Craigdry,
I pulled the latest spec from the Intel website: http://download.intel.com...atashts/29064523.pdf
It confirms that A20 is the highest address bit for the 32Mbit parts which is the part should be populated on the M5474LITEKIT. Also, during production the product goes through an In Circuit Test as well as a Functional Test which both detect the parts geometry and verifies the known usable address space.

Some things you might want to check if you are using non Logic provided software:
- The flash part is 16bit wide so the address lines out of the processor are shifted up by 1 address line to address it properly. To assert the A20 pin on the Flash chip, AD21 from the processor must be asserted.
- Make sure software correctly sets the BAM bits for chip select zero.
- Make sure software sets the ASET bits = 0x01 for all chip select areas to ensure proper address latch timing.
- Have you confirmed the manufacture's part number is correct? It should be a (TE28F320C3xxxx).

KurtL,

Thank you for your reply. Apparently my document is out of date(1998) and/or the part number doesn't match exactly(28F320S5). It is document number 290609-004. I apologize for that. In my quest to understand the problem, it seemed the perfect explanation since I can address the first 2 megabytes of the chip using CF Flahser, but not the last 2 megabytes. I also wrote a program that sequentially erased sectors. It too worked right up to the 2MB limit and then hung when trying to flash beyond that. So both programs(CF Flasher) and my own are not able to access beyond the first 2 MB. The A20 line is attached to R427. I thought this might be a problem as well. Even the dBUG program access 0's (rather than FFFF) when trying to access beyond the first 2 Meg.

CraigDry,
CF Flasher which is provided by Freescale should work if you have version 3.1 or later. Earlier releases had problems which prevented them from working on all EVB hardware.
Dbug that ships on our kits is also supplied by Freescale and we do update Dbug before shipping kits when updates are available. So I expect that Dbug should work.
Have you tried accessing that area of flash using Lolo? It might help to try it just to get a baseline.

On the ETX board there should be a white sticker with a part number and revision number or letter. Post that here and I can see what else I can learn out about your specific hardware.

One other thing you might want to check is R427. Make sure it does not look physically damaged. Use an ohm meter to check to make sure it is making connection between the two pads that it connects on the PCB. It is possible the part may be damaged.

KurtL,

CF Flasher is version 3.1.

The white sticker on the small daughter card is 1000164REV1.

There is a larger white sticker on the daughter card with a bar
code and 4504M00034.

It measures .3 ohm across the resistor at R427

Thanks for your help.

Have you confirmed the flash part number is TE28F320C3BD70?
Have you tried Lolo to excercise the full range of the flash chip?
Do you see the A21 address line toggle using an oscilloscope when you attempt to read/write to that area?

Do you have CSAR0, CSMR0, and CSCR0 setup properly for the 4Mbyte chip? How are they set now?

Thanks.