This thread has been locked.

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.

TMS320F28377D: External SRAM Interface Design

Part Number: TMS320F28377D
Other Parts Discussed in Thread: C2000WARE

Hello there,

We're trying to get a 10ns SRAM( ISSI  IS64WV25616BLL-10CTL ) working at the highest possible frequency with the 320F28377D EMIF1. Considering that the EMIF clock can  be set to be equal with thew CPU clock, that is 200MHz, we were hoping to achieve somewhere around 20 ns read/write cycles but  on our custom designed  board the NEM1CS3, NEM1WE and NEM1OE signals need about 50ns to switch from 3.3V to less than 0.4V for a logic "0" and about the same for the rising edge. In order to ensure the SRAM will be able to detect a correct low level (< 0.4V) I had to set the RSTROBE and WSTROBE to 11 (55ns) which is much slower that what we would expect.

On our board the hardware designers used a 33 ohms in series with all control, address and data lines connecting the DSP with the SRAM, the longest trace is about 4 cm.

For comparison I tried running the "emif1_16bit_asram_cpu0" project provided in the C2000Ware_1_00_01_00\device_support\f2837xd\examples on the F2837x controlCARD v1.3 board (without actually having the SRAM connected to the DSP) and noticed that without any load the NEM1CS3, NEM1WE and NEM1OE signals need about 35-40ns for switching from high to low and the same for low to high.

Except for adding signal buffers is it possible to speed up the 320F28377D external memory interface?

Thanks a lot,

Doru

  • Hello,
    I am writing to let you know that a C2000 team member has been assigned to this post and should be answering shortly.

    Regards
    Baskaran
  • Hi,

    35-40ns switching time look too much. What kind of probe you are using to check this? Also could you share the scope snapshot.

    Regards,
    Vivek Singh
  • Hi Vivek,

    Thanks for your reply. The probes that I'm using are Tektronix P2220 (200MHz/6MHz, 10Mohms/ 1Mohms, 16pF/95pF, 10X/1X).  Due to your question I realized I have an embarrassing problem, one of the probes is not behaving correctly, it behaves as it was set on 1X and therefore it limits the signals' bandwidth. With the working probe the rise and fall edges are around 3-4 ns which is much closer to what it should be.

    Thanks a lot for taking your time to answer, the real issue is our probes.

    Best regards,

    Doru