Forum Discussion

SZack's avatar
SZack
Icon for Occasional Contributor rankOccasional Contributor
7 years ago

Does the A5 version of a MAX V have different timing than an I5?

For the MAX V (EPM570 in this case) does the automotive part (A5) have different timing specs than the industrial temp part (I5)?

I have a customer doing a design in a MAX V and it passes timing if he chooses the A5 part but fails timing if he chooses the I5 part.

5 Replies

  • Hi SZack,

    Are you sure it is MAX V device?

    I see EPM'570 belongs to MAX II Family Signature.

    Kindly reconfirm your device by refering to MAX II and MAX V handbook below from Packaging Ordering Information.

    https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/hb/max2/max2_mii5v1.pdf - (MAX II - Page 101)

    https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/hb/max-v/max5_handbook.pdf - (MAX V - Page 12)

    There is no 'Timing' difference in automotive part and industrial part but has difference in 'Operating Temperature'.

    Check the link below - Table 2 (Page 3)

    https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/an/an428.pdf

    Did you test in different temperature room? Because difference in temperature could affect electrical components in FPGA's degrade in time.

    Check the link below, it might be helpful for you.

    https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/wp/wp-01139-timing-model.pdf

    Thank you.

  • SZack's avatar
    SZack
    Icon for Occasional Contributor rankOccasional Contributor
    Hi Calvin, My bad – the part is a MAX V – the 5M570, not EPM570. The device is failing timing in TimeQuest, not the actual board when the I5 part is specified, but it passes timing when the A5 part is chosen. If you want me to send you the project I can. You’ll see the difference in the TimeQuest results depending on whether the I5 or A5 part is chosen. Steve
  • SZack's avatar
    SZack
    Icon for Occasional Contributor rankOccasional Contributor
    Hi Calvin, My bad – the part is a MAX V – the 5M570, not EPM570. The device is failing timing in TimeQuest, not the actual board when the I5 part is specified, but it passes timing when the A5 part is chosen. If you want me to send you the project I can. You’ll see the difference in the TimeQuest results depending on whether the I5 or A5 part is chosen. I sent this email on 9/30 but haven’t seen a response so maybe you didn’t see it. Steve
  • SZack's avatar
    SZack
    Icon for Occasional Contributor rankOccasional Contributor

    Calvin,

    Can you (or someone else) explain why my customer (and I) get different results from TimeQuest depending on whether we choose an I5 or an A5 device? I would think that Quartus would provide the exact same place and route. However, when the customer compiled the same design for 4 different MAX V 570 devices (different speed grade/temp range) here are his results:

    I took the design and compiled it under the 4 different speed grades and looked at FMax. From what I can tell the I5/C5 have same timing, A5 is faster, and C4 is fastest. I know that they keep implying that the A5 is just a "binned" I5/C5 but that is certainly not what the Timing Analyzer seems to be reporting.

    FMAX: A5 C4 C5 I5

    -------------- ----------- ----------- - ---------- -----------

    smclk (24MHz): 30.24 MHz 51.1 MHz 25.48 MHz 25.48 MHz

    lclk (66MHz): 65.33 MHz 128.5 MHz 65.33 MHz 65.33 MHz

    Steve

  • SZack's avatar
    SZack
    Icon for Occasional Contributor rankOccasional Contributor

    Hi Calvin,

    Can you explain the difference in timing reported by TimeQuest when specifying either an I5 or A5 part? When I look in the documentation I would expect them to have the same timing results.