If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Off board SPI communication over 2 meter Length Bus

Hello ,

I have few queries on my SPI interface application.

Specifications:

Sensor: LDC1000. Controller: F28M35H52C

Data Out: Over SPI bus, f = 4MHz. SPI Bus Length = 2 meters.

Raising edge= Data out, Falling edge= Data In.

I want to use Differential drivers to have noise immunity.

After going through the online documents on extended SPI communication, i have done below calculation.

f=4MHz => T=250ns. After Falling edge of SCLK, available time to get data from slave is

250ns/2 = 125ns. Otherwise it will be a transmit error (correct me if i am wrong ).

Propagation delay calculation:

1. Delay due to 2meter SPI bus = 10ns (signal velocity= 1m in 5ns)

2. Differential driver delay = 12ns. Differential receiver delay=27ns.

3. Sensor response time = 20ns

So total propagation delay between SCLK and SOMI is

2*(10ns+12ns+27ns)+ 20ns = 118ns.

available Extra time = 125ns - 118ns = 7ns.

Queries :

1. Is this calculation correct/accurate??

2. If yes, are there any risks in practical implementation??

3. What are the other better techniques ??

Note:

1. I have to use SPI interface only as per system requirements.

2. I am trying to avoid extra SPI interface as suggested in TI APN:

Regards,

Govind.

4 Replies

• Govind,

The problem is the SPI interface is clocked by master. So delay is big problem for slave transmit. Dual SPI is immune to delay (other than differential delay of clock/data) because only the master transmits.

For long distance I see no other way. Your math show the problem with bidirectional data with unidirectional clock.

Regards,
Ronald Michallick
Linear Applications

TI assumes no liability for applications assistance or customer product design. Customer is fully responsible for all design decisions and engineering with regard to its products, including decisions relating to application of TI products. By providing technical information, TI does not intend to offer or provide engineering services or advice concerning Customer's design. If Customer desires engineering services, the Customer should rely on its retained employees and consultants and/or procure engineering services from a licensed professional engineer (LPE).

• In reply to Ron Michallick:

Hello Mr.Ron,

Thanks for feedback. If i reduce the Clock frequncy to 3MHz, available time increased to 48ns. Is this time enough to have proper communication??

• In reply to Govind K:

Hello Mr. Ron,

Is it possible to impliment this interface without differential line drivers by using proper termination networks??

As per Controller Data sheet, SPI driver out put impedance is 165 Ohms. If i use 100 ohm Cat 5 cable (2 meter lenght),  How to match the impedance??

• In reply to Govind K:

Govind,

The simple answer is to add 253 ohms in parallel with the transmission line. Now the source looks like 100 ohms assuming the source' s 165 ohms is resistive. 165 || 253 = 100.

The voltage gain without the cable is 253*(253+165)=60%. (40% reduction)

Regards,
Ronald Michallick
Linear Applications

TI assumes no liability for applications assistance or customer product design. Customer is fully responsible for all design decisions and engineering with regard to its products, including decisions relating to application of TI products. By providing technical information, TI does not intend to offer or provide engineering services or advice concerning Customer's design. If Customer desires engineering services, the Customer should rely on its retained employees and consultants and/or procure engineering services from a licensed professional engineer (LPE).