Fast External Memory

[prstein]

Actually I could use SHIFTIN and SHIFTOUT for accessing the F-RAM memory, as you told, it is a "comedy" wiring error, it seems like totally reversed. But for hardware SPI, I think the MISO,MOSI connection should be followed, right? In your code, it is connected like this.

The only thing I don't like about SHIFTIN/SHIFTOUT is that they are so slow. But following advice from another thread (several threads actually) about SPI I used them to make sure everything was working in some fashion before I stepped up to the hardware SPI. Reversing the MISO/MOSI connections was just one of the many errors I started with. (I also spent an hour trying to figure out why I couldn't get *any* response from the F-RAM; turns out I hadn't actually installed the chip yet!)

Best Regards,
Paul

[Normnet]

The fastest memory I know of would be SRAM parallel.
Just inc the PIC port in parallel with the SRAM to inc the memory address.
Fastest PIC interface is setting a port, faster than SPI etc.

Just search Digikey for SRAM then Parallel for ns transfer time memory.

Norm

[hankshone]

The only thing I don't like about SHIFTIN/SHIFTOUT is that they are so slow. But following advice from another thread (several threads actually) about SPI I used them to make sure everything was working in some fashion before I stepped up to the hardware SPI. Reversing the MISO/MOSI connections was just one of the many errors I started with. (I also spent an hour trying to figure out why I couldn't get *any* response from the F-RAM; turns out I hadn't actually installed the chip yet!)

Best Regards,
Paul

As I found out the reads are all zeros, I use os-scope to track the signals and found out that there is totally no signals out in SCK, therefore no SDI because the master is not generating clocks. the SDO has some data out which I could not understand...So I am think if the MCU hardware got any problem.. or the MSSP module was not running?...it is really a torture...
.....
Vinson

[prstein]

As I found out the reads are all zeros, I use os-scope to track the signals and found out that there is totally no signals out in SCK, therefore no SDI because the master is not generating clocks. the SDO has some data out which I could not understand...So I am think if the MCU hardware got any problem.. or the MSSP module was not running?...it is really a torture...
.....
Vinson

Torture? If it were easy it wouldn't be nearly as much fun...

So no clock signals. SCK is on PORTC.3, which the datasheet says defaults to an input. You'll need to add "SCK_TRIS=0".

One other thing that jumped out at me is that the 16LF877 is only rated for 4MHz and you are using 10MHz according to your defines. Could that also be a source of trouble?

I hope you also took languer's suggestions. I'm pretty sure he's sharper than me...

Best Regards,
Paul

[prstein]

Whoops, looks like you got it first! The rated speed (4MHz) might still be an issue...

[hankshone]

Whoops, looks like you got it first! The rated speed (4MHz) might still be an issue...

It works fine at 10MHz. So I don't think that will introduce problem. the memory writing speed goes up a lot! I will try his codes tomorrow..also looks interesting. anyway thanks for u guys' help!

[hankshone]

The only thing I don't like about SHIFTIN/SHIFTOUT is that they are so slow. But following advice from another thread (several threads actually) about SPI I used them to make sure everything was working in some fashion before I stepped up to the hardware SPI. Reversing the MISO/MOSI connections was just one of the many errors I started with. (I also spent an hour trying to figure out why I couldn't get *any* response from the F-RAM; turns out I hadn't actually installed the chip yet!)

Best Regards,
Paul

Just follow up with the my test. It is interesting that because I didn't specify the SCK's direction which killed the engine for SPI. Right now it is working. A little tricky problem, and anyway I appreciate your help very much!!!