IMX6Q SPI Speed

キャンセル
次の結果を表示 
表示  限定  | 次の代わりに検索 
もしかして: 

IMX6Q SPI Speed

4,406件の閲覧回数
brli
Contributor III

Hi,

Could you suggest how to improve SPI speed?

We are connecting LCD(320x240) device to IMX6Q SPI interface, but its data transfer speed is slow. detail as attached

Free for 9.5 us per 8 bytes during transferring 320*240*2 bytes data size.

Free for 245 ns per 8 bits during transferring 320*240*2 bytes data size.

How can i reduce these free time?

 

Thanks!

Br Li 

ラベル(1)
タグ(1)
0 件の賞賛
返信
5 返答(返信)

4,103件の閲覧回数
igorpadykov
NXP Employee
NXP Employee

Hi br

max. ecspi frequency can be found in sect.4.12.2.1 ECSPI Master Mode Timing

i.MX 6Dual/6Quad Applications Processors for Consumer Products - Data Sheet

So in software one can try to set its clock to max. available values and use

bursts configured by BURST_LENGTH field ECSPIx_CONREG register.

Best regards
igor
-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

0 件の賞賛
返信

4,005件の閲覧回数
angelo_d
Senior Contributor I

Note that there are different speed limit if you only write on TX fifo or only read from RX,
at least so it seems (IMX6DQIEC.pdf 4.12.2.1)

RX  max speed seems 25Mhz for fast pad group, 18Mhz for slow group.

While for TX seems possible to reach 66Mhz.

Is this assumption correct ?

 

 

 

0 件の賞賛
返信

4,103件の閲覧回数
brli
Contributor III

Hi Igorpadykov,

Where can i set the BURST_LENGTH in Linux source code?

Thanks!

Br Li

0 件の賞賛
返信

4,103件の閲覧回数
igorpadykov
NXP Employee
NXP Employee

one can look at ecspi driver

spi-imx.c\spi\drivers - linux-imx - i.MX Linux kernel 

Best regards
igor

0 件の賞賛
返信

4,103件の閲覧回数
brli
Contributor III

Hi Igorpadykov,

Modify the burst length can reduce the the 245ns free time per 8 bits,but my LCD device protocol specified is 8 bits transfer. otherwise it works abnormal.

Can we possible to reduce the 9.5 us free time per 8 bytes, or increase the 8 bytes to 16/24/32 bytes?

Thanks!

Br Li

0 件の賞賛
返信