exxos's Atari, Amiga & retro forum

I wonder if we shouldn’t just switch to the same circuit as MiST for the next revision?

terriblefire wrote: ↑Tue Jan 28, 2020 9:53 am I wonder if we shouldn’t just switch to the same circuit as MiST for the next revision?

I have been reading a little bit more reference manual on host usb and structure of this peripherial, and it has some limitations.
I think moving to max3421e would make porting code from mist much easier.

arkadiusz.makarenko wrote: ↑Tue Jan 28, 2020 11:14 am
I have been reading a little bit more reference manual on host usb and structure of this peripherial, and it has some limitations.
I think moving to max3421e would make porting code from mist much easier.

I guess the question then is if we should keep the STM32 or move to the SAM chip. Personally prefer the STM32 if the firmware will fit. Comms to the usb chip is SPi so that’s easy either way.

terriblefire wrote: ↑Tue Jan 28, 2020 12:20 pm

arkadiusz.makarenko wrote: ↑Tue Jan 28, 2020 11:14 am
I have been reading a little bit more reference manual on host usb and structure of this peripherial, and it has some limitations.
I think moving to max3421e would make porting code from mist much easier.

I guess the question then is if we should keep the STM32 or move to the SAM chip. Personally prefer the STM32 if the firmware will fit. Comms to the usb chip is SPi so that’s easy either way.

Personal preference, don't know much about SAM chip, so STM32.

There is plenty of different STM32 families and versions. It is easy to find faster or bigger ram/flash options.
Another advantage is quite a lot of resources online, even if reference manual is preatty dry, there are plenty tutorials, forums and few books to get some advice in case of issues.

Yeah. It should be possible to hook up the rev0 to a max usb dev board

I just thought I would inject this post into the CD32 Riser thread as it will help some people with some issues

So over on Amibay `The Corfiot`(Bas) bought a TF330 from me and when he used it he was having graphical issues , glitches and the like,
Obviously I thoroughly test all my boards before shipping so was a little surprised so after some emails I decided to send a replacement so we can narrow down the problem.

Anyway Bas luckily being a tech wizzard carried on diagnosing the problem and with some good skills solve his own problem, just to add it was not the TF330 at all thankfully ..but something else

I have copied and pasted Bas`s great email to me as the fault and solution he found, so in his own words below

Hi Mark

Had a bit of a breakthrough

My graphics issue was loss of SYNC...

I tried the other CD32 motherboard and it worked perfectly...

Then I plugged back into my motherboard and it also worked perfectly.

Having been lazy I had hooked up both motherboards using Composite video...
As soon as I switched to an Amiga 23pin RGB cable both motherboards misbehaved.

So using the scope I checked the CSYNC signal on kippers riser and was shocked to find it hitting 5V TTL level.
A1200 Motherboard provides 1.2V

Now this is not a fault of kippers riser as he is using the CSYNC off the edge connector but the CD32 has a wierd video buffering and switching system due to being specified for the mpeg FMV module and it is producing TTL CYNC which of course my monitor is struggling to cope with.

So... I added a 330R resistor inline to the CSYNC signal on my RGB cable and the SYNC level is now as it should be and the graphical glitches are gone.

Most off the shelf Amiga RGB cables tend to have a 330R resistor on CSYNC already as the A4000 also produces TTL CSYNC.. A600 and A1200 dont.. and of course my homemade cable didnt have as I only use a1200/A600

So, that's sorted.

Cheers
Bas

Thanks for that..

Interesting point i've noticed.

The STM32 doesnt seem to allow slow enough speeds for the keyboard.

I tried marshalling SPI to the keyboard and it managed to crash the amiga completely.. v weird.

The crash aside the slowest i can get the SPI port to talk on the ARM is 140Kbps and the amiga need 17kbps.

terriblefire wrote: ↑Wed Jan 29, 2020 9:19 pm Thanks for that..

Interesting point i've noticed.

The STM32 doesnt seem to allow slow enough speeds for the keyboard.

I tried marshalling SPI to the keyboard and it managed to crash the amiga completely.. v weird.

The crash aside the slowest i can get the SPI port to talk on the ARM is 140Kbps and the amiga need 17kbps.

Sorry for ignorant question, but are you using _kb_data and _kb_clock to send keyboard data?

arkadiusz.makarenko wrote: ↑Thu Jan 30, 2020 4:19 pm

terriblefire wrote: ↑Wed Jan 29, 2020 9:19 pm Thanks for that..

Interesting point i've noticed.

The STM32 doesnt seem to allow slow enough speeds for the keyboard.

I tried marshalling SPI to the keyboard and it managed to crash the amiga completely.. v weird.

The crash aside the slowest i can get the SPI port to talk on the ARM is 140Kbps and the amiga need 17kbps.

Sorry for ignorant question, but are you using _kb_data and _kb_clock to send keyboard data?

Yeah. Just doesnt like the speed i'm sending it at.

I have been playing with timers, and time interrupts to do it.
I had code counting it from 0 to 255 in right time intervals.
Haven't tested it on real Amiga yet.