AM335x GPMC ACCESSTIME

Kzk

Hello,

I'd like to know AccessTime of GPMC for asynchronous NOR. The tCE of NOR is 85ns, and FCLK is 104MHz (= 9.615ns). ACCESSTIME is calculated by the following formula. ACCESSTIME = round max (tCE) = 85ns / 9.615ns = 8.84 -> 9 (= 86.538ns). But in this case, I think there is no setup time (86.538ns - 85ns = 1.538ns). Should ACCESSTIME be set 10 as plus 1?

Regards,
Kazu

over 10 years ago

0 Biser Gatchev-XID over 10 years ago

TI__Guru**** 393215 points

Hi Kazu,

ACCESSTIME is calculated from the beginning of the access cycle. See section 7.1.5.1.4 in the AM335X TRM Rev. L for details.

0 Kzk over 10 years ago in reply to Biser Gatchev-XID

Genius 5785 points

Hello Biser,

Yes, I've seen the description of TRM and below.
e2e.ti.com/.../870009

If ACCESSTIME is set same value for each tCE, the difference of setup time occurs. I think 1.535ns is too short for setup time.

[tCE=80] 9 * 9.615 - 80 = 86.535 - 80 = 6.535 ns
[tCE=85] 9 * 9.615 - 85 = 86.535 - 85 = 1.535 ns

Please give me some advice.

Regards,
Kazu

0 Biser Gatchev-XID over 10 years ago in reply to Kzk

TI__Guru**** 393215 points

If your GPMC_FCLK period is 9.615ns and tCE is 80ns then: ACCESSTIME = 80 / 9.615 = 8.32 rounded up to 9 GPMC_FCLK cycles.

If your GPMC_FCLK period is 9.615ns and tCE is 85ns then: ACCESSTIME = 85 / 9.615 = 8.84 again rounded up to 9 GPMC_FCLK cycles.

I don't know where you get this setup time from.

0 Kzk over 10 years ago in reply to Biser Gatchev-XID

Genius 5785 points

Hello Biser,

I calculated actual setup time as below. Is my understanding uncorrect?

Regards,
Kazu

0 Biser Gatchev-XID over 10 years ago in reply to Kzk

TI__Guru**** 393215 points

I see what you mean now. However data is latched from the bus on OEOFFTIME rising edge, so there is one extra GPMC_FLCK cycle to compensate for this.

Processors

Processors forum

AM335x GPMC ACCESSTIME