SPI SS Issue

[sstt1976]

I've got successful SPI communication between an 18F2550(master) and a 16F887(slave) with the slave SS pin enabled. Here is the code for my master and slave and an analyzer screenshot of successful communication with the master sending 28 and the slave returning 99, as expected:

7474



'SPI MASTER CODE for 18F2550

DEFINE OSC 20

pause 5000

cs VAR PORTB.7
cs2 VAR PORTB.6
SSPEN VAR SSPCON1.5
CKP VAR SSPCON1.4
SMP VAR SSPSTAT.7
CKE VAR SSPSTAT.6
BF VAR SSPSTAT.0
sig VAR byte
sig2 VAR byte

TRISA = %11111111 ' Set PORTA to all input
TRISB = %00000001
TRISC.7 = 0

cs = 1
cs2 = 1

SSPEN = 1

SSPCON1.0 = 0
SSPCON1.1 = 1
SSPCON1.2 = 0
SSPCON1.3 = 0
CKP = 0
CKE = 0
SMP = 0

mainloop:

cs = 0
SSPBUF = 28
while BF = 0 : wend
sig = SSPBUF
cs = 1
pauseus 100
goto mainloop

END





'SPI SLAVE CODE for 16F887

DEFINE OSC 20

SSPEN VAR SSPCON.5
CKP VAR SSPCON.4
SMP VAR SSPSTAT.7
CKE VAR SSPSTAT.6
BF VAR SSPSTAT.0

TRISA = %11111111
TRISB = %11111111
TRISC = %11011111
TRISD = %00000000

SSPCON = %00000100

CKP = 0
CKE = 0
SMP = 0

SSPEN = 1

mainloop:

while BF=0: wend
SSPBUF = 99
goto mainloop

END


The problem is that when I test this without dropping the SS line low to enable I still see the MISO transmission from my slave. I thought that if SS does not go low then the slave chip would not be enabled? Why would the slave respond even when its SS pin is high as seen in the screenshot below from my analyzer?

It seems that my slave chip does not know it is in "SPI Slave mode, clock = SCK pin, SS pin control enabled". I'd like to work this out because when I add more slaves with their own SS lines from my master they all respond at the same time on the MISO line and the data is incorrect because they aren't individually enabling/disabling as expected.

Any insight would be greatly appreciated.

7475

[sstt1976]

This has been solved. The issue was that on the PIC16F887, SS is on RA5 which is also AN4, and for that reason I had to write to a register to enable the digital input buffer because I am using that pin as a digital input. In this case it is ANSEL. It had to be cleared for that pin. Many thanks to Ric.