg 0 1 2 3 4 5 6 7 8 9 A B C D E F
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

If the code is now changed to;

GOSUB RFM98Print 'print values at reset

PIN(NSS) = 0
SPI WRITE 4, &H86, &H6C, &H99, &H99
PIN(NSS) = 1

SPI CLOSE
SPI OPEN 1000000, 3, 8

GOSUB RFM98Print 'check frequency is programmed

The result is;

RFM98Print:
Reg 0 1 2 3 4 5 6 7 8 9 A B C D E F
00 00 09 1A 0B 00 52 6C 99 99 4F 09 2B 20 08 02 0A

Which is as expected.

It appears to me that the SPI WRITE command is working correctly, but something is preventing subsequent reads from working, unless the SPI interface is closed and opened after the SPI WRITE.

Same symptoms with an RFM22B.

Edited by srnet 2015-02-02

$50SAT is Silent but probably still working.
For information on LoRa visit http://www.loratracker.uk/

WhiteWizzard
Guru

Joined: 05/04/2013
Location: United Kingdom
Posts: 2932

Posted: 05:01am 01 Feb 2015

Copy link to clipboard

Print this post

Could be a timing issue!
Please give the code for your RF98Print sub and this may provide more clues . . .

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10197

Posted: 05:14am 01 Feb 2015

Copy link to clipboard

Print this post

I'm pretty certain this is because the SPI WRITE firmware isn't also reading bytes which means the SPI receive buffer is filling up. I had this problem when coding direct to the SPI registers in my Cfunctions.

Geoff, looking at the SPI read code, I think there needs to be a statement:

junk=SpiChnGetC(SPI_Channel);

inside the end of the "while(nbr--)" loop to stop the receive buffer overflowing.

As a workround, try using i=SPI() to write and read the bytes 1 at a time rather than closing and opening the port

Edited by matherp 2015-02-02

srnet
Senior Member

Joined: 08/08/2014
Location: United Kingdom
Posts: 164

Posted: 05:59am 01 Feb 2015

Copy link to clipboard

Print this post

Here is the whole program, I would be amazed if its a timing issue, the print routine behaves properly if I use the WriteRFM SUB to write the registers one at a time;

'OPTION AUTORUN ON

DIM Integer LOOPV1
DIM Integer LOOPV2
DIM Integer Var1
DIM Integer RFMDATA
DIM Integer RFMREG

CONST NSS = 26
CONST RFM98RESET = 15

CONST trTXfrMSB = &H6C 'MSB of frequency setting for 434.400 Mhz, tracker TX
CONST trTXfrMid = &H99 'Mid of frequency setting for 434.400 Mhz, tracker TX
CONST trTXfrLSB = &H99 '(&99 or 153) LSB of frequency setting for 434.400 Mhz, , tracker TX

CONST RegFrMSB = &H06
CONST RegFrMid = &H07
CONST RegFrLSB = &H08

CONST WRegFrMSB = &H86

gosub startup

GOSUB RFM98Print

'WriteRFM &H86,trTXfrMSB
'WriteRFM &H87,trTXfrMid
'WriteRFM &H88,trTXfrLSB

PIN(NSS) = 0
SPI WRITE 4, &H86, &H6C, &H99, &H99
PIN(NSS) = 1

'SPI CLOSE
'SPI OPEN 1000000, 3, 8

GOSUB RFM98Print

PRINT "Finished"

END

RFM98Print:
'prints the contents of RFM98 registers to serial terminal
PRINT ""
PRINT "RFM98Print:"
PRINT "Reg 0 1 2 3 4 5 6 7 8 9 A B C D E F"

RFMREG = 0 'point to first register of RFM98

for loopv1 = 0 to 7 '8 lines
PRINT hex$(loopv1,1);
PRINT "0 ";

for loopv2 = 0 to 15 '16 bytes per line
PRINT hex$(READREG(RFMREG),2);
PRINT " ";
RFMREG = RFMREG + 1
next loopv2

PRINT ""
next loopv1
PRINT ""
return

STARTUP:
PRINT "RFM98Print Test"
SPI OPEN 1000000, 3, 8
SETPIN NSS, DOUT
PIN(NSS) = 1
SETPIN RFM98RESET, DOUT
GOSUB RESETRFM98
return

FUNCTION READREG(RFMREG) 'uses rfmreg to get regdata
PIN(NSS) = 0
READREG = SPI(RFMREG) 'send register address for a read 0-&7F
READREG = SPI(0) 'and read it in
PIN(NSS) = 1
END FUNCTION

RESETRFM98:
PIN(RFM98RESET) = 1
PAUSE 5
PIN(RFM98RESET) = 0
return

SUB WriteRFM(RFMREG,RFMDATA)
RFMREG = RFMREG OR &H80
PRINT "Writing $";
PRINT hex$(RFMREG,2);
PRINT ",$";
PRINT hex$(RFMDATA,2)
PIN(NSS) = 0
VAR1 = SPI(RFMREG) 'send register address for a write $80 to $FF
VAR1 = SPI(RFMDATA) 'and write data out
PIN(NSS) = 1
END SUB

$50SAT is Silent but probably still working.
For information on LoRa visit http://www.loratracker.uk/

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 10197

Posted: 09:36am 01 Feb 2015

Copy link to clipboard

Print this post

We have confirmed that there is a bug in SPI WRITE which causes problems for subsequent SPI READs. I'll PM Geoff to get it on the list.

Print this page

To reply to this topic, you need to log in.