SPI Interrupt and other ?'s from SPI Newbie

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
lock attach
Attachments are accessible only for community members.
joli_601446
Level 5
Level 5
10 likes given 10 sign-ins 5 sign-ins

Hello, I'm starting out with a simple SPI program in order to learn SPI. I'm using PSoc5 kit to program an SPI Master and SPI Slave. I'm getting data from the Slave but I have few questions.

1.  When I use the Function call SPIS_WriteTXDataZero() this puts an extra byte in the TX FIFO and the Master SPI must send an extra word to get all of the data in the Slave TX FIFO clocked out.  Doesn't this cause some problems, in managing what's in the FIFO and getting things aligned properly. I guess I could have the Master SPI send and extra word on the first SPI Master write. Comments?

2. I'm writing 8-bytes to the Master TX FIFO but on the scope I see a 50usec gap between the first four bytes and the remaining 4 bytes. Why?

Transmit_Gap.jpg

3. Last, I'm tying to get an interrupt on the SPI Slave to work. I have tried to implement an Interrupt Service Routine but the data from the Slave RX FIFO seems to be missing some words:

SlaveFifo_Array.jpg

I would expect to see 8-bytes in the dsp_spi_rcv_buffer but I'm only getting 4. Why?

Can anyone help me out please.

Thanks,

Joe

0 Likes
1 Solution
MotooTanaka
Level 9
Level 9
Distributor - Marubun (Japan)
First comment on blog Beta tester First comment on KBA

Dear Joe-san,

I have written at least several programs using SPI slave device(s) with MCUs including Cypress, NXP,

and I have never seen a slave device which requires an extra byte for sending data.

And as I wrote in my previous reply, PSoC's SPIS also does not require an extra byte,

at least it was receiving the data into its rx buffer.

So I don't think you need to worry about the extra byte sending with XRA1403.

moto

View solution in original post

0 Likes
6 Replies
lock attach
Attachments are accessible only for community members.
MotooTanaka
Level 9
Level 9
Distributor - Marubun (Japan)
First comment on blog Beta tester First comment on KBA

Hi,

First of all, you can refer to the component's datasheet from its configure Dialog.

datasheet_button.JPG

Meantime, although written for PSoC4, following 2 application notes seem to be very helpful information.

(I've tried polling and dma samples and they worked fine.)

=============================

CE224339 - PSoC 4 SPI Master

http://www.cypress.com/documentation/code-examples/ce224339-psoc-4-spi-master

CE224463 - PSoC 4 SPI Slave

http://www.cypress.com/documentation/code-examples/ce224463-psoc-4-spi-slave

=============================

1. Recently I encountered similar question.

  The SPIS writes 0 in the first byte of the buffer.

So if we want to send 7 bytes, let's say data[1]~data[7].

We need to first call SPIS_WriteTXDataZero() with data[1].

Then we need to SPIS_WriteTxData for data[2], ... data[7].

Note: Please refer to the datasheet of SPIS

So I would try

==================

SPIS_WriteTxDataZero(0x1u) ;

SPIS_WriteTxData(0x2u) ;

SPIS_WriteTxData(0x3u) ;

SPIS_WriteTxData(0x4u) ;

SPIS_WriteTxData(0x5u) ;

SPIS_WriteTxData(0x6u) ;

SPIS_WriteTxData(0x7u) ;

==================

With the application notes mentioned above,

I did not notice that we need extra cycle for the master to

clock out the last data, but as far as I tried today,

yes, I needed the "one more" byte cycle.

2. According to the datasheet of SPIM, FIFO is implemented

using 4-byte/word FIFO, and I suppose this is causing interval(s)

between each 4bytes.

3. As you showed in 2. the transaction(s) took place in 2 parts,

and I suppose the ISR was called at the first transaction.

As far as I tried with CY8CKIT-046, ISR was called multiple times

Attached is my trial with CY8CKIT-046 based on your project.

Hopefully the "main.c" will work for PSoC 5...

teraterm_log.JPG

moto

0 Likes

Moto,

Hello, thank you responding to my message. Do you know if actual SPI Slave Devices  that have SPI Mode CPHA(0),CPOL(0) require an extra byte from the Master? Or, is that just present with PSoC SPI modules? I'm using the PSoC5 to interface with a XRA1403 GPIO expander. The dataheet does not mention this extra byte, so I'm contacting the company.

Last, when I send the extra byte  I get all 8 words in Slave receiver buffer in the ISR.

Joe

0 Likes

Another thing, I changed the Interrupt "Interrupt Type" to Level. After making this switch the I was able to read all 8-bytes in the Slave Receive buffer. Perhaps I will missing an interrupt before and this change fixed things.

Joe

0 Likes
lock attach
Attachments are accessible only for community members.
MotooTanaka
Level 9
Level 9
Distributor - Marubun (Japan)
First comment on blog Beta tester First comment on KBA

Dear Joe-san,

With your reply, I played with my code a little more and

I added a debug line to display remaining slave's rx buffer size,

and I noticed that there was 1 buffer left unread.

Then I added some lines to salvage remaining rx buffer of Slave to the dsp_spi_rcv_buffer.

After this without sending one more byte cycle my program can receive all 7 bytes.

So my conclusion is it was matter of interrupt timing and using SPIS_WriteTxDataZero()

does not add requirement of additional byte clock.

000-teraterm_log.JPG

Attached is my updated project for CY8CKIT-046 and I "hope" that the main.c should be usable for 5LP, too.

moto

0 Likes
MotooTanaka
Level 9
Level 9
Distributor - Marubun (Japan)
First comment on blog Beta tester First comment on KBA

Dear Joe-san,

I have written at least several programs using SPI slave device(s) with MCUs including Cypress, NXP,

and I have never seen a slave device which requires an extra byte for sending data.

And as I wrote in my previous reply, PSoC's SPIS also does not require an extra byte,

at least it was receiving the data into its rx buffer.

So I don't think you need to worry about the extra byte sending with XRA1403.

moto

0 Likes

Thanks Moto!

0 Likes