SPI Bus board for RC2014

[Jay Cotton]

I have pushed a beta version of my SPI bus board and driver code to:

https://sourceforge.net/projects/spi-bus-z80-rc2014/

The board is at OSHPARK.COM   https://www.oshpark.com/shared_projects/K5Ffb6B9

This drop includes a driver for the SPI bus board and a few lame support programs that

allow directory viewing and copying of files.  The file code is based on FatFS, and is compiled

using z88dk.

These both will need a lot of work to be truly useful.

It is my hope that in time I can get the driver working with RomWBW,  thus closing the loop

....  spinning rust, CF cards and SD cards.

My driver is probably not nearly ready for RomWBW yet.

I'd like to recognize J.B. Langsdon for the inspiration for this project and some of his code

can be found in the sdir.c program.

Notes about performance.

The spi bus spec says it will work at 20mhz or better.  Not much need to worry about hitting that

limit with the Z80.  The good news is, the Z80 supplies the data clock, and since I bit bang the

data, we are good and slow.

I have copied 2 mb files from the sd card to the cf card using scopy.  This takes quite some time 

to complete.  No idea yet what the throughput is, but its low.

I suspect that some optimizations can be applied to the FatFS code that will make it smaller and

somewhat faster.  My SPI.C driver is a combination of C and asm code that I optimized by looking

for low hanging water melons.  

Well, I hope people find this to be useful.

jc

[J.B. Langston]

Cool project! I had thought about doing something similar at one point.  There are also people who have used shift registers with the Z80 to make SPI faster than bit banging the data and clock lines... maybe an idea for future optimization.  To give proper credit, that dir function actually came from the fatfs AVR demo by ChaN... I only slightly adapted it to my project.

[Jay Cotton]

TNX JB.

O.k. so I modified your modifications to ChaN's code.  Just making sure every one gets

credit.

mmm  shift registers.  Well that's a good idea, one I completely missed.  I think I'll stick with the bit banger for now, its most

malleable, unlike h/w.

jc

[Jay Cotton]

The boards came in today.  

I had to make a small rework do to a routing error (on my part).

The board is now running.  I'll post some updates to the project

in a few days, and call it done.

jc

[Jay Cotton]

sourceforge updated with new board layout.  Small bug in the layout that was easy to fix.

Here is a link to the new board.  Old one is deleted.

https://www.oshpark.com/shared_projects/IBGLloBl

including a screen shot of the code running.

[Jay Cotton]

Since size matters with the z80, sdir is currently 36k

and scopy is 40k.  

On Thursday, June 7, 2018 at 8:29:20 PM UTC-7, Jay Cotton wrote:

[Jay Cotton]

Updated the board again.  Never trust autoroute.

Added multi disk support (still needs some work).

Been trying to read a USB stick... no joy yet.

jc

On Thursday, June 7, 2018 at 8:29:20 PM UTC-7, Jay Cotton wrote:

[Jay Cotton]

USB.  Needs a different fatfs handler.   :(

[Jay Cotton]

Well, I am making progress on the usb via spi driver.  Had it reading the status from the chip last night, then it fritz'd out.

I suspect a compile bug (compiler not doing the right thing) we will see about that....

Getting very close to reading the usb drive....  can't wait.

jc

[Jay Cotton]

Self inflicted bug.  Now reading registers on the controller and so on.  Not seeing the usb bus 

activity I expected.  

One thing that really held me up was the lack of a pull down resister on the miso line.  Added a 10k and that

got the miso working, btw a 10k pull up did not work...  

jc

[Jay Cotton]

flaky code or something.

today I was able to detect a keyboard plug in.  Now, I can seem to repeat the

success.    various combinations of power cycle, reset and swearing does

not seem to work now.

On Thursday, June 7, 2018 at 8:29:20 PM UTC-7, Jay Cotton wrote:

[Jay Cotton]

I'm taking a break from the USB board.  Now I have this nifty 8 digit led board that is spi buss based.

I have a driver working for it and it seems very stable.

I think at this point, my spi board is working and will only need a code cleanup.  So I will work on that

this week and push the v1.0 code.

I looked at using the jtlpcb site for the final v1 board, its a bit fiddly, but for the cost it should be worth

the effort to get the board from them.

guess Ill get back to usb latter next week..